Beej’s Guide to C Programming
Brian “Beej Jorgensen” Hall
v0.6.35, Copyright © September 30, 2021 Contents
1 Foreword 1 1.1 Audience ...... 1 1.2 Platform and Compiler ...... 2 1.3 Official Homepage ...... 2 1.4 Email Policy ...... 2 1.5 Mirroring ...... 2 1.6 Note for Translators ...... 2 1.7 Copyright and Distribution ...... 3
2 Hello, World! 4 2.1 What to Expect from C ...... 4 2.2 Hello, World! ...... 5 2.3 Compilation Details ...... 6 2.4 Building with gcc ...... 7 2.5 Building with clang ...... 7 2.6 Building from IDEs ...... 7 2.7 C Versions ...... 8
3 Variables and Statements 9 3.1 Variables ...... 9 3.1.1 Variable Names ...... 9 3.1.2 Variable Types ...... 10 3.1.3 Boolean Types ...... 11 3.2 Operators and Expressions ...... 12 3.2.1 Arithmetic ...... 12 3.2.2 Ternary Operator ...... 12 3.2.3 Pre-and-Post Increment-and-Decrement ...... 13 3.2.4 The Comma Operator ...... 13 3.2.5 Conditional Operators ...... 14 3.2.6 Boolean Operators ...... 14 3.2.7 The sizeof Operator ...... 15 3.3 Flow Control ...... 15 3.3.1 The if-else statement ...... 16 3.3.2 The while statement ...... 17 3.3.3 The do-while statement ...... 18 3.3.4 The for statement ...... 19 3.3.5 The switch Statement ...... 19
4 Functions 22 4.1 Passing by Value ...... 23 4.2 Function Prototypes ...... 24 4.3 Empty Parameter Lists ...... 25
5 Pointers—Cower In Fear! 26 5.1 Memory and Variables ...... 26 5.2 Pointer Types ...... 28 5.3 Dereferencing ...... 29 5.4 Passing Pointers as Arguments ...... 29
i CONTENTS ii
5.5 The NULL Pointer ...... 31 5.6 A Note on Declaring Pointers ...... 31 5.7 sizeof and Pointers ...... 32
6 Arrays 33 6.1 Easy Example ...... 33 6.2 Getting the Length of an Array ...... 34 6.3 Array Initializers ...... 34 6.4 Out of Bounds! ...... 36 6.5 Multidimensional Arrays ...... 36 6.6 Arrays and Pointers ...... 37 6.6.1 Getting a Pointer to an Array ...... 37 6.6.2 Passing Single Dimensional Arrays to Functions ...... 38 6.6.3 Changing Arrays in Functions ...... 39 6.6.4 Passing Multidimensional Arrays to Functions ...... 40
7 Strings 41 7.1 String Literals ...... 41 7.2 String Variables ...... 41 7.3 String Variables as Arrays ...... 41 7.4 String Initializers ...... 42 7.5 Getting String Length ...... 43 7.6 String Termination ...... 43 7.7 Copying a String ...... 44
8 Structs 46 8.1 Declaring a Struct ...... 46 8.2 Struct Initializers ...... 47 8.3 Passing Structs to Functions ...... 47 8.4 The Arrow Operator ...... 48 8.5 Copying and Returning structs ...... 49
9 File Input/Output 50 9.1 The FILE* Data Type ...... 50 9.2 Reading Text Files ...... 51 9.3 End of File: EOF ...... 51 9.3.1 Reading a Line at a Time ...... 52 9.4 Formatted Input ...... 53 9.5 Writing Text Files ...... 53 9.6 Binary File I/O ...... 54 9.6.1 struct and Number Caveats ...... 55
10 typedef: Making New Types 57 10.1 typedef in Theory ...... 57 10.1.1 Scoping ...... 57 10.2 typedef in Practice ...... 57 10.2.1 typedef and structs ...... 57 10.2.2 typedef and Other Types ...... 59 10.2.3 typedef and Pointers ...... 59 10.2.4 typedef and Capitalization ...... 59 10.3 Arrays and typedef ...... 60
11 Pointers II: Arithmetic 61 11.1 Pointer Arithmetic ...... 61 11.1.1 Adding to Pointers ...... 61 11.1.2 Changing Pointers ...... 62 11.1.3 Subtracting Pointers ...... 63 11.2 Array/Pointer Equivalence ...... 63 11.2.1 Array/Pointer Equivalence in Function Calls ...... 64 CONTENTS iii
11.3 void Pointers ...... 65
12 Manual Memory Allocation 69 12.1 Allocating and Deallocating, malloc() and free() ...... 69 12.2 Error Checking ...... 70 12.3 Allocating Space for an Array ...... 70 12.4 An Alternative: calloc() ...... 71 12.5 Changing Allocated Size with realloc() ...... 72 12.5.1 Reading in Lines of Arbitrary Length ...... 73 12.5.2 realloc() with NULL ...... 74 12.6 Aligned Allocations ...... 75
13 Scope 77 13.1 Block Scope ...... 77 13.1.1 Where To Define Variables ...... 77 13.1.2 Variable Hiding ...... 78 13.2 File Scope ...... 78 13.3 for-loop Scope ...... 79 13.4 A Note on Function Scope ...... 79
14 Types II: Way More Types! 80 14.1 Signed and Unsigned Integers ...... 80 14.2 Character Types ...... 81 14.3 More Integer Types: short, long, long long ...... 82 14.4 More Float: double and long double ...... 84 14.4.1 How Many Decimal Digits? ...... 85 14.4.2 Converting to Decimal and Back ...... 86 14.5 Constant Numeric Types ...... 87 14.5.1 Hexadecimal and Octal ...... 87 14.5.2 Integer Constants ...... 88 14.5.3 Floating Point Constants ...... 89
15 Types III: Conversions 91 15.1 String Conversions ...... 91 15.1.1 Numeric Value to String ...... 91 15.1.2 String to Numeric Value ...... 92 15.2 Numeric Conversions ...... 94 15.2.1 Boolean ...... 94 15.2.2 Integer to Integer Conversions ...... 94 15.2.3 Integer and Floating Point Conversions ...... 95 15.3 Implicit Conversions ...... 95 15.3.1 The Integer Promotions ...... 95 15.3.2 The Usual Arithmetic Conversions ...... 95 15.3.3 void* ...... 96 15.4 Explicit Conversions ...... 96 15.4.1 Casting ...... 96
16 Types IV: Qualifiers and Specifiers 98 16.1 Type Qualifiers ...... 98 16.1.1 const ...... 98 16.1.2 restrict ...... 100 16.1.3 volatile ...... 101 16.1.4 _Atomic ...... 101 16.2 Storage-Class Specifiers ...... 101 16.2.1 auto ...... 101 16.2.2 static ...... 101 16.2.3 extern ...... 102 16.2.4 register ...... 103 16.2.5 _Thread_local ...... 104 CONTENTS iv
17 Multifile Projects 105 17.1 Includes and Function Prototypes ...... 105 17.2 Dealing with Repeated Includes ...... 107 17.3 static and extern ...... 108 17.4 Compiling with Object Files ...... 108
18 The Outside Environment 109 18.1 Command Line Arguments ...... 109 18.1.1 The Last argv is NULL ...... 111 18.1.2 The Alternate: char **argv ...... 111 18.1.3 Fun Facts ...... 112 18.2 Exit Status ...... 112 18.2.1 Other Exit Status Values ...... 114 18.3 Environment Variables ...... 114 18.3.1 Setting Environment Variables ...... 115 18.3.2 Unix-like Alternative Environment Variables ...... 115
19 The C Preprocessor 117 19.1 #include ...... 117 19.2 Simple Macros ...... 118 19.3 Conditional Compilation ...... 118 19.3.1 If Defined, #ifdef and #endif ...... 118 19.3.2 If Not Defined, #ifndef ...... 119 19.3.3 #else ...... 120 19.3.4 General Conditional: #if, #elif ...... 120 19.3.5 Losing a Macro: #undef ...... 121 19.4 Built-in Macros ...... 122 19.4.1 Mandatory Macros ...... 122 19.4.2 Optional Macros ...... 123 19.5 Macros with Arguments ...... 123 19.5.1 Macros with One Argument ...... 123 19.5.2 Macros with More than One Argument ...... 124 19.5.3 Macros with Variable Arguments ...... 125 19.5.4 Stringification ...... 126 19.5.5 Concatenation ...... 126 19.6 Multiline Macros ...... 126 19.7 Example: An Assert Macro ...... 127 19.8 The #error Directive ...... 128 19.9 The #pragma Directive ...... 128 19.9.1 Non-Standard Pragmas ...... 129 19.9.2 Standard Pragmas ...... 129 19.9.3 _Pragma Operator ...... 129 19.10 The #line Directive ...... 130 19.11 The Null Directive ...... 130
20 structs II: More Fun with structs 131 20.1 Initializers of Nested structs and Arrays ...... 131 20.2 Anonymous structs ...... 133 20.3 Self-Referential structs ...... 133 20.4 Flexible Array Members ...... 134 20.5 Padding Bytes ...... 136 20.6 offsetof ...... 136 20.7 Fake OOP ...... 137 20.8 Bit-Fields ...... 138 20.8.1 Non-Adjacent Bit-Fields ...... 139 20.8.2 Signed or Unsigned ints ...... 139 20.8.3 Unnamed Bit-Fields ...... 139 20.8.4 Zero-Width Unnamed Bit-Fields ...... 140 20.9 Unions ...... 140 CONTENTS v
20.9.1 Pointers to unions ...... 141
21 Characters and Strings II 143 21.1 Escape Sequences ...... 143 21.1.1 Frequently-used Escapes ...... 143 21.1.2 Rarely-used Escapes ...... 144 21.1.3 Numeric Escapes ...... 145
22 Enumerated Types: enum 147 22.1 Behavior of enum ...... 147 22.1.1 Numbering ...... 147 22.1.2 Trailing Commas ...... 148 22.1.3 Scope ...... 148 22.1.4 Style ...... 148 22.2 Your enum is a Type ...... 148
23 Pointers III: Pointers to Pointers and More 150 23.1 Pointers to Pointers ...... 150 23.1.1 Pointer Pointers and const ...... 152 23.2 Multibyte Values ...... 153 23.3 The NULL Pointer and Zero ...... 155 23.4 Pointers as Integers ...... 155 23.5 Pointer Differences ...... 155 23.6 Pointers to Functions ...... 156
24 Bitwise Operations 159 24.1 Bitwise AND, OR, XOR, and NOT ...... 159 24.2 Bitwise Shift ...... 159
25 Variadic Functions 161 25.1 Ellipses in Function Signatures ...... 161 25.2 Getting the Extra Arguments ...... 162 25.3 va_list Functionality ...... 163 25.4 Library Functions That Use va_lists ...... 163
26 Locale and Internationalization 165 26.1 Setting the Localization, Quick and Dirty ...... 165 26.2 Getting the Monetary Locale Settings ...... 166 26.2.1 Monetary Digit Grouping ...... 167 26.2.2 Separators and Sign Position ...... 167 26.2.3 Example Values ...... 168 26.3 Localization Specifics ...... 168
27 Unicode, Wide Characters, and All That 170 27.1 What is Unicode? ...... 170 27.2 Code Points ...... 170 27.3 Encoding ...... 171 27.4 Source and Execution Character Sets ...... 172 27.5 Unicode in C ...... 172 27.6 A Quick Note on UTF-8 Before We Swerve into the Weeds ...... 173 27.7 Different Character Types ...... 174 27.7.1 Multibyte Characters ...... 174 27.7.2 Wide Characters ...... 175 27.8 Using Wide Characters and wchar_t ...... 175 27.8.1 Multibyte to wchar_t Conversions ...... 175 27.9 Wide Character Functionality ...... 177 27.9.1 wint_t ...... 177 27.9.2 I/O Stream Orientation ...... 177 27.9.3 I/O Functions ...... 177 CONTENTS vi
27.9.4 Type Conversion Functions ...... 178 27.9.5 String and Memory Copying Functions ...... 178 27.9.6 String and Memory Comparing Functions ...... 178 27.9.7 String Searching Functions ...... 178 27.9.8 Length/Miscellaneous Functions ...... 179 27.9.9 Character Classification Functions ...... 179 27.10 Parse State, Restartable Functions ...... 179 27.11 Unicode Encodings and C ...... 181 27.11.1 UTF-8 ...... 181 27.11.2 UTF-16, UTF-32, char16_t, and char32_t ...... 182 27.11.3 Multibyte Conversions ...... 182 27.11.4 Third-Party Libraries ...... 183
28 Exiting a Program 184 28.1 Normal Exits ...... 184 28.1.1 Returning From main() ...... 184 28.1.2 exit() ...... 184 28.1.3 Setting Up Exit Handlers with atexit() ...... 185 28.2 Quicker Exits with quick_exit() ...... 185 28.3 Nuke it from Orbit: _Exit() ...... 186 28.4 Exiting Sometimes: assert() ...... 186 28.5 Abnormal Exit: abort() ...... 186
29 Signal Handling 188 29.1 What Are Signals? ...... 188 29.2 Handling Signals with signal() ...... 188 29.3 Writing Signal Handlers ...... 189 29.4 What Can We Actually Do? ...... 191 29.5 Friends Don’t Let Friends signal() ...... 192
30 Variable-Length Arrays (VLAs) 193 30.1 The Basics ...... 193 30.2 sizeof and VLAs ...... 194 30.3 Multidimensional VLAs ...... 195 30.4 Passing One-Dimensional VLAs to Functions ...... 195 30.5 Passing Multi-Dimensional VLAs to Functions ...... 196 30.5.1 Partial Multidimensional VLAs ...... 196 30.6 Compatibility with Regular Arrays ...... 197 30.7 typedef and VLAs ...... 197 30.8 Jumping Pitfalls ...... 198 30.9 General Issues ...... 198
31 goto 199 31.1 A Simple Example ...... 199 31.2 Labeled continue ...... 200 31.3 Bailing Out ...... 201 31.4 Labeled break ...... 201 31.5 Multi-level Cleanup ...... 202 31.6 Restarting Interrupted System Calls ...... 202 31.7 goto and Variable Scope ...... 203 31.8 goto and Variable-Length Arrays ...... 204
32 Types Part V: Compound Literals and Generic Selections 205 32.1 Compound Literals ...... 205 32.1.1 Passing Unnamed Objects to Functions ...... 205 32.1.2 Unnamed structs ...... 206 32.1.3 Pointers to Unnamed Objects ...... 207 32.1.4 Unnamed Objects and Scope ...... 207 32.1.5 Silly Unnamed Object Example ...... 208 CONTENTS vii
32.2 Generic Selections ...... 208
33 Arrays Part II 211 33.1 Type Qualifiers for Arrays in Parameter Lists ...... 211 33.2 static for Arrays in Parameter Lists ...... 211 33.3 Equivalent Initializers ...... 212
34 Long Jumps with setjmp, longjmp 215 34.1 Using setjmp and longjmp ...... 215 34.2 Pitfalls ...... 216 34.2.1 The Values of Local Variables ...... 216 34.2.2 How Much State is Saved? ...... 217 34.2.3 You Can’t Name Anything setjmp ...... 217 34.2.4 You Can’t setjmp() in a Larger Expression ...... 217 34.2.5 When Can’t You longjmp()? ...... 218 34.2.6 You Can’t Pass 0 to longjmp() ...... 218 34.2.7 longjmp() and Variable Length Arrays ...... 218
35 Incomplete Types 219 35.1 Use Case: Self-Referential Structures ...... 219 35.2 Incomplete Type Error Messages ...... 220 35.3 Other Incomplete Types ...... 220 35.4 Use Case: Arrays in Header Files ...... 220 35.5 Completing Incomplete Types ...... 221
36 Complex Numbers 223 36.1 Complex Types ...... 223 36.2 Assigning Complex Numbers ...... 224 36.3 Constructing, Deconstructing, and Printing ...... 224 36.4 Complex Arithmetic and Comparisons ...... 225 36.5 Complex Math ...... 226 36.5.1 Trigonometry Functions ...... 226 36.5.2 Exponential and Logarithmic Functions ...... 227 36.5.3 Power and Absolute Value Functions ...... 227 36.5.4 Manipulation Functions ...... 227
37 Fixed Width Integer Types 228 37.1 The Bit-Sized Types ...... 228 37.2 Maximum Integer Size Type ...... 229 37.3 Using Fixed Size Constants ...... 229 37.4 Limits of Fixed Size Integers ...... 229 37.5 Format Specifiers ...... 230
38 Date and Time Functionality 232 38.1 Quick Terminology and Information ...... 232 38.2 Date Types ...... 232 38.3 Initialization and Conversion Between Types ...... 233 38.3.1 Converting time_t to struct tm ...... 233 38.3.2 Converting struct tm to time_t ...... 234 38.4 Formatted Date Output ...... 234 38.5 More Resolution with timespec_get() ...... 236 38.6 Differences Between Times ...... 236
39 Multithreading 238 39.1 Background ...... 238 39.2 Things You Can Do ...... 238 39.3 Data Races and the Standard Library ...... 239 39.4 Creating and Waiting for Threads ...... 239 39.5 Detaching Threads ...... 243 CONTENTS viii
39.6 Thread Local Data ...... 244 39.6.1 _Thread_local Storage-Class ...... 245 39.6.2 Another Option: Thread-Specific Storage ...... 246 39.7 Mutexes ...... 247 39.7.1 Different Mutex Types ...... 250 39.8 Condition Variables ...... 251 39.8.1 Timed Condition Wait ...... 254 39.8.2 Broadcast: Wake Up All Waiting Threads ...... 255 39.9 Running a Function One Time ...... 255
40
41
42
42.6 fpclassify() ...... 314 42.7 isfinite(), isinf(), isnan(), isnormal() ...... 316 42.8 signbit() ...... 317 42.9 acos(), acosf(), acosl() ...... 317 42.10 asin(), asinf(), asinl() ...... 318 42.11 atan(), atanf(), atanl(), atan2(), atan2f(), atan2l() ...... 319 42.12 cos(), cosf(), cosl() ...... 320 42.13 sin(), sinf(), sinl() ...... 320 42.14 tan(), tanf(), tanl() ...... 321 42.15 acosh(), acoshf(), acoshl() ...... 322 42.16 asinh(), asinhf(), asinhl() ...... 322 42.17 atanh(), atanhf(), atanhl() ...... 323 42.18 cosh(), coshf(), coshl() ...... 324 42.19 sinh(), sinhf(), sinhl() ...... 324 42.20 tanh(), tanhf(), tanhl() ...... 325 42.21 exp(), expf(), expl() ...... 325 42.22 exp2(), exp2f(), exp2l() ...... 326 42.23 expm1(), expm1f(), expm1l() ...... 327 42.24 frexp(), frexpf(), frexpl() ...... 327 42.25 ilogb(), ilogbf(), ilogbl() ...... 328 42.26 ldexp(), ldexpf(), ldexpl() ...... 329 42.27 log(), logf(), logl() ...... 330 42.28 log10(), log10f(), log10l() ...... 331 42.29 logp1(), logp1f(), logp1l() ...... 331 42.30 log2(), log2f(), log2l() ...... 332 42.31 logb(), logbf(), logbl() ...... 333 42.32 modf(), modff(), modfl() ...... 334 42.33 scalbn(), scalbnf(), scalbnl() scalbln(), scalblnf(), scalblnl() . . . . . 335 42.34 cbrt(), cbrtf(), cbrtl() ...... 336 42.35 fabs(), fabsf(), fabsl() ...... 337 42.36 hypot(), hypotf(), hypotl() ...... 337 42.37 pow(), powf(), powl() ...... 338 42.38 sqrt() ...... 339 42.39 erf(), erff(), erfl() ...... 340 42.40 erfc(), erfcf(), erfcl() ...... 341 42.41 lgamma(), lgammaf(), lgammal() ...... 342 42.42 tgamma(), tgammaf(), tgammal() ...... 343 42.43 ceil(), ceilf(), ceill() ...... 343 42.44 floor(), floorf(), floorl() ...... 344 42.45 nearbyint(), nearbyintf(), nearbyintl() ...... 345 42.46 rint(), rintf(), rintl() ...... 346 42.47 lrint(), lrintf(), lrintl(), llrint(), llrintf(), llrintl() ...... 347 42.48 round(), roundf(), roundl() ...... 348 42.49 lround(), lroundf(), lroundl() llround(), llroundf(), llroundl() . . . . . 348 42.50 trunc(), truncf(), truncl() ...... 349 42.51 fmod(), fmodf(), fmodl() ...... 350 42.52 remainder(), remainderf(), remainderl() ...... 351 42.53 remquo(), remquof(), remquol() ...... 352 42.54 copysign(), copysignf(), copysignl() ...... 353 42.55 nan(), nanf(), nanl() ...... 354 42.56 nextafter(), nextafterf(), nextafterl() ...... 355 42.57 nexttoward(), nexttowardf(), nexttowardl() ...... 355 42.58 fdim(), fdimf(), fdiml() ...... 356 42.59 fmax(), fmax(), fmax(), fmin(), fmin(), fmin() ...... 357 42.60 fma(), fmaf(), fmal() ...... 357 42.61 isgreater(), isgreaterequal(), isless(), islessequal() ...... 358 42.62 islessgreater() ...... 359 42.63 isunordered() ...... 360 CONTENTS x
43
44
45
46
46.3 cnd_destroy() ...... 422 46.4 cnd_init() ...... 424 46.5 cnd_signal() ...... 425 46.6 cnd_timedwait() ...... 427 46.7 cnd_wait() ...... 429 46.8 mtx_destroy() ...... 430 46.9 mtx_init() ...... 431 46.10 mtx_lock() ...... 433 46.11 mtx_timedlock() ...... 434 46.12 mtx_trylock() ...... 436 46.13 mtx_unlock() ...... 438 46.14 thrd_create() ...... 439 46.15 thrd_current() ...... 441 46.16 thrd_detach() ...... 442 46.17 thrd_equal() ...... 443 46.18 thrd_exit() ...... 445 46.19 thrd_join() ...... 446 46.20 thrd_sleep() ...... 447 46.21 thrd_yield() ...... 448 46.22 tss_create() ...... 450 46.23 tss_delete() ...... 452 46.24 tss_get() ...... 454 46.25 tss_set() ...... 455
47
48
48.31 wctombr() ...... 493 48.32 mbsrtowcs() ...... 495 48.33 wcsrtombs() ...... 497
49
50
51
52
53
53.2 OS X issue ...... 551 53.3 mbrtoc16() mbrtoc32() ...... 552 53.4 c16rtomb() c32rtomb() ...... 554
54
55
56
57
58
59
60
61
62
62.1 ptrdiff_t ...... 601 62.2 size_t ...... 601 62.3 max_align_t ...... 602 62.4 wchar_t ...... 602 62.5 offsetof ...... 602
63
64
Foreword
No point in wasting words here, folks, let’s jump straight into the C code: E((ck?main((z?(stat(M,&t)?P+=a+'{'?0:3: execv(M,k),a=G,i=P,y=G&255, sprintf(Q,y/'@'-3?A(*L(V(%d+%d)+%d,0)
And they lived happily ever after. The End. What’s this? You say something’s still not clear about this whole C programming language thing? Well, to be quite honest, I’m not even sure what the above code does. It’s a snippet from one of the entries in the 2001 International Obfuscated C Code Contest1, a wonderful competition wherein the entrants attempt to write the most unreadable C code possible, with often surprising results. The bad news is that if you’re a beginner in this whole thing, all C code you see probably looks obfuscated! The good news is, it’s not going to be that way for long. What we’ll try to do over the course of this guide is lead you from complete and utter sheer lost confusion on to the sort of enlightened bliss that can only be obtained though pure C programming. Right on.
1.1 Audience
This guide assumes that you’ve already got some programming knowledge under your belt from another language, such as Python2, JavaScript3, Java4, Rust5, Go6, Swift7, etc. (Objective-C8 devs will have a particularly easy time of it!) We’re going to assume you know what variables are, what loops do, how functions work, and so on. If that’s not you for whatever reason the best I can hope to provide is some pastey entertainment for your reading pleasure. The only thing I can reasonably promise is that this guide won’t end on a cliffhanger… or will it? 1https://www.ioccc.org/ 2https://en.wikipedia.org/wiki/Python_(programming_language) 3https://en.wikipedia.org/wiki/JavaScript 4https://en.wikipedia.org/wiki/Java_(programming_language) 5https://en.wikipedia.org/wiki/Rust_(programming_language) 6https://en.wikipedia.org/wiki/Go_(programming_language) 7https://en.wikipedia.org/wiki/Swift_(programming_language) 8https://en.wikipedia.org/wiki/Objective-C
1 Chapter 1. Foreword 2
1.2 Platform and Compiler
I’ll try to stick to Plain Ol’-Fashioned ISO-standard C9. Well, for the most part. Here and there I might go crazy and start talking about POSIX10 or something, but we’ll see. Unix users (e.g. Linux, BSD, etc.) try running cc or gcc from the command line–you might already have a compiler installed. If you don’t, search your distribution for installing gcc or clang. Windows users should check out Visual Studio Community11. Or, if you’re looking for a more Unix-like experience (recommended!), install WSL12 and gcc. Mac users will want to install XCode13, and in particular the command line tools. There are a lot of compilers out there, and virtually all of them will work for this book. And a C++ compiler will compile a lot of (but not all!) C code. Best use a proper C compiler if you can.
1.3 Official Homepage
This official location of this document is https://beej.us/guide/bgc/14. Maybe this’ll change in the future, but it’s more likely that all the other guides are migrated off Chico State computers.
1.4 Email Policy
I’m generally available to help out with email questions so feel free to write in, but I can’t guarantee a response. I lead a pretty busy life and there are times when I just can’t answer a question you have. When that’s the case, I usually just delete the message. It’s nothing personal; I just won’t ever have the time to give the detailed answer you require. As a rule, the more complex the question, the less likely I am to respond. If you can narrow down your question before mailing it and be sure to include any pertinent information (like platform, compiler, error messages you’re getting, and anything else you think might help me troubleshoot), you’re much more likely to get a response. If you don’t get a response, hack on it some more, try to find the answer, and if it’s still elusive, then write me again with the information you’ve found and hopefully it will be enough for me to help out. Now that I’ve badgered you about how to write and not write me, I’d just like to let you know that I fully appreciate all the praise the guide has received over the years. It’s a real morale boost, and it gladdens me to hear that it is being used for good! :-) Thank you!
1.5 Mirroring
You are more than welcome to mirror this site, whether publicly or privately. If you publicly mirror the site and want me to link to it from the main page, drop me a line at [email protected].
1.6 Note for Translators
If you want to translate the guide into another language, write me at [email protected] and I’ll link to your translation from the main page. Feel free to add your name and contact info to the translation. Please note the license restrictions in the Copyright and Distribution section, below.
9https://en.wikipedia.org/wiki/ANSI_C 10https://en.wikipedia.org/wiki/POSIX 11https://visualstudio.microsoft.com/vs/community/ 12https://docs.microsoft.com/en-us/windows/wsl/install-win10 13https://developer.apple.com/xcode/ 14https://beej.us/guide/bgc/ Chapter 1. Foreword 3
1.7 Copyright and Distribution
Beej’s Guide to C is Copyright © 2021 Brian “Beej Jorgensen” Hall. With specific exceptions for source code and translations, below, this work is licensed under the Cre- ative Commons Attribution-Noncommercial-No Derivative Works 3.0 License. To view a copy of this license, visit https://creativecommons.org/licenses/by-nc-nd/3.0/ or send a letter to Creative Commons, 171 Second Street, Suite 300, San Francisco, California, 94105, USA. One specific exception to the “No Derivative Works” portion of the license is as follows: this guide may be freely translated into any language, provided the translation is accurate, and the guide is reprinted in its entirety. The same license restrictions apply to the translation as to the original guide. The translation may also include the name and contact information for the translator. The C source code presented in this document is hereby granted to the public domain, and is completely free of any license restriction. Educators are freely encouraged to recommend or supply copies of this guide to their students. Contact [email protected] for more information. Chapter 2
Hello, World!
2.1 What to Expect from C
“Where do these stairs go?” “They go up.” —Ray Stantz and Peter Venkman, Ghostbusters C is a low-level language. It didn’t used to be. Back in the day when people carved punch cards out of granite, C was an incredible way to be free of the drudgery of lower-level languages like assembly1. But now in these modern times, current-generation languages offer all kinds of features that didn’t exist in 1972 when C was invented. This means C is a pretty basic language with not a lot of features. It can do anything, but it can make you work for it. So why would we even use it today? • As a learning tool: not only is C a venerable piece of computing history, but it is connected to the bare metal2 in a way that present-day languages are not. When you learn C, you learn about how software interfaces with computer memory at a low level. There are no seatbelts. You’ll write software that crashes, I assure you. And that’s all part of the fun! • As a useful tool: C still is used for certain applications, such as building operating systems3 or in embedded systems4. (Though the Rust5 programming language is eyeing both these fields!) If you’re familiar with another language, a lot of things about C are easy. C inspired many other languages, and you’ll see bits of it in Go, Rust, Swift, Python, JavaScript, Java, and all kinds of other languages. Those parts will be familiar. The one thing about C that hangs people up is pointers. Virtually everything else is familiar, but pointers are the weird one. The concept behind pointers is likely one you already know, but C forces you to be explicit about it, using operators you’ve likely never seen before. It’s especially insidious because once you grok6 pointers, they’re suddenly easy. But up until that moment, they’re slippery eels. Everything else in C is just memorizing another way (or sometimes the same way!) of doing something you’ve done already. Pointers are the weird bit. And, arguably, even pointers are variations on a theme you’re probably familiar with.
1https://en.wikipedia.org/wiki/Assembly_language 2https://en.wikipedia.org/wiki/Bare_machine 3https://en.wikipedia.org/wiki/Operating_system 4https://en.wikipedia.org/wiki/Embedded_system 5https://en.wikipedia.org/wiki/Rust_(programming_language) 6https://en.wikipedia.org/wiki/Grok
4 Chapter 2. Hello, World! 5
So get ready for a rollicking adventure as close to the core of the computer as you can get without assembly, in the most influential computer language of all time7. Hang on!
2.2 Hello, World!
This is the canonical example of a C program. Everyone uses it. (Note that the numbers to the left are for reader reference only, and are not part of the source code.)
1 /*Helloworldprogram*/
2
3 #include
4
5 int main(void)
6 {
7 printf("Hello, World!\n"); //Actuallydotheworkhere
8 }
We’re going to don our long-sleeved heavy-duty rubber gloves, grab a scalpel, and rip into this thing to see what makes it tick. So, scrub up, because here we go. Cutting very gently… Let’s get the easy thing out of the way: anything between the digraphs /* and */ is a comment and will be completely ignored by the compiler. Same goes for anything on a line after a //. This allows you to leave messages to yourself and others, so that when you come back and read your code in the distant future, you’ll know what the heck it was you were trying to do. Believe me, you will forget; it happens. Now, what is this #include? GROSS! Well, it tells the C Preprocessor to pull the contents of another file and insert it into the code right there. Wait—what’s a C Preprocessor? Good question. There are two stages8 to compilation: the preprocessor and the compiler. Anything that starts with pound sign, or “octothorpe”, (#) is something the preprocessor operates on before the compiler even gets started. Common preprocessor directives, as they’re called, are #include and #define. More on that later. Before we go on, why would I even begin to bother pointing out that a pound sign is called an octothorpe? The answer is simple: I think the word octothorpe is so excellently funny, I have to gratuitously spread its name around whenever I get the opportunity. Octothorpe. Octothorpe, octothorpe, octothorpe. So anyway. After the C preprocessor has finished preprocessing everything, the results are ready for the compiler to take them and produce assembly code9, machine code10, or whatever it’s about to do. Machine code is the “language” the CPU understands, and it can understand it very rapidly. This is one of the reasons C programs tend to be quick. Don’t worry about the technical details of compilation for now; just know that your source runs through the preprocessor, then the output of that runs through the compiler, then that produces an executable for you to run. What about the rest of the line? What’s
7I know someone will fight me on that, but it’s gotta be at least in the top three, right? 8Well, technically there are more than two, but hey, let’s pretend there are two—ignorance is bliss, right? 9https://en.wikipedia.org/wiki/Assembly_language 10https://en.wikipedia.org/wiki/Machine_code 11Technically, it contains preprocessor directives and function prototypes (more on that later) for common input and output needs. Chapter 2. Hello, World! 6
Holy moly. That was all to cover the first line! But, let’s face it, it has been completely dissected. No mystery shall remain! So take a breather…look back over the sample code. Only a couple easy lines to go. Welcome back from your break! I know you didn’t really take a break; I was just humoring you. The next line is main(). This is the definition of the function main(); everything between the squirrelly braces ({ and }) is part of the function definition. (How do you call a different function, anyway? The answer lies in the printf() line, but we’ll get to that in a minute.) Now, the main function is a special one in many ways, but one way stands above the rest: it is the function that will be called automatically when your program starts executing. Nothing of yours gets called before main(). In the case of our example, this works fine since all we want to do is print a line and exit. Oh, that’s another thing: once the program executes past the end of main(), down there at the closing squirrelly brace, the program will exit, and you’ll be back at your command prompt. So now we know that that program has brought in a header file, stdio.h, and declared a main() function that will execute when the program is started. What are the goodies in main()? I am so happy you asked. Really! We only have the one goodie: a call to the function printf(). You can tell this is a function call and not a function definition in a number of ways, but one indicator is the lack of squirrelly braces after it. And you end the function call with a semicolon so the compiler knows it’s the end of the expression. You’ll be putting semicolons after almost everything, as you’ll see. You’re passing one argument to the function printf(): a string to be printed when you call it. Oh, yeah—we’re calling a function! We rock! Wait, wait—don’t get cocky. What’s that crazy \n at the end of the string? Well, most characters in the string will print out just like they are stored. But there are certain characters that you can’t print on screen well that are embedded as two-character backslash codes. One of the most popular is \n (read “backslash-N”) that corresponds to the newline character. This is the character that causes further printing to continue at the beginning of the next line instead of the current. It’s like hitting return at the end of the line. So copy that code into a file called hello.c and build it. On a Unix-like platform (e.g. Linux, BSD, Mac, or WSL), from the command line you’ll build with a command like so: gcc -o hello hello.c
(This means “compile hello.c, and output an executable called hello”.) After that’s done, you should have a file called hello that you can run with this command: ./hello
(The leading ./ tells the shell to “run from the current directory”.) And see what happens: Hello, World!
It’s done and tested! Ship it!
2.3 Compilation Details
Let’s talk a bit more about how to build C programs, and what happens behind the scenes there. Like other languages, C has source code. But, depending on what language you’re coming from, you might never have had to compile your source code into an executable. Compilation is the process of taking your C source code and turning it into a program that your operating system can execute. JavaScript and Python devs aren’t used to a separate compilation step at all–though behind the scenes it’s happening! Python compiles your source code into something called bytecode that the Python virtual Chapter 2. Hello, World! 7 machine can execute. Java devs are used to compilation, but that produces bytecode for the Java Virtual Machine. When compiling C, machine code is generated. This is the 1s and 0s that can be executed directly and speedily by the CPU. Languages that typically aren’t compiled are called interpreted languages. But as we men- tioned with Java and Python, they also have a compilation step. And there’s no rule saying that C can’t be interpreted. (There are C interpreters out there!) In short, it’s a bunch of gray areas. Compilation in general is just taking source code and turning it into another, more easily-executed form. The C compiler is the program that does the compilation. As we’ve already said, gcc is a compiler that’s installed on a lot of Unix-like operating systems12. And it’s commonly run from the command line in a terminal, but not always. You can run it from your IDE, as well. So how do we do command line builds?
2.4 Building with gcc
If you have a source file called hello.c in the current directory, you can build that into a program called hello with this command typed in a terminal: gcc -o hello hello.c
The -o means “output to this file”13. And there’s hello.c at the end, the name of the file we want to compile. If your source is broken up into multiple files, you can compile them all together (almost as if they were one file, but the rules are actually more complex than that) by putting all the .c files on the command line: gcc -o awesomegame ui.c characters.c npc.c items.c and they’ll all get built together into a big executable. That’s enough to get started—later we’ll talk details about multiple source files, object files, and all kinds of fun stuff.
2.5 Building with clang
On Macs, the compiler isn’t gcc—it’s clang. But a wrapper is also installed so you can run gcc and have it still work.
2.6 Building from IDEs
If you’re using an Integrated Development Environment (IDE), you probably don’t have to build from the command line. With Visual Studio, CTRL-F7 will build, and CTRL-F5 will run. With VS Code, things are more complex, but you can hit F5 to run via the debugger. (You’ll have to install the C/C++ Extension.) With XCode, you can build with COMMAND-B and run with COMMAND-R. To get the command line tools, Google for “XCode command line tools” and you’ll find instructions for installing them. For getting started, I encourage you to also try to build from the command line—it’s history!
12https://en.wikipedia.org/wiki/Unix 13If you don’t give it an output filename, it will export to a file called a.out by default—this filename has its roots deep in Unix history. Chapter 2. Hello, World! 8
2.7 C Versions
C has come a long way over the years, and it had many named version numbers to describe which dialect of the language you’re using. These generally refer to the year of the specification. The most famous are C89, C99, C11, and C2x. We’ll focus on the latter in this book. But here’s a more complete table:
Version Description K&R C 1978, the original. Named after Brian Kernighan and Dennis Ritchie. Ritchie designed and coded the language, and Kernighan co-authored the book on it. You rarely see original K&R code today. If you do, it’ll look odd, like Middle English looks odd to modern English readers. C89, ANSI C, C90 In 1989, the American National Standards Institute (ANSI) produced a C language specification that set the tone for C that persists to this day. A year later, the reins were handed to the International Organization for Standardization (ISO) that produced the identical C90. C95 A rarely-mentioned addition to C89 that included wide character support. C99 The first big overhaul with lots of language additions. The thing most people remember is the addition of //-style comments. This is the most popular version of C in use as of this writing. C11 This major version update includes Unicode support and multi-threading. Be advised that if you start using these language features, you might be sacrificing portability with places that are stuck in C99 land. But, honestly, 1999 is getting to be a while back now. C17, C18 Bugfix update to C11. C17 seems to be the official name, but the publication was delayed until 2018. As far as I can tell, these two are interchangeable, with C17 being preferred. C2x What’s coming next! Expected to eventually become C21.
You can force GCC to use one of these standards with the -std= command line argument. If you want it to be picky about the standard, add -pedantic. For example: gcc -std=c11 -pedantic foo.c
For this book, I compile programs for C2x with all warnings set: gcc -Wall -Wextra -std=c2x -pedantic foo.c Chapter 3
Variables and Statements
“It takes all kinds to make a world, does it not, Padre?” “So it does, my son, so it does.” —Pirate Captain Thomas Bartholomew Red to the Padre, Pirates There sure can be lotsa stuff in a C program. Yup. And for various reasons, it’ll be easier for all of us if we classify some of the types of things you can find in a program, so we can be clear what we’re talking about.
3.1 Variables
It’s said that “variables hold values”. But another way to think about it is that a variable is a human- readable name that refers to some data in memory. We’re going to take a second here and take a peek down the rabbit hole that is pointers. Don’t worry about it. You can think of memory as a big array of bytes1 Data is stored in this “array”2. If a number is larger than a single byte, it is stored in multiple bytes. Because memory is like an array, each byte of memory can be referred to by its index. This index into memory is also called an address, or a location, or a pointer. When you have a variable in C, the value of that variable is in memory somewhere, at some address. Of course. After all, where else would it be? But it’s a pain to refer to a value by its numeric address, so we make a name for it instead, and that’s what the variable is. The reason I’m bringing all this up is twofold: 1. It’s going to make it easier to understand pointer variables later—they’re variables that hold the address of other variables! 2. Also, it’s going to make it easier to understand pointers later. So a variable is a name for some data that’s stored in memory at some address.
3.1.1 Variable Names You can use any characters in the range 0-9, A-Z, a-z, and underscore for variable names, with the follow- ing rules: • You can’t start a variable with a digit 0-9. • You can’t start a variable name with two underscores. • You can’t start a variable name with an underscore followed by a capital A-Z.
1A “byte” is an 8-bit binary number. Think of it as an integer that can only hold the values from 0 to 255, inclusive. 2I’m seriously oversimplifying how modern memory works, here. But the mental model works, so please forgive me.
9 Chapter 3. Variables and Statements 10
For Unicode, just try it. There are some rules in the spec in §D.2 that talk about which Unicode codepoint ranges are allowed in which parts of identifiers, but that’s too much to write about here and is probably something you’ll never have to think about anyway.
3.1.2 Variable Types Depending on which languages you already have in your toolkit, you might or might not be familiar with the idea of types. But C’s kinda picky about them, so we should do a refresher. Some example types, some of the most basic:
Type Example C Type Integer 3490 int Floating point 3.14159 float Character (single) 'c' char String "Hello, world!" char *3
C makes an effort to convert automatically between most numeric types when you ask it to. But other than that, all conversions are manual, notably between string and numeric. Almost all of the types in C are variants on these types. Before you can use a variable, you have to declare that variable and tell C what type the variable holds. Once declared, the type of variable cannot be changed later at runtime. What you set it to is what it is until it falls out of scope and is reabsorbed into the universe. Let’s take our previous “Hello, world” code and add a couple variables to it:
1 #include
2
3 int main(void)
4 {
5 int i; // Holds signed integers, e.g. -3, -2, 0, 1, 10
6 float f; // Holds signed floating point numbers, e.g. -3.1416
7
8 printf("Hello, World!\n"); // Ah, blessed familiarity
9 }
There! We’ve declared a couple of variables. We haven’t used them yet, and they’re both uninitialized. One holds an integer number, and the other holds a floating point number (a real number, basically, if you have a math background). Uninitialized variables have indeterminate value4. They have to be initialized or else you must assume they contain some nonsense number. This is one of the places C can “get you”. Much of the time, in my experience, the indeter- minate value is zero… but it can vary from run to run! Never assume the value will be zero, even if you see it is. Always explicitly initialize variables to some value before you use them5. What’s this? You want to store some numbers in those variables? Insanity! Let’s go ahead and do that:
1 int main(void)
2 {
3 int i;
4
3Read this as “pointer to a char” or “char pointer”. “Char” for character. Though I can’t find a study, it seems anecdotally most people pronounce this as “char”, a minority say “car”, and a handful say “care”. We’ll talk more about pointers later. 4Colloquially, we say they have “random” values, but they aren’t truly—or even pseudo-truly—random numbers. 5This isn’t strictly 100% true. When we get to learning about static storage duration, you’ll find the some variables are initialized to zero automatically. But the safe thing to do is always initialize them. Chapter 3. Variables and Statements 11
5 i = 2; // Assign the value 2 into the variable i
6
7 printf("Hello, World!\n");
8 }
Killer. We’ve stored a value. Let’s print it. We’re going to do that by passing two amazing parameters to the printf() function. The first argument is a string that describes what to print and how to print it (called the format string), and the second is the value to print, namely whatever is in the variable i. printf() hunts through the format string for a variety of special sequences which start with a percent sign (%) that tell it what to print. For example, if it finds a %d, it looks to the next parameter that was passed, and prints it out as an integer. If it finds a %f, it prints the value out as a float. If it finds a %s, it prints a string. As such, we can print out the value of various types like so:
1 int main(void)
2 {
3 int i = 2;
4 float f = 3.14;
5 char *s = "Hello, world!"; // char * ("char pointer") is the string type
6
7 printf("%s i = %d and f = %f!\n", s, i, f);
8 }
And the output will be: Hello, world! i = 2 and f = 3.14!
In this way, printf() might be similar to various types of format strings or parameterized strings in other languages you’re familiar with.
3.1.3 Boolean Types C has Boolean types, true or false? 1! Historically, C didn’t have a Boolean type, and some might argue it still doesn’t. In C, 0 means “false”, and non-zero means “true”. So 1 is true. And -37 is true. And 0 is false. You can just declare Boolean types as ints: int x = 1;
if (x) { printf("x is true!\n"); }
If you #include
1 #include
2 #include
3
4 int main(void) {
5 bool x = true;
6
7 if (x) {
8 printf("x is true!\n"); Chapter 3. Variables and Statements 12
9 }
10 }
But these are identical to using integer values for true and false. They’re just a facade to make things look nice.
3.2 Operators and Expressions
C operators should be familiar to you from other languages. Let’s blast through some of them here. (There are a bunch more details than this, but we’re going to do enough in this section to get started.)
3.2.1 Arithmetic Hopefully these are familiar: i = i + 3; // Addition (+) and assignment (=) operators, add 3 to i i = i - 8; // Subtraction, subtract 8 from i i = i * 9; // Multiplication i = i / 2; // Division i = i % 5; // Modulo (division remainder)
There are shorthand variants for all of the above. Each of those lines could more tersely be written as: i += 3; // Same as "i = i + 3", add 3 to i i -= 8; // Same as "i = i - 8" i *= 9; // Same as "i = i * 9" i /= 2; // Same as "i = i / 2" i %= 5; // Same as "i = i % 5"
There is no exponentiation. You’ll have to use one of the pow() function variants from math.h. Let’s get into some of the weirder stuff you might not have in your other languages!
3.2.2 Ternary Operator C also includes the ternary operator. This is an expression whose value depends on the result of a condi- tional embedded in it. // If x > 10, add 17 to y. Otherwise add 37 to y.
y += x > 10? 17: 37;
What a mess! You’ll get used to it the more you read it. To help out a bit, I’ll rewrite the above expression using if statements: // This expression:
y += x > 10? 17: 37;
// is equivalent to this non-expression:
if (x > 10) y += 17; else y += 37;
Compare those two until you see each of the components of the ternary operator. Or, another example that prints if a number stored in x is odd or even: printf("The number %d is %s.\n", x, x % 2 == 0? "even": "odd") Chapter 3. Variables and Statements 13
The %s format specifier in printf() means print a string. If the expression x % 2 evaluates to 0, the value of the entire ternary expression evaluates to the string "even". Otherwise it evaluates to the string "odd". Pretty cool! It’s important to note that the ternary operator isn’t flow control like the if statement is. It’s just an expression that evaluates to a value.
3.2.3 Pre-and-Post Increment-and-Decrement Now, let’s mess with another thing that you might not have seen. These are the legendary post-increment and post-decrement operators: i++; // Add one to i (post-increment) i--; // Subtract one from i (post-decrement)
Very commonly, these are just used as shorter versions of: i += 1; // Add one to i i -= 1; // Subtract one from i but they’re more subtly different than that, the clever scoundrels. Let’s take a look at this variant, pre-increment and pre-decrement: ++i; // Add one to i (pre-increment) --i; // Subtract one from i (pre-decrement)
With pre-increment and pre-decrement, the value of the variable is incremented or decremented before the expression is evaluated. Then the expression is evaluated with the new value. With post-increment and post-decrement, the value of the expression is first computed with the value as-is, and then the value is incremented or decremented after the value of the expression has been determined. You can actually embed them in expressions, like this: i = 10; j = 5 + i++; // Compute 5 + i, _then_ increment i printf("%d, %d\n", i, j); // Prints 11, 15
Let’s compare this to the pre-increment operator: i = 10; j = 5 + ++i; // Increment i, _then_ compute 5 + i printf("%d, %d\n", i, j); // Prints 11, 16
This technique is used frequently with array and pointer access and manipulation. It gives you a way to use the value in a variable, and also increment or decrement that value before or after it is used. But by far the most common place you’ll see this is in a for loop: for (i = 0; i < 10; i++) printf("i is %d\n", i);
But more on that later.
3.2.4 The Comma Operator This is an uncommonly-used way to separated expressions that will run left to right: x = 10, y = 20; // First assign 10 to x, then 20 to y
Seems a bit silly, since you could just replace the comma with a semicolon, right? Chapter 3. Variables and Statements 14 x = 10; y = 20; // First assign 10 to x, then 20 to y
But that’s a little different. The latter is two separate expressions, while the former is a single expression! With the comma operator, the value of the comma expression is the value of the rightmost expression: x = 1, 2, 3; printf("x is %d\n", x); // Prints 3, because 3 is rightmost in the comma list
But even that’s pretty contrived. One common place the comma operator is used is in for loops to do multiple things in each section of the statement: for (i = 0, j = 10; i < 100; i++, j++) printf("%d, %d\n", i, j);
We’ll revisit that later.
3.2.5 Conditional Operators For Boolean values, we have a raft of standard operators: a == b; // True if a is equivalent to b a != b; // True if a is not equivalent to b a < b; // True if a is less than b a > b; // True if a is greater than b a <= b; // True if a is less than or equal to b a >= b; // True if a is greater than or equal to b
Don’t mix up assignment = with comparison ==! Use two equals to compare, one to assign. We can use the comparison expressions with if statements: if (a <= 10) printf("Success!\n");
3.2.6 Boolean Operators We can chain together or alter conditional expressions with Boolean operators for and, or, and not.
Operator Boolean meaning && and || or ! not
An example of Boolean “and”: // Do something if x less than 10 and y greater than 20: if (x < 10 && y > 20) printf("Doing something!\n");
An example of Boolean “not”: if (!(x < 12)) printf("x is not less than 12\n");
! has higher precedence than the other Boolean operators, so we have to use parentheses in that case. Of course, that’s just the same as: if (x >= 12) printf("x is not less than 12\n"); Chapter 3. Variables and Statements 15 but I needed the example!
3.2.7 The sizeof Operator This operator tells you the size (in bytes) that a particular variable or data type uses in memory. More particularly, it tells you the size (in bytes) that the type of a particular expression (which might be just a single variable) uses in memory. This can be different on different systems, except for char and its variants (which are always 1 byte). And this might not seem very useful now, but we’ll be making references to it here and there, so it’s worth covering. Since this computes the number of bytes needed to store a type, you might think it would return an int. Or… since the size can’t be negative, maybe an unsigned? But it turns out C has a special type to represent the return value from sizeof. It’s size_t, pronounced “size tee”6. All we know is that it’s an unsigned integer type that can hold the size in bytes of anything you can give to sizeof. size_t shows up a lot of different places where counts of things are passed or returned. Think of it as a value that represents a count. You can take the sizeof a variable or expression: int a = 999;
// %zu is the format specifier for type size_t printf("%zu\n", sizeof a); // Prints 4 on my system printf("%zu\n", sizeof(2 + 7)); // Prints 4 on my system printf("%zu\n", sizeof 3.14); // Prints 8 on my system
// If you need to print out negative size_t values, use %zd
Remember: it’s the size in bytes of the type of the expression, not the size of the expression itself. That’s why the size of 2+7 is the same as the size of a—they’re both type int. We’ll revisit this number 4 in the very next block of code… …Where we’ll see you can take the sizeof a type (note the parentheses are required around a type name, unlike an expression): printf("%zu\n", sizeof(int)); // Prints 4 on my system printf("%zu\n", sizeof(char)); // Prints 1 on all systems
It’s important to note that sizeof is a compile-time operation7. The result of the expression is determined entirely at compile-time, not at runtime. We’ll make use of this later on.
3.3 Flow Control
Booleans are all good, but of course we’re nowhere if we can’t control program flow. Let’s take a look at a number of constructs: if, for, while, and do-while. First, a general forward-looking note about statements and blocks of statements brought to you by your local friendly C developer: After something like an if or while statement, you can either put a single statement to be executed, or a block of statements to all be executed in sequence. Let’s start with a single statement:
6The _t is short for type. 7Except for with variable length arrays—but that’s a story for another time. Chapter 3. Variables and Statements 16 if (x == 10) printf("x is 10\n");
This is also sometimes written on a separate line. (Whitespace is largely irrelevant in C—it’s not like Python.) if (x == 10) printf("x is 10\n");
But what if you want multiple things to happen due to the conditional? You can use squirrelly braces to mark a block or compound statement. if (x == 10) { printf("x is 10\n"); printf("And also this happens when x is 10\n"); }
It’s a really common style to always use squirrelly braces even if they aren’t necessary: if (x == 10) { printf("x is 10\n"); }
Some devs feel the code is easier to read and avoids errors like this where things visually look like they’re in the if block, but actually they aren’t. // BAD ERROR EXAMPLE if (x == 10) printf("This happens if x is 10\n"); printf("This happens ALWAYS\n"); // Surprise!! Unconditional! while and for and the other looping constructs work the same way as the examples above. If you want to do multiple things in a loop or after an if, wrap them up in squirrelly braces. In other words, the if is going to run the one thing after the if. And that one thing can be a single statement or a block of statements.
3.3.1 The if-else statement
We’ve already been using if for multiple examples, since it’s likely you’ve seen it in a language before, but here’s another: int i = 10; if (i > 10) { printf("Yes, i is greater than 10.\n"); printf("And this will also print if i is greater than 10.\n"); } if (i <= 10) printf("i is less than or equal to 10.\n");
In the example code, the message will print if i is greater than 10, otherwise execution continues to the next line. Notice the squirrley braces after the if statement; if the condition is true, either the first statement or expression right after the if will be executed, or else the collection of code in the squirlley braces after the if will be executed. This sort of code block behavior is common to all statements. Of course, because C is fun this way, you can also do something if the condition is false with an else clause on your if: int i = 99; if (i == 10) printf("i is 10!\n"); Chapter 3. Variables and Statements 17 else { printf("i is decidedly not 10.\n"); printf("Which irritates me a little, frankly.\n"); }
And you can even cascade these to test a variety of conditions, like this: int i = 99; if (i == 10) printf("i is 10!\n"); else if (i == 20) printf("i is 20!\n"); else if (i == 99) { printf("i is 99! My favorite\n"); printf("I can't tell you how happy I am.\n"); printf("Really.\n"); } else printf("i is some crazy number I've never heard of.\n");
Though if you’re going that route, be sure to check out the switch statement for a potentially better solution. The catch is switch only works with equality comparisons with constant numbers. The above if-else cascade could check inequality, ranges, variables, or anything else you can craft in a conditional expression.
3.3.2 The while statement while is your average run-of-the-mill looping construct. Do a thing while a condition expression is true. Let’s do one! // Print the following output: // // i is now 0! // i is now 1! // [ more of the same between 2 and 7 ] // i is now 8! // i is now 9! i = 0; while (i < 10) { printf("i is now %d!\n", i); i++; } printf("All done!\n");
That gets you a basic loop. C also has a for loop which would have been cleaner for that example. A not-uncommon use of while is for infinite loops where you repeat while true: while (1) { printf("1 is always true, so this repeats forever.\n"); } Chapter 3. Variables and Statements 18
3.3.3 The do-while statement
So now that we’ve gotten the while statement under control, let’s take a look at its closely related cousin, do-while. They are basically the same, except if the loop condition is false on the first pass, do-while will execute once, but while won’t execute at all. In other words, the test to see whether or not to execute the block happens at the end of the block with do-while. It happens at the beginning of the block with while. Let’s see by example: // Using a while statement:
i = 10;
// this is not executed because i is not less than 10: while(i < 10) { printf("while: i is %d\n", i); i++; }
// Using a do-while statement:
i = 10;
// this is executed once, because the loop condition is not checked until // after the body of the loop runs:
do { printf("do-while: i is %d\n", i); i++; } while (i < 10);
printf("All done!\n");
Notice that in both cases, the loop condition is false right away. So in the while, the loop fails, and the following block of code is never executed. With the do-while, however, the condition is checked after the block of code executes, so it always executes at least once. In this case, it prints the message, increments i, then fails the condition, and continues to the “All done!” output. The moral of the story is this: if you want the loop to execute at least once, no matter what the loop condition, use do-while. All these examples might have been better done with a for loop. Let’s do something less deterministic— repeat until a certain random number comes up!
1 #include
2 #include
3
4 int main(void)
5 {
6 int r;
7
8 do {
9 r = rand() % 100; // Get a random number between 0 and 99
10 printf("%d\n", r);
11 } while (r != 37); // Repeat until 37 comes up
12 }
Side note: did you run that more than once? If you did, did you notice the same sequence of numbers came up again. And again. And again? This is because rand() is a pseudorandom number generator that must be seeded with a different number in order to generate a different sequence. Look up the srand() Chapter 3. Variables and Statements 19 function for more details.
3.3.4 The for statement
Welcome to one of the most popular loops in the world! The for loop! This is a great loop if you know the number of times you want to loop in advance. You could do the same thing using just a while loop, but the for loop can help keep the code cleaner. Here are two pieces of equivalent code—note how the for loop is just a more compact representation: // Print numbers between 0 and 9, inclusive...
// Using a while statement: i = 0; while (i < 10) { printf("i is %d\n", i); i++; }
// Do the exact same thing with a for-loop: for (i = 0; i < 10; i++) { printf("i is %d\n", i); }
That’s right, folks—they do exactly the same thing. But you can see how the for statement is a little more compact and easy on the eyes. (JavaScript users will fully appreciate its C origins at this point.) It’s split into three parts, separated by semicolons. The first is the initialization, the second is the loop condition, and the third is what should happen at the end of the block if the loop condition is true. All three of these parts are optional. for (initialize things; loop if this is true; do this after each loop)
Note that the loop will not execute even a single time if the loop condition starts off false. for-loop fun fact! You can use the comma operator to do multiple things in each clause of the for loop! for (i = 0, j = 999; i < 10; i++, j--) { printf("%d, %d\n", i, j); }
An empty for will run forever: for(;;) { // "forever" printf("I will print this again and again and again\n" ); printf("for all eternity until the heat-death of the universe.\n");
printf("Or until you hit CTRL-C.\n"); }
3.3.5 The switch Statement
Depending on what languages you’re coming from, you might or might not be familiar with switch, or C’s version might even be more restrictive than you’re used to. This is a statement that allows you to take a variety of actions depending on the value of an integer expression. Basically, it evaluates an expression to an integer value, jumps to the case that corresponds to that value. Execution resumes from that point. If a break statement is encountered, then execution jumps out of the Chapter 3. Variables and Statements 20
switch. Let’s do an example where the user enters a number of goats and we print out a gut-feel of how many goats that is.
1 #include
2
3 int main(void)
4 {
5 int goat_count;
6
7 printf("Enter a goat count: ");
8 scanf("%d", &goat_count); // Read an integer from the keyboard
9
10 switch (goat_count) {
11 case 0:
12 printf("You have no goats.\n");
13 break;
14
15 case 1:
16 printf("You have a singular goat.\n");
17 break;
18
19 case 2:
20 printf("You have a brace of goats.\n");
21 break;
22
23 default:
24 printf("You have a bona fide plethora of goats!\n");
25 break;
26 }
27 }
In that example, if the user enters, say, 2, the switch will jump to the case 2 and execute from there. When (if) it hits a break, it jumps out of the switch. Also, you might see that default label there at the bottom. This is what happens when no cases match. Every case, including default, is optional. And they can occur in any order, but it’s really typical for default, if any, to be listed last. So the whole thing acts like an if-else cascade: if (goat_count == 0) printf("You have no goats.\n"); else if (goat_count == 1) printf("You have a singular goat.\n"); else if (goat_count == 2) printf("You have a brace of goats.\n"); else printf("You have a bona fide plethora of goats!\n");
With some key differences: • switch is often faster to jump to the correct code (though the spec makes no such guarantee). • if-else can do things like relational conditionals like < and >= and floating point and other types, while switch cannot. There’s one more neat thing about switch that you sometimes see that is quite interesting: fall through. Remember how break causes us to jump out of the switch? Well, what happens if we don’t break? Chapter 3. Variables and Statements 21
Turns out we just keep on going into the next case! Demo! switch (x) { case 1: printf("1\n"); // Fall through! case 2: printf("2\n"); break; case 3: printf("3\n"); break; }
If x == 1, this switch will first hit case 1, it’ll print the 1, but then it just continues on to the next line of code… which prints 2! And then, at last, we hit a break so we jump out of the switch. if x == 2, then we just hit the case 2, print 2, and break as normal. Not having a break is called fall through. ProTip: ALWAYS put a comment in the code where you intend to fall through, like I did above. It will save other programmers from wondering if you meant to do that. In fact, this is one of the common places to introduce bugs in C programs: forgetting to put a break in your case. You gotta do it if you don’t want to just roll into the next case8. Earlier I said that switch works with integer types—keep it that way. Don’t use floating point or string types in there. One loophole-ish thing here is that you can use character types because those are secretly integers themselves. So this is perfectly acceptable: char c = 'b'; switch (c) { case 'a': printf("It's 'a'!\n"); break;
case 'b': printf("It's 'b'!\n"); break;
case 'c': printf("It's 'c'!\n"); break; }
Finally, you can use enums in switch since they are also integer types. But more on that in the enum chapter.
8This was considered such a hazard that the designers of the Go Programming Language made break the default; you have to explicitly use Go’s fallthrough statement if you want to fall into the next case. Chapter 4
Functions
“Sir, not in an environment such as this. That’s why I’ve also been programmed for over thirty secondary functions that—” —C3PO, before being rudely interrupted, reporting a now-unimpressive number of additional functions, Star Wars script Very much like other languages you’re used to, C has the concept of functions. Functions can accept a variety of arguments and return a value. One important thing, though: the argu- ments and return value types are predeclared—because that’s how C likes it! Let’s take a look at a function. This is a function that takes an int as an argument, and returns an int.
1 int plus_one(int n) // The "definition"
2 {
3 return n + 1;
4 }
5
The int before the plus_one indicates the return type. The int n indicates that this function takes one int argument, stored in parameter n. A parameter is a special type of local variable into which the arguments are copied. I’m going to drive home the point that the arguments are copied into the parameters, here. Lots of things in C are easier to understand if you know that the parameter is a copy of the argument, not the argument itself. More on that in a minute. Continuing the program down into main(), we can see the call to the function, where we assign the return value into local variable j:
6 int main(void)
7 {
8 int i = 10, j;
9
10 j = plus_one(i); // The "call"
11
12 printf("i + 1 is %d\n", j);
13 }
Before I forget, notice that I defined the function before I used it. If I hadn’t done that, the compiler wouldn’t know about it yet when it compiles main() and it would have given an unknown function call error. There is a more proper way to do the above code with function prototypes, but we’ll talk about that later. Also notice that main() is a function! It returns an int.
22 Chapter 4. Functions 23
But what’s this void thing? This is a keyword that’s used to indicate that the function accepts no argu- ments. You can also return void to indicate that you don’t return a value:
1 // This function takes no arguments and returns no value:
2
3 void hello(void)
4 {
5 printf("Hello, world!\n");
6 }
7
8 int main(void)
9 {
10 hello(); // Prints "Hello, world!"
11 }
4.1 Passing by Value
I’d mentioned earlier that when you pass an argument to a function, a copy of that argument gets made and stored in the corresponding parameter. If the argument is a variable, a copy of the value of that variable gets made and stored in the parameter. More generally, the entire argument expression is evaluated and its value determined. That value is copied to the parameter. In any case, the value in the parameter is its own thing. It is independent of whatever values or variables you used as arguments when you made the function call. So let’s look at an example here. Study it and see if you can determine the output before running it:
1 void increment(int a)
2 {
3 a++;
4 }
5
6 int main(void)
7 {
8 int i = 10;
9
10 increment(i);
11
12 printf("i == %d\n", i); // What does this print?
13 }
At first glance, it looks like i is 10, and we pass it to the function increment(). There the value gets incremented, so when we print it, it must be 11, right? “Get used to disappointment.” Dread Pirate Roberts, The Princess Bride But it’s not 11—it prints 10! How? It’s all about the fact that the expressions you pass to functions get copied onto their corresponding pa- rameters. The parameter is a copy, not the original. So i is 10 out in main(). And we pass it to increment(). The corresponding parameter is called a in that function. And the copy happens, as if by assignment. Loosely, a = i. So at that point, a is 10. And out in main(), i is also 10. Then we increment a to 11. But we’re not touching i at all! It remains 10. Chapter 4. Functions 24
Finally, the function is complete. All its local variables are discarded (bye, a!) and we return to main(), where i is still 10. And we print it, getting 10, and we’re done. This is why in the previous example with the plus_one() function, we returned the locally modified value so that we could see it again in main(). Seems a little bit restrictive, huh? Like you can only get one piece of data back from a function, is what you’re thinking. There is, however, another way to get data back; C folks call it passing by reference and that’s a story we’ll tell another time. But no fancy-schmancy name will distract you from the fact that EVERYTHING you pass to a function WITHOUT EXCEPTION is copied into its corresponding parameter, and the function operates on that local copy, NO MATTER WHAT. Remember that, even when we’re talking about this so-called passing by reference.
4.2 Function Prototypes
So if you recall back in the ice age a few sections ago, I mentioned that you had to define the function before you used it, otherwise the compiler wouldn’t know about it ahead of time, and would bomb out with an error. This isn’t quite strictly true. You can notify the compiler in advance that you’ll be using a function of a certain type that has a certain parameter list. That way the function can be defined anywhere (even in a different file), as long as the function prototype has been declared before you call that function. Fortunately, the function prototype is really quite easy. It’s merely a copy of the first line of the function definition with a semicolon tacked on the end for good measure. For example, this code calls a function that is defined later, because a prototype has been declared first:
1 int foo(void); // This is the prototype!
2
3 int main(void)
4 {
5 int i;
6
7 // We can call foo() here before it's definition because the
8 // prototype has already been declared, above!
9
10 i = foo();
11 }
12
13 int foo(void) // This is the definition, just like the prototype!
14 {
15 return 3490;
16 }
If you don’t declare your function before you use it (either with a prototype or its definition), you’re performing something called an implicit declaration. This was allowed in the first C standard (C89), and that standard has rules about it, but is no longer allowed today. And there is no legitimate reason to rely on it in new code. You might notice something about the sample code we’ve been using… That is, we’ve been using the good old printf() function without defining it or declaring a prototype! How do we get away with this lawlessness? We don’t, actually. There is a prototype; it’s in that header file stdio.h that we included with #include, remember? So we’re still legit, officer! Chapter 4. Functions 25
4.3 Empty Parameter Lists
You might see these from time to time in older code, but you shouldn’t ever code one up in new code. Always use void to indicate that a function takes no parameters. There’s never1 a reason to do this in modern code. If you’re good just remembering to put void in for empty parameter lists in functions and prototypes, you can skip the rest of this section. There are two contexts for this: • Omitting all parameters where the function is defined • Omitting all parameters in a prototype Let’s look at a potential function definition first: void foo() // Should really have a `void` in there { printf("Hello, world!\n"); }
While the spec spells out that the behavior in this instance is as-if you’d indicated void (C11 §6.7.6.3¶14), the void type is there for a reason. Use it. But in the case of a function prototype, there is a significant difference between using void and not: void foo(); void foo(void); // Not the same!
Leaving void out of the prototype indicates to the compiler that there is no additional information about the parameters to the function. It effectively turns off all that type checking. With a prototype definitely use void when you have an empty parameter list.
1Never say “never”. Chapter 5
Pointers—Cower In Fear!
“How do you get to Carnegie Hall?” “Practice!” —20th-century joke of unknown origin Pointers are one of the most feared things in the C language. In fact, they are the one thing that makes this language challenging at all. But why? Because they, quite honestly, can cause electric shocks to come up through the keyboard and physically weld your arms permanently in place, cursing you to a life at the keyboard in this language from the 70s! Really? Well, not really. I’m just trying to set you up for success. Depending on what language you came from, you might already understand the concept of references, where a variable refers to an object of some type. This is very much the same, except we have to be more explicit with C about when we’re talking about the reference or the thing it refers to.
5.1 Memory and Variables
Computer memory holds data of all kinds, right? It’ll hold floats, ints, or whatever you have. To make memory easy to cope with, each byte of memory is identified by an integer. These integers increase sequentially as you move up through memory1. You can think of it as a bunch of numbered boxes, where each box holds a byte2 of data. Or like a big array where each element holds a byte, if you come from a language with arrays. The number that represents each box is called its address. Now, not all data types use just a byte. For instance, an int is often four bytes, as is a float, but it really depends on the system. You can use the sizeof operator to determine how many bytes of memory a certain type uses. // %zu is the format specifier for type size_t printf("an int uses %zu bytes of memory\n", sizeof(int));
// That prints "4" for me, but can vary by system.
Memory Fun Facts: When you have a data type that uses more than a byte of memory, the bytes that make up the data are always adjacent to one another in memory. Sometimes they’re in order, and sometimes they’re not3, but that’s platform-dependent, and often taken care of for you without you needing to worry about pesky byte orderings.
1Typically. I’m sure there are exceptions out there in the dark corridors of computing history. 2A byte is a number made up of no more than 8 binary digits, or bits for short. This means in decimal digits just like grandma used to use, it can hold an unsigned number between 0 and 255, inclusive. 3The order that bytes come in is referred to as the endianness of the number. Common ones are big endian and little endian. This usually isn’t something you need to worry about.
26 Chapter 5. Pointers—Cower In Fear! 27
So anyway, if we can get on with it and get a drum roll and some foreboding music playing for the definition of a pointer, a pointer is a variable that holds an address. Imagine the classical score from 2001: A Space Odyssey at this point. Ba bum ba bum ba bum BAAAAH! Ok, so maybe a bit overwrought here, yes? There’s not a lot of mystery about pointers. They are the address of data. Just like an int variable can hold the value 12, a pointer variable can hold the address of data. This means that all these things mean the same thing, i.e. a number that represents a point in memory: • Index into memory (if you’re thinking of memory like a big array) • Address • Location I’m going to use these interchangeably. And yes, I just threw location in there because you can never have enough words that mean the same thing. And a pointer variable holds that address number. Just like a float variable might hold 3.14159. Imagine you have a bunch of Post-it® notes all numbered in sequence with their address. (The first one is at index numbered 0, the next at index 1, and so on.) In addition to the number representing their positions, you can also write another number of your choice on each. It could be the number of dogs you have. Or the number of moons around Mars… …Or, it could be the index of another Post-it note! If you have written the number of dogs you have, that’s just a regular variable. But if you wrote the index of another Post-it in there, that’s a pointer. It points to the other note! Another analogy might be with house addresses. You can have a house with certain qualities, yard, metal roof, solar, etc. Or you could have the address of that house. The address isn’t the same as the house itself. One’s a full-blown house, and the other is just a few lines of text. But the address of the house is a pointer to that house. It’s not the house itself, but it tells you where to find it. And we can do the same thing in the computer with data. You can have a data variable that’s holding some value. And that value is in memory at some address. And you could have a different pointer variable hold the address of that data variable. It’s not the data variable itself, but, like with a house address, it tells us where to find it. When we have that, we say we have a “pointer to” that data. And we can follow the pointer to access the data itself. (Though it doesn’t seem particularly useful yet, this all becomes indispensable when used with function calls. Bear with me until we get there.) So if we have an int, say, and we want a pointer to it, what we want is some way to get the address of that int, right? After all, the pointer just holds the address of the data. What operator do you suppose we’d use to find the address of the int? Well, by a shocking surprise that must come as something of a shock to you, gentle reader, we use the address-of operator (which happens to be an ampersand: “&”) to find the address of the data. Amper- sand. So for a quick example, we’ll introduce a new format specifier for printf() so you can print a pointer. You know already how %d prints a decimal integer, yes? Well, %p prints a pointer. Now, this pointer is going to look like a garbage number (and it might be printed in hexadecimal4 instead of decimal), but it is merely the index into memory the data is stored in. (Or the index into memory that the first byte of data is stored in, if the data is multi-byte.) In virtually all circumstances, including this one, the actual value of the number printed is unimportant to you, and I show it here only for demonstration of the address-of operator.
1 #include
2
4That is, base 16 with digits 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, and F. Chapter 5. Pointers—Cower In Fear! 28
3 int main(void)
4 {
5 int i = 10;
6
7 printf("The value of i is %d, and its address is %p\n", i, &i);
8 }
On my computer, this prints: The value of i is 10, and its address is 0x7ffda2546fc4
If you’re curious, that hexadecimal number is 140,727,326,896,068 in decimal (base 10 just like Grandma used to use). That’s the index into memory where the variable i’s data is stored. It’s the address of i. It’s the location of i. It’s a pointer to i. It’s a pointer because it lets you know where i is in memory. Like a home address written on a scrap of paper tells you where you can find a particular house, this number indicates to us where in memory we can find the value of i. It points to i. Again, we don’t really care what the address’s exact number is, generally. We just care that it’s a pointer to i.
5.2 Pointer Types
So… this is all well and good. You can now successfully take the address of a variable and print it on the screen. There’s a little something for the ol’ resume, right? Here’s where you grab me by the scruff of the neck and ask politely what the frick pointers are good for. Excellent question, and we’ll get to that right after these messages from our sponsor.
ACME ROBOTIC HOUSING UNIT CLEANING SERVICES. YOUR HOMESTEAD WILL BE DRAMATICALLY IMPROVED OR YOU WILL BE TERMINATED. MESSAGE ENDS. Welcome back to another installment of Beej’s Guide. When we met last we were talking about how to make use of pointers. Well, what we’re going to do is store a pointer off in a variable so that we can use it later. You can identify the pointer type because there’s an asterisk (*) before the variable name and after its type:
1 int main(void)
2 {
3 int i; // i's type is "int"
4 int *p; // p's type is "pointer to an int", or "int-pointer"
5 }
Hey, so we have here a variable that is a pointer type, and it can point to other ints. That it, is can hold the address of other ints. We know it points to ints, since it’s of type int* (read “int-pointer”). When you do an assignment into a pointer variable, the type of the right hand side of the assignment has to be the same type as the pointer variable. Fortunately for us, when you take the address-of a variable, the resultant type is a pointer to that variable type, so assignments like the following are perfect: int i; int *p; // p is a pointer, but is uninitialized and points to garbage
p = &i; // p is assigned the address of i--p now "points to" i
On the left of the assignment, we have a variable of type pointer-to-int (int*), and on the right side, we have expression of type pointer-to-int since i is an int (because address-of int gives you a pointer to int). The address of a thing can be stored in a pointer to that thing. Get it? I know it still doesn’t quite make much sense since you haven’t seen an actual use for the pointer variable, but we’re taking small steps here so that no one gets lost. So now, let’s introduce you to the anti-address-of operator. It’s kind of like what address-of would be like in Bizarro World. Chapter 5. Pointers—Cower In Fear! 29
5.3 Dereferencing
A pointer variable can be thought of as referring to another variable by pointing to it. It’s rare you’ll hear anyone in C land talking about “referring” or “references”, but I bring it up just so that the name of this operator will make a little more sense. When you have a pointer to a variable (roughly “a reference to a variable”), you can use the original variable through the pointer by dereferencing the pointer. (You can think of this as “de-pointering” the pointer, but no one ever says “de-pointering”.) Back to our analogy, this is vaguely like looking at a home address and then going to that house. Now, what do I mean by “get access to the original variable”? Well, if you have a variable called i, and you have a pointer to i called p, you can use the dereferenced pointer p exactly as if it were the original variable i! You almost have enough knowledge to handle an example. The last tidbit you need to know is actually this: what is the dereference operator? It is the asterisk, again: *. Now, don’t get this confused with the asterisk you used in the pointer declaration, earlier. They are the same character, but they have different meanings in different contexts5. Here’s a full-blown example:
1 #include
2
3 int main(void)
4 {
5 int i;
6 int *p; // this is NOT a dereference--this is a type "int*"
7
8 p = &i; // p now points to i, p holds address of i
9
10 i = 10; // i is now 10
11 *p = 20; // the thing p points to (namely i!) is now 20!!
12
13 printf("i is %d\n", i); // prints "20"
14 printf("i is %d\n", *p); // "20"! dereference-p is the same as i!
15 }
Remember that p holds the address of i, as you can see where we did the assignment to p on line 8. What the dereference operator does is tells the computer to use the object the pointer points to instead of using the pointer itself. In this way, we have turned *p into an alias of sorts for i. Great, but why? Why do any of this?
5.4 Passing Pointers as Arguments
Right about now, you’re thinking that you have an awful lot of knowledge about pointers, but absolutely zero application, right? I mean, what use is *p if you could just simply say i instead? Well, my friend, the real power of pointers comes into play when you start passing them to functions. Why is this a big deal? You might recall from before that you could pass all kinds of arguments to functions and they’d be dutifully copied into parameters, and then you could manipulate local copies of those variables from within the function, and then you could return a single value. What if you wanted to bring back more than one single piece of data from the function? I mean, you can only return one thing, right? What if I answered that question with another question? …Er, two questions? What happens when you pass a pointer as an argument to a function? Does a copy of the pointer get put into its corresponding parameter? You bet your sweet peas it does. Remember how earlier I rambled on
5That’s not all! It’s used in /*comments*/ and multiplication and in function prototypes with variable length arrays! It’s all the same *, but the context gives it different meaning. Chapter 5. Pointers—Cower In Fear! 30
and on about how EVERY SINGLE ARGUMENT gets copied into parameters and the function uses a copy of the argument? Well, the same is true here. The function will get a copy of the pointer. But, and this is the clever part: we will have set up the pointer in advance to point at a variable… and then the function can dereference its copy of the pointer to get back to the original variable! The function can’t see the variable itself, but it can certainly dereference a pointer to that variable! This is analogous to writing a home address on a piece of paper, and then copying that onto another piece of paper. You now have two pointers to that house, and both are equally good at getting you to the house itself. In the case of a function call. one of the copies is stored in a pointer variable out in the calling scope, and the other is stored in a pointer variable that is the parameter of the function. Example! Let’s revisit our old increment() function, but this time let’s make it so that it actually incre- ments the value out in the caller.
1 #include
2
3 void increment(int *p) // note that it accepts a pointer to an int
4 {
5 *p = *p + 1; // add one to the thing p points to
6 }
7
8 int main(void)
9 {
10 int i = 10;
11 int *j = &i; // note the address-of; turns it into a pointer to i
12
13 printf("i is %d\n", i); // prints "10"
14 printf("i is also %d\n", *j); // prints "10"
15
16 increment(j); // j is an int*--to i
17
18 printf("i is %d\n", i); // prints "11"!
19 }
Ok! There are a couple things to see here… not the least of which is that the increment() function takes an int* as an argument. We pass it an int* in the call by changing the int variable i to an int* using the address-of operator. (Remember, a pointer holds an address, so we make pointers to variables by running them through the address-of operator.) The increment() function gets a copy of the pointer. Both the original pointer j (in main()) and the copy of that pointer p (the parameter in increment()) point to the same address, namely the one holding the value i. (Again, by analogy, like two pieces of paper with the same home address written on them.) Dereferencing either will allow you to modify the original variable i! The function can modify a variable in another scope! Rock on! The above example is often more concisely written in the call just by using address-of right in the argument list: printf("i is %d\n", i); // prints "10" increment(&i); printf("i is %d\n", i); // prints "11"!
Pointer enthusiasts will recall from early on in the guide, we used a function to read from the keyboard, scanf()… and, although you might not have recognized it at the time, we used the address-of to pass a pointer to a value to scanf(). We had to pass a pointer, see, because scanf() reads from the keyboard (typically) and stores the result in a variable. The only way it can see that variable out in the calling function’s scope is if we pass a pointer to that variable: int i = 0; Chapter 5. Pointers—Cower In Fear! 31 scanf("%d", &i); // pretend you typed "12" printf("i is %d\n", i); // prints "i is 12"
See, scanf() dereferences the pointer we pass it in order to modify the variable it points to. And now you know why you have to put that pesky ampersand in there!
5.5 The NULL Pointer
Any pointer variable of any pointer type can be set to a special value called NULL. This indicates that this pointer doesn’t point to anything. int *p; p = NULL;
Since it doesn’t point to a value, dereferencing it is undefined behavior, and probably will result in a crash: int *p = NULL;
*p = 12; // CRASH or SOMETHING PROBABLY BAD. BEST AVOIDED.
Despite being called the billion dollar mistake by its creator6, the NULL pointer is a good sentinel value7 and general indicator that a pointer hasn’t yet been initialized. (Of course, like other variables, the pointer points to garbage unless you explicitly assign it to point to an address or NULL.)
5.6 A Note on Declaring Pointers
The syntax for declaring a pointer can get a little weird. Let’s look at this example: int a; int b;
We can condense that into a single line, right? int a, b; // Same thing
So a and b are both ints. No problem. But what about this? int a; int *p;
Can we make that into one line? We can. But where does the * go? The rule is that the * goes in front of any variable that is a pointer type. That is. the * is not part of the int in this example. it’s a part of variable p. With that in mind, we can write this: int a, *p; // Same thing
It’s important to note that the following line does not declare two pointers: int *p, q; // p is a pointer to an int; q is just an int.
This can be particularly insidious-looking if the programmer writes this following (valid) line of code which is functionally identical to the one above.
6https://en.wikipedia.org/wiki/Null_pointer#History 7https://en.wikipedia.org/wiki/Sentinel_value Chapter 5. Pointers—Cower In Fear! 32 int* p, q; // p is a pointer to an int; q is just an int.
So take a look at this and determine which variables are pointers and which are not: int *a, b, c, *d, e, *f, g, h, *i;
I’ll drop the answer in a footnote8.
5.7 sizeof and Pointers
Just a little bit of syntax here that might be confusing and you might see from time to time. Recall that sizeof operates on the type of the expression. int *p; sizeof(int); // Returns size of an `int` sizeof p // p is type int*, so returns size of `int*` sizeof *p // *p is type int, so returns size of `int`
You might see code with that last sizeof in there. Just remember that sizeof is all about the type of the expression, not the variables in the expression themselves.
8The pointer type variables are a, d, f, and i, because those are the ones with * in front of them. Chapter 6
Arrays
“Should array indices start at 0 or 1? My compromise of 0.5 was rejected without, I thought, proper consideration.” —Stan Kelly-Bootle, computer scientist Luckily, C has arrays. I mean, I know it’s considered a low-level language1 but it does at least have the concept of arrays built-in. And since a great many languages drew inspiration from C’s syntax, you’re probably already familiar with using [ and ] for declaring and using arrays. But C only barely has arrays! As we’ll find out later, arrays are just syntactic sugar in C—they’re actually all pointers and stuff deep down. Freak out! But for now, let’s just use them as arrays. Phew.
6.1 Easy Example
Let’s just crank out an example:
1 #include
2
3 int main(void)
4 {
5 int i;
6 float f[4]; // Declare an array of 4 floats
7
8 f[0] = 3.14159; // Indexing starts at 0, of course.
9 f[1] = 1.41421;
10 f[2] = 1.61803;
11 f[3] = 2.71828;
12
13 // Print them all out:
14
15 for (i = 0; i < 4; i++) {
16 printf("%f\n", f[i]);
17 }
18 }
When you declare an array, you have to give it a size. And the size has to be fixed2. In the above example, we made an array of 4 floats. The value in the square brackets in the declaration lets us know that. Later on in subsequent lines, we access the values in the array, setting them or getting them, again with square brackets.
1These days, anyway. 2Again, not really, but variable-length arrays—of which I’m not really a fan—are a story for another time.
33 Chapter 6. Arrays 34
Hopefully this looks familiar from languages you already know!
6.2 Getting the Length of an Array
You can’t…ish. C doesn’t record this information3. You have to manage it separately in another variable. When I say “can’t”, I actually mean there are some circumstances when you can. There is a trick to get the number of elements in an array in the scope in which an array is declared. But, generally speaking, this won’t work the way you want if you pass the array to a function4. Let’s take a look at this trick. The basic idea is that you take the sizeof the array, and then divide that by the size of each element to get the length. For example, if an int is 4 bytes, and the array is 32 bytes 32 long, there must be room for 4 or 8 ints in there. int x[12]; // 12 ints
printf("%zu\n", sizeof x); // 48 total bytes printf("%zu\n", sizeof(int)); // 4 bytes per int
printf("%zu\n", sizeof x / sizeof(int)); // 48/4 = 12 ints!
If it’s an array of chars, then sizeof the array is the number of elements, since sizeof(char) is defined to be 1. For anything else, you have to divide by the size of each element. But this trick only works in the scope in which the array was defined. If you pass the array to a function, it doesn’t work. Even if you make it “big” in the function signature: void foo(int x[12]) { printf("%zu\n", sizeof x); // 8?! What happened to 48? printf("%zu\n", sizeof(int)); // 4 bytes per int
printf("%zu\n", sizeof x / sizeof(int)); // 8/4 = 2 ints?? WRONG. }
This is because when you “pass” arrays to functions, you’re only passing a pointer to the first element, and that’s what sizeof measures. More on this in the Passing Single Dimensional Arrays to Functions section, below. One more thing you can do with sizeof and arrays is get the size of an array of a fixed number of elements without declaring the array. This is like how you can get the size of an int with sizeof(int). For example, to see how many bytes would be needed for an array of 48 doubles, you can do this: sizeof(double [48]);
6.3 Array Initializers
You can initialize an array with constants ahead of time:
1 #include
2
3 int main(void)
4 {
5 int i;
6 int a[5] = {22, 37, 3490, 18, 95}; // Initialize with these values
7
8 for (i = 0; i < 5; i++) {
3Since arrays are just pointers to the first element of the array under the hood, there’s no additional information recording the length. 4Because when you pass an array to a function, you’re actually just passing a pointer to the first element of that array, not the “entire” array. Chapter 6. Arrays 35
9 printf("%d\n", a[i]);
10 }
11 }
Catch: initializer values must be constant terms. Can’t throw variables in there. Sorry, Illinois! You should never have more items in your initializer than there is room for in the array, or the compiler will get cranky: foo.c: In function ‘main’: foo.c:6:39: warning: excess elements in array initializer 6 | int a[5] = {22, 37, 3490, 18, 95, 999}; | ^~~ foo.c:6:39: note: (near initialization for ‘a’)
But (fun fact!) you can have fewer items in your initializer than there is room for in the array. The remaining elements in the array will be automatically initialized with zero. This is true in general for all types of array initializers: if you have an initializer, anything not explicitly set to a value will be set to zero. int a[5] = {22, 37, 3490};
// is the same as:
int a[5] = {22, 37, 3490, 0, 0};
It’s a common shortcut to see this in an initializer when you want to set an entire array to zero: int a[100] = {0};
Which means, “Make the first element zero, and then automatically make the rest zero, as well.” You can set specific array elements in the initializer, as well, by specifying an index for the value! When you do this, C will happily keep initializing subsequent values for you until the initializer runs out, filling everything else with 0. To do this, put the index in square brackets with an = after, and then set the value. Here’s an example where we build an array: int a[10] = {0, 11, 22, [5]=55, 66, 77};
Because we listed index 5 as the start for 55, the resulting data in the array is: 0 11 22 0 0 55 66 77 0 0
You can put simple constant expressions in there, as well. #define COUNT 5
int a[COUNT] = {[COUNT-3]=3, 2, 1};
which gives us: 0 0 3 2 1
Lastly, you can also have C compute the size of the array from the initializer, just by leaving the size off: int a[3] = {22, 37, 3490};
// is the same as:
int a[] = {22, 37, 3490}; // Left the size off! Chapter 6. Arrays 36
6.4 Out of Bounds!
C doesn’t stop you from accessing arrays out of bounds. It might not even warn you. Let’s steal the example from above and keep printing off the end of the array. It only has 5 elements, but let’s try to print 10 and see what happens:
1 #include
2
3 int main(void)
4 {
5 int i;
6 int a[5] = {22, 37, 3490, 18, 95};
7
8 for (i = 0; i < 10; i++) { // BAD NEWS: printing too many elements!
9 printf("%d\n", a[i]);
10 }
11 }
Running it on my computer prints: 22 37 3490 18 95 32765 1847052032 1780534144 -56487472 21890
Yikes! What’s that? Well, turns out printing off the end of an array results in what C developers call undefined behavior. We’ll talk more about this beast later, but for now it means, “You’ve done something bad, and anything could happen during your program run.” And by anything, I mean typically things like finding zeroes, finding garbage numbers, or crashing. But really the C spec says in this circumstance the compiler is allowed to emit code that does anything5. Short version: don’t do anything that causes undefined behavior. Ever6.
6.5 Multidimensional Arrays
You can add as many dimensions as you want to your arrays. int a[10]; int b[2][7]; int c[4][5][6];
These are stored in memory in row-major order7. This means with a 2D array, the first index listed indicates the row, and the second the column. You an also use initializers on multidimensional arrays by nesting them:
1 #include
2
5In the good old MS-DOS days before memory protection was a thing, I was writing some particularly abusive C code that deliberately engaged in all kinds of undefined behavior. But I knew what I was doing, and things were working pretty well. Until I made a misstep that caused a lockup and, as I found upon reboot, nuked all my BIOS settings. That was fun. (Shout-out to @man for those fun times.) 6There are a lot of things that cause undefined behavior, not just out-of-bounds array accesses. This is what makes the C language so exciting. 7https://en.wikipedia.org/wiki/Row-_and_column-major_order Chapter 6. Arrays 37
3 int main(void)
4 {
5 int row, col;
6
7 int a[2][5] = { // Initialize a 2D array
8 {0, 1, 2, 3, 4},
9 {5, 6, 7, 8, 9}
10 };
11
12 for (row = 0; row < 2; row++) {
13 for (col = 0; col < 5; col++) {
14 printf("(%d,%d) = %d\n", row, col, a[row][col]);
15 }
16 }
17 }
For output of: (0,0) = 0 (0,1) = 1 (0,2) = 2 (0,3) = 3 (0,4) = 4 (1,0) = 5 (1,1) = 6 (1,2) = 7 (1,3) = 8 (1,4) = 9
And you can initialize with explicit indexes: // Make a 3x3 identity matrix
int a[3][3] = {[0][0]=1, [1][1]=1, [2][2]=1};
which builds a 2D array like this: 1 0 0 0 1 0 0 0 1
6.6 Arrays and Pointers
[Casually] So… I kinda might have mentioned up there that arrays were pointers, deep down? We should take a shallow dive into that now so that things aren’t completely confusing. Later on, we’ll look at what the real relationship between arrays and pointers is, but for now I just want to look at passing arrays to functions.
6.6.1 Getting a Pointer to an Array I want to tell you a secret. Generally speaking, when a C programmer talks about a pointer to an array, they’re talking about a pointer to the first element of the array8. So let’s get a pointer to the first element of an array.
1 #include
2
3 int main(void)
8This is technically incorrect, as a pointer to an array and a pointer to the first element of an array have different types. But we can burn that bridge when we get to it. Chapter 6. Arrays 38
4 {
5 int a[5] = {11, 22, 33, 44, 55};
6 int *p;
7
8 p = &a[0]; // p points to the array
9 // Well, to the first element, actually
10
11 printf("%d\n", *p); // Prints "11"
12 }
This is so common to do in C that the language allows us a shorthand:
1 p = &a[0]; // p points to the array
2
3 // is the same as:
4
5 p = a; // p points to the array, but much nicer-looking!
Just referring to the array name in isolation is the same as getting a pointer to the first element of the array! We’re going to use this extensively in the upcoming examples. But hold on a second–isn’t p an int*? And *p gives us 11, same as a[0]? Yessss. You’re starting to get a glimpse of how arrays and pointers are related in C.
6.6.2 Passing Single Dimensional Arrays to Functions Let’s do an example with a single dimensional array. I’m going to write a couple functions that we can pass the array to that do different things. Prepare for some mind-blowing function signatures!
1 #include
2
3 // Passing as a pointer to the first element
4 void times2(int *a, int len)
5 {
6 for (int i = 0; i < len; i++)
7 printf("%d\n", a[i] * 2);
8 }
9
10 // Same thing, but using array notation
11 void times3(int a[], int len)
12 {
13 for (int i = 0; i < len; i++)
14 printf("%d\n", a[i] * 3);
15 }
16
17 // Same thing, but using array notation with size
18 void times4(int a[5], int len)
19 {
20 for (int i = 0; i < len; i++)
21 printf("%d\n", a[i] * 4);
22 }
23
24 int main(void)
25 {
26 int x[5] = {11, 22, 33, 44, 55};
27
28 times2(x, 5);
29 times3(x, 5); Chapter 6. Arrays 39
30 times4(x, 5);
31 }
All those methods of listing the array as a parameter in the function are identical. void times2(int *a, int len) void times3(int a[], int len) void times4(int a[5], int len)
In usage by C regulars, the first is the most common, by far. And, in fact, in the latter situation, the compiler doesn’t even care what number you pass in (other than it has to be greater than zero9). It doesn’t enforce anything at all. Now that I’ve said that, the size of the array in the function declaration actually does matter when you’re passing multidimensional arrays into functions, but let’s come back to that.
6.6.3 Changing Arrays in Functions We’ve said that arrays are just pointers in disguise. This means that if you pass an array to a function, you’re likely passing a pointer to the first element in the array. But if the function has a pointer to the data, it is able to manipulate that data! So changes that a function makes to an array will be visible back out in the caller. Here’s an example where we pass a pointer to an array to a function, the function manipulates the values in that array, and those changes are visible out in the caller.
1 #include
2
3 void double_array(int *a, int len)
4 {
5 // Multiple each element by 2
6 //
7 // This doubles the values in x in main() since x and a both point
8 // to the same array in memory!
9
10 for (int i = 0; i < len; i++)
11 a[i] *= 2;
12 }
13
14 int main(void)
15 {
16 int x[5] = {1, 2, 3, 4, 5};
17
18 double_array(x, 5);
19
20 for (int i = 0; i < 5; i++)
21 printf("%d\n", x[i]); // 2, 4, 6, 8, 10!
22 }
Even though we passed the array in as parameter a which is type int*, look at how we access it using array notation with a[i]! Whaaaat. This is totally allowed. Later when we talk about the equivalence between arrays and pointers, we’ll see how this makes a lot more sense. For now, it’s enough to know that functions can make changes to arrays that are visible out in the caller. 9C11 §6.7.6.2¶1 requires it be greater than zero. But you might see code out there with arrays declared of zero length at the end of structs and GCC is particularly lenient about it unless you compile with -pedantic. This zero-length array was a hackish mechanism for making variable-length structures. Unfortunately, it’s technically undefined behavior to access such an array even though it basically worked everywhere. C99 codified a well-defined replacement for it called flexible array members, which we’ll chat about later. Chapter 6. Arrays 40
6.6.4 Passing Multidimensional Arrays to Functions The story changes a little when we’re talking about multidimensional arrays. C needs to know all the dimensions (except the first one) so it has enough information to know where in memory to look to find a value. Here’s an example where we’re explicit with all the dimensions:
1 #include
2
3 void print_2D_array(int a[2][3])
4 {
5 for (int row = 0; row < 2; row++) {
6 for (int col = 0; col < 3; col++)
7 printf("%d ", a[row][col]);
8 printf("\n");
9 }
10 }
11
12 int main(void)
13 {
14 int x[2][3] = {
15 {1, 2, 3},
16 {4, 5, 6}
17 };
18
19 print_2D_array(x);
20 }
But in this case, these two10 are equivalent: void print_2D_array(int a[2][3]) void print_2D_array(int a[][3])
The compiler really only needs the second dimension so it can figure out how far in memory to skip for each increment of the first dimension. In general, it needs to know all the dimensions except the first one. Also, remember that the compiler does minimal compile-time bounds checking (if you’re lucky), and C does zero runtime checking of bounds. No seat belts! Don’t crash by accessing array elements out of bounds!
10This is also equivalent: void print_2D_array(int (*a)[3]), but that’s more than I want to get into right now. Chapter 7
Strings
Finally! Strings! What could be simpler? Well, turns out strings aren’t actually strings in C. That’s right! They’re pointers! Of course they are! Much like arrays, strings in C barely exist. But let’s check it out—it’s not really such a big deal.
7.1 String Literals
Before we start, let’s talk about string literals in C. These are sequences of characters in double quotes ("). (Single quotes enclose characters, and are a different animal entirely.) Examples: "Hello, world!\n" "This is a test." "When asked if this string had quotes in it, she replied, \"It does.\""
The first one has a newline at the end—quite a common thing to see. The last one has quotes embedded within it, but you see each is preceded by (we say “escaped by”) a backslash (\) indicating that a literal quote belongs in the string at this point. This is how the C compiler can tell the difference between printing a double quote and the double quote at the end of the string.
7.2 String Variables
Now that we know how to make a string literal, let’s assign it to a variable so we can do something with it. char *s = "Hello, world!";
Check out that type: pointer to a char. The string variable s is actually a pointer to the first character in that string, namely the H. And we can print it with the %s (for “string”) format specifier: char *s = "Hello, world!"; printf("%s\n", s); // "Hello, world!"
7.3 String Variables as Arrays
Another option is this, nearly equivalent to the above char* usage:
41 Chapter 7. Strings 42
char s[14] = "Hello, world!";
// or, if we were properly lazy and have the compiler // figure the length for us:
char s[] = "Hello, world!";
This means you can use array notation to access characters in a string. Let’s do exactly that to print all the characters in a string on the same line:
1 #include
2
3 int main(void)
4 {
5 char s[] = "Hello, world!";
6
7 for (int i = 0; i < 13; i++)
8 printf("%c\n", s[i]);
9 }
Note that we’re using the format specifier %c to print a single character. Also, check this out. The program will still work fine if we change the definition of s to be a char* type:
1 #include
2
3 int main(void)
4 {
5 char *s = "Hello, world!"; // char* here
6
7 for (int i = 0; i < 13; i++)
8 printf("%c\n", s[i]); // But still use arrays here...?
9 }
And we still can use array notation to get the job done when printing it out! This is surprising, but is still only because we haven’t talked about array/pointer equivalence yet. But this is yet another hint that arrays and pointers are the same thing, deep down.
7.4 String Initializers
We’ve already seen some examples with initializing string variables with string literals: char *s = "Hello, world!"; char t[] = "Hello, again!";
But these two are subtly different. This one is a pointer to a string literal (i.e. a pointer to the first character in a string): char *s = "Hello, world!";
If you try to mutate that string with this: char *s = "Hello, world!";
s[0] = 'z'; // BAD NEWS: tried to mutate a string literal!
The behavior is undefined. Probably, depending on your system, a crash will result. But declaring it as an array is different. This one is a mutable copy of the string that we can change at will: Chapter 7. Strings 43
char t[] = "Hello, again!"; // t is an array copy of the string t[0] = 'z'; // No problem
printf("%s\n", t); // "zello, again!"
So remember: if you have a pointer to a string literal, don’t try to change it! And if you use a string in double quotes to initialize an array, that’s not actually a string literal.
7.5 Getting String Length
You can’t, since C doesn’t track it for you. And when I say “can’t”, I actually mean “can”1. There’s a function in
1 #include
2 #include
3
4 int main(void)
5 {
6 char *s = "Hello, world!";
7
8 printf("The string is %zu bytes long.\n", strlen(s));
9 }
The strlen() function returns type size_t, which is an integer type so you can use it for integer math. We print size_t with %zu. The above program prints: The string is 13 bytes long.
Great! So it is possible to get the string length! But… if C doesn’t track the length of the string anywhere, how does it know how long the string is?
7.6 String Termination
C does strings a little differently than many programming languages, and in fact differently than almost every modern programming language. When you’re making a new language, you have basically two options for storing a string in memory: 1. Store the bytes of the string along with a number indicating the length of the string. 2. Store the bytes of the string, and mark the end of the string with a special byte called the terminator. If you want strings longer than 255 characters, option 1 requires at least two bytes to store the length. Whereas option 2 only requires one byte to terminate the string. So a bit of savings there. Of course, these days it seems ridiculous to worry about saving a byte (or 3—lots of languages will happily let you have strings that are 4 gigabytes in length). But back in the day, it was a bigger deal. So C took approach #2. In C, a “string” is defined by two basic characteristics: • A pointer to the first character in the string. • A zero-valued byte (or NUL character3) somewhere in memory after the pointer that indicates the end of the string. A NUL character can be written in C code as \0, though you don’t often have to do this.
1Though it is true that C doesn’t track the length of strings. 2If you’re using the basic character set or an 8-bit character set, you’re used to one character being one byte. This isn’t true in all character encodings, though. 3This is different than the NULL pointer, and I’ll abbreviate it NUL when talking about the character versus NULL for the pointer. Chapter 7. Strings 44
When you include a string in double quotes in your code, the NUL character is automatically, implicitly included. char *s = "Hello!"; // Actually "Hello!\0" behind the scenes
So with this in mind, let’s write our own strlen() function that counts chars in a string until it finds a NUL. The procedure is to look down the string for a single NUL character, counting as we go4: int my_strlen(char *s) { int count = 0;
while (s[count] != '\0') // Single quotes for single char count++;
return count; }
And that’s basically how the built-in strlen() gets the job done.
7.7 Copying a String
You can’t copy a string through the assignment operator (=). All that does is make a copy of the pointer to the first character… so you end up with two pointers to the same string:
1 #include
2
3 int main(void)
4 {
5 char s[] = "Hello, world!";
6 char *t;
7
8 // This makes a copy of the pointer, not a copy of the string!
9 t = s;
10
11 // We modify t
12 t[0] = 'z';
13
14 // But printing s shows the modification!
15 // Because t and s point to the same string!
16
17 printf("%s\n", s); // "zello, world!"
18 }
If you want to make a copy of a string, you have to copy it a byte at a time—but this is made easier with the strcpy() function5. Before you copy the string, make sure you have room to copy it into, i.e. the destination array that’s going to hold the characters needs to be at least as long as the string you’re copying.
1 #include
2 #include
3
4 int main(void)
5 {
6 char s[] = "Hello, world!";
7 char t[100]; // Each char is one byte, so plenty of room
4Later we’ll learn a neater way to do it with pointer arithmetic. 5There’s a safer function called strncpy() that you should probably use instead, but we’ll get to that later. Chapter 7. Strings 45
8
9 // This makes a copy of the string!
10 strcpy(t, s);
11
12 // We modify t
13 t[0] = 'z';
14
15 // And s remains unaffected because it's a different string
16 printf("%s\n", s); // "Hello, world!"
17
18 // But t has been changed
19 printf("%s\n", t); // "zello, world!"
20 }
Notice with strcpy(), the destination pointer is the first argument, and the source pointer is the second. A mnemonic I use to remember this is that it’s the order you would have put t and s if an assignment = worked for strings, with the source on the right and the destination on the left. Chapter 8
Structs
In C, we have something called a struct, which is a user-definable type that holds multiple pieces of data, potentially of different types. It’s a convenient way to bundle multiple variables into a single one. This can be beneficial for passing variables to functions (so you just have to pass one instead of many), and useful for organizing data and making code more readable. If you’ve come from another language, you might be familiar with the idea of classes and objects. These don’t exist in C, natively1. You can think of a struct as a class with only data members, and no methods.
8.1 Declaring a Struct
You can declare a struct in your code like so: struct car { char *name; float price; int speed; };
This is often done at the global scope outside any functions so that the struct is globally available. When you do this, you’re making a new type. The full type name is struct car. (Not just car—that won’t work.) There aren’t any variables of that type yet, but we can declare some: struct car saturn; // Variable "saturn" of type "struct car"
And now we have an uninitialized variable saturn2 of type struct car. We should initialize it! But how do we set the values of those individual fields? Like in many other languages that stole it from C, we’re going to use the dot operator (.) to access the individual fields. saturn.name = "Saturn SL/2"; saturn.price = 15999.99; saturn.speed = 175; printf("Name: %s\n", saturn.name);
1Although in C individual items in memory like ints are referred to as “objects”, they’re not objects in an object-oriented programming sense. 2The Saturn was a popular brand of economy car in the United States until it was put out of business by the 2008 crash, sadly so to us fans.
46 Chapter 8. Structs 47 printf("Price (USD): %f\n", saturn.price); printf("Top Speed (km): %d\n", saturn.speed);
There on the first lines, we set the values in the struct car, and then in the next bit, we print those values out.
8.2 Struct Initializers
That example in the previous section was a little unwieldy. There must be a better way to initialize that struct variable! You can do it with an initializer by putting values in for the fields in the order they appear in the struct when you define the variable. (This won’t work after the variable has been defined—it has to happen in the definition). struct car { char *name; float price; int speed; };
// Now with an initializer! Same field order as in the struct declaration: struct car saturn = {"Saturn SL/2", 16000.99, 175}; printf("Name: %s\n", saturn.name); printf("Price: %f\n", saturn.price); printf("Top Speed: %d km\n", saturn.speed);
The fact that the fields in the initializer need to be in the same order is a little freaky. If someone changes the order in struct car, it could break all the other code! We can be more specific with our initializers: struct car saturn = {.speed=172, .name="Saturn SL/2"};
Now it’s independent of the order in the struct declaration. Which is safer code, for sure. Similar to array initializers, any missing field designators are initialized to zero (in this case, that would be .price, which I’ve omitted).
8.3 Passing Structs to Functions
You can do a couple things to pass a struct to a function. 1. Pass the struct. 2. Pass a pointer to the struct. Recall that when you pass something to a function, a copy of that thing gets made for the function to operate on, whether it’s a copy of a pointer, an int, a struct, or anything. There are basically two cases when you’d want to pass a pointer to the struct: 1. You need the function to be able to make changes to the struct that was passed in, and have those changes show in the caller. 2. The struct is somewhat large and it’s more expensive to copy that onto the stack than it is to just copy a pointer3 For those two reasons, it’s far more common to pass a pointer to a struct to a function, though its by no means illegal to pass the struct itself.
3A pointer is likely 8 bytes on a 64-bit system. Chapter 8. Structs 48
Let’s try passing in a pointer, making a function that will allow you to set the .price field of the struct car:
1 struct car {
2 char *name;
3 float price;
4 int speed;
5 };
6
7 int main(void)
8 {
9 struct car saturn = {.speed=175, .name="Saturn SL/2"};
10
11 // Pass a pointer to this struct car, along with a new,
12 // more realistic, price:
13 set_price(&saturn, 799.99);
14
15 // ... code continues ...
You should be able to come up with the function signature for set_price() just by looking at the types of the arguments we have there. saturn is a struct car, so &saturn must be the address of the struct car, AKA a pointer to a struct car, namely a struct car*. And 799.99 is a float. So the function declaration must look like this: void set_price(struct car *c, float new_price)
We just need to write the body. One attempt might be: void set_price(struct car *c, float new_price) { c.price = new_price; // ERROR!! }
That won’t work because the dot operator only works on structs… it doesn’t work on pointers to structs. Ok, so we can dereference the struct to de-pointer it to get to the struct itself. Dereferencing a struct car* results in the struct car that the pointer points to, which we should be able to use the dot operator on: void set_price(struct car *c, float new_price) { (*c).price = new_price; // Works, but is ugly and non-idiomatic :( }
And that works! But it’s a little clunky to type all those parens and the asterisk. C has some syntactic sugar called the arrow operator that helps with that.
8.4 The Arrow Operator
void set_price(struct car *c, float new_price) { // (*c).price = new_price; // Works, but non-idiomatic :( // // The line above is 100% equivalent to the one below:
c->price = new_price; // That's the one! }
The arrow operator helps refer to fields in pointers to structs. Chapter 8. Structs 49
So when accessing fields, when do we use dot and when do we use arrow? • If you have a struct, use dot (.). • If you have a pointer to a struct, use arrow (->).
8.5 Copying and Returning structs
Here’s an easy one for you! Just assign from one to the other! struct car a, b; b = a; // Copy the struct
And returning a struct (as opposed to a pointer to one) from a function also makes a similar copy to the receiving variable. This is not a “deep copy”4. All fields are copied as-is, including pointers to things.
4A deep copy follows pointer in the struct and copies the data they point to, as well. A shallow copy just copies the pointers, but not the things they point to. C doesn’t come with any built-in deep copy functionality. Chapter 9
File Input/Output
We’ve already seen a couple examples of I/O with scanf() and printf() for doing I/O at the console (screen/keyboard). But we’ll push those concepts a little farther this chapter.
9.1 The FILE* Data Type
When we do any kind of I/O in C, we do so though a piece of data that you get in the form of a FILE* type. This FILE* holds all the information needed to communicate with the I/O subsystem about which file you have open, where you are in the file, and so on. The spec refers to these as streams, i.e. a stream of data from a file or from any source. I’m going to use “files” and “streams” interchangeably, but really you should think of a “file” as a special case of a “stream”. There are other ways to stream data into a program than just reading from a file. We’ll see in a moment how to go from having a filename to getting an open FILE* for it, but first I want to mention three streams that are already open for you and ready for use.
FILE* name Description stdin Standard Input, generally the keyboard by default stdout Standard Output, generally the screen by default stderr Standard Error, generally the screen by default, as well
We’ve actually been using these implicitly already, it turns out. For example, these two calls are the same: printf("Hello, world!\n"); fprintf(stdout, "Hello, world!\n"); // printf to a file
But more on that later. Also you’ll notice that both stdout and stderr go to the screen. While this seems at first either like an oversight or redundancy, it actually isn’t. Typical operating systems allow you to redirect the output of either of those into different files, and it can be convenient to be able to separate error messages from regular non-error output. For example, in a POSIX shell (like sh, ksh, bash, zsh, etc.) on a Unix-like system, we could run a program and send just the non-error (stdout) output to one file, and all the error (stderr) output to another file. ./foo > output.txt 2> errors.txt # This command is Unix-specific
For this reason, you should send serious error messages to stderr instead of stdout. More on how to do that later.
50 Chapter 9. File Input/Output 51
9.2 Reading Text Files
Streams are largely categorized two different ways: text and binary. Text streams are allowed to do significant translation of the data, most notably translations of newlines to their different representations1. Text files are logically a sequence of lines separated by newlines. To be portable, your input data should always end with a newline. But the general rule is that if you’re able to edit the file in a regular text editor, it’s a text file. Otherwise, it’s binary. More on binary later. So let’s get to work—how do we open a file for reading, and pull data out of it? Let’s create a file called hello.txt that has just this in it: Hello, world!
And let’s write a program to open the file, read a character out of it, and then close the file when we’re done. That’s the game plan!
1 #include
2
3 int main(void)
4 {
5 FILE *fp; // Variable to represent open file
6
7 fp = fopen("hello.txt", "r"); // Open file for reading
8
9 char c = fgetc(fp); // Read a single character
10 printf("%c\n", c); // Print char to stdout
11
12 fclose(fp); // Close the file when done
13 }
See how when we opened the file with fopen(), it returned the FILE* to us so we could use it later. (I’m leaving it out for brevity, but fopen() will return NULL if something goes wrong, like file-not-found, so you should really error check it!) Also notice the "r" that we passed in—this means “open a text stream for reading”. (There are various strings we can pass to fopen() with additional meaning, like writing, or appending, and so on.) After that, we used the fgetc() function to get a character from the stream. Finally, we close the stream when we’re done with it. All streams are automatically closed when the program exits, but it’s good form and good housekeeping to explicitly close any files yourself when done with them. The FILE* keeps track of our position in the file. So subsequent calls to fgetc() would get the next character in the file, and then the next, until the end. But that sounds like a pain. Let’s see if we can make it easier.
9.3 End of File: EOF
There is a special character defined as a macro: EOF. This is what fgetc() will return when the end of the file has been reached and you’ve attempted to read another character. We can use this to read the whole file in a loop.
1We used to have three different newlines in broad effect: Carriage Return (CR, used on old Macs), Linefeed (LF, used on Unix systems), and Carriage Return/Linefeed (CRLF, used on Windows systems). Thankfully the introduction of OS X, being Unix-based, reduced this number to two. Chapter 9. File Input/Output 52
1 #include
2
3 int main(void)
4 {
5 FILE *fp;
6 int c;
7
8 fp = fopen("hello.txt", "r");
9
10 while ((c = fgetc(fp)) != EOF)
11 printf("%c", c);
12
13 fclose(fp);
14 }
(If line 10 is too weird, just break it down starting with the innermost-nested parens. The first thing we do is assign the result of fgetc() into c, and then we compare that against EOF. We’ve just crammed it into a single line. This might look hard to read, but study it—it’s idiomatic C.) And running this, we see: Hello, world!
But still, we’re operating a character at a time, and lots of text files make more sense at the line level. Let’s switch to that.
9.3.1 Reading a Line at a Time
So how can we get an entire line at once? fgets() to the rescue! For arguments, it takes a pointer to a char buffer to hold bytes, a maximum number of bytes to read, and a FILE* to read from. It returns NULL on end-of-file or error. fgets() is even nice enough to NUL-terminate the string when its done2. Let’s do a similar loop as before, except let’s have a multiline file and read it in a line at a time. Here’s a file quote.txt: A wise man can learn more from a foolish question than a fool can learn from a wise answer. --Bruce Lee
And here’s some code that reads that file a line at a time and prints out a line number before each one:
1 #include
2
3 int main(void)
4 {
5 FILE *fp;
6 char s[1024]; // Big enough for any line this program will encounter
7 int linecount = 0;
8
9 fp = fopen("quote.txt", "r");
10
11 while (fgets(s, sizeof s, fp) != NULL)
12 printf("%d: %s", ++linecount, s);
13
14 fclose(fp);
15 }
Which gives the output:
2If the buffer’s not big enough to read in an entire line, it’ll just stop reading mid-line, and the next call to fgets() will continue reading the line. Chapter 9. File Input/Output 53
1: A wise man can learn more from 2: a foolish question than a fool 3: can learn from a wise answer. 4: --Bruce Lee
9.4 Formatted Input
You know how you can get formatted output with printf() (and, thus, fprintf() like we’ll see, below)? You can do the same thing with fscanf(). Let’s have a file with a series of data records in it. In this case, whales, with name, length in meters, and weight in tonnes. whales.txt: blue 29.9 173 right 20.7 135 gray 14.9 41 humpback 16.0 30
Yes, we could read these with fgets() and then parse the string with sscanf() (and in some ways that’s more resilient against corrupted files), but in this case, let’s just use fscanf() and pull it in directly. The fscanf() function skips leading whitespace when reading, and returns EOF on end-of-file or error.
1 #include
2
3 int main(void)
4 {
5 FILE *fp;
6 char name[1024]; // Big enough for any line this program will encounter
7 float length;
8 int mass;
9
10 fp = fopen("whales.txt", "r");
11
12 while (fscanf(fp, "%s %f %d", name, &length, &mass) != EOF)
13 printf("%s whale, %d tonnes, %.1f meters\n", name, mass, length);
14
15 fclose(fp);
16 }
Which gives the result: blue whale, 173 tonnes, 29.9 meters right whale, 135 tonnes, 20.7 meters gray whale, 41 tonnes, 14.9 meters humpback whale, 30 tonnes, 16.0 meters
9.5 Writing Text Files
In much the same way we can use fgetc(), fgets(), and fscanf() to read text streams, we can use fputc(), fputs(), and fprintf() to write text streams. To do so, we have to fopen() the file in write mode by passing "w" as the second argument. Opening an existing file in "w" mode will instantly truncate that file to 0 bytes for a full overwrite. We’ll put together a simple program that outputs a file output.txt using a variety of output functions.
1 #include
2
3 int main(void) Chapter 9. File Input/Output 54
4 {
5 FILE *fp;
6 int x = 32;
7
8 fp = fopen("output.txt", "w");
9
10 fputc('B', fp);
11 fputc('\n', fp); // newline
12 fprintf(fp, "x = %d\n", x);
13 fputs("Hello, world!\n", fp);
14
15 fclose(fp);
16 }
And this produces a file, output.txt, with these contents: B x = 32 Hello, world!
Fun fact: since stdout is a file, you could replace line 8 with: fp = stdout;
and the program would have outputted to the console instead of to a file. Try it!
9.6 Binary File I/O
So far we’ve just been talking text files. But there’s that other beast we mentioned early on called binary files, or binary streams. These work very similarly to text files, except the I/O subsystem doesn’t perform any translations on the data like it might with a text file. With binary files, you get a raw stream of bytes, and that’s all. The big difference in opening the file is that you have to add a "b" to the mode. That is, to read a binary file, open it in "rb" mode. To write a file, open it in "wb" mode. Because it’s streams of bytes, and streams of bytes can contain NUL characters, and the NUL character is the end-of-string marker in C, it’s rare that people use the fprintf()-and-friends functions to operate on binary files. Instead the most common functions are fread() and fwrite(). The functions read and write a specified number of bytes to the stream. To demo, we’ll write a couple programs. One will write a sequence of byte values to disk all at once. And the second program will read a byte at a time and print them out3.
1 #include
2
3 int main(void)
4 {
5 FILE *fp;
6 unsigned char bytes[6] = {5, 37, 0, 88, 255, 12};
7
8 fp = fopen("output.bin", "wb"); // wb mode for "write binary"!
9
10 // In the call to fwrite, the arguments are:
11 //
12 // * Pointer to data to write
13 // * Size of each "piece" of data
3Normally the second program would read all the bytes at once, and then print them out in a loop. That would be more efficient. But we’re going for demo value, here. Chapter 9. File Input/Output 55
14 // * Count of each "piece" of data
15 // * FILE*
16
17 fwrite(bytes, sizeof(char), 6, fp);
18
19 fclose(fp);
20 }
Those two middle arguments to fwrite() are pretty odd. But basically what we want to tell the function is, “We have items that are this big, and we want to write that many of them.” This makes it convenient if you have a record of a fixed length, and you have a bunch of them in an array. You can just tell it the size of one record and how many to write. In the example above, we tell it each record is the size of a char, and we have 6 of them. Running the program gives us a file output.bin, but opening it in a text editor doesn’t show anything friendly! It’s binary data—not text. And random binary data I just made up, at that! If I run it through a hex dump4 program, we can see the output as bytes: 05 25 00 58 ff 0c
And those values in hex do match up to the values (in decimal) that we wrote out. But now let’s try to read them back in with a different program. This one will open the file for binary reading ("rb" mode) and will read the bytes one at a time in a loop. fread() has the neat feature where it returns the number of bytes read, or 0 on EOF. So we can loop until we see that, printing numbers as we go.
1 #include
2
3 int main(void)
4 {
5 FILE *fp;
6 unsigned char c;
7
8 fp = fopen("output.bin", "rb"); // rb for "read binary"!
9
10 while (fread(&c, sizeof(char), 1, fp) > 0)
11 printf("%d\n", c);
12 }
And, running it, we see our original numbers! 5 37 0 88 255 12
Woo hoo!
9.6.1 struct and Number Caveats
As we saw in the structs section, the compiler is free to add padding to a struct as it sees fit. And different compilers might do this differently. And the same compiler on different architectures could do it differently. And the same compiler on the same architectures could do it differently. What I’m getting at is this: it’s not portable to just fwrite() an entire struct out to a file when you don’t know where the padding will end up.
4https://en.wikipedia.org/wiki/Hex_dump Chapter 9. File Input/Output 56
How do we fix this? Hold that thought—we’ll look at some ways to do this after looking at another related problem. Numbers! Turns out all architectures don’t represent numbers in memory the same way. Let’s look at a simple fwrite() of a 2-byte number. We’ll write it in hex so each byte is clear. The most significant byte will have the value 0x12 and the least significant will have the value 0x34. unsigned short v = 0x1234; // Two bytes, 0x12 and 0x34 fwrite(&v, sizeof v, 1, fp);
What ends up in the stream? Well, it seems like it should be 0x12 followed by 0x34, right? But if I run this on my machine and hex dump the result, I get: 34 12
They’re reversed! What gives? This has something to do with what’s called the endianess5 of the architecture. Some write the most significant bytes first, and some the least significant bytes first. This means that if you write a multibyte number out straight from memory, you can’t do it in a portable way6. A similar problem exists with floating point. Most systems use the same format for their floating point numbers, but some do not. No guarantees! So… how can we fix all these problems with numbers and structs to get our data written in a portable way? The summary is to serialize the data, which is a general term that means to take all the data and write it out in a format that you control, that is well-known, and programmable to work the same way on all platforms. As you might imagine, this is a solved problem. There are a bunch of serialization libraries you can take advantage of, such as Google’s protocol buffers7, out there and ready to use. They will take care of all the gritty details for you, and even will allow data from your C programs to interoperate with other languages that support the same serialization methods. Do yourself and everyone a favor! Serialize your binary data when you write it to a stream! This will keep things nice and portable, even if you transfer data files from one architecture to another.
5https://en.wikipedia.org/wiki/Endianess 6And this is why I used individual bytes in my fwrite() and fread() examples, above, shrewdly. 7https://en.wikipedia.org/wiki/Protocol_buffers Chapter 10 typedef: Making New Types
Well, not so much making new types as getting new names for existing types. Sounds kinda pointless on the surface, but we can really use this to make our code cleaner.
10.1 typedef in Theory
Basically, you take an existing type and you make an alias for it with typedef. Like this: typedef int antelope; // Make "antelope" an alias for "int" antelope x = 10; // Type "antelope" is the same as type "int"
You can take any existing type and do it. You can even make a number of types with a comma list: typedef int antelope, bagel, mushroom; // These are all "int"
That’s really useful, right? That you can type mushroom instead of int? You must be super excited about this feature! OK, Professor Sarcasm—we’ll get to some more common applications of this in a moment.
10.1.1 Scoping typedef follows regular scoping rules. For this reason, it’s quite common to find typedef at file scope (“global”) so that all functions can use the new types at will.
10.2 typedef in Practice
So renaming int to something else isn’t that exciting. Let’s see where typedef commonly makes an appearance.
10.2.1 typedef and structs
Sometimes a struct will be typedef’d to a new name so you don’t have to type the word struct over and over. struct animal { char *name; int leg_count, speed; };
57 Chapter 10. typedef: Making New Types 58
// original name new name // | | // v v // |------| |----| typedef struct animal animal; struct animal y; // This works animal z; // This also works because "animal" is an alias
Personally, I don’t care for this practice. I like the clarity the code has when you add the word struct to the type; programmers know what they’re getting. But it’s really common so I’m including it here. Now I want to run the exact same example in a way that you might commonly see. We’re going to put the struct animal in the typedef. You can mash it all together like this: // original name // | // v // |------| typedef struct animal { char *name; int leg_count, speed; } animal; // <-- new name struct animal y; // This works animal z; // This also works because "animal" is an alias
That’s exactly the same as the previous example, just more concise. But that’s not all! There’s another common shortcut that you might see in code using what are called anonymous structures1. It turns out you don’t actually need to name the structure in a variety of places, and with typedef is one of them. Let’s do the same example with an anonymous structure: // Anonymous struct! It has no name! // | // v // |----| typedef struct { char *name; int leg_count, speed; } animal; // <-- new name
//struct animal y; // ERROR: this no longer works--no such struct! animal z; // This works because "animal" is an alias
As another example, we might find something like this: typedef struct { int x, y; } point; point p = {.x=20, .y=40}; printf("%d, %d\n", p.x, p.y); // 20, 40
1We’ll talk more about these later. Chapter 10. typedef: Making New Types 59
10.2.2 typedef and Other Types
It’s not that using typedef with a simple type like int is completely useless… it helps you abstract the types to make it easier to change them later. For example, if you have float all over your code in 100 zillion places, it’s going to be painful to change them all to double if you find you have to do that later for some reason. But if you prepared a little with: typedef float app_float;
// and app_float f1, f2, f3;
Then if later you want to change to another type, like long double, you just need to change the typedef: // voila! // |------| typedef long double app_float;
// and no need to change this line: app_float f1, f2, f3; // Now these are all long doubles
10.2.3 typedef and Pointers You can make a type that is a pointer. typedef int *intptr; int a = 10; intptr x = &a; // "intptr" is type "int*"
I really don’t like this practice. It hides the fact that x is a pointer type because you don’t see a * in the declaration. IMHO, it’s better to explicitly show that you’re declaring a pointer type so that other devs can clearly see it and don’t mistake x for having a non-pointer type. But at last count, say, 832,007 people had a different opinion.
10.2.4 typedef and Capitalization
I’ve seen all kinds of capitalization on typedef. typedef struct { int x, y; } my_point; // lower snake case typedef struct { int x, y; } MyPoint; // CamelCase typedef struct { int x, y; } Mypoint; // Leading uppercase typedef struct { int x, y; } MY_POINT; // UPPER SNAKE CASE Chapter 10. typedef: Making New Types 60
The C11 specification doesn’t dictate one way or another, and shows examples in all uppercase and all lowercase. K&R2 uses leading uppercase predominantly, but show some examples in uppercase and snake case (with _t). If you have a style guide in use, stick with it. If you don’t, grab one and stick with it.
10.3 Arrays and typedef
The syntax is a little weird, and this is rarely seen in my experience, but you can typedef an array of some number of items. // Make type five_ints an array of 5 ints typedef int five_ints[5]; five_ints x = {11, 22, 33, 44, 55};
I don’t like it because it hides the array nature of the variable, but it’s possible to do. Chapter 11
Pointers II: Arithmetic
Time to get more into it with a number of new pointer topics! If you’re not up to speed with pointers, check out the first section in the guide on the matter.
11.1 Pointer Arithmetic
Turns out you can do math on pointers, notably addition and subtraction. But what does it mean when you do that? In short, if you have a pointer to a type, adding one to the pointer moves to the next item of that type directly after it in memory. It’s important to remember that as we move pointers around and look at different places in memory, we need to make sure that we’re always pointing to a valid place in memory before we dereference. If we’re off in the weeds and we try to see what’s there, the behavior is undefined and a crash is a common result. This is a little chicken-and-eggy with Array/Pointer Equivalence, below, but we’re going to give it a shot, anyway.
11.1.1 Adding to Pointers First, let’s take an array of numbers. int a[5] = {11, 22, 33, 44, 55};
Then let’s get a pointer to the first element in that array: int a[5] = {11, 22, 33, 44, 55}; int *p = &a[0]; // Or "int *p = a;" works just as well
The let’s print the value there by dereferencing the pointer: printf("%d\n", *p); // Prints 11
Now let’s use pointer arithmetic to print the next element in the array, the one at index 1: printf("%d\n", *(p + 1)); // Prints 22!!
What happened there? C knows that p is a pointer to an int. So it knows the sizeof an int1 and it knows to skip that many bytes to get to the next int after the first one! In fact, the prior example could be written these two equivalent ways: printf("%d\n", *p); // Prints 11 printf("%d\n", *(p + 0)); // Prints 11
1Recall that the sizeof operator tells you the size in bytes of an object in memory.
61 Chapter 11. Pointers II: Arithmetic 62 because adding 0 to a pointer results in the same pointer. Let’s think of the upshot here. We can iterate over elements of an array this way instead of using an array: int a[5] = {11, 22, 33, 44, 55}; int *p = &a[0]; // Or "int *p = a;" works just as well for (int i = 0; i < 5; i++) { printf("%d\n", *(p + i)); // Same as p[i]! }
And that works the same as if we used array notation! Oooo! Getting closer to that array/pointer equiva- lence thing! More on this later in this chapter. But what’s actually happening, here? How does it work? Remember from early on that memory is like a big array, where a byte is stored at each array index? And the array index into memory has a few names: • Index into memory • Location • Address • Pointer! So a point is an index into memory, somewhere. For a random example, say that a number 3490 was stored at address (“index”) 23,237,489,202. If we have an int pointer to that 3490, that value of that pointer is 23,237,489,202… because the pointer is the memory address. Different words for the same thing. And now let’s say we have another number, 4096, stored right after the 3490 at address 23,237,489,210 (8 higher than the 3490 because each int in this example is 8 bytes long). If we add 1 to that pointer, it actually jumps ahead sizeof(int) bytes to the next int. It knows to jump that far ahead because it’s an int pointer. If it were a float pointer, it’d jump sizeof(float) bytes ahead to get to the next float! So you can look at the next int, by adding 1 to the pointer, the one after that by adding 2 to the pointer, and so on.
11.1.2 Changing Pointers We saw how we could add an integer to a pointer in the previous section. This time, let’s modify the pointer, itself. You can just add (or subtract) integer values directly to (or from) any pointer! Let’s do that example again, except with a couple changes. First, I’m going to add a 999 to the end of our numbers to act as a sentinel value. This will let us know where the end of the data is. int a[] = {11, 22, 33, 44, 55, 999}; // Add 999 here as a sentinel int *p = &a[0]; // p points to the 11
And we also have p pointing to the element at index 0 of a, namely 11, just like before. Now—let’s starting incrementing p so that it points at subsequent elements of the array. We’ll do this until p points to the 999; that is, we’ll do it until *p == 999: while (*p != 999) { // While the thing p points to isn't 999 printf("%d\n", *p); // Print it p++; // Move p to point to the next int! }
Pretty crazy, right? Chapter 11. Pointers II: Arithmetic 63
When we give it a run, first p points to 11. Then we increment p, and it points to 22, and then again, it points to 33. And so on, until it points to 999 and we quit.
11.1.3 Subtracting Pointers You can subtract a value from a pointer to get to earlier address, as well, just like we were adding to them before. But we can also subtract two pointers to find the difference between them, e.g. we can calculate how many ints there are between two int*s. The catch is that this only works within a single array2—if the pointers point to anything else, you get undefined behavior. Remember how strings are char*s in C? Let’s see if we can use this to write another variant of strlen() to compute the length of a string that utilizes pointer subtraction. The idea is that if we have a pointer to the beginning of the string, we can find a pointer to the end of the string by scanning ahead for the NUL character. And if we have a pointer to the beginning of the string, and we computed the pointer to the end of the string, we can just subtract the two pointers to come up with the length!
1 #include
2
3 int my_strlen(char *s)
4 {
5 // Start scanning from the beginning of the string
6 char *p = s;
7
8 // Scan until we find the NUL character
9 while (*p != '\0')
10 p++;
11
12 // Return the difference in pointers
13 return p - s;
14 }
15
16 int main(void)
17 {
18 printf("%d\n", my_strlen("Hello, world!")); // Prints "13"
19 }
Remember that you can only use pointer subtraction between two pointers that point to the same array!
11.2 Array/Pointer Equivalence
We’re finally ready to talk about this! We’ve seen plenty of examples of places where we’ve intermixed array notation, but let’s give out the fundamental formula of array/pointer equivalence: a[b] == *(a + b)
Study that! Those are equivalent and can be used interchangeably! I’ve oversimplified a bit, because in my above example a and b can both be expressions, and we might want a few more parentheses to force order of operations in case the expressions are complex. The spec is specific, as always, declaring (in C11 §6.5.2.1¶2): E1[E2] is identical to (*((E1)+(E2))) but that’s a little harder to grok. Just make sure you include parentheses if the expressions are complicated so all your math happens in the right order.
2Or string, which is really an array of chars. Somewhat peculiarly, you can also have a pointer that references one past the end of the array without a problem and still do math on it. You just can’t dereference it when it’s out there. Chapter 11. Pointers II: Arithmetic 64
This means we can decide if we’re going to use array or pointer notation for any array or pointer (assuming it points to an element of an array). Let’s use an array and pointer with both array and pointer notation:
1 #include
2
3 int main(void)
4 {
5 int a[] = {11, 22, 33, 44, 55}; // Add 999 here as a sentinel
6
7 int *p = a; // p points to the first element of a, 11
8
9 // Print all elements of the array a variety of ways:
10
11 for (int i = 0; i < 5; i++)
12 printf("%d\n", a[i]); // Array notation with a
13
14 for (int i = 0; i < 5; i++)
15 printf("%d\n", p[i]); // Array notation with p
16
17 for (int i = 0; i < 5; i++)
18 printf("%d\n", *(a + i)); // Pointer notation with a
19
20 for (int i = 0; i < 5; i++)
21 printf("%d\n", *(p + i)); // Pointer notation with p
22
23 for (int i = 0; i < 5; i++)
24 printf("%d\n", *(p++)); // Moving pointer p
25 //printf("%d\n", *(a++)); // Moving array variable a--ERROR!
26 }
So you can see that in general, if you have an array variable, you can use pointer or array notion to access elements. Same with a pointer variable. The one big difference is that you can modify a pointer to point to a different address, but you can’t do that with an array variable.
11.2.1 Array/Pointer Equivalence in Function Calls This is where you’ll encounter this concept the most, for sure. If you have a function that takes a pointer argument, e.g.: int my_strlen(char *s)
this means you can pass either an array or a pointer to this function and have it work! char s[] = "Antelopes"; char *t = "Wombats";
printf("%d\n", my_strlen(s)); // Works! printf("%d\n", my_strlen(t)); // Works, too!
And it’s also why these two function signatures are equivalent: int my_strlen(char *s) // Works! int my_strlen(char s[]) // Works, too! Chapter 11. Pointers II: Arithmetic 65
11.3 void Pointers
You’ve already seen the void keyword used with functions, but this is an entirely separate, unrelated animal. Sometimes it’s useful to have a pointer to a thing that you don’t know the type of. I know. Bear with me just a second. There are basically two use cases for this. 1. A function is going to operate on something byte-by-byte. For example, memcpy() copies bytes of memory from one pointer to another, but those pointers can point to any type. memcpy() takes advantage of the fact that if you iterate through char*s, you’re iterating through the bytes of an object no matter what type the object is. More on this in the Multibyte Values subsection. 2. Another function is calling a function you passed to it (a callback), and it’s passing you data. You know the type of the data, but the function calling you doesn’t. So it passes you void*s—’cause it doesn’t know the type—and you convert those to the type you need. The built-in qsort() and bsearch() use this technique. Let’s look at an example, the built-in memcpy() function: void *memcpy(void *s1, void *s2, size_t n);
This function copies n bytes of memory starting from address s1 into the memory starting at address s2. But look! s1 and s2 are void*s! Why? What does it mean? Let’s run more examples to see. For instance, we could copy a string with memcpy() (though strcpy() is more appropriate for strings):
1 #include
2 #include
3
4 int main(void)
5 {
6 char s[] = "Goats!";
7 char t[100];
8
9 memcpy(t, s, 7); // Copy 7 bytes--including the NUL terminator!
10
11 printf("%s\n", t); // "Goats!"
12 }
Or we can copy some ints:
1 #include
2 #include
3
4 int main(void)
5 {
6 int a[] = {11, 22, 33};
7 int b[3];
8
9 memcpy(b, a, 3 * sizeof(int)); // Copy 3 ints of data
10
11 printf("%d\n", b[1]); // 22
12 }
That one’s a little wild—you see what we did there with memcpy()? We copied the data from a to b, but we had to specify how many bytes to copy, and an int is more than one byte. OK, then—how many bytes does an int take? Answer: depends on the system. But we can tell how many bytes any type takes with the sizeof operator. So there’s the answer: an int takes sizeof(int) bytes of memory to store. Chapter 11. Pointers II: Arithmetic 66
And if we have 3 of them in our array, like we did in that example, the entire space used for the 3 ints must be 3 * sizeof(int). (In the string example, earlier, it would have been more technically accurate to copy 7 * sizeof(char) bytes. But chars are always one byte large, by definition, so that just devolves into 7 * 1.) We could even copy a float or a struct with memcpy()! (Though this is abusive—we should just use = for that): struct antelope my_antelope; struct antelopy my_clone_antelope;
// ... memcpy(&my_clone, &my_antelope, sizeof my_antelope);
Look at how versatile memcpy() is! If you have a pointer to a source and a pointer to a destination, and you have the number of bytes you want to copy, you can copy any type of data. Imagine if we didn’t have void*. We’d have to write specialized memcpy() functions for each type: memcpy_int(int *a, int *b, int count); memcpy_float(float *a, float *b, int count); memcpy_double(double *a, double *b, int count); memcpy_char(char *a, char *b, int count); memcpy_unsigned_char(unsigned char *a, unsigned char *b, int count);
// etc... blech!
Much better to just use void* and have one function that can do it all. That’s the power of void*. You can write functions that don’t care about the type and is still able to do things with it. But with great power comes great responsibility. Maybe not that great in this case, but there are some limits. 1. You cannot do pointer arithmetic on a void*. 2. You cannot dereference a void*. 3. You cannot use the arrow operator on a void*, since it’s also a dereference. 4. You cannot use array notation on a void*, since it’s also a dereference, as well3. And if you think about it, these rules make sense. All those operations rely on knowing the sizeof the type of data pointed to, and with void*, we don’t know the size of the data being pointed to—it could be anything! But wait—if you can’t dereference a void* what good can it ever do you? Like with memcpy(), it helps you write generic functions that can handle multiple types of data. But the secret is that, deep down, you convert the void* to another type before you use it! And conversion is easy: you can just assign into a variable of the desired type4. char a = 'X'; // A single char void *p = &a; // p points to the 'X' char *q = p; // q also points to the 'X' printf("%c\n", *p); // ERROR--cannot dereference void*! printf("%c\n", *q); // Prints "X"
Let’s write our own memcpy() to try this out. We can copy bytes (chars), and we know the number of bytes because it’s passed in.
3Because remember that array notation is just a dereference and some pointer math, and you can’t dereference a void*! 4You can also cast the void* to another type, but we haven’t gotten to casts yet. Chapter 11. Pointers II: Arithmetic 67
void *my_memcpy(void *dest, void *src, int byte_count) { // Convert void*s to char*s char *s = src, *d = dest;
// Now that we have char*s, we can dereference and copy them while (byte_count--) { *d++ = *s++; }
// Most of these functions return the destination, just in case // that's useful to the caller. return dest; }
Right there at the beginning, we copy the void*s into char*s so that we can use them as char*s. It’s as easy as that. Then some fun in a while loop, where we decrement byte_count until it becomes false (0). Remember that with post-decrement, the value of the expression is computed (for while to use) and then the variable is decremented. And some fun in the copy, where we assign *d = *s to copy the byte, but we do it with post-increment so that both d and s move to the next byte after the assignment is made. Lastly, most memory and string functions return a copy of a pointer to the destination string just in case the caller wants to use it. Now that we’ve done that, I just want to quickly point out that we can use this technique to iterate over the bytes of any object in C, floats, structs, or anything! Let’s run one more real-world example with the built-in qsort() routine that can sort anything thanks to the magic of void*s. (In the following example, you can ignore the word const, which we haven’t covered yet.)
1 #include
2 #include
3
4 // The type of structure we're going to sort
5 struct animal {
6 char *name;
7 int leg_count;
8 };
9
10 // This is a comparison function called by qsort() to help it determine
11 // what exactly to sort by. We'll use it to sort an array of struct
12 // animals by leg_count.
13 int compar(const void *elem1, const void *elem2)
14 {
15 // We know we're sorting struct animals, so let's make both
16 // arguments pointers to struct animals
17 const struct animal *animal1 = elem1;
18 const struct animal *animal2 = elem2;
19
20 // Return <0 =0 or >0 depending on whatever we want to sort by.
21
22 // Let's sort ascending by leg_count, so we'll return the difference
23 // in the leg_counts
24 return animal1->leg_count - animal2->leg_count;
25 } Chapter 11. Pointers II: Arithmetic 68
26
27 int main(void)
28 {
29 // Let's build an array of 4 struct animals with different
30 // characteristics. This array is out of order by leg_count, but
31 // we'll sort it in a second.
32 struct animal a[4] = {
33 {.name="Dog", .leg_count=4},
34 {.name="Monkey", .leg_count=2},
35 {.name="Antelope", .leg_count=4},
36 {.name="Snake", .leg_count=0}
37 };
38
39 // Call qsort() to sort the array. qsort() needs to be told exactly
40 // what to sort this data by, and we'll do that inside the compar()
41 // function.
42 //
43 // This call is saying: qsort array a, which has 4 elements, and
44 // each element is sizeof(struct animal) bytes big, and this is the
45 // function that will compare any two elements.
46 qsort(a, 4, sizeof(struct animal), compar);
47
48 // Print them all out
49 for (int i = 0; i < 4; i++) {
50 printf("%d: %s\n", a[i].leg_count, a[i].name);
51 }
52 }
As long as you give qsort() a function that can compare two items that you have in your array to be sorted, it can sort anything. And it does this without needing to have the types of the items hardcoded in there anywhere. qsort() just rearranges blocks of bytes based on the results of the compar() function you passed in. Chapter 12
Manual Memory Allocation
This is one of the big areas where C likely diverges from languages you already know: manual memory management. Other languages uses reference counting, garbage collection, or other means to determine when to allocate new memory for some data—and when to deallocate it when no variables refer to it. And that’s nice. It’s nice to be able to not worry about it, to just drop all the references to an item and trust that at some point the memory associated with it will be freed. But C’s not like that, entirely. Of course, in C, some variables are automatically allocated and deallocated when they come into scope and leave scope. We call these automatic variables. They’re your average run-of-the-mill block scope “local” variables. No problem. But what if you want something to persist longer than a particular block? This is where manual memory management comes into play. You can tell C explicitly to allocate for you a certain number of bytes that you can use as you please. And these bytes will remain allocated until you explicitly free that memory1. It’s important to free the memory you’re done with! If you don’t, we call that a memory leak and your process will continue to reserve that memory until it exits. If you manually allocated it, you have to manually free it when you’re done with it. So how do we do this? We’re going to learn a couple new functions, and make use of the sizeof operator to help us learn how many bytes to allocate. In common C parlance, devs say that automatic local variables are allocated “on the stack”, and manually- allocated memory is “on the heap”. The spec doesn’t talk about either of those things, but all C devs will know what you’re talking about if you bring them up. All functions we’re going to learn in this chapter can be found in
12.1 Allocating and Deallocating, malloc() and free()
The malloc() function accepts a number of bytes to allocate, and returns a void pointer to that block of newly-allocated memory. Since it’s a void*, you can assign it into whatever pointer type you want… normally this will correspond in some way to the number of bytes you’re allocating.
1Or until the program exits, in which case all the memory allocated by it is freed. Asterisk: some systems allow you to allocate memory that persists after a program exits, but it’s system dependent, out of scope for this guide, and you’ll certainly never do it on accident.
69 Chapter 12. Manual Memory Allocation 70
So… how many bytes should I allocate? We can use sizeof to help with that. If we want to allocate enough room for a single int, we can use sizeof(int) and pass that to malloc(). After we’re done with some allocated memory, we can call free() to indicate we’re done with that memory and it can be used for something else. As an argument, you pass the same pointer you got from malloc() (or a copy of it). It’s undefined behavior to use a memory region after you free() it. Let’s try. We’ll allocate enough memory for an int, and then store something there, and the print it. // Allocate space for a single int (sizeof(int) bytes-worth): int *p = malloc(sizeof(int));
*p = 12; // Store something there printf("%d\n", *p); // Print it: 12 free(p); // All done with that memory
//*p = 3490; // ERROR: undefined behavior! Use after free()!
Now, in that contrived example, there’s really no benefit to it. We could have just used an automatic int and it would have worked. But we’ll see how the ability to allocate memory this way has its advantages, especially with more complex data structures. One more thing you’ll commonly see takes advantage of the fact that sizeof can give you the size of the result type of any constant expression. So you could put a variable name in there, too, and use that. Here’s an example of that, just like the previous one: int *p = malloc(sizeof *p); // *p is an int, so same as sizeof(int)
12.2 Error Checking
All the allocation functions return a pointer to the newly-allocated stretch of memory, or NULL if the memory cannot be allocated for some reason. Some OSes like Linux can be configured in such a way that malloc() never returns NULL, even if you’re out of memory. But despite this, you should always code it up with protections in mind. int *x; x = malloc(sizeof(int) * 10); if (x == NULL) { printf("Error allocating 10 ints\n"); // do something here to handle it }
Here’s a common pattern that you’ll see, where we do the assignment and the condition on the same line: int *x; if ((x = malloc(sizeof(int) * 10)) == NULL) printf("Error allocating 10 ints\n"); // do something here to handle it }
12.3 Allocating Space for an Array
We’ve seen how to allocate space for a single thing; now what about for a bunch of them in an array? In C, an array is a bunch of the same thing back-to-back in a contiguous stretch of memory. Chapter 12. Manual Memory Allocation 71
We can allocate a contiguous stretch of memory—we’ve seen how to do that. If we wanted 3490 bytes of memory, we could just ask for it: char *p = malloc(3490); // Voila
And—indeed!—that’s an array of 3490 chars (AKA a string!) since each char is 1 byte. In other words, sizeof(char) is 1. Note: there’s no initialization done on the newly-allocated memory—it’s full of garbage. Clear it with memset() if you want to, or see calloc(), below. But we can just multiply the size of the thing we want by the number of elements we want, and then access them using either pointer or array notation. Example!
1 #include
2 #include
3
4 int main(void)
5 {
6 // Allocate space for 10 ints
7 int *p = malloc(sizeof(int) * 10);
8
9 // Assign them values 0-45:
10 for (int i = 0; i < 10; i++)
11 p[i] = i * 5;
12
13 // Print all values 0, 5, 10, 15, ..., 40, 45
14 for (int i = 0; i < 10; i++)
15 printf("%d\n", p[i]);
16
17 // Free the space
18 free(p);
19 }
The key’s in that malloc() line. If we know each int takes sizeof(int) bytes to hold it, and we know we want 10 of them, we can just allocate exactly that many bytes with: sizeof(int) * 10
And this trick works for every type. Just pass it to sizeof and multiply by the size of the array.
12.4 An Alternative: calloc()
This is another allocation function that works similarly to malloc(), with two key differences: • Instead of a single argument, you pass the size of one element, and the number of elements you wish to allocate. It’s like it’s made for allocating arrays. • It clears the memory to zero. You still use free() to deallocate memory obtained through calloc(). Here’s a comparison of calloc() and malloc(). // Allocate space for 10 ints with calloc(), initialized to 0: int *p = calloc(sizeof(int), 10);
// Allocate space for 10 ints with malloc(), initialized to 0: int *q = malloc(sizeof(int) * 10); memset(q, 0, sizeof(int) * 10); // set to 0
Again, the result is the same for both except malloc() doesn’t zero the memory by default. Chapter 12. Manual Memory Allocation 72
12.5 Changing Allocated Size with realloc()
If you’ve already allocated 10 ints, but later you decide you need 20, what can you do? One option is to allocate some new space, and then memcpy() the memory over… but it turns out that sometimes you don’t need to move anything. And there’s one function that’s just smart enough to do the right thing in all the right circumstances: realloc(). It takes a pointer to some previously-allocted memory (by malloc() or calloc()) and a new size for the memory region to be. It then grows or shrinks that memory, and returns a pointer to it. Sometimes it might return the same pointer (if the data didn’t have to be copied elsewhere), or it might return a different one (if the data did have to be copied). Be sure when you call realloc(), you specify the number of bytes to allocate, and not just the number of array elements! That is: num_floats *= 2;
np = realloc(p, num_floats); // WRONG: need bytes, not number of elements!
np = realloc(p, num_floats * sizeof(float)); // Better!
Let’s allocate an array of 20 floats, and then change our mind and make it an array of 40. We’re going to assign the return value of realloc() into another pointer just to make sure it’s not NULL. If it’s not, then we can reassign it into our original pointer. (If we just assigned the return value directly into the original pointer, we’d lose that pointer if the function returned NULL and we’d have no way to get it back.)
1 #include
2 #include
3
4 int main(void)
5 {
6 // Allocate space for 20 floats
7 float *p = malloc(sizeof *p * 20); // sizeof *p same as sizeof(float)
8
9 // Assign them fractional values 0.0-1.0:
10 for (int i = 0; i < 20; i++)
11 p[i] = i / 20.0;
12
13 // But wait! Let's actually make this an array of 40 elements
14 float *new_p = realloc(p, sizeof *p * 40);
15
16 // Check to see if we successfully reallocated
17 if (new_p == NULL) {
18 printf("Error reallocing\n");
19 return 1;
20 }
21
22 // If we did, we can just reassign p
23 p = new_p;
24
25 // And assign the new elements values in the range 1.0-2.0
26 for (int i = 20; i < 40; i++)
27 p[i] = 1.0 + (i - 20) / 20.0;
28
29 // Print all values 0.0-2.0 in the 40 elements:
30 for (int i = 0; i < 40; i++)
31 printf("%f\n", p[i]); Chapter 12. Manual Memory Allocation 73
32
33 // Free the space
34 free(p);
35 }
Notice in there how we took the return value from realloc() and reassigned it into the same pointer variable p that we passed in. That’s pretty common to do. Also if line 7 is looking weird, with that sizeof *p in there, remember that sizeof works on the size of the type of the expression. And the type of *p is float, so that line is equivalent to sizeof(float).
12.5.1 Reading in Lines of Arbitrary Length I want to demonstrate two things with this full-blown example. 1. Use of realloc() to grow a buffer as we read in more data. 2. Use of realloc() to shrink the buffer down to the perfect size after we’ve completed the read. What we see here is a loop that calls fgetc() over and over to append to a buffer until we see that the last character is a newline. Once it finds the newline, it shrinks the buffer to just the right size and returns it.
1 #include
2 #include
3
4 // Read a line of arbitrary size from a file
5 //
6 // Returns a pointer to the line.
7 // Returns NULL on EOF or error.
8 //
9 // It's up to the caller to free() this pointer when done with it.
10 //
11 // Note that this strips the newline from the result. If you need
12 // it in there, probably best to switch this to a do-while.
13
14 char *readline(FILE *fp)
15 {
16 int offset = 0; // Index next char goes in the buffer
17 int bufsize = 4; // Preferably power of 2 initial size
18 char *buf; // The buffer
19 int c; // The character we've read in
20
21 buf = malloc(bufsize); // Allocate initial buffer
22
23 if (buf == NULL) // Error check
24 return NULL;
25
26 // Main loop--read until newline or EOF
27 while (c = fgetc(fp), c != '\n' && c != EOF) {
28
29 // Check if we're out of room in the buffer accounting
30 // for the extra byte for the NUL terminator
31 if (offset == bufsize - 1) { // -1 for the NUL terminator
32 bufsize *= 2; // 2x the space
33
34 char *new_buf = realloc(buf, bufsize);
35
36 if (new_buf == NULL) {
37 free(buf); // On error, free and bail
38 return NULL; Chapter 12. Manual Memory Allocation 74
39 }
40
41 buf = new_buf; // Successful realloc
42 }
43
44 buf[offset++] = c; // Add the byte onto the buffer
45 }
46
47 // We hit newline or EOF...
48
49 // If at EOF and we read no bytes, free the buffer and
50 // return NULL to indicate we're at EOF:
51 if (c == EOF && offset == 0) {
52 free(buf);
53 return NULL;
54 }
55
56 // Shrink to fit
57 if (offset < bufsize - 1) { // If we're short of the end
58 char *new_buf = realloc(buf, offset + 1); // +1 for NUL terminator
59
60 // If successful, point buf to new_buf;
61 // otherwise we'll just leave buf where it is
62 if (new_buf != NULL)
63 buf = new_buf;
64 }
65
66 // Add the NUL terminator
67 buf[offset] = '\0';
68
69 return buf;
70 }
71
72 int main(void)
73 {
74 FILE *fp = fopen("foo.txt", "r");
75
76 char *line;
77
78 while ((line = readline(fp)) != NULL) {
79 printf("%s\n", line);
80 free(line);
81 }
82
83 fclose(fp);
84 }
When growing memory like this, it’s common (though hardly a law) to double the space needed each step just to minimize the number of realloc()s that occur. Finally you might note that readline() returns a pointer to a malloc()d buffer. As such, it’s up to the caller to explicitly free() that memory when it’s done with it.
12.5.2 realloc() with NULL Trivia time! These two lines are equivalent: char *p = malloc(3490); char *p = realloc(NULL, 3490); Chapter 12. Manual Memory Allocation 75
That could be convenient if you have some kind of allocation loop and you don’t want to special-case the first malloc(). int *p = NULL; int length = 0;
while (!done) { // Allocate 10 more ints: length += 10; p = realloc(p, sizeof *p * length);
// Do amazing things // ... }
In that example, we didn’t need an initial malloc() since p was NULL to start.
12.6 Aligned Allocations
You probably aren’t going to need to use this. And I don’t want to get too far off in the weeds talking about it right now, but there’s this thing called memory alignment, which has to do with the memory address (pointer value) being a multiple of a certain number. For example, a system might require that 16-bit values begin on memory addresses that are multiples of 2. Or that 64-bit values begin on memory addresses that are multiples of 2, 4, or 8, for example. It depends on the CPU. Some systems require this kind of alignment for fast memory access, or some even for memory access at all. Now, if you use malloc(), calloc(), or realloc(), C will give you a chunk of memory that’s well- aligned for any value at all, even structs. Works in all cases. But there might be times that you know that some data can be aligned at a smaller boundary, or must be aligned at a larger one for some reason. I imagine this is more common with embedded systems programming. In those cases, you can specify an alignment with aligned_alloc(). The alignment is an integer power of two greater than zero, so 2, 4, 8, 16, etc. and you give that to aligned_alloc() before the number of bytes you’re interested in. The other restriction is that the number of bytes you allocate needs to be a multiple of the alignment. But this might be changing. See C Defect Report 4602 Let’s do an example, allocating on a 64-byte boundary:
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 // Allocate 256 bytes aligned on a 64-byte boundary
8 char *p = aligned_alloc(64, 256); // 256 == 64 * 4
9
10 // Copy a string in there and print it
11 strcpy(p, "Hello, world!");
12 printf("%s\n", p);
13
2http://www.open-std.org/jtc1/sc22/wg14/www/docs/summary.htm#dr_460 Chapter 12. Manual Memory Allocation 76
14 // Free the space
15 free(p);
16 }
I want to throw a note here about realloc() and aligned_alloc(). realloc() doesn’t have any alignment guarantees, so if you need to get some aligned reallocated space, you’ll have to do it the hard way with memcpy(). Here’s a non-standard aligned_realloc() function, if you need it: void *aligned_realloc(void *ptr, size_t old_size, size_t alignment, size_t size) { char *new_ptr = aligned_alloc(alignment, size);
if (new_ptr == NULL) return NULL;
size_t copy_size = old_size < size? old_size: size; // get min
if (ptr != NULL) memcpy(new_ptr, ptr, copy_size);
free(ptr);
return new_ptr; }
Note that it always copies data, taking time, while real realloc() will avoid that if it can. So this is hardly efficient. Avoid needing to reallocate custom-aligned data. Chapter 13
Scope
Scope is all about what variables are visible in what contexts.
13.1 Block Scope
This is the scope of almost all the variables devs define. It includes what other languages might call “function scope”, i.e. variables that are declared inside functions. The basic rule is that if you’ve declared a variable in a block delimited by squirrelly braces, the scope of that variable is that block. If there’s a block inside a block, then variables declared in the inner block are local to that block, and cannot be seen in the outer scope. Once a variable’s scope ends, that variable can no longer be referenced, and you can consider its value to be gone into the great bit bucket1 in the sky. An example with nested scope:
1 int main(void)
2 {
3 int a = 12; // Local to outer block, but visible in inner block
4
5 if (a == 12) {
6 int b = 99; // Local to inner block, not visible in outer block
7
8 printf("%d %d\n", a, b); // OK: "12 99"
9 }
10
11 printf("%d\n", a); // OK, we're still in a's scope
12
13 printf("%d\n", b); // ILLEGAL, out of b's scope
14 }
13.1.1 Where To Define Variables Another fun fact is that you can define variables anywhere in the block, within reason—they have the scope of that block, but cannot be used before they are defined.
1 #include
2
3 int main(void)
4 {
1https://en.wikipedia.org/wiki/Bit_bucket
77 Chapter 13. Scope 78
5 int i = 0;
6
7 printf("%d\n", i); // OK: "0"
8
9 //printf("%d\n", j); // ILLEGAL--can't use j before it's defined
10
11 int j = 5;
12
13 printf("%d %d\n", i, j); // OK: "0 5"
14 }
Historically, C required all the variables be defined before any code in the block, but this is no longer the case in the C99 standard.
13.1.2 Variable Hiding If you have a variable named the same thing at an inner scope as one at an outer scope, the one at the inner scope takes precedence at long as you’re running in the inner scope. That is, it hides the one at outer scope for the duration of its lifetime.
1 #include
2
3 int main(void)
4 {
5 int i = 10;
6
7 {
8 int i = 20;
9
10 printf("%d\n", i); // Inner scope i, 20 (outer i is hidden)
11 }
12
13 printf("%d\n", i); // Outer scope i, 10
14 }
You might have noticed in that example that I just threw a block in there at line 7, not so much as a for or if statement to kick it off! This is perfectly legal. Sometimes a dev will want to group a bunch of local variables together for a quick computation and will do this, but it’s rare to see.
13.2 File Scope
If you define a variable outside of a block, that variable has file scope. It’s visible in all functions in the file that come after it, and shared between them. (An exception is if a block defines a variable of the same name, it would hide the one at file scope.) This is closest to what you would consider to be “global” scope in another language. For example:
1 #include
2
3 int shared = 10; // File scope! Visible to the whole file after this!
4
5 void func1(void)
6 {
7 shared += 100; // Now shared holds 110
8 }
9
10 void func2(void)
11 { Chapter 13. Scope 79
12 printf("%d\n", shared); // Prints "110"
13 }
14
15 int main(void)
16 {
17 func1();
18 func2();
19 }
Note that if shared were declared at the bottom of the file, it wouldn’t compile. It has to be declared before any functions use it. There are ways to further modify items at file scope, namely with (static)[#static] and (extern)[#extern], but we’ll talk more about those later.
13.3 for-loop Scope
I really don’t know what to call this, as C11 §6.8.5.3¶1 doesn’t give it a proper name. We’ve done it already a few times in this guide, as well. It’s when you declare a variable inside the first clause of a for-loop: for (int i = 0; i < 10; i++) printf("%d\n", i);
printf("%d\n", i); // ILLEGAL--i is only in scope for the for-loop
In that example, i’s lifetime begins the moment it is defined, and continues for the duration of the loop. If the loop body is enclosed in a block, the variables defined in the for-loop are visible from that inner scope. Unless, of course, that inner scope hides them. This crazy example prints 999 five times:
1 #include
2
3 int main(void)
4 {
5 for (int i = 0; i < 5; i++) {
6 int i = 999; // Hides the i in the for-loop scope
7 printf("%d\n", i);
8 }
9 }
13.4 A Note on Function Scope
The C spec does refer to function scope, but it’s used exclusively with labels, something we haven’t discussed yet. More on that another day. Chapter 14
Types II: Way More Types!
We’re used to char, int, and float types, but it’s now time to take that stuff to the next level and see what else we have out there in the types department!
14.1 Signed and Unsigned Integers
So far we’ve used int as a signed type, that is, a value that can be either negative or positive. But C also has specific unsigned integer types that can only hold positive numbers. These types are prefaced by the keyword unsigned. int a; // signed signed int a; // signed signed a; // signed, "shorthand" for "int" or "signed int", rare unsigned int b; // unsigned unsigned c; // unsigned, shorthand for "unsigned int"
Why? Why would you decide you only wanted to hold positive numbers? Answer: you can get larger numbers in an unsigned variable than you can in a signed ones. But why is that? You can think of integers being represented by a certain number of bits1. On my computer, an int is represented by 64 bits. And each permutation of bits that are either 1 or 0 represents a number. We can decide how to divvy up these numbers. With signed numbers, we use (roughly) half the permutations to represent negative numbers, and the other half to represent positive numbers. With unsigned, we use all the permutations to represent positive numbers. On my computer with 64-bit ints using two’s complement2 to represent unsigned numbers, I have the following limits on integer range:
Type Minimum Maximum int -9,223,372,036,854,775,808 9,223,372,036,854,775,807 unsigned int 0 18,446,744,073,709,551,615
Notice that the largest positive unsigned int is approximately twice as large as the largest positive int.
1“Bit” is short for binary digit. Binary is just another way of representing numbers. Instead of digits 0-9 like we’re used to, it’s digits 0-1. 2https://en.wikipedia.org/wiki/Two%27s_complement
80 Chapter 14. Types II: Way More Types! 81
So you can get some flexibility there.
14.2 Character Types
Remember char? The type we can use to hold a single character? char c = 'B'; printf("%c\n", c); // "B"
I have a shocker for you: it’s actually an integer. char c = 'B';
// Change this from %c to %d: printf("%d\n", c); // 66 (!!)
Deep down, char is just a small int, namely an integer that uses just a single byte of space, limiting its range to… Here the C spec gets just a little funky. It assures us that a char is a single byte, i.e. sizeof(char) == 1. But then in C11 §3.6¶3 it goes out of its way to say: A byte is composed of a contiguous sequence of bits, the number of which is implementation- defined. Wait—what? Some of you might be used to the notion that a byte is 8 bits, right? I mean, that’s what it is, right? And the answer is, “Almost certainly.”3 But C is an old language, and machines back in the day had, shall we say, a more relaxed opinion over how many bits were in a byte. And through the years, C has retained this flexibility. But assuming your bytes in C are 8 bits, like they are for virtually all machines in the world that you’ll ever see, the range of a char is… —So before I can tell you, it turns out that chars might be signed or unsigned depending on your compiler. Unless you explicitly specify. In many cases, just having char is fine because you don’t care about the sign of the data. But if you need signed or unsigned chars, you must be specific: char a; // Could be signed or unsigned signed char b; // Definitely signed unsigned char c; // Definitely unsigned
OK, now, finally, we can figure out the range of numbers if we assume that a char is 8 bits and your system uses the virtually universal two’s complement representation for signed and unsigned4. So, assuming those constraints, we can finally figure our ranges:
char type Minimum Maximum signed char -128 127 unsigned char 0 255
And the ranges for char are implementation-defined. Let me get this straight. char is actually a number, so can we do math on it? Yup! Just remember to keep things in the range of a char!
3The industry term for a sequence of exactly, indisputably 8 bits is an octet. 4In general, f you have an 푛 bit two’s complement number, the signed range is −2푛−1 to 2푛−1 − 1. And the unsigned range is 0 to 2푛 − 1. Chapter 14. Types II: Way More Types! 82
1 #include
2
3 int main(void)
4 {
5 char a = 10, b = 20;
6
7 printf("%d\n", a + b); // 30!
8 }
What about those constant characters in single quotes, like 'B'? How does that have a numeric value? The spec is also hand-wavey here, since C isn’t designed to run on a single type of underlying system. But let’s just assume for the moment that your character set is based on ASCII5 for at least the first 128 characters. In that case, the character constant will be converted to a char whose value is the same as the ASCII value of the character. That was a mouthful. Let’s just have an example:
1 #include
2
3 int main(void)
4 {
5 char a = 10;
6 char b = 'B'; // ASCII value 66
7
8 printf("%d\n", a + b); // 76!
9 }
This depends on your execution environment and the character set used6. One of the most popular charac- ter sets today is Unicode7 (which is a superset of ASCII), so for your basic 0-9, A-Z, a-z and punctuation, you’ll almost certainly get the ASCII values out of them.
14.3 More Integer Types: short, long, long long
So far we’ve just generally been using two integer types: • char • int and we recently learned about the unsigned variants of the integer types. And we learned that char was secretly a small int in disguise. So we know the ints can come in multiple bit sizes. But there are a couple more integer types we should look at, and the minimum minimum and maximum values they can hold. Yes, I said “minimum” twice. The spec says that these types will hold numbers of at least these sizes, so your implementation might be different. The header file
// And so are these:
5https://en.wikipedia.org/wiki/ASCII 6https://en.wikipedia.org/wiki/List_of_information_system_character_sets 7https://en.wikipedia.org/wiki/Unicode Chapter 14. Types II: Way More Types! 83 short int x; short x;
Let’s take a look at the integer data types and sizes in ascending order, grouped by signedness.
Type Minimum Bytes Minimum Value Maximum Value char 1 -127 or 0 127 or 2558 signed char 1 -127 127 short 2 -32767 32767 int 2 -32767 32767 long 4 -2147483647 2147483647 long long 8 -9223372036854775807 9223372036854775807 unsigned char 1 0 255 unsigned short 2 0 65535 unsigned int 2 0 65535 unsigned long 4 0 44294967295 unsigned long long 8 0 9223372036854775807
There is no long long long type. You can’t just keep adding longs like that. Don’t be silly. Two’s complement fans might have noticed something funny about those numbers. Why does, for example, the signed char stop at -127 instead of -128? Remember: these are only the minimums required by the spec. Some number representations (like sign and magnitude9) top off at ±127. Let’s run the same table on my 64-bit, two’s complement system and see what comes out:
Type My Bytes Minimum Value Maximum Value char 1 -128 12710 signed char 1 -128 127 short 2 -32768 32767 int 4 -2147483648 2147483647 long 8 -9223372036854775808 9223372036854775807 long long 8 -9223372036854775808 9223372036854775807 unsigned char 1 0 255 unsigned short 2 0 65535 unsigned int 4 0 4294967295 unsigned long 8 0 18446744073709551615 unsigned long long 8 0 18446744073709551615
That’s a little more sensible, but we can see how my system has larger limits than the minimums in the specification. So what are the macros in
Type Min Macro Max Macro char CHAR_MIN CHAR_MAX signed char SCHAR_MIN SCHAR_MAX short SHRT_MIN SHRT_MAX int INT_MIN INT_MAX long LONG_MIN LONG_MAX long long LLONG_MIN LLONG_MAX unsigned char 0 UCHAR_MAX
8Depends on if a char defaults to signed char or unsigned char 9https://en.wikipedia.org/wiki/Signed_number_representations#Signed_magnitude_representation 10My char is signed. Chapter 14. Types II: Way More Types! 84
Type Min Macro Max Macro unsigned short 0 USHRT_MAX unsigned int 0 UINT_MAX unsigned long 0 ULONG_MAX unsigned long long 0 ULLONG_MAX
Notice there’s a way hidden in there to determine if a system uses signed or unsigned chars. If CHAR_MAX == UCHAR_MAX, it must be unsigned. Also notice there’s no minimum macro for the unsigned variants—they’re just 0.
14.4 More Float: double and long double
Let’s see what the spec has to say about floating point numbers in §5.2.4.2.2¶1-2: The following parameters are used to define the model for each floating-point type:
Parameter Definition 푠 sign (±1) 푏 base or radix of exponent representation (an integer > 1) 푒 exponent (an integer between a minimum 푒푚푖푛 and a maximum 푒푚푎푥) 푝 precision (the number of base-푏 digits in the significand) 푓푘 nonnegative integers less than 푏 (the significand digits)
A floating-point number (푥) is defined by the following model: 푝 푒 −푘 푥 = 푠푏 ∑ 푓푘푏 , 푒푚푖푛 ≤ 푒 ≤ 푒푚푎푥 푘=1 I hope that cleared it right up for you. Okay, fine. Let’s step back a bit and see what’s practical. Note: we refer to a bunch of macros in this section. They can be found in the header
11https://en.wikipedia.org/wiki/IEEE_754 Chapter 14. Types II: Way More Types! 85
But more particularly, only a certain number of significant decimal digits can be represented accurately. How can you get more? You can use larger data types! And we have a couple of them. We know about float already, but for more precision we have double. And for even more precision, we have long double (unrelated to long int except by name). The spec doesn’t go into how many bytes of storage each type should take, but on my system, we can see the relative size increases:
Type sizeof float 4 double 8 long double 16
So each of the types (on my system) uses those additional bits for more precision. But how much precision are we talking, here? How many decimal numbers can be represented by these values? Well, C provides us with a bunch of macros in
14.4.1 How Many Decimal Digits? The million dollar question is, “How many significant decimal digits can I store in a given floating point type so that I get out the same decimal number when I print it?” The number of decimal digits you can store in a floating point type and surely get the same number back out when you print it is given by these macros:
Type Decimal Digits You Can Store Minimum float FLT_DIG 6 double DBL_DIG 10 long double LDBL_DIG 10
On my system, FLT_DIG is 6, so I can be sure that if I print out a 6 digit float, I’ll get the same thing back. (It could be more digits—some numbers will come back correctly with more digits. But 6 is definitely coming back.) For example, printing out floats following this pattern of increasing digits, we apparently make it to 8 digits before something goes wrong, but after that we’re back to 7 correct digits. 0.12345 0.123456 0.1234567 0.12345678 0.123456791 <-- Things start going wrong 0.1234567910
Let’s do another demo. In this code we’ll have two floats that both hold numbers that have FLT_DIG significant decimal digits12. Then we add those together, for what should be 12 significant decimal digits. But that’s more than we can store in a float and correctly recover as a string—so we see when we print it out, things start going wrong after the 7th significant digit.
12This program runs as its comments indicate on a system with FLT_DIG of 6 that uses IEEE-754 base-2 floating point numbers. Otherwise, you might get different output. Chapter 14. Types II: Way More Types! 86
1 #include
2 #include
3
4 int main(void)
5 {
6 // Both these numbers have 6 significant digits, so they can be
7 // stored accurately in a float:
8
9 float f = 3.14159f;
10 float g = 0.00000265358f;
11
12 printf("%.5f\n", f); // 3.14159 -- correct!
13 printf("%.11f\n", g); // 0.00000265358 -- correct!
14
15 // Now add them up
16 f += g; // 3.14159265358 is what f _should_ be
17
18 printf("%.11f\n", f); // 3.14159274101 -- wrong!
19 }
(The above code has an f after the numeric constants—this indicates that the constant is type float, as opposed to the default of double. More on this later.) Remember that FLT_DIG is the safe number of digits you can store in a float and retrieve correctly. Sometimes you might get one or two more out of it. But sometimes you’ll only get FLT_DIG digits back. The sure thing: if you store any number of digits up to and including FLT_DIG in a float, you’re sure to get them back correctly. So that’s the story. FLT_DIG. The End. …Or is it?
14.4.2 Converting to Decimal and Back But storing a base 10 number in a floating point number and getting it back out is only half the story. Turns out floating point numbers can encode numbers that require more decimal places to print out com- pletely. It’s just that your big decimal number might not map to one of those numbers. That is, when you look at floating point numbers from one to the next, there’s a gap. If you try to encode a decimal number in that gap, it’ll use the closest floating point number. That’s why you can only encode FLT_DIG for a float. But what about those floating point numbers that aren’t in the gap? How many places do you need to print those out accurately? Another way to phrase this question is for any given floating point number, how many decimal digits do I have to preserve if I want to convert the decimal number back into an identical floating point number? That is, how many digits do I have to print in base 10 to recover all the digits in base 2 in the original number? Sometimes it might only be a few. But to be sure, you’ll want to convert to decimal with a certain safe number of decimal places. That number is encoded in the following macros:
Macro Description FLT_DECIMAL_DIG Number of decimal digits encoded in a float. DBL_DECIMAL_DIG Number of decimal digits encoded in a double. LDBL_DECIMAL_DIG Number of decimal digits encoded in a long double. DECIMAL_DIG Same as the widest encoding, LDBL_DECIMAL_DIG. Chapter 14. Types II: Way More Types! 87
Let’s see an example where DBL_DIG is 15 (so that’s all we can have in a constant), but DBL_DECIMAL_DIG is 17 (so we have to convert to 17 decimal numbers to preserve all the bits of the original double). Let’s assign the 15 significant digit number 0.123456789012345 to x, and let’s assign the 1 significant digit number 0.0000000000000006 to y. x is exact: 0.12345678901234500 Printed to 17 decimal places y is exact: 0.00000000000000060
But let’s add them together. This should give 0.1234567890123456, but that’s more than DBL_DIG, so strange things might happen… let’s look: x + y not quite right: 0.12345678901234559 Should end in 4560!
That’s what we get for printing more than DBL_DIG, right? But check this out… that number, above, is exactly representable as it is! If we assign 0.12345678901234559 (17 digits) to z and print it, we get: z is exact: 0.12345678901234559 17 digits correct! More than DBL_DIG!
If we’d truncated z down to 15 digits, it wouldn’t have been the same number. That’s why to preserve all the bits of a double, we need DBL_DECIMAL_DIG and not just the lesser DBL_DIG. All that being said, it’s clear that when we’re messing with decimal numbers in general, it’s not safe to print more than FLT_DIG, DBL_DIG, or LDBL_DIG digits to be sensible in relation to the original base 10 numbers and any subsequent math. But when converting from float to a decimal representation and back to float, definitely use FLT_DECIMAL_DIG to do that so that all the bits are preserved exactly.
14.5 Constant Numeric Types
When you write down a constant number, like 1234, it has a type. But what type is it? Let’s look at the how C decides what type the constant is, and how to force it to choose a specific type.
14.5.1 Hexadecimal and Octal In addition to good ol’ decimal like Grandma used to bake, C also supports constants of different bases. If you lead a number with 0x, it is read as a hex number: int a = 0x1A2B; // Hexadecimal int b = 0x1a2b; // Case doesn't matter for hex digits printf("%x", a); // Print a hex number, "1a2b"
If you lead a number with a 0, it is read as an octal number: int a = 012; printf("%o\n", a); // Print an octal number, "12"
This is particularly problematic for beginner programmers who try to pad decimal numbers on the left with 0 to line things up nice and pretty, inadvertently changing the base of the number: int x = 11111; // Decimal 11111 int y = 00111; // Decimal 73 (Octal 111) int z = 01111; // Decimal 585 (Octal 1111)
14.5.1.1 A Note on Binary
An unofficial extension13 in many C compilers allows you to represent a binary number with a 0b prefix:
13It’s really surprising to me that C doesn’t have this in the spec yet. In the C99 Rationale document, they write, “A proposal to add binary constants was rejected due to lack of precedent and insufficient utility.” Which seems kind of silly in light of some of Chapter 14. Types II: Way More Types! 88 int x = 0b101010; // Binary 101010 printf("%d\n", x); // Prints 42 decimal
There’s no printf() format specifier for printing a binary number. You have to do it a character at a time with bitwise operators.
14.5.2 Integer Constants You can force a constant integer to be a certain type by appending a suffix to it that indicates the type. We’ll do some assignments to demo, but most often devs leave off the suffixes unless needed to be precise. The compiler is pretty good at making sure the types are compatible. int x = 1234; long int x = 1234L; long long int x = 1234LL unsigned int x = 1234U; unsigned long int x = 1234UL; unsigned long long int x = 1234ULL;
The suffix can be uppercase or lowercase. And the U and L or LL can appear either one first.
Type Suffix int None long int L long long int LL unsigned int U unsigned long int UL unsigned long long int ULL
I mentioned in the table that “no suffix” means int… but it’s actually more complex than that. So what happens when you have an unsuffixed number like: int x = 1234;
What type is it? What C will generally do is choose the smallest type from int up that can hold the value. But specifically, that depends on the number’s base (decimal, hex, or octal), as well. The spec has a great table indicating which type gets used for what unsuffixed value. In fact, I’m just going to copy it wholesale right here. C11 §6.4.4.1¶5 reads, “The type of an integer constant is the first of the first of the corresponding list in which its value can be represented.” And then goes on to show this table:
the other features they kitchen-sinked in there! I’ll bet one of the next releases has it. Chapter 14. Types II: Way More Types! 89
Octal or Hexadecimal Suffix Decimal Constant Constant none int int long int unsigned int long int unsigned long int long long int unsigned long long int
u or U unsigned int unsigned int unsigned long int unsigned long int unsigned long long int unsigned long long int
l or L long int long int long long int unsigned long int long long int unsigned long long int
Both u or U unsigned long int unsigned long int and l or L unsigned long long int unsigned long long int
ll or LL long long int long long int unsigned long long int
Both u or U unsigned long long int unsigned long long int and ll or LL
What that’s saying is that, for example, if you specify a number like 123456789U, first C will see if it can be unsigned int. If it doesn’t fit there, it’ll try unsigned long int. And then unsigned long long int. It’ll use the smallest type that can hold the number.
14.5.3 Floating Point Constants
You’d think that a floating point constant like 1.23 would have a default type of float, right? Surprise! Turns out unsuffiexed floating point numbers are type double! Happy belated birthday! You can force it to be of type float by appending an f (or F—it’s case-insensitive). You can force it to be of type long double by appending l (or L).
Type Suffix float F double None long double L
For example: float x = 3.14f; double x = 3.14; long double x = 3.14L;
This whole time, though, we’ve just been doing this, right? float x = 3.14;
Isn’t the left a float and the right a double? Yes! But C’s pretty good with automatic numeric conver- sions, so it’s more common to have an unsuffixed floating point constant than not. More on that later. Chapter 14. Types II: Way More Types! 90
14.5.3.1 Scientific Notation Remember earlier when we talked about how a floating point number can be represented by a significand, base, and exponent? Well, there’s a common way of writing such a number, shown here followed by it’s more recognizable equivalent which is what you get when you actually run the math: 1.2345 × 103 = 1234.5 Writing numbers in the form 푠×푏푒 is called scientific notation14. In C, these are written using “E notation”, so these are equivalent:
Scientific Notation E notation 1.2345 × 10−3 = 12.345 1.2345e-3 1.2345 × 104 = 123450000 1.2345e+4
You can print a number in this notation with %e: printf("%e\n", 123456.0); // Prints 1.234560e+05
A couple little fun facts about scientific notation: • You don’t have to write them with a single leading digit before the decimal point. Any number of numbers can go in front. double x = 123.456e+3; // 123456
However, when you print it, it will change the exponent so there is only one digit in front of the decimal point. • The plus can be left off the exponent, as it’s default, but this is uncommon in practice from what I’ve seen. 1.2345e10 == 1.2345e+10
• You can apply the F or L suffixes to E-notation constants: 1.2345e10F 1.2345e10L
14.5.3.2 Hexadecimal Floating Point Constants But wait, there’s more floating to be done! Turns out there are hexadecimal floating point constants, as well! These work similar to decimal floating point numbers, but they begin with a 0x just like integer numbers. The catch is that you must specify an exponent, and this exponent produces a power of 2. That is: 2푥. And then you use a p instead of an e when writing the number: So 0xa.1p3 is 10.0625 × 23 == 80.5. When using floating point hex constants, We can print hex scientific notation with %a: double x = 0xa.1p3; printf("%a\n", x); // 0x1.42p+6 printf("%f\n", x); // 80.500000
14https://en.wikipedia.org/wiki/Scientific_notation Chapter 15
Types III: Conversions
In this chapter, we want to talk all about converting from one type to another. C has a variety of ways of doing this, and some might be a little different that you’re used to in other languages. Before we talk about how to make conversions happen, let’s talk about how they work when they do happen.
15.1 String Conversions
Unlike many languages, C doesn’t do string-to-number (and vice-versa) conversions in quite as stream- lined a manner as it does numeric conversions. For these, we’ll have to call functions to do the dirty work.
15.1.1 Numeric Value to String
When we want to convert a number to a string, we can use either sprintf() (pronounced SPRINT-f) or snprintf() (s-n-print-f)1 These basically work like printf(), except they output to a string instead, and you can print that string later, or whatever. For example, turning part of the value π into a string:
1 #include
2
3 int main(void)
4 {
5 char s[10];
6 float f = 3.14159;
7
8 // Convert "f" to string, storing in "s", writing at most 10 characters
9 // including the NUL terminator
10
11 snprintf(s, 10, "%f", f);
12
13 printf("String value: %s\n", s); // String value: 3.141590
14 }
If we wanted to convert a double, we’d use %lf. Or a long double, %Lf.
1They’re the same except snprintf() allows you to specify a maximum number of bytes to output, preventing the overrunning of the end of your string. So it’s safer.
91 Chapter 15. Types III: Conversions 92
15.1.2 String to Numeric Value
There are a couple families of functions to do this in C. We’ll call these the atoi (pronounced a-to-i) family and the strtol (stir-to-long) family. For basic conversion from a string to a number, try the atoi functions from
Function Description atoi String to int atof String to float atol String to long int atoll String to long long int
Though the spec doesn’t cop to it, the a at the beginning of the function stands for ASCII2, so really atoi() is “ASCII-to-integer”, but saying so today is a bit ASCII-centric. Here’s an example converting a string to a float:
1 #include
2 #include
3
4 int main(void)
5 {
6 char *pi = "3.14159";
7 float f;
8
9 f = atof(pi);
10
11 printf("%f\n", f);
12 }
But, like I said, we get undefined behavior from weird things like this: int x = atoi("what"); // "What" ain't no number I ever heard of
(When I run that, I get 0 back, but you really shouldn’t count on that in any way. You could get something completely different.) For better error handling characteristics, let’s check out all those strtol functions, also in
Function Description strtol String to long int strtoll String to long long int strtoul String to unsigned long int strtoull String to unsigned long long int strtof String to float strtod String to double strtold String to long double
These functions all follow a similar pattern of use, and are a lot of people’s first experience with pointers to pointers! But never fret—it’s easier than it looks. Let’s do an example where we convert a string to an unsigned long, discarding error information (i.e. in- formation about bad characters in the input string):
2https://en.wikipedia.org/wiki/ASCII Chapter 15. Types III: Conversions 93
1 #include
2 #include
3
4 int main(void)
5 {
6 char *s = "3490";
7
8 // Convert string s, a number in base 10, to an unsigned long int.
9 // NULL means we don't care to learn about any error information.
10
11 unsigned long int x = strtoul(s, NULL, 10);
12
13 printf("%lu\n", x); // 3490
14 }
Notice a couple things there. Even though we didn’t deign to capture any information about error charac- ters in the string, strtoul() won’t give us undefined behavior; it will just return 0. Also, we specified that this was a decimal (base 10) number. Does this mean we can convert numbers of different bases? Sure! Let’s do binary!
1 #include
2 #include
3
4 int main(void)
5 {
6 char *s = "101010"; // What's the meaning of this number?
7
8 // Convert string s, a number in base 2, to an unsigned long int.
9
10 unsigned long int x = strtoul(s, NULL, 2);
11
12 printf("%lu\n", x); // 42
13 }
OK, that’s all fun and games, but what’s with that NULL in there? What’s that for? That helps us figure out if an error occurred in the processing of the string. It’s a pointer to a pointer to a char, which sounds scary, but isn’t once you wrap your head around it. Let’s do an example where we feed in a deliberately bad number, and we’ll see how strtol() lets us know where the first invalid digit is.
1 #include
2 #include
3
4 int main(void)
5 {
6 char *s = "34x90"; // "x" is not a valid digit in base 10!
7 char *badchar;
8
9 // Convert string s, a number in base 10, to an unsigned long int.
10
11 unsigned long int x = strtoul(s, &badchar, 10);
12
13 // It tries to convert as much as possible, so gets this far:
14
15 printf("%lu\n", x); // 34
16
17 // But we can see the offending bad character because badchar Chapter 15. Types III: Conversions 94
18 // points to it!
19
20 printf("Invalid character: %c\n", *badchar); // "x"
21 }
So there we have strtoul() modifying what badchar points to in order to show us where things went wrong3. But what if nothing goes wrong? In that case, badchar will point to the NUL terminator at the end of the string. So we can test for it:
1 #include
2 #include
3
4 int main(void)
5 {
6 char *s = "3490"; // "x" is not a valid digit in base 10!
7 char *badchar;
8
9 // Convert string s, a number in base 10, to an unsigned long int.
10
11 unsigned long int x = strtoul(s, &badchar, 10);
12
13 // Check if things went well
14
15 if (*badchar == '\0') {
16 printf("Success! %lu\n", x);
17 } else {
18 printf("Partial conversion: %lu\n", x);
19 printf("Invalid character: %c\n", *badchar);
20 }
21 }
So there you have it. The atoi()-style functions are good in a controlled pinch, but the strtol()-style functions give you far more control over error handling and the base of the input.
15.2 Numeric Conversions
15.2.1 Boolean
If you convert a zero to bool, the result is 0. Otherwise it’s 1.
15.2.2 Integer to Integer Conversions If an integer type is converted to unsigned and doesn’t fit in it, the unsigned result wraps around odometer- style until it fits in the unsigned4. If an integer type is converted to a signed number and doesn’t fit, the result is implementation-defined! Something documented will happen, but you’ll have to look it up5
3We have to pass a pointer to badchar to strtoul() or it won’t be able to modify it in any way we can see, analogous to why you have to pass a pointer to an int to a function if you want that function to be able to change that value of that int. 4In practice, what’s probably happening on your implementation is that the high-order bits are just being dropped from the result, so a 16-bit number 0x1234 being converted to an 8-bit number ends up as 0x0034, or just 0x34. 5Again, in practice, what will likely happen on your system is that the bit pattern for the original will be truncated and then just used to represent the signed number, two’s complement. For example, my system takes an unsigned char of 192 and converts it to signed char -64. In two’s complement, the bit pattern for both these numbers is binary 11000000. Chapter 15. Types III: Conversions 95
15.2.3 Integer and Floating Point Conversions If a floating point type is converted to an integer type, the fractional part is discarded with prejudice6. But—and here’s the catch—if the number is too large to fit in the integer, you get undefined behavior. So don’t do that. Going From integer or floating point to floating point, C makes a best effort to find the closest floating point number to the integer that it can. Again, though, if the original value can’t be represented, it’s undefined behavior.
15.3 Implicit Conversions
These are conversions the compiler does automatically for you when you mix and match types.
15.3.1 The Integer Promotions
In a number of places, if a int can be used to represent a value from char or short (signed or unsigned), that value is promoted up to int. If it doesn’t fit in an int, it’s promoted to unsigned int. This is how we can do something like this: char x = 10, y = 20; int i = x + y;
In that case, x and y get promoted to int by C before the math takes place. The integer promotions take place during The Usual Arithmetic Conversions, with variadic functions7, unary + and - operators, or when passing values to functions without prototypes8.
15.3.2 The Usual Arithmetic Conversions These are automatic conversions that C does around numeric operations that you ask for. (That’s actually what they’re called, by the way, by C11 §6.3.1.8.) Note that for this section, we’re just talking about numeric types—strings will come later. These conversions answer questions about what happens when you mix types, like this: int x = 3 + 1.2; // Mixing int and double // 4.2 is converted to int // 4 is stored in x float y = 12 * 2; // Mixing float and int // 24 is converted to float // 24.0 is stored in y
Do they become ints? Do they become floats? How does it work? Here are the steps, paraphrased for easy consumption. 1. If one thing in the expression is a floating type, convert the other things to that floating type. 2. Otherwise, if both types are integer types, perform the integer promotions on each, then make the operand types as big as they need to be hold the common largest value. Sometimes this involves changing signed to unsigned. If you want to know the gritty details, check out C11 §6.3.1.8. But you probably don’t. Just generally remember that int types become float types if there’s a floating point type anywhere in there, and the compiler makes an effort to make sure mixed integer types don’t overflow.
6Not really—it’s just discarded regularly. 7Functions with a variable number of arguments. 8This is rarely done because the compiler will complain and having a prototype is the Right Thing to do. I think this still works for historic reasons, before prototypes were a thing. Chapter 15. Types III: Conversions 96
Finally, if you convert from one floating point type to another, the compiler will try to make an exact conversion. If it can’t, it’ll do the best approximation it can. If the number is too large to fit in the type you’re converting into, boom: undefined behavior!
15.3.3 void*
The void* type is interesting because it can be converted from or to any pointer type. int x = 10; void *p = &x; // &x is type int*, but we store it in a void* int *q = p; // p is void*, but we store it in an int*
15.4 Explicit Conversions
These are conversions from type to type that you have to ask for; the compiler won’t do it for you. You can convert from one type to another by assigning one type to another with an =. You can also convert explicitly with a cast.
15.4.1 Casting You can explicitly change the type of an expression by putting a new type in parentheses in front of it. Some C devs frown on the practice unless absolutely necessary, but it’s likely you’ll come across some C code with these in it. Let’s do an example where we want to convert an int into a long so that we can store it in a long. Note: this example is contrived and the cast in this case is completely unnecessary because the x + 12 expression would automatically be changed to long int to match the wider type of y. int x = 10; long int y = (long int)x + 12;
In that example, even those x was type int before, the expression (long int)x has type long int. We say, “We cast x to long int.” More commonly, you might see a cast being used to convert a void* into a specific pointer type so it can be dereferenced. A callback from the built-in qsort() function might display this behavior since it has void*s passed into it: int compar(const void *elem1, const void *elem2) { return *((const int*)elem2) - *((const int*)elem1); }
But you could also clearly write it with an assignment: int compar(const void *elem1, const void *elem2) { const int *e1 = elem1; const int *e2 = elem2;
return *e2 - *e1; }
One place you’ll see casts more commonly is to avoid a warning when printing pointer values with the rarely-used %p which gets picky with anything other than a void*: Chapter 15. Types III: Conversions 97 int x = 3490; int *p = &x; printf("%p\n", p); generates this warning: warning: format ‘%p’ expects argument of type ‘void *’, but argument 2 has type ‘int *’
You can fix it with a cast: printf("%p\n", (void *)p);
Another place is with explicit pointer changes, if you don’t want to use an intervening void*, but these are also pretty uncommon: long x = 3490; long *p = &x; unsigned char *c = (unsigned char *)p;
Again, casting is rarely needed in practice. If you find yourself casting, there might be another way to do the same thing, or maybe you’re casting unnecessarily. Or maybe it is necessary. Personally, I try to avoid it, but am not afraid to use it if I have to. Chapter 16
Types IV: Qualifiers and Specifiers
Now that we have some more types under our belts, turns out we can give these types some additional attributes that control their behavior. These are the type qualifiers and storage-class specifiers.
16.1 Type Qualifiers
These are going to allow you to declare constant values, and also to give the compiler optimization hints that it can use.
16.1.1 const This is the most common type qualifier you’ll see. It means the variable is constant, and any attempt to modify it will result in a very angry compiler. const int x = 2; x = 4; // COMPILER PUKING SOUNDS, can't assign to a constant
You can’t change a const value. Often you see const in parameter lists for functions: void foo(const int x) { printf("%d\n", x + 30); // OK, doesn't modify "x" }
16.1.1.1 const and Pointers This one gets a little funky, because there are two usages that have two meanings when it comes to pointers. For one thing, we can make it so you can’t change the thing the pointer points to. You do this by putting the const up front with the type name (before the asterisk) in the type declaration. int x[] = {10, 20}; const int *p = x; p++; // We can modify p, no problem
*p = 30; // Compiler error! Can't change what it points to
Somewhat confusingly, these two things are equivalent: const int *p; // Can't modify what p points to int const *p; // Can't modify what p points to, just like the previous line
98 Chapter 16. Types IV: Qualifiers and Specifiers 99
Great, so we can’t change the thing the pointer points to, but we can change the pointer itself. What if we want the other way around? We want to be able to change what the pointer points to, but not the pointer itself? Just move the const after the asterisk in the declaration: int *const p; // We can't modify "p" with pointer arithmetic p++; // Compiler error!
But we can modify what they point to: int x = 10; int *const p = &x;
*p = 20; // Set "x" to 20, no problem
You can also do make both things const: const int *const p; // Can't modify p or *p!
Finally, if you have multiple levels of indirection, you should const the appropriate levels. Just because a pointer is const, doesn’t mean the pointer it points to must also be. You can explicitly set them like in the following examples: char **p; p++; // OK! (*p)++; // OK! char **const p; p++; // Error! (*p)++; // OK! char *const *p; p++; // OK! (*p)++; // Error! char *const *const p; p++; // Error! (*p)++; // Error!
16.1.1.2 const Correctness One more thing I have to mention is that the compiler will warn on something like this: const int x = 20; int *p = &x; saying something to the effect of: initialization discards 'const' qualifier from pointer type target
What’s happening there? Well, we need to look at the types on either side of the assignment: const int x = 20; int *p = &x; // ^ ^ // | | // int* const int*
The compiler is warning us that the value on the right side of the assignment is const, but the one of the left is not. And the compiler is letting us know that it is discarding the “const-ness” of the expression on the right. Chapter 16. Types IV: Qualifiers and Specifiers 100
That is, we can still try to do the following, but it’s just wrong. The compiler will warn, and it’s undefined behavior: const int x = 20; int *p = &x;
*p = 40; // Undefined behavior--maybe it modifies "x", maybe not!
printf("%d\n", x); // 40, if you're lucky
16.1.2 restrict TLDR: you never have to use this and you can ignore it every time you see it. restrict is a hint to the compiler that a particular piece of memory will only be accessed by one pointer and never another. If a developer declares a pointer to be restrict and then accesses the object it points to in another way, the behavior is undefined. Basically you’re telling C, “Hey—I guarantee that this one single pointer is the only way I access this memory, and if I’m lying, you can pull undefined behavior on me.” And C uses that information to perform certain optimizations. For example, let’s write a function to swap two variables, and we’ll use the restrict keyword to assure C that we’ll never pass in pointers to the same thing. And then let’s blow it an try passing in pointers to the same thing.
1 void swap(int *restrict a, int *restrict b)
2 {
3 int t;
4
5 t = *a;
6 *a = *b;
7 *b = t;
8 }
9
10 int main(void)
11 {
12 int x = 10, y = 20;
13
14 swap(&x, &y); // OK! "a" and "b", above, point to different things
15
16 swap(&x, &x); // Undefined behavior! "a" and "b" point to the same thing
17 }
If we were to take out the restrict keywords, above, that would allow both calls to work safely. But then the compiler might not be able to optimize. restrict has block scope, that is, the restriction only lasts for the scope its used. If it’s in a parameter list for a function, it’s in the block scope of that function. If the restricted pointer points to an array, the restriction covers the entire array. If it’s outside any function in file scope, the restriction covers the entire program. You’re likely to see this in library functions like printf(): int printf(const char * restrict format, ...);
Again, that’s just telling the compiler that inside the printf() function, there will be only one pointer that refers to any part of that format string. Chapter 16. Types IV: Qualifiers and Specifiers 101
16.1.3 volatile You’re unlikely to see or need this unless you’re dealing with hardware directly. volatile tells the compiler that a value might change behind its back and should be looked up every time. An example might be where the compiler is looking in memory at an address that continuously updates behind the scenes, e.g. some kind of hardware timer. If the compiler decides to optimize that and store the value in a register for a protracted time, the value in memory will update and won’t be reflected in the register. By declaring something volatile, you’re telling the compiler, “Hey, the thing this points at might change at any time for reasons outside this program code.” volatile int *p;
16.1.4 _Atomic This is an optional C feature that we’ll talk about another time.
16.2 Storage-Class Specifiers
Storage-class specifiers are similar to type quantifiers. They give the compiler more information about the type of a variable.
16.2.1 auto
You barely ever see this keyword, since auto is the default for block scope variables. It’s implied. These are the same: { int a; // auto is the default... auto int a; // So this is redundant }
The auto keyword indicates that this object has automatic storage duration. That is, it exists in the scope in which it is defined, and is automatically deallocated when the scope is exited. One gotcha about automatic variables is that their value is indeterminate until you explicitly initialize them. We say they’re full of “random” or “garbage” data, though neither of those really makes me happy. In any case, you won’t know what’s in it unless you initialize it. Always initialize all automatic variables before use!
16.2.2 static This keyword has two meanings, depending on if the variable is file scope or block scope. Let’s start with block scope.
16.2.2.1 static in Block Scope In this case, we’re basically saying, “I just want a single instance of this variable to exist, shared between calls.” That is, its value will persist between calls. static in block scope with an initializer will only be initialized one time on program startup, not each time the function is called. Let’s do an example: Chapter 16. Types IV: Qualifiers and Specifiers 102
1 #include
2
3 void counter(void)
4 {
5 static int count = 1; // This is initialized one time
6
7 printf("This has been called %d time(s)\n", count);
8
9 count++;
10 }
11
12 int main(void)
13 {
14 counter(); // "This has been called 1 time(s)"
15 counter(); // "This has been called 2 time(s)"
16 counter(); // "This has been called 3 time(s)"
17 counter(); // "This has been called 4 time(s)"
18 }
See how the value of count persists between calls? One thing of note is that static block scope variables are initialized to 0 by default. static int foo; // Default starting value is `0`... static int foo = 0; // So the `0` assignment is redundant
Finally, be advised that if you’re writing multithreaded programs, you have to be sure you don’t let multiple threads trample the same variable.
16.2.2.2 static in File Scope When you get out to file scope, outside any blocks, the meaning rather changes. Variables at file scope already persist between function calls, so that behavior is already there. Instead what static means in this context is that this variable isn’t visible outside of this particular source file. Kinda like “global”, but only in this file. More on that in the section about building with multiple source files.
16.2.3 extern
The extern storage-class specifier gives us a way to refer to objects in other source files. Let’s say, for example, the file bar.c had the following as its entirety:
1 // bar.c
2
3 int a = 37;
Just that. Declaring a new int a in file scope. But what if we had another source file, foo.c, and we wanted to refer to the a that’s in bar.c? It’s easy with the extern keyword:
1 // foo.c
2
3 extern int a;
4
5 int main(void)
6 {
7 printf("%d\n", a); // 37, from bar.c!
8 Chapter 16. Types IV: Qualifiers and Specifiers 103
9 a = 99;
10
11 printf("%d\n", a); // Same "a" from bar.c, but it's now 99
12 }
We could have also made the extern int a in block scope, and it still would have referred to the a in bar.c:
1 // foo.c
2
3 int main(void)
4 {
5 extern int a;
6
7 printf("%d\n", a); // 37, from bar.c!
8
9 a = 99;
10
11 printf("%d\n", a); // Same "a" from bar.c, but it's now 99
12 }
Now, if a in bar.c had been marked static. this wouldn’t have worked. static variables at file scope are not visible outside that file. A final note about extern on functions. For functions, extern is the default, so it’s redundant. You can declare a function static if you only want it visible in a single source file.
16.2.4 register This is a keyword to hint to the compiler that this variable is frequently-used, and should be made as fast as possible to access. The compiler is under no obligation to agree to it. Now, modern C compiler optimizers are pretty effective at figuring this out themselves, so it’s rare to see these days. But if you must:
1 #include
2
3 int main(void)
4 {
5 register int a; // Make "a" as fast to use as possible.
6
7 for (a = 0; a < 10; a++)
8 printf("%d\n", a);
9 }
It does come at a price, however. You can’t take the address of a register: register int a; int *p = &a; // COMPILER ERROR! Can't take address of a register
The same applies to any part of an array: register int a[] = {11, 22, 33, 44, 55}; int *p = a; // COMPILER ERROR! Can't take address of a[0]
Or dereferencing part of an array: register int a[] = {11, 22, 33, 44, 55};
int a = *(a + 2); // COMPILER ERROR! Address of a[0] taken
Interestingly, for the equivalent with array notation, gcc only warns: Chapter 16. Types IV: Qualifiers and Specifiers 104 register int a[] = {11, 22, 33, 44, 55}; int a = a[2]; // COMPILER WARNING! with: warning: ISO C forbids subscripting ‘register’ array
The fact that you can’t take the address of a register variable frees the compiler up to make optimizations around that assumption if it hasn’t figured them out already. Also adding register to a const variable prevents one from accidentally passing its pointer to another function that willfully ignore its constness1. A bit of historic backstory, here: deep inside the CPU are little dedicated “variables” called registers2. They are super fast to access compared to RAM, so using them gets you a speed boost. But they’re not in RAM, so they don’t have an associated memory address (which is why you can’t take the address-of or get a pointer to them). But, like I said, modern compilers are really good at producing optimal code, using registers whenever possible regardless of whether or not you specified the register keyword. Not only that, but the spec allows them to just treat it as if you’d typed auto, if they want. So no guarantees.
16.2.5 _Thread_local
When you’re using multiple threads and you have some variables in either global or static block scope, this is a way to make sure that each thread gets its own copy of the variable. This’ll help you avoid race conditions and threads stepping on each other’s toes. If you’re in block scope, you have to use this along with either extern or static. Also, if you include
1https://gustedt.wordpress.com/2010/08/17/a-common-misconsception-the-register-keyword/ 2https://en.wikipedia.org/wiki/Processor_register Chapter 17
Multifile Projects
So far we’ve been looking at toy programs that for the most part fit in a single file. But complex C programs are made up of many files that are all compiled and linked together into a single executable. In this chapter we’ll check out some of the common patterns and practices for putting together larger projects.
17.1 Includes and Function Prototypes
A really common situation is that some of your functions are defined in one file, and you want to call them from another. This actually works out of the box with a warning… let’s first try it and then look at the right way to fix the warning. For these examples, we’ll put the filename as the first comment in the source. To compile them, you’ll need to specify all the sources on the command line: # output file source files # v v # |----| |------| gcc -o foo foo.c bar.c
In that examples, foo.c and bar.c get built into the executable named foo. So let’s take a look at the source file bar.c:
1 // File bar.c
2
3 int add(int x, int y)
4 {
5 return x + y;
6 }
And the file foo.c with main in it:
1 // File foo.c
2
3 #include
4
5 int main(void)
6 {
7 printf("%d\n", add(2, 3)); // 5!
8 }
105 Chapter 17. Multifile Projects 106
See how from main() we call add()—but add() is in a completely different source file! It’s in bar.c, while the call to it is in foo.c! If we build this with: gcc -o foo foo.c bar.c
we get this error: error: implicit declaration of function 'add' is invalid in C99
(Or you might get a warning. Which you should not ignore. Never ignore warnings in C; address them all.) If you recall from the section on prototypes, implicit declarations are banned in modern C and there’s no legitimate reason to introduce them into new code. We should fix it. What implicit declaration means is that we’re using a function, namely add() in this case, without letting C know anything about it ahead of time. C wants to know what it returns, what types it takes as arguments, and things such as that. We saw how to fix that earlier with a function prototype. Indeed, if we add one of those to foo.c before we make the call, everything works well:
1 // File foo.c
2
3 #include
4
5 int add(int, int); // Add the prototype
6
7 int main(void)
8 {
9 printf("%d\n", add(2, 3)); // 5!
10 }
No more error! But that’s a pain—needing to type in the prototype every time you want to use a function. I mean, we used printf() right there and didn’t need to type in a prototype; what gives? If you remember from what back with hello.c at the beginning of the book, we actually did include the prototype for printf()! It’s in the file stdio.h! And we included that with #include! Can we do the same with our add() function? Make a prototype for it and put it in a header file? Sure! Header files in C have a .h extension by default. And they often, but not always, have the same name as their corresponding .c file. So let’s make a bar.h file for our bar.c file, and we’ll stick the prototype in it:
1 // File bar.h
2
3 int add(int, int);
And now let’s modify foo.c to include that file. Assuming it’s in the same directory, we include it inside double quotes (as opposed to angle brackets):
1 // File foo.c
2
3 #include
4
5 #include "bar.h" // Include from current directory
6
7 int main(void)
8 { Chapter 17. Multifile Projects 107
9 printf("%d\n", add(2, 3)); // 5!
10 }
Notice how we don’t have the prototype in foo.c anymore—we included it from bar.h. Now any file that wants that add() functionality can just #include "bar.h" to get it, and you don’t need to worry about typing in the function prototype. As you might have guessed, #include literally includes the named file right there in your source code, just as if you’d typed it in. And building and running: ./foo 5
Indeed, we get the result of 2 + 3! Yay! But don’t crack open your drink of choice quite yet. We’re almost there! There’s just one more piece of boilerplate we have to add.
17.2 Dealing with Repeated Includes
It’s not uncommon that a header file will itself #include other headers needed for the functionality of its corresponding C files. I mean, why not? And it could be that you have a header #included multiple times from different places. Maybe that’s no problem, but maybe it would cause compiler errors. And we can’t control how many places #include it! Even, worse we might get into a crazy situation where header a.h includes header b.h, and b.h includes a.h! It’s an #include infinite cycle! Trying to build such a thing gives an error: error: #include nested depth 200 exceeds maximum of 200
What we need to do is make it so that if a file gets included once, subsequent #includes for that file are ignored. The stuff that we’re about to do is so common that you should just automatically do it every time you make a header file! And the common way to do this is with a preprocessor variable that we set the first time we #include the file. And then for subsequent #includes, we first check to make sure that the variable isn’t defined. For that variable name, it’s super common to take the name of the header file, like bar.h, make it upper- case, and replace the period with an underscore: BAR_H. So put a check at the very, very top of the file where you see if it’s already been included, and effectively comment the whole thing out if it has. (Don’t put a leading underscore (because a leading underscore followed by a capital letter is reserved) or a double leading underscore (because that’s also reserved.))
1 #ifndef BAR_H // If BAR_H isn't defined...
2 #define BAR_H // Define it (with no particular value)
3
4 // File bar.h
5
6 int add(int, int);
7
8 #endif // End of the #ifndef BAR_H
This will effectively cause the header file to be included only a single time, no matter how many places try to #include it. Chapter 17. Multifile Projects 108
17.3 static and extern
When it comes to multifile projects, you can make sure file-scope variables and functions are not visible from other source files with the static keyword. And you can refer to objects in other files with extern. For more info, check out the sections in the book on the static and extern storage-class specifiers.
17.4 Compiling with Object Files
This isn’t part of the spec, but it’s 99.999% common in the C world. You can compile C files into an intermediate representation called object files. These are compiled ma- chine code that hasn’t been put into an executable yet. Object files in Windows have a .OBJ extension; in Unix-likes, they’re .o. In gcc, we can build some like this, with the -c (compile only!) flag: gcc -c foo.c # produces foo.o gcc -c bar.c # produces bar.o
And then we can link those together into a single executable: gcc -o foo foo.o bar.o
Voila, we’ve produced an executable foo from the two object files. But you’re thinking, why bother? Can’t we just: gcc -o foo foo.c bar.c and kill two boids1 with one stone? For little programs, that’s fine. I do it all the time. But for larger programs, we can take advantage of the fact that compiling from source to object files is relatively slow, and linking together a bunch of object files is relatively fast. This really shows with the make utility that only rebuilds sources that are newer than their outputs. Let’s say you had a thousand C files. You could compile them all to object files to start (slowly) and then combine all those object files into an executable (fast). Now say you modified just one of those C source files—here’s the magic: you only have to rebuild that one object file for that source file! And then you rebuild the executable (fast). All the other C files don’t have to be touched. In other words, by only rebuilding the object files we need to, we cut down on compilation times radically. (Unless of course you’re doing a “clean” build, in which case all the object files have to be created.)
1https://en.wikipedia.org/wiki/Boids Chapter 18
The Outside Environment
When you run a program, it’s actually you talking to the shell, saying, “Hey, please run this thing.” And the shell says, “Sure,” and then tells the operating system, “Hey, could you please make a new process and run this thing?” And if all goes well, the OS complies and your program runs. But there’s a whole world outside your program in the shell that can be interacted with from within C. We’ll look at a few of those in this chapter.
18.1 Command Line Arguments
Many command line utilities accept command line arguments. For example, if we want to see all files that end in .txt, we can type something like this on a Unix-like system: ls *.txt
(or dir instead of ls on a Windows system). In this case, the command is ls, but it arguments are all all files that end with .txt1. So how can we see what is passed into program from the command line? Say we have a program called add that adds all numbers passed on the command line and prints the result: ./add 10 30 5 45
That’s gonna pay the bills for sure! But seriously, this is a great tool for seeing how to get those arguments from the command line and break them down. First, let’s see how to get them at all. For this, we’re going to need a new main()! Here’s a program that prints out all the command line arguments. For example, if we name the executable foo, we can run it like this: ./foo i like turtles and we’ll see this output: arg 0: ./foo arg 1: i arg 2: like arg 3: turtles
1Historially, MS-DOS and Windows programs would do this differently than Unix. In Unix, the shell would expand the wildcard into all matching files before your program saw it, whereas the Microsoft variants would pass the wildcard expression into the program to deal with. In any case, there are arguments that get passed into the program.
109 Chapter 18. The Outside Environment 110
It’s a little weird, because the zeroth argument is the name of the executable, itself. But that’s just some- thing to get used to. The arguments themselves follow directly. Source:
1 #include
2
3 int main(int argc, char *argv[])
4 {
5 for (int i = 0; i < argc; i++) {
6 printf("arg %d: %s\n", i, argv[i]);
7 }
8 }
Whoa! What’s going on with the main() function signature? What’s argc and argv2 (pronounced arg-c and arg-v)? Let’s start with the easy one first: argc. This is the argument count, including the program name, itself. If you think of all the arguments as an array of strings, which is exactly what they are, then you can think of argc as the length of that array, which is exactly what it is. And so what we’re doing in that loop is going through all the argvs and printing them out one at a time, so for a given input: ./foo i like turtles
we get a corresponding output: arg 0: ./foo arg 1: i arg 2: like arg 3: turtles
With that in mind, we should be good to go with our adder program. Our plan: • Look at all the command line arguments (past argv[0], the program name) • Convert them to integers • Add them to a running total • Print the result Let’s get to it!
1 #include
2 #include
3
4 int main(int argc, char **argv)
5 {
6 int total = 0;
7
8 for (int i = 1; i < argc; i++) { // Start at 1, the first argument
9 int value = atoi(argv[i]); // Use strtol() for better error handling
10
11 total += value;
12 }
13
14 printf("%d\n", total);
15 }
Sample runs:
2Since they’re just regular parameter names, you don’t actually have to call them argc and argv. But it’s so very idiomatic to use those names, if you get creative, other C programmers will look at you with a suspicious eye, indeed! Chapter 18. The Outside Environment 111
$ ./add 0 $ ./add 1 1 $ ./add 1 2 3 $ ./add 1 2 3 6 $ ./add 1 2 3 4 10
Of course, it might puke if you pass in a non-integer, but hardening against that is left as an exercise to the reader.
18.1.1 The Last argv is NULL
One bit of fun trivia about argv is that after the last string is a pointer to NULL. That is: argv[argc] == NULL
is always true! This might seem pointless, but it turns out to be useful in a couple places; we’ll take a look at one of those right now.
18.1.2 The Alternate: char **argv Remember that when you call a function, C doesn’t differentiate between array notation and pointer nota- tion in the function signature. That is, these are the same: void foo(char a[]) void foo(char *a)
Now, it’s been convenient to think of argv as an array of strings, i.e. an array of char*s, so this made sense: int main(int argc, char *argv[])
but because of the equivalence, you could also write: int main(int argc, char **argv)
Yeah, that’s a pointer to a pointer, all right! If it makes it easier, think of it as a pointer to a string. But really, it’s a pointer to a value that points to a char. Also recall that these are equivalent: argv[i] *(argv + i)
which means you can do pointer arithmetic on argv. So an alternate way to consume the command line arguments might be to just walk along the argv array by bumping up a pointer until we hit that NULL at the end. Let’s modify our adder to do that:
1 #include
2 #include
3
4 int main(int argc, char **argv)
5 { Chapter 18. The Outside Environment 112
6 int total = 0;
7
8 // Cute trick to get the compiler to stop warning about the
9 // unused variable argc:
10 (void)argc;
11
12 for (char **p = argv; *p != NULL; p++) {
13 int value = atoi(*p); // Use strtol() for better error handling
14
15 total += value;
16 }
17
18 printf("%d\n", total);
19 }
Personally, I use array notation to access argv, but have seen this style floating around, as well.
18.1.3 Fun Facts
Just a few more things about argc and argv. • Some environments might not set argv[0] to the program name. If it’s not available, argv[0] will be an empty string. I’ve never seen this happen. • The spec is actually pretty liberal with what an implementation can do with argv and where those values come from. But every system I’ve been on works the same way, as we’ve discussed in this section. • You can modify argc, argv, or any of the strings that argv points to. (Just don’t make those strings longer than they already are!) • On some Unix-like systems, modifying the string argv[0] results in the output of ps changing3. Normally, if you have a program called foo that you’ve run with ./foo, you might see this in the output of ps: 4078 tty1 S 0:00 ./foo
But if you modify argv[0] like so, being careful that the new string "Hi! " is the same length as the old one "./foo": strcpy(argv[0], "Hi! ");
and then run ps while the program ./foo is still executing, we’ll see this instead: 4079 tty1 S 0:00 Hi!
This behavior is not in the spec and is highly system-dependent.
18.2 Exit Status
Did you notice that the function signatures for main() have it returning type int? What’s that all about? It has to do with a thing called the exit status, which is an integer that can be returned to the program that launched yours to let it know how things went. Now, there are a number of ways a program can exit in C, including returning from main(), or calling one of the exit() variants. All of these methods accept an int as an argument. Side note: did you see that in basically all my examples, even though main() is supposed to return an int, I don’t actually return anything? In any other function, this would be illegal, but there’s a special case
3ps, Process Status, is a Unix command to see what processes are running at the moment. Chapter 18. The Outside Environment 113
in C: if execution reaches the end of main() without finding a return, it automatically does a return 0. But what does the 0 mean? What other numbers can we put there? And how are they used? The spec is both clear and vague on the matter, as is common. Clear because it spells out what you can do, but vague in that it doesn’t particularly limit it, either. Nothing for it but to forge ahead and figure it out! Let’s get Inception4 for a second: turns out that when you run your program, you’re running it from another program. Usually this other program is some kind of shell5 that doesn’t do much on its own except launch other programs. But this is a multi-phase process, especially visible in command-line shells: 1. The shell launches your program 2. The shell typically goes to sleep (for command-line shells) 3. Your program runs 4. Your program terminates 5. The shell wakes up and waits for another command Now, there’s a little piece of communication that takes place between steps 4 and 5: the program can return a status value that the shell can interrogate. Typically, this value is used to indicate the success or failure of your program, and, if a failure, what type of failure. This value is what we’ve been returning from main(). That’s the status. Now, the C spec allows for two different status values, which have macro names defined in
Status Description EXIT_SUCCESS or 0 Program terminated successfully. EXIT_FAILURE Program terminated with an error.
Let’s write a short program that multiplies two numbers from the command line. We’ll require that you specify exactly two values. If you don’t, we’ll print an error message, and exit with an error status.
1 #include
2 #include
3
4 int main(int argc, char **argv)
5 {
6 if (argc != 3) {
7 printf("usage: mult x y\n");
8 return EXIT_FAILURE; // Indicate to shell that it didn't work
9 }
10
11 printf("%d\n", atoi(argv[1]) * atoi(argv[2]));
12
13 return 0; // same as EXIT_SUCCESS, everything was good.
14 }
Now if we try to run this, we get the expected effect until we specify exactly the right number of command- line arguments: $ ./mult usage: mult x y
$ ./mult 3 4 5
4https://en.wikipedia.org/wiki/Inception 5https://en.wikipedia.org/wiki/Shell_(computing) Chapter 18. The Outside Environment 114 usage: mult x y
$ ./mult 3 4 12
But that doesn’t really show the exit status that we returned, does it? We can get the shell to print it out, though. Assuming you’re running Bash or another POSIX shell, you can use echo $? to see it6. Let’s try: $ ./mult usage: mult x y $ echo $? 1
$ ./mult 3 4 5 usage: mult x y $ echo $? 1
$ ./mult 3 4 12 $ echo $? 0
Interesting! We see that on my system, EXIT_FAILURE is 1. The spec doesn’t spell this out, so it could be any number. But try it; it’s probably 1 on your system, too.
18.2.1 Other Exit Status Values
The status 0 most definitely means success, but what about all the other integers, even negative ones? Here we’re going off the C spec and into Unix land. In general, while 0 means success, a positive non-zero number means failure. So you can only have one type of success, and multiple types of failure. Bash says the exit code should be between 0 and 255, though a number of codes are reserved. In short, if you want to indicate different error exit statuses in a Unix environment, you can start with 1 and work your way up. On Linux, if you try any code outside the range 0-255, it will bitwise AND the code with 0xff, effectively clamping it to that range. You can script the shell to later use these status codes to make decisions about what to do next.
18.3 Environment Variables
Before I get into this, I need to warn you that C doesn’t specify what an environment variable is. So I’m going to describe the environment variable system that works on every major platform I’m aware of. Basically, the environment is the program that’s going to run your program, e.g. the bash shell. And it might have some bash variables defined. In case you didn’t know, the shell can make its own variables. Each shell is different, but in bash you can just type set and it’ll show you all of them. Here’s an excerpt from the 61 variables that are defined in my bash shell: HISTFILE=/home/beej/.bash_history HISTFILESIZE=500 HISTSIZE=500 HOME=/home/beej HOSTNAME=FBILAPTOP
6In Windows cmd.exe, type echo %errorlevel%. In PowerShell, type $LastExitCode. Chapter 18. The Outside Environment 115
HOSTTYPE=x86_64 IFS=$' \t\n'
Notice they are in the form of key/value pairs. For example, one key is HOSTTYPE and its value is x86_64. From a C perspective, all values are strings, even if they’re numbers7. So, anyway! Long story short, it’s possible to get these values from inside your C program. Let’s write a program that uses the standard getenv() function to look up a value that you set in the shell. getenv() will return a pointer to the value string, or else NULL if the environment variable doesn’t exist.
1 #include
2 #include
3
4 int main(void)
5 {
6 char *val = getenv("FROTZ"); // Try to get the value
7
8 // Check to make sure it exists
9 if (val == NULL) {
10 printf("Cannot find the FROTZ environment variable\n");
11 return EXIT_FAILURE;
12 }
13
14 printf("Value: %s\n", val);
15 }
If I run this directly, I get this: $ ./foo Cannot find the FROTZ environment variable
which makes since, since I haven’t set it yet. In bash, I can set it to something with8: $ export FROTZ="C is awesome!"
Then if I run it, I get: $ ./foo Value: C is awesome!
In this way, you can set up data in environment variables, and you can get it in your C code and modify your behavior accordingly.
18.3.1 Setting Environment Variables This isn’t standard, but a lot of systems provide ways to set environment variables. If on a Unix-like, look up the documentation for putenv(), setenv(), and unsetenv(). On Windows, see _putenv().
18.3.2 Unix-like Alternative Environment Variables If you’re on a Unix-like system, odds are you have another couple ways of getting access to environment variables. Note that although the spec points this out as a common extension, it’s not truly part of the C standard. It is, however, part of the POSIX standard. One of these is a variable called environ that must be declared like so:
7If you need a numeric value, convert the string with something like atoi() or strtol(). 8In Windows CMD.EXE, use set FROTZ=value. In PowerShell, use $Env:FROTZ=value. Chapter 18. The Outside Environment 116
extern char **environ;
It’s an array of strings terminated with a NULL pointer. You should declare it yourself before you use it, or you might find it in the non-standard
1 #include
2
3 extern char **environ; // MUST be extern AND named "environ"
4
5 int main(void)
6 {
7 for (char **p = environ; *p != NULL; p++) {
8 printf("%s\n", *p);
9 }
10
11 // Or you could do this:
12 for (int i = 0; environ[i] != NULL; i++) {
13 printf("%s\n", environ[i]);
14 }
15 }
For a bunch of output that looks like this: SHELL=/bin/bash COLORTERM=truecolor TERM_PROGRAM_VERSION=1.53.2 LOGNAME=beej HOME=/home/beej ... etc ...
Use getenv() if at all possible because it’s more portable. But if you have to iterate over environment variables, using environ might be the way to go. Another non-standard way to get the environment variables is as a parameter to main(). It works much the same way, but you avoid needing to add your extern environ variable. Not even the POSIX spec supports this9 as far as I can tell, but it’s common in Unix land.
1 #include
2
3 int main(int argc, char **argv, char **env) // <-- env!
4 {
5 (void)argc; (void)argv; // Suppress unused warnings
6
7 for (char **p = env; *p != NULL; p++) {
8 printf("%s\n", *p);
9 }
10
11 // Or you could do this:
12 for (int i = 0; env[i] != NULL; i++) {
13 printf("%s\n", env[i]);
14 }
15 }
Just like using environ but even less portable. It’s good to have goals.
9https://en.wikipedia.org/wiki/https://pubs.opengroup.org/onlinepubs/9699919799/functions/exec.html Chapter 19
The C Preprocessor
Before your program gets compiled, it actually runs through a phase called preprocessing. It’s almost like there’s a language on top of the C language that runs first. And it outputs the C code, which then gets compiled. We’ve already seen this to an extent with #include! That’s the C Preprocessor! Where it sees that directive, it includes the named file right there, just as if you’d typed it in there. And then the compiler builds the whole thing. But it turns out it’s a lot more powerful than just being able to include things. You can define macros that are substituted… and even macros that take arguments!
19.1 #include
Let’s start with the one we’ve already seen a bunch. This is, of course, a way to include other sources in your source. Very commonly used with header files. While the spec allows for all kinds of behavior with #include, we’re going to take a more pragmatic approach and talk about the way it works on every system I’ve ever seen. We can split header files into two categories: system and local. Things that are built-in, like stdio.h, stdlib.h, math.h, and so on, you can include with angle brackets: #include
The angle brackets tell C, “Hey, don’t look in the current directory for this header file—look in the system- wide include directory instead.” Which, of course, implies that there must be a way to include local files from the current directory. And there is: with double quotes: #include "myheader.h"
Or you can very probably look in relative directories using forward slashes and dots, like this: #include "mydir/myheader.h" #include "../someheader.py"
Don’t use a backslash (\) for your path separators in your #include! It’s undefined behavior! Use forward slash (/) only, even on Windows. In summary, used angle brackets (< and >) for the system includes, and use double quotes (") for your personal includes.
117 Chapter 19. The C Preprocessor 118
19.2 Simple Macros
A macro is an identifier that gets expanded to another piece of code before the compiler even sees it. Think of it like a placeholder—when the preprocessor sees one of those identifiers, it replaces it with another value that you’ve defined. We do this with #define (often read “pound define”). Here’s an example:
1 #include
2
3 #define HELLO "Hello, world"
4 #define PI 3.14159
5
6 int main(void)
7 {
8 printf("%s, %f\n", HELLO, PI);
9 }
On lines 3 and 4 we defined a couple macros. Wherever these appear elsewhere in the code (line 8), they’ll be substituted with the defined values. From the C compiler’s perspective, it’s exactly as if we’d written this, instead:
1 #include
2
3 int main(void)
4 {
5 printf("%s, %f\n", "Hello, world", 3.14159);
6 }
See how HELLO was replaced with "Hello, world" and PI was replaced with 3.14159? From the compiler’s perspective, it’s just like those values had appeared right there in the code. Note that the macros don’t have a specific type, per se. Really all that happens is they get replaced wholesale with whatever they’re #defined as. If the resulting C code is invalid, the compiler will puke. You can also define a macro with no value: #define EXTRA_HAPPY
in that case, the macro exists and is defined, but is defined to be nothing. So anyplace it occurs in the text will just be replaced with nothing. We’ll see a use for this later. It’s conventional to write macro names in ALL_CAPS even though that’s not technically required. Overall, this gives you a way to define constant values that are effectively global and can be used any place. Even in those places where a const variable won’t work, e.g. in switch cases and fixed array lengths. That said, the debate rages online whether a typed const variable is better than #define macro in the general case. It can also be used to replace or modify keywords, a concept completely foreign to const, though this practice should be used sparingly.
19.3 Conditional Compilation
It’s possible to get the preprocessor to decide whether or not to present certain blocks of code to the compiler, or just remove them entirely before compilation. We do that by basically wrapping up the code in conditional blocks, similar to if-else statements.
19.3.1 If Defined, #ifdef and #endif First of all, let’s try to compile specific code depending on whether or not a macro is even defined. Chapter 19. The C Preprocessor 119
1 #include
2
3 #define EXTRA_HAPPY
4
5 int main(void)
6 {
7
8 #ifdef EXTRA_HAPPY
9 printf("I'm extra happy!\n");
10 #endif
11
12 printf("OK!\n");
13 }
In that example, we define EXTRA_HAPPY (to be nothing, but it is defined), then on line 8 we check to see if it is defined with an #ifdef directive. If it is defined, the subsequent code will be included up until the #endif. So because it is defined, the code will be included for compilation and the output will be: I'm extra happy! OK!
If we were to comment out the #define, like so: //#define EXTRA_HAPPY
then it wouldn’t be defined, and the code wouldn’t be included in compilation. And the output would just be: OK!
It’s important to remember that these decisions happen at compile time! The code actually get compiled or removed depending on the condition. This is in contrast to a standard if statement that gets evaluated while the program is running.
19.3.2 If Not Defined, #ifndef
There’s also the negative sense of “if defined”: “if not defined”, or #ifndef. We could change the previous example to read to output different things based on whether or not something was defined:
8 #ifdef EXTRA_HAPPY
9 printf("I'm extra happy!\n");
10 #endif
11
12 #ifndef EXTRA_HAPPY
13 printf("I'm just regular\n");
14 #endif
We’ll see a cleaner way to do that in the next section. Tying it all back in to header files, we’ve seen how we can cause header files to only be included one time by wrapping them in preprocessor directives like this: #ifndef MYHEADER_H // First line of myheader.h #define MYHEADER_H
int x = 12;
#endif // Last line of myheader.h
This demonstrates how a macro persists across files and multiple #includes. If it’s not yet defined, let’s define it and compile the whole header file. Chapter 19. The C Preprocessor 120
But the next time it’s included, we see that MYHEADER_H is defined, so we don’t send the header file to the compiler—it gets effectively removed.
19.3.3 #else
But that’s not all we can do! There’s also an #else that we can throw in the mix. Let’s mod the previous example:
8 #ifdef EXTRA_HAPPY
9 printf("I'm extra happy!\n");
10 #else
11 printf("I'm just regular\n");
12 #endif
Now if EXTRA_HAPPY is not defined, it’ll hit the #else clause and print: I'm just regular
19.3.4 General Conditional: #if, #elif
This works very much like the #ifdef and #ifndef directives in that you can also have an #else and the whole thing wraps up with #endif. The only difference is that the constant expression after the #if must evaluate to true (non-zero) for the code in the #if to be compiled. So instead of whether or not something is defined, we want an expression that evaluates to true.
1 #include
2
3 #define HAPPY_FACTOR 1
4
5 int main(void)
6 {
7
8 #if HAPPY_FACTOR == 0
9 printf("I'm not happy!\n");
10 #elif HAPPY_FACTOR == 1
11 printf("I'm just regular\n");
12 #else
13 printf("I'm extra happy!\n");
14 #endif
15
16 printf("OK!\n");
17 }
Again, for the unmatched #if clauses, the compiler won’t even see those lines. For the above code, after the preprocessor gets finished with it, all the compiler sees is:
1 #include
2
3 int main(void)
4 {
5
6 printf("I'm just regular\n");
7
8 printf("OK!\n");
9 }
One hackish thing this is used for is to comment out large numbers of lines quickly1.
1You can’t always just wrap the code in /* */ comments because those won’t nest. Chapter 19. The C Preprocessor 121
If you put an #if 0 (“if false”) at the front of the block to be commented out and an #endif at the end, you can get this effect: #if 0 printf("All this code"); /* is effectively */ printf("commented out"); // by the #if 0 #endif
You might have noticed that there’s no #elifdef or #elifndef directives. How can we get the same effect with #if? That is, what if I wanted this: #ifdef FOO x = 2; #elifdef BAR // ERROR: Not supported by standard C x = 3; #endif
How could I do it? Turns out there’s a preprocessor operator called defined that we can use with an #if statement. These are equivalent: #ifdef FOO #if defined FOO #if defined(FOO) // Parentheses optional
As are these: #ifndef FOO #if !defined FOO #if !defined(FOO) // Parentheses optional
Notice how we can use the standard logical NOT operator (!) for “not defined”. So now we’re back in #if land and we can use #elif with impunity! This broken code: #ifdef FOO x = 2; #elifdef BAR // ERROR: Not supported by standard C x = 3; #endif
can be replaced with: #if defined FOO x = 2; #elif defined BAR x = 3; #endif
19.3.5 Losing a Macro: #undef
If you’ve defined something but you don’t need it any longer, you can undefine it with #undef.
1 #include
2
3 int main(void)
4 {
5 #define GOATS
6
7 #ifdef GOATS
8 printf("Goats detected!\n"); // prints
9 #endif Chapter 19. The C Preprocessor 122
10
11 #undef GOATS // Make GOATS no longer defined
12
13 #ifdef GOATS
14 printf("Goats detected, again!\n"); // doesn't print
15 #endif
16 }
19.4 Built-in Macros
The standard defines a lot of built-in macros that you can test and use for conditional compilation. Let’s look at those here.
19.4.1 Mandatory Macros These are all defined:
Macro Description __DATE__ The date of compilation—like when you’re compiling this file—in Mmm dd yyyy format __TIME__ The time of compilation in hh:mm:ss format __FILE__ A string containing this file’s name __LINE__ The line number of the file this macro appears on __func__ The name of the function this appears in, as a string2 __STDC__ Defined with 1 if this is a standard C compiler __STDC_HOSTED__ This will be 1 if the compiler is a hosted implementation3, otherwise 0 __STDC_VERSION__ This version of C, a constant long int in the form yyyymmL, e.g. 201710L
Let’s put these together.
1 #include
2
3 int main(void)
4 {
5 printf("This function: %s\n", __func__);
6 printf("This file: %s\n", __FILE__);
7 printf("This line: %d\n", __LINE__);
8 printf("Compiled on: %s %s\n", __DATE__, __TIME__);
9 printf("C Version: %ld\n", __STDC_VERSION__);
10 }
The output on my system is: This function: main This file: foo.c This line: 7 Compiled on: Nov 23 2020 17:16:27 C Version: 201710
__FILE__, __func__ and __LINE__ are particularly useful to report error conditions in messages to
2This isn’t really a macro—it’s technically an identifier. But it’s the only predefined identifier and it feels very macro-like, so I’m including it here. Like a rebel. 3A hosted implementation basically means you’re running the full C standard, probably on an operating system of some kind. Which you probably are. If you’re running on bare metal in some kind of embedded system, you’re probably on a standalone implementation. Chapter 19. The C Preprocessor 123
developers. The assert() macro in
19.4.2 Optional Macros Your implementation might define these, as well. Or it might not.
Macro Description __STDC_ISO_10646__ If defined, wchar_t holds Unicode values, otherwise something else __STDC_MB_MIGHT_NEQ_WC__ A 1 indicates that the values in multibyte characters might not map equally to values in wide characters __STDC_UTF_16__ A 1 indicates that the system uses UTF-16 encoding in type char16_t __STDC_UTF_32__ A 1 indicates that the system uses UTF-32 encoding in type char32_t __STDC_ANALYZABLE__ A 1 indicates the code is analyzable4 __STDC_IEC_559__ 1 if IEEE-754 (aka IEC 60559) floating point is supported __STDC_IEC_559_COMPLEX__ 1 if IEC 60559 complex floating point is supported __STDC_LIB_EXT1__ 1 if this implementation supports a variety of “safe” alternate standard library functions (they have _s suffixes on the name) __STDC_NO_ATOMICS__ 1 if this implementation does not support _Atomic or
19.5 Macros with Arguments
Macros are more powerful than simple substitution, though. You can set them up to take arguments that are substituted in, as well. A question often arises for when to use parameterized macros versus functions. Short answer: use func- tions. But you’ll see lots of macros in the wild and in the standard library. People tend to use them for short, mathy things, and also for features that might change from platform to platform. You can define different keywords for one platform or another.
19.5.1 Macros with One Argument Let’s start with a simple one that squares a number:
1 #include
2
3 #define SQR(x) x * x // Not quite right, but bear with me
4
5 int main(void)
6 {
4OK, I know that was a cop-out answer. Basically there’s an optional extension compilers can implement wherein they agree to limit certain types of undefined behavior so that the C code is more amenable to static code analysis. It is unlikely you’ll need to use this. Chapter 19. The C Preprocessor 124
7 printf("%d\n", SQR(12)); // 144
8 }
What that’s saying is “everywhere you see SQR with some value, replace it with that value times itself”. So line 7 will be changed to:
7 printf("%d\n", 12 * 12); // 144
which C comfortably converts to 144. But we’ve made an elementary error in that macro, one that we need to avoid. Let’s check it out. What if we wanted to compute SQR(3 + 4)? Well, 3 + 4 = 7, so we must want to compute 72 = 49. That’s it; 49—final answer. Let’s drop it in our code and see that we get… 19?
7 printf("%d\n", SQR(3 + 4)); // 19!!??
What happened? If we follow the macro expansion, we get
7 printf("%d\n", 3 + 4 * 3 + 4); // 19!
Oops! Since multiplication takes precedence, we do the 4 × 3 = 12 first, and get 3 + 12 + 4 = 19. Not what we were after. So we have to fix this to make it right. This is so common that you should automatically do it every time you make a parameterized math macro! The fix is easy: just add some parentheses!
3 #define SQR(x) (x) * (x) // Better... but still not quite good enough!
And now our macro expands to:
7 printf("%d\n", (3 + 4) * (3 + 4)); // 49! Woo hoo!
But we actually still have the same problem which might manifest if we have a higher-precedence operator than multiply (*) nearby. So the safe, proper way to put the macro together is to wrap the whole thing in additional parentheses, like so:
3 #define SQR(x) ((x) * (x)) // Good!
Just make it a habit to do that when you make a math macro and you can’t go wrong.
19.5.2 Macros with More than One Argument You can stack these things up as much as you want: #define TRIANGLE_AREA(w, h) (0.5 * (w) * (h))
Let’s do some macros that solve for 푥 using the quadratic formula. Just in case you don’t have it on the top of your head, it says for equations of the form: 푎푥2 + 푏푥 + 푐 = 0 you can solve for 푥 with the quadratic formula: √ − 푏 ± 푏2 − 4푎푐 푥 = 2푎 Which is crazy. Also notice the plus-or-minus (±) in there, indicating that there are actually two solutions. So let’s make macros for both: Chapter 19. The C Preprocessor 125
#define QUADP(a, b, c) ((-(b) + sqrt((b) * (b) - 4 * (a) * (c))) / (2 * (a))) #define QUADM(a, b, c) ((-(b) - sqrt((b) * (b) - 4 * (a) * (c))) / (2 * (a)))
So that gets us some math. But let’s define one more that we can use as arguments to printf() to print both answers. // macro replacement // |------| |------| #define QUAD(a, b, c) QUADP(a, b, c), QUADM(a, b, c)
That’s just a couple values separated by a comma—and we can use that as a “combined” argument of sorts to printf() like this: printf("x = %f or x = %f\n", QUAD(2, 10, 5));
Let’s put it together into some code:
1 #include
2 #include
3
4 #define QUADP(a, b, c) ((-(b) + sqrt((b) * (b) - 4 * (a) * (c))) / (2 * (a)))
5 #define QUADM(a, b, c) ((-(b) - sqrt((b) * (b) - 4 * (a) * (c))) / (2 * (a)))
6 #define QUAD(a, b, c) QUADP(a, b, c), QUADM(a, b, c)
7
8 int main(void)
9 {
10 printf("2*x^2 + 10*x + 5 = 0\n");
11 printf("x = %f or x = %f\n", QUAD(2, 10, 5));
12 }
And this gives us the output: 2*x^2 + 10*x + 5 = 0 x = -0.563508 or x = -4.436492
Plugging in either of those values gives us roughly zero (a bit off because the numbers aren’t exact): 2 × −0.5635082 + 10 × −0.563508 + 5 ≈ 0.000003
19.5.3 Macros with Variable Arguments
There’s also a way to have a variable number of arguments passed to a macro, using ellipses (...) after the known, named arguments. When the macro is expanded, all of the extra arguments will be in a comma- separated list in the __VA_ARGS__ macro, and can be replaced from there:
1 #include
2
3 // Combine the first two arguments to a single number,
4 // then have a commalist of the rest of them:
5
6 #define X(a, b, ...) (10*(a) + 20*(b)), __VA_ARGS__
7
8 int main(void)
9 {
10 printf("%d %f %s %d\n", X(5, 4, 3.14, "Hi!", 12));
11 }
The substitution that takes place on line 10 would be:
10 printf("%d %f %s %d\n", (10*(5) + 20*(4)), 3.14, "Hi!", 12);
for output: Chapter 19. The C Preprocessor 126
130 3.140000 Hi! 12
You can also “stringify” __VA_ARGS__ by putting a # in front of it: #define X(...) #__VA_ARGS__
printf("%s\n", X(1,2,3)); // Prints "1, 2, 3"
19.5.4 Stringification
Already mentioned, just above, you can turn any argument into a string by preceding it with a # in the replacement text. For example, we could print anything as a string with this macro and printf(): #define STR(x) #x
printf("%s\n", STR(3.14159));
In that case, the substitution leads to: printf("%s\n", "3.14159");
Let’s see if we can use this to greater effect so that we can pass any int variable name into a macro, and have it print out it’s name and value.
1 #include
2
3 #define PRINT_INT_VAL(x) printf("%s = %d\n", #x, x)
4
5 int main(void)
6 {
7 int a = 5;
8
9 PRINT_INT_VAL(a); // prints "a = 5"
10 }
On line 9, we get the following macro replacement:
1 printf("%s = %d\n", "a", 5);
19.5.5 Concatenation
We can concatenate two arguments together with ##, as well. Fun times! #define CAT(a, b) a ## b
printf("%f\n", CAT(3.14, 1592)); // 3.141592
19.6 Multiline Macros
It’s possible to continue a macro to multiple lines if you escape the newline with a backslash (\). Let’s write a multiline macro that prints numbers from 0 to the product of the two arguments passed in.
1 #include
2
3 #define PRINT_NUMS_TO_PRODUCT(a, b) { \
4 int product = (a) * (b); \
5 for (int i = 0; i < product; i++) { \
6 printf("%d\n", i); \
7 }\ Chapter 19. The C Preprocessor 127
8 }
9
10 int main(void)
11 {
12 PRINT_NUMS_TO_PRODUCT(2, 4); // Outputs numbers from 0 to 7
13 }
A couple things to note there: • Escapes at the end of every line except the last one to indicate that the macro continues. • Though not strictly necessary, I wrapped the whole thing in curly braces. This did two things: 1. Made it look nice. 2. Made a new block scope for my product variable so it wouldn’t conflict with any other existing variables at the outer block scope.
19.7 Example: An Assert Macro
Adding asserts to your code is a good way to catch conditions that you think shouldn’t happen. C provides assert() functionality. It checks a condition, and if it’s false, the program bombs out telling you the file and line number on which the assertion failed. But this is wanting. 1. First of all, you can’t specify an additional message with the assert. 2. Secondly, there’s no easy on-off switch for all the asserts. We can address the first with macros. Basically, when I have this code: ASSERT(x < 20, "x must be under 20");
I want something like this to happen (assuming the ASSERT() is on line 220 of foo.c): if (!(x < 20)) { fprintf(stderr, "foo.c:220: assertion x < 20 failed: "); fprintf(stderr, "x must be under 20\n"); exit(1); }
We can get the filename out of the __FILE__ macro, and the line number from __LINE__. The message is already a string, but x < 20 is not, so we’ll have to stringify it with #. We can make a multiline macro by using backslash escapes at the end of the line. #define ASSERT(c, m) \ {\ if (!(c)) { \ fprintf(stderr, __FILE__ ":%d: assertion %s failed: %s\n", \ __LINE__, #c, m); \ exit(1); \ }\ }
(It looks a little weird with __FILE__ out front like that, but remember it is a string literal, and string literals next to each other are automagically concatenated. __LINE__ on the other hand, it’s just an int.) And that works! If I run this: int x = 30;
ASSERT(x < 20, "x must be under 20");
I get this output: Chapter 19. The C Preprocessor 128
foo.c:23: assertion x < 20 failed: x must be under 20 Very nice! The only thing left is a way to turn it on and off, and we could do that with conditional compilation. Here’s the complete example:
1 #include
2 #include
3
4 #define ASSERT_ENABLED 1
5
6 #if ASSERT_ENABLED
7 #define ASSERT(c, m) \
8 {\
9 if (!(c)) { \
10 fprintf(stderr, __FILE__ ":%d: assertion %s failed: %s\n", \
11 __LINE__, #c, m); \
12 exit(1); \
13 }\
14 }
15 #else
16 #define ASSERT(c, m) // Empty macro if not enabled
17 #endif
18
19 int main(void)
20 {
21 int x = 30;
22
23 ASSERT(x < 20, "x must be under 20");
24 }
This has the output: foo.c:23: assertion x < 20 failed: x must be under 20
19.8 The #error Directive
This directive causes the compiler to error out as soon as it sees it. Commonly, this is used inside a conditional to prevent compilation unless some prerequisites are met: #ifndef __STDC_IEC_559__ #error I really need IEEE-754 floating point to compile. Sorry! #endif
Some compilers have a non-standard complementary #warning directive that will output a warning but not stop compilation, but this is not in the C11 spec.
19.9 The #pragma Directive
This is one funky directive, short for “pragmatic”. You can use it to do… well, anything your compiler supports you doing with it. Basically the only time you’re going to add this to your code is if some documentation tells you to do so. Chapter 19. The C Preprocessor 129
19.9.1 Non-Standard Pragmas
Here’s one non-standard example of using #pragma to cause the compiler to execute a for loop in parallel with multiple threads (if the compiler supports the OpenMP5 extension): #pragma omp parallel for for (int i = 0; i < 10; i++) { ... }
There are all kinds of #pragma directives documented across all four corners of the globe. All unrecognized #pragmas are ignored by the compiler.
19.9.2 Standard Pragmas
There are also a few standard ones, and these start with STDC, and follow the same form: #pragma STDC pragma_name on-off
The on-off portion can be either ON, OFF, or DEFAULT. And the pragma_name can be one of these:
Pragma Name Description FP_CONTRACT Allow floating point expressions to be contracted into a single operation to avoid rounding errors that might occur from multiple operations. FENV_ACCESS Set to ON if you plan to access the floating point status flags. If OFF, the compiler might perform optimizations that cause the values in the flags to be inconsistent or invalid. CX_LIMITED_RANGE Set to ON to allow the compiler to skip overflow checks when performing complex arithmetic. Defaults to OFF.
For example: #pragma STDC FP_CONTRACT OFF #pragma STDC CX_LIMITED_RANGE ON
As for CX_LIMITED_RANGE, the spec points out: The purpose of the pragma is to allow the implementation to use the formulas: (푥 + 푖푦) × (푢 + 푖푣) = (푥푢 − 푦푣) + 푖(푦푢 + 푥푣) (푥 + 푖푦)/(푢 + 푖푣) = [(푥푢 + 푦푣) + 푖(푦푢 − 푥푣)]/(푢2 + 푣2) |푥 + 푖푦| = √푥2 + 푦2 where the programmer can determine they are safe.
19.9.3 _Pragma Operator This is another way to declare a pragma that you could use in a macro. These are equivalent: #pragma "Unnecessary" quotes _Pragma("\"Unnecessary\" quotes")
This can be used in a macro, if need be:
5https://www.openmp.org/ Chapter 19. The C Preprocessor 130
#define PRAGMA(x) _Pragma(#x)
19.10 The #line Directive
This allows you to override the values for __LINE__ and __FILE__. If you want. I’ve never wanted to do this, but in K&R2, they write: For the benefit of other preprocessors that generate C programs […] So maybe there’s that. To override the line number to, say 300: #line 300 and __LINE__ will keep counting up from there. To override the line number and the filename: #line 300 "newfilename"
19.11 The Null Directive
A # on a line by itself is ignored by the preprocessor. Now, to be entirely honest, I don’t know what the use case is for this. I’ve seen examples like this: #ifdef FOO # #else printf("Something"); #endif which is just cosmetic; the line with the solitary # can be deleted with no ill effect. Or maybe for cosmetic consistency, like this: # #ifdef FOO x = 2; #endif # #if BAR == 17 x = 12; #endif #
But, with respect to cosmetics, that’s just ugly. Another post mentions elimination of comments—that in GCC, a comment after a # will not be seen by the compiler. Which I don’t doubt, but the specification doesn’t seem to say this is standard behavior. My searches for rationale aren’t bearing much fruit. So I’m going to just say this is some good ol’ fashioned C esoterica. Chapter 20
structs II: More Fun with structs
Turns out there’s a lot more you can do with structs than we’ve talked about, but it’s just a big pile of miscellaneous things. So we’ll throw them in this chapter. If you’re good with struct basics, you can round out your knowledge here.
20.1 Initializers of Nested structs and Arrays
Remember how you could initialize structure members along these lines? struct foo x = {.a=12, .b=3.14};
Turns out we have more power in these initializers than we’d originally shared. Exciting! For one thing, if you have a nested substructure like the following, you can initialize members of that substructure by following the variable names down the line: struct foo x = {.a.b.c=12};
Let’s look at an example:
1 #include
2
3 struct cabin_information {
4 int window_count;
5 int o2level;
6 };
7
8 struct spaceship {
9 char *manufacturer;
10 struct cabin_information ci;
11 };
12
13 int main(void)
14 {
15 struct spaceship s = {
16 .manufacturer="General Products",
17 .ci.window_count = 8, // <-- NESTED INITIALIZER!
18 .ci.o2level = 21
19 };
20
21 printf("%s: %d seats, %d%% oxygen\n",
22 s.manufacturer, s.ci.window_count, s.ci.o2level);
23 }
131 Chapter 20. structs II: More Fun with structs 132
Check out lines 16-17! That’s where we’re initializing members of the struct cabin_information in the definition of s, our struct spaceship. And here is another option for that same initializer—this time we’ll do something more standard-looking, but either approach works:
15 struct spaceship s = {
16 .manufacturer="General Products",
17 .ci={
18 .window_count = 8,
19 .o2level = 21
20 }
21 };
Now, as if the above information isn’t spectacular enough, we can also mix in array initializers in there, too. Let’s change this up to get an array of passenger information in there, and we can check out how the initializers work in there, too.
1 #include
2
3 struct passenger {
4 char *name;
5 int covid_vaccinated; // Boolean
6 };
7
8 #define MAX_PASSENGERS 8
9
10 struct spaceship {
11 char *manufacturer;
12 struct passenger passenger[MAX_PASSENGERS];
13 };
14
15 int main(void)
16 {
17 struct spaceship s = {
18 .manufacturer="General Products",
19 .passenger = {
20 // Initialize a field at a time
21 [0].name = "Gridley, Lewis",
22 [0].covid_vaccinated = 0,
23
24 // Or all at once
25 [7] = {.name="Brown, Teela", .covid_vaccinated=1},
26 }
27 };
28
29 printf("Passengers for %s ship:\n", s.manufacturer);
30
31 for (int i = 0; i < MAX_PASSENGERS; i++)
32 if (s.passenger[i].name != NULL)
33 printf(" %s (%svaccinated)\n",
34 s.passenger[i].name,
35 s.passenger[i].covid_vaccinated? "": "not ");
36 } Chapter 20. structs II: More Fun with structs 133
20.2 Anonymous structs
These are “the struct with no name”. We also mention these in the typedef section, but we’ll refresh here. Here’s a regular struct: struct animal { char *name; int leg_count, speed; };
And here’s the anonymous equivalent: struct { // <-- No name! char *name; int leg_count, speed; };
Okaaaaay. So we have a struct, but it has no name, so we have no way of using it later? Seems pretty pointless. Admittedly, in that example, it is. But we can still make use of it a couple ways. One is rare, but since the anonymous struct represents a type, we can just put some variable names after it and use them. struct { // <-- No name! char *name; int leg_count, speed; } a, b, c; // 3 variables of this struct type a.name = "antelope"; c.leg_count = 4; // for example
But that’s still not that useful. Far more common is use of anonymous structs with a typedef so that we can use it later (e.g. to pass variables to functions). typedef struct { // <-- No name! char *name; int leg_count, speed; } animal; // New type: animal animal a, b, c; a.name = "antelope"; c.leg_count = 4; // for example
Personally, I don’t use many anonymous structs. I think it’s more pleasant to see the entire struct animal before the variable name in a declaration. But that’s just, like, my opinion, man.
20.3 Self-Referential structs
For any graph-like data structure, it’s useful to be able to have pointers to the connected nodes/vertices. But this means that in the definition of a node, you need to have a pointer to a node. It’s chicken and eggy! But it turns out you can do this in C with no problem whatsoever. For example, here’s a linked list node: Chapter 20. structs II: More Fun with structs 134
struct node { int data; struct node *next; };
It’s important to note that next is a pointer. This is what allows the whole thing to even build. Even though the compiler doesn’t know what the entire struct node looks like yet, all pointers are the same size. Here’s a cheesy linked list program to test it out:
1 #include
2 #include
3
4 struct node {
5 int data;
6 struct node *next;
7 };
8
9 int main(void)
10 {
11 struct node *head;
12
13 // Hackishly set up a linked list (11)->(22)->(33)
14 head = malloc(sizeof(struct node));
15 head->data = 11;
16 head->next = malloc(sizeof(struct node));
17 head->next->data = 22;
18 head->next->next = malloc(sizeof(struct node));
19 head->next->next->data = 33;
20 head->next->next->next = NULL;
21
22 // Traverse it
23 for (struct node *cur = head; cur != NULL; cur = cur->next) {
24 printf("%d\n", cur->data);
25 }
26 }
Running that prints: 11 22 33
20.4 Flexible Array Members
Back in the good old days, when people carved C code out of wood, some folks thought would be neat if they could allocate structs that had variable length arrays at the end of them. I want to be clear that the first part of the section is the old way of doing things, and we’re going to do things the new way after that. For example, maybe you could define a struct for holding strings and the length of that string. It would have a length and an array to hold the data. Maybe something like this: struct len_string { int length; char data[8]; };
But that has 8 hardcoded as the maximum length of a string, and that’s not much. What if we did something Chapter 20. structs II: More Fun with structs 135 clever and just malloc()d some extra space at the end after the struct, and then let the data overflow into that space? Let’s do that, and then allocate another 40 bytes on top of it: struct len_string *s = malloc(sizeof *s + 40);
Because data is the last field of the struct, if we overflow that field, it runs out into space that we already allocated! For this reason, this trick only works if the short array is the last field in the struct. // Copy more than 8 bytes! strcpy(s->data, "Hello, world!"); // Won't crash. Probably.
In fact, there was a common compiler workaround for doing this, where you’d allocate a zero length array at the end: struct len_string { int length; char data[0]; };
And then every extra byte you allocated was ready for use in that string. Because data is the last field of the struct, if we overflow that field, it runs out into space that we already allocated! // Copy more than 8 bytes! strcpy(s->data, "Hello, world!"); // Won't crash. Probably.
But, of course, actually accessing the data beyond the end of that array is undefined behavior! In these modern times, we no longer deign to resort to such savagery. Luckily for us, we can still get the same effect with C99 and later, but now it’s legal. Let’s just change our above definition to have no size for the array1: struct len_string { int length; char data[]; };
Again, this only works if the flexible array member is the last field in the struct. And then we can allocate all the space we want for those strings by malloc()ing larger than the struct len_string, as we do in this example that makes a new struct len_string from a C string: struct len_string *len_string_from_c_string(char *s) { int len = strlen(s);
// Allocate "len" more bytes than we'd normally need struct len_string *ls = malloc(sizeof *ls + len);
ls->length = len;
// Copy the string into those extra bytes memcpy(ls->data, s, len);
return ls; }
1Technically we say that it has an incomplete type. Chapter 20. structs II: More Fun with structs 136
20.5 Padding Bytes
Beware that C is allowed to add padding bytes within or after a struct as it sees fit. You can’t trust that they will be directly adjacent in memory2. Let’s take a look at this program. We output two numbers. One is the sum of the sizeofs the individual field types. The other is the sizeof the entire struct. One would expect them to be the same. The size of the total is the size of the sum of its parts, right?
1 #include
2
3 struct foo {
4 int a;
5 char b;
6 int c;
7 char d;
8 };
9
10 int main(void)
11 {
12 printf("%zu\n", sizeof(int) + sizeof(char) + sizeof(int) + sizeof(char));
13 printf("%zu\n", sizeof(struct foo));
14 }
But on my system, this outputs: 10 16
They’re not the same! The compiler has added 6 bytes of padding to help it be more performant. Maybe you got different output with your compiler, but unless you’re forcing it, you can’t be sure there’s no padding.
20.6 offsetof
In the previous section, we saw that the compiler could inject padding bytes at will inside a structure. What if we needed to know where those were? We can measure it with offsetof, defined in
1 #include
2 #include
3
4 struct foo {
5 int a;
6 char b;
7 int c;
8 char d;
9 };
10
11 int main(void)
12 {
13 printf("%zu\n", offsetof(struct foo, a));
14 printf("%zu\n", offsetof(struct foo, b));
15 printf("%zu\n", offsetof(struct foo, c));
16 printf("%zu\n", offsetof(struct foo, d));
17 }
2Though some compilers have options to force this to occur—search for __attribute__((packed)) to see how to do this with GCC. Chapter 20. structs II: More Fun with structs 137
For me, this outputs: 0 4 8 12
indicating that we’re using 4 bytes for each of the fields. It’s a little weird, because char is only 1 byte, right? The compiler is putting 3 padding bytes after each char so that all the fields are 4 bytes long. Presumably this will run faster on my CPU.
20.7 Fake OOP
There’s a slightly abusive thing that’s sort of OOP-like that you can do with structs. Since the pointer to the struct is the same as a pointer to the first element of the struct, you can freely cast a pointer to the struct to a pointer to the first element. What this means is that we can set up a situation like this: struct parent { int a, b; };
struct child { struct parent super; // MUST be first int c, d; };
Then we are able to pass a pointer to a struct child to a function that expects either that or a pointer to a struct parent! Because struct parent super is the first item in the struct child, a pointer to any struct child is the same as a pointer to that super field3. Let’s set up an example here. We’ll make structs as above, but then we’ll pass a pointer to a struct child to a function that needs a pointer to a struct parent… and it’ll still work.
1 #include
2
3 struct parent {
4 int a, b;
5 };
6
7 struct child {
8 struct parent super; // MUST be first
9 int c, d;
10 };
11
12 // Making the argument `void*` so we can pass any type into it
13 // (namely a struct parent or struct child)
14 void print_parent(void *p)
15 {
16 // Expects a struct parent--but a struct child will also work
17 // because the pointer points to the struct parent in the first
18 // field:
19 struct parent *self = p;
20
21 printf("Parent: %d, %d\n", self->a, self->b);
22 }
3super isn’t a keyword, incidentally. I’m just stealing some OOP terminology. Chapter 20. structs II: More Fun with structs 138
23
24 void print_child(struct child *self)
25 {
26 printf("Child: %d, %d\n", self->c, self->d);
27 }
28
29 int main(void)
30 {
31 struct child c = {.super.a=1, .super.b=2, .c=3, .d=4};
32
33 print_child(&c);
34 print_parent(&c); // Also works even though it's a struct child!
35 }
See what we did on the last line of main()? We called print_parent() but passed a struct child* as the argument! Even though print_parent() needs the argument to point to a struct parent, we’re getting away with it because the first field in the struct child is a struct parent. Again, this works because a pointer to a struct has the same value as a pointer to the first field in that struct.
20.8 Bit-Fields
In my experience, these are rarely used, but you might see them out there from time to time, especially in lower-level applications that pack bits together into larger spaces. Let’s take a look at some code to demonstrate a use case:
1 #include
2
3 struct foo {
4 unsigned int a;
5 unsigned int b;
6 unsigned int c;
7 unsigned int d;
8 };
9
10 int main(void)
11 {
12 printf("%zu\n", sizeof(struct foo));
13 }
For me, this prints 16. Which makes sense, since unsigneds are 4 bytes on my system. But what if we knew that all the values that were going to be stored in a and b could be stored in 5 bits, and the values in c, and d could be stored in 3 bits? That’s only a total 16 bits. Why have 128 bits reserved for them if we’re only going to use 16? Well, we can tell C to pretty-please try to pack these values in. We can specify the maximum number of bits that values can take (from 1 up the size of the containing type). We do this by putting a colon after the field name, followed by the field width in bits.
3 struct foo {
4 unsigned int a:5;
5 unsigned int b:5;
6 unsigned int c:3;
7 unsigned int d:3;
8 };
Now when I ask C how big my struct foo is, it tells me 4! It was 16 bytes, but now it’s only 4. It has “packed” those 4 values down into 4 bytes, which is a four-fold memory savings. Chapter 20. structs II: More Fun with structs 139
The tradeoff is, of course, that the 5-bit fields can only hold values from 0-31 and the 3-bit fields can only hold values from 0-7. But life’s all about compromise, after all.
20.8.1 Non-Adjacent Bit-Fields A gotcha: C will only combine adjacent bit-fields. If they’re interrupted by non-bit-fields, you get no savings: struct foo { // sizeof(struct foo) == 16 (for me) unsigned int a:1; // since a is not adjacent to c. unsigned int b; unsigned int c:1; unsigned int d; };
In that example, since a is not adjacent to c, they are both “packed” in their own ints. So we have one int each for a, b, c, and d. Since my ints are 4 bytes, that’s a grand total of 16 bytes. A quick rearrangement yields some space savings from 16 bytes down to 12 bytes (on my system): struct foo { // sizeof(struct foo) == 12 (for me) unsigned int a:1; unsigned int c:1; unsigned int b; unsigned int d; };
And now, since a is next to c, the compiler puts them together into a single int. So we have one int for a combined a and c, and one int each for b and d. For a grand total of 3 ints, or 12 bytes. Put all your bitfields together to get the compiler to combine them.
20.8.2 Signed or Unsigned ints
If you just declare a bit-field to be int, the different compilers will treat it as signed or unsigned. Just like the situation with char. Be specific about the signedness when using bit-fields.
20.8.3 Unnamed Bit-Fields In some specific circumstances, you might need to reserve some bits for hardware reasons, but not need to use them in code. For example, let’s say you have a byte where the top 2 bits have a meaning, the bottom 1 bit has a meaning, but the middle 5 bits do not get used by you4. We could do something like this: struct foo { unsigned char a:2; unsigned char dummy:5; unsigned char b:1; };
And that works—in our code we use a and b, but never dummy. It’s just there to eat up 5 bits to make sure a and b are in the “required” (by this contrived example) positions within the byte. C allows us a way to clean this up: unnamed bit-fields. You can just leave the name (dummy) out in this case, and C is perfectly happy for the same effect:
4Assuming 8-bit chars, i.e. CHAR_BIT == 8. Chapter 20. structs II: More Fun with structs 140 struct foo { unsigned char a:2; unsigned char :5; // <-- unnamed bit-field! unsigned char b:1; };
20.8.4 Zero-Width Unnamed Bit-Fields
Some more esoterica out here… Let’s say you were packing bits into an unsigned int, and you needed some adjacent bit-fields to pack into the next unsigned int. That is, if you do this: struct foo { unsigned int a:1; unsigned int b:2; unsigned int c:3; unsigned int d:4; }; the compiler packs all those into a single unsigned int. But what if you needed a and b in one int, and c and d in a different one? There’s a solution for that: put an unnamed bit-field of width 0 where you want the compiler to start anew with packing bits in a different int: struct foo { unsigned int a:1; unsigned int b:2; unsigned int :0; // <--Zero-width unnamed bit-field unsigned int c:3; unsigned int d:4; };
It’s analogous to an explicit page break in a word processor. You’re telling the compiler, “Stop packing bits in this unsigned, and start packing them in the next one.” By adding the zero-width unnamed bit field in that spot, the compiler puts a and b in one unsigned int, and c and d in another unsigned int. Two total, for a size of 8 bytes on my system (unsigned ints are 4 bytes each).
20.9 Unions
These are basically just like structs, except the fields overlap in memory. The union will be only large enough for the largest field, and you can only use one field at a time. It’s a way to reuse the same memory space for different types of data. You declare them just like structs, except it’s union. Take a look at this: union foo { int a, b, c, d, e, f; float g, h; char i, j, k, l; };
Now, that’s a lot of fields. If this were a struct, my system would tell me it took 36 bytes to hold it all. But it’s a union, so all those fields overlap in the same stretch of memory. The biggest one is int (or float), taking up 4 bytes on my system. And, indeed, if I ask for the sizeof the union foo, it tells me 4! Chapter 20. structs II: More Fun with structs 141
The tradeoff is that you can only portably use one of those fields at a time. If you try to read from a field that was not the last one written to, the behavior is unspecified. Let’s take that crazy union and first store an int in it, then a float. Then we’ll print out the int again to see what’s in there—even though, since it wasn’t the last value we stored, the result is unspecified.
1 #include
2
3 union foo {
4 int a, b, c, d, e, f;
5 float g, h;
6 char i, j, k, l;
7 };
8
9 int main(void)
10 {
11 union foo x;
12
13 x.a = 12;
14 printf("%d\n", x.a); // OK--x.a was the last thing we stored into
15
16 x.g = 3.141592;
17 printf("%f\n", x.g); // OK--x.g was the last thing we stored into
18
19 printf("%d\n", x.a); // Unspecified behavior!
20 }
On my machine, this prints: 12 3.141592 1078530008
Probably deep down the decimal value 1078530008 is probably the same pattern of bits as 3.141592, but the spec makes no guarantees about this.
20.9.1 Pointers to unions
If you have a pointer to a union, you can cast that pointer to any of the types of the fields in that union and get the values out that way. In this example, we see that the union has ints and floats in it. And we get pointers to the union, but we cast them to int* and float* types (the cast silences compiler warnings). And then if we dereference those, we see that they have the values we stored directly in the union.
1 #include
2
3 union foo {
4 int a, b, c, d, e, f;
5 float g, h;
6 char i, j, k, l;
7 };
8
9 int main(void)
10 {
11 union foo x;
12
13 int *foo_int_p = (int *)&x;
14 float *foo_float_p = (float *)&x;
15
16 x.a = 12; Chapter 20. structs II: More Fun with structs 142
17 printf("%d\n", x.a); // 12
18 printf("%d\n", *foo_int_p); // 12, again
19
20 x.g = 3.141592;
21 printf("%f\n", x.g); // 3.141592
22 printf("%f\n", *foo_float_p); // 3.141592, again
23 }
The reverse is also true. If we have a pointer to a type inside the union, we can cast that to a pointer to the union and access its members. union foo x; int *foo_int_p = (int *)&x; // Pointer to int field union foo *p = (union foo *)foo_int_p; // Back to pointer to union
p->a = 12; // This line the same as... x.a = 12; // this one.
All this just lets you know that, under the hood, all these values in a union start at the same place in memory, and that’s the same as where the entire union is. Chapter 21
Characters and Strings II
We’ve talked about how char types are actually just small integer types… but it’s the same for a character in single quotes. But a string in double quotes is type const char *. Turns out there are few more types of strings and characters, and it leads down one of the most infamous rabbit holes in the language: the whole multibyte/wide/Unicode/localization thingy. We’re going to peer into that rabbit hole, but not go in. …Yet!
21.1 Escape Sequences
We’re used to strings and characters with regular letters, punctuation, and numbers: char *s = "Hello!"; char t = 'c';
But what if we want some special characters in there that we can’t type on the keyboard because they don’t exist (e.g. “€”), or even if we want a character that’s a single quote? We clearly can’t do this: char t = ''';
To do these things, we use something called escape sequences. These are the backslash character (\) followed by another character. The two (or more) characters together have special meaning. For our single quote character example, we can put an escape (that is, \) in front of the central single quote to solve it: char t = '\'';
Now C knows that \' means just a regular quote we want to print, not the end of the character sequence. You can say either “backslash” or “escape” in this context (“escape that quote”) and C devs will know what you’re talking about. Also, “escape” in this context is different than your Esc key or the ASCII ESC code.
21.1.1 Frequently-used Escapes In my humble opinion, these escape characters make up 99.2%1 of all escapes.
Code Description \n Newline character—when printing, continue subsequent output on the next line \' Single quote—used for a single quote character constant \" Double quote—used for a double quote in a string literal
1I just made up that number, but it’s probably not far off
143 Chapter 21. Characters and Strings II 144
Code Description \\ Backslash—used for a literal \ in a string or character
Here are some examples of the escapes and what they output when printed. printf("Use \\n for newline\n"); // Use \n for newline printf("Say \"hello\"!\n"); // Say "hello"! printf("%c\n", '\''); // '
21.1.2 Rarely-used Escapes But there are more escapes! You just don’t see these as often.
Code Description \a Alert. This makes the terminal make a sound or flash, or both! \b Backspace. Moves the cursor back a character. Doesn’t delete the character. \f Formfeed. This moves to the next “page”, but that doesn’t have much modern meaning. On my system, this behaves like \v. \r Return. Move to the beginning of the same line. \t Horizontal tab. Moves to the next horizontal tab stop. On my machine, this lines up on columns that are multiples of 8, but YMMV. \v Vertical tab. Moves to the next vertical tab stop. On my machine, this moves to the same column on the next line. \? Literal question mark. Sometimes you need this to avoid trigraphs, as shown below.
21.1.2.1 Single Line Status Updates
A use case for \b or \r is to show status updates that appear on the same line on the screen and don’t cause the display to scroll. Here’s an example that does a countdown from 10. (Note this makes use of the non-standard POSIX function sleep() from
1 #include
2 #include
3
4 int main(void)
5 {
6 for (int i = 10; i >= 0; i--) {
7 printf("\rT minus %d second%s... \b", i, i != 1? "s": "");
8
9 fflush(stdout); // Force output to update
10
11 sleep(1); // Delay 1 second
12 }
13
14 printf("\rLiftoff! \n");
15 }
Quite a few things are happening on line 7. First of all, we lead with a \r to get us to the beginning of the current line, then we overwrite whatever’s there with the current countdown. (There’s ternary operator out there to make sure we print 1 second instead of 1 seconds.) Also, there’s a space after the ... That’s so that we properly overwrite the last . when i drops from 10 to 9 and we get a column narrower. Try it without the space to see what I mean. And we wrap it up with a \b to back up over that space so the cursor sits at the exact end of the line in an aesthetically-pleasing way. Chapter 21. Characters and Strings II 145
Note that line 14 also has a lot of spaces at the end to overwrite the characters that were already there from the countdown. Finally, we have a weird fflush(stdout) in there, whatever that means. Short answer is that most termi- nals are line buffered by default, meaning they don’t actually display anything until a newline character is encountered. Since we don’t have a newline (we just have \r), without this line, the program would just sit there until Liftoff! and then print everything all in one instant. fflush() overrides this behavior and forces output to happen right now.
21.1.2.2 The Question Mark Escape Why bother with this? After all, this works just fine: printf("Doesn't it?\n");
And it works fine with the escape, too: printf("Doesn't it\?\n"); // Note \?
So what’s the point??! Let’s get more emphatic with another question mark and an exclamation point: printf("Doesn't it??!\n");
When I compile this, I get this warning: foo.c: In function ‘main’: foo.c:5:23: warning: trigraph ??! converted to | [-Wtrigraphs] 5 | printf("Doesn't it??!\n"); |
And running it gives this unlikely result: Doesn't it|
So trigraphs? What the heck is this??! I’m sure we’ll revisit this dusty corner of the language later, but the short of it is the compiler looks for certain triplets of characters starting with ?? and it substitutes other characters in their place. So if you’re on some ancient terminal without a pipe symbol (|) on the keyboard, you can type ??! instead. You can fix this by escaping the second question mark, like so: printf("Doesn't it?\?!\n");
And then it compiles and works as-expected. These days, of course, no one ever uses trigraphs. But that whole ??! does sometimes appear if you decide to use it in a string for emphasis.
21.1.3 Numeric Escapes In addition, there are ways to specify numeric constants or other character values inside strings or character constants. If you know an octal or hexadecimal representation of a byte, you can include that in a string or character constant. The following table has example numbers, but any hex or octal numbers may be used. Pad with leading zeros if necessary to read the proper digit count.
Code Description \123 Embed the byte with octal value 123, 3 digits exactly. \x4D Embed the byte with hex value 4D, 2 digits. \u2620 Embed the Unicode character at code point with hex value 2620, 4 digits. Chapter 21. Characters and Strings II 146
Code Description \U0001243F Embed the Unicode character at code point with hex value 1243F, 8 digits.
Here’s an example of the less-commonly used octal notation to represent the letter B in between A and C. Normally this would be used for some kind of special unprintable character, but we have other ways to do that, below, and this is just an octal demo: printf("A\102C\n"); // 102 is `B` in ASCII/UTF-8
Note there’s no leading zero on the octal number when you include it this way. But it does need to be three characters, so pad with leading zeros if you need to. But far more common is to use hex constants these days. Here’s a demo that you shouldn’t use, but it demos embedding the UTF-8 bytes 0xE2, 0x80, and 0xA2 in a string, which corresponds to the Unicode “bullet” character (•). printf("\xE2\x80\xA2 Bullet 1\n"); printf("\xE2\x80\xA2 Bullet 2\n"); printf("\xE2\x80\xA2 Bullet 3\n");
Produces the following output if you’re on a UTF-8 console (or probably garbage if you’re not): • Bullet 1 • Bullet 2 • Bullet 3
But that’s a crummy way to do Unicode. You can use the escapes \u (16-bit) or \U (32-bit) to just refer to Unicode by code point number. The bullet is 2022 (hex) in Unicode, so you can do this and get more portable results: printf("\u2022 Bullet 1\n"); printf("\u2022 Bullet 2\n"); printf("\u2022 Bullet 3\n");
Be sure to pad \u with enough leading zeros to get to four characters, and \U with enough to get to eight. For example, that bullet could be done with \U and four leading zeros: printf("\U00002022 Bullet 1\n");
But who has time to be that verbose? Chapter 22
Enumerated Types: enum
C offers us another way to have constant integer values by name: enum. For example: enum { ONE=1, TWO=2 }; printf("%d %d", ONE, TWO); // 1 2
In some ways, it can be better—or different—than using a #define. Key differences: • enums can only be integer types. • #define can define anything at all. • enums are often shown by their symbolic identifier name in a debugger. • #defined numbers just show as raw numbers which are harder to know the meaning of while debugging. Since they’re integer types, they can be used any place integers can be used, including in array dimensions and case statements. Let’s tear into this more.
22.1 Behavior of enum
22.1.1 Numbering enums are automatically numbered unless you override them. They start at 0, and autoincrement up from there, by default: enum { SHEEP, // Value is 0 WHEAT, // Value is 1 WOOD, // Value is 2 BRICK, // Value is 3 ORE // Value is 4 }; printf("%d %d\n", SHEEP, BRICK); // 0 3
You can force particular integer values, as we saw earlier: enum { X=2,
147 Chapter 22. Enumerated Types: enum 148
Y=18, Z=-2 };
Duplicates are not a problem: enum { X=2, Y=2, Z=2 }; if values are omitted, numbering continues counting in the positive direction from whichever value was last specified. For example: enum { A, // 0, default starting value B, // 1 C=4, // 4, manually set D, // 5 E, // 6 F=3 // 3, manually set G, // 4 H // 5 }
22.1.2 Trailing Commas This is perfectly fine, if that’s your style: enum { X=2, Y=18, Z=-2, // <-- Trailing comma };
It’s gotten more popular in languages of the recent decades so you might be pleased to see it.
22.1.3 Scope enums scope as you’d expect. If at file scope, the whole file can see it. If in a block, it’s local to that block. It’s really common for enums to be defined in header files so they can be #included at file scope.
22.1.4 Style
As you’ve noticed, it’s common to declare the enum symbols in uppercase (with underscores). This isn’t a requirement, but is a very, very common idiom.
22.2 Your enum is a Type
This is an important thing to know about enum: they’re a type, analogous to how a struct is a type. You can give them a tag name so you can refer to the type later and declare variables of that type. Now, since enums are integer types, why not just use int? In C, the best reason for this is code clarity–it’s a nice, typed way to describe your thinking in code. C (unlike C++) doesn’t actually enforce any values being in range for a particular enum. Let’s do an example where we declare a variable r of type enum resource that can hold those values: Chapter 22. Enumerated Types: enum 149
// Named enum, type is "enum resource" enum resource { SHEEP, WHEAT, WOOD, BRICK, ORE };
// Declare a variable "r" of type "enum resource" enum resource r = BRICK; if (r == BRICK) { printf("I'll trade you a brick for two sheep.\n"); }
You can also typedef these, of course, though I personally don’t like to. typedef enum { SHEEP, WHEAT, WOOD, BRICK, ORE } RESOURCE;
RESOURCE r = BRICK;
Another shortcut that’s legal but rare is to declare variables when you declare the enum: // Declare an enum and some initialized variables of that type: enum { SHEEP, WHEAT, WOOD, BRICK, ORE } r = BRICK, s = WOOD;
You can also give the enum a name so you can use it later, which is probably what you want to do in most cases: // Declare an enum and some initialized variables of that type: enum resource { // <-- type is now "enum resource" SHEEP, WHEAT, WOOD, BRICK, ORE } r = BRICK, s = WOOD;
In short, enums are a great way to write nice, scoped, typed, clean code. Chapter 23
Pointers III: Pointers to Pointers and More
Here’s where we cover some intermediate and advanced pointer usage. If you don’t have pointers down well, review the previous chapters on pointers and pointer arithmetic before starting on this stuff.
23.1 Pointers to Pointers
If you can have a pointer to a variable, and a variable can be a pointer, can you have a pointer to a variable that it itself a pointer? Yes! This is a pointer to a pointer, and it’s held in variable of type pointer-pointer. Before we tear into that, I want to try for a gut feel for how pointers to pointers work. Remember that a pointer is just a number. It’s a number that represents an index in computer memory, typically one that holds a value we’re interested in for some reason. That pointer, which is a number, has to be stored somewhere. And that place is memory, just like every- thing else1. But because it’s stored in memory, it must have an index it’s stored at, right? The pointer must have an index in memory where it is stored. And that index is a number. It’s the address of the pointer. It’s a pointer to the pointer. Let’s start with a regular pointer to an int, back from the earlier chapters:
1 #include
2
3 int main(void)
4 {
5 int x = 3490; // Type: int
6 int *p = &x; // Type: pointer to an int
7
8 printf("%d\n", *p); // 3490
9 }
Straightforward enough, right? We have two types represented: int and int*, and we set up p to point to x. Then we can dereference p on line 8 and print out the value 3490. But, like we said, we can have a pointer to any variable… so does that mean we can have a pointer to p? In other words, what type is this expression?
1There’s some devil in the details with values that are stored in registers only, but we can safely ignore that for our purposes here. Also the C spec makes no stance on these “register” things beyond the register keyword, the description for which doesn’t mention registers.
150 Chapter 23. Pointers III: Pointers to Pointers and More 151
int x = 3490; // Type: int int *p = &x; // Type: pointer to an int
&p // <-- What type is the address of p? AKA a pointer to p?
If x is an int, then &x is a pointer to an int that we’ve stored in p which is type int*. Follow? (Repeat this paragraph until you do!) And therefore &p is a pointer to an int*, AKA a “pointer to a pointer to an int”. AKA “int-pointer- pointer”. Got it? (Repeat the previous paragraph until you do!) We write this type with two asterisks: int **. Let’s see it in action.
1 #include
2
3 int main(void)
4 {
5 int x = 3490; // Type: int
6 int *p = &x; // Type: pointer to an int
7 int **q = &p; // Type: pointer to pointer to int
8
9 printf("%d %d\n", *p, **q); // 3490 3490
10 }
Let’s make up some pretend addresses for the above values as examples and see what these three variables might look like in memory. The address values, below are just made up by me for example purposes:
Variable Stored at Address Value Stored There x 28350 3490—the value from the code p 29122 28350—the address of x! q 30840 29122—the address of p!
Indeed, let’s try it for real on my computer2 and print out the pointer values with %p and I’ll do the same table again with actual references (printed in hex).
Variable Stored at Address Value Stored There x 0x7ffd96a07b94 3490—the value from the code p 0x7ffd96a07b98 0x7ffd96a07b94—the address of x! q 0x7ffd96a07ba0 0x7ffd96a07b98—the address of p!
You can see those addresses are the same except the last byte, so just focus on those. On my system, ints are 4 bytes, which is why we’re seeing the address go up by 4 from x to p3 and then goes up by 8 from p to q. On my system, all pointers are 8 bytes. Does it matter if it’s an int* or an int**? Is one more bytes than the other? Nope! Remember that all pointers are addresses, that is indexes into memory. And on my machine you can represent an index with 8 bytes… doesn’t matter what’s stored at that index. Now check out what we did there on line 9 of the previous example: we double dereferenced q to get back to our 3490. This is the important bit about pointers and pointers to pointers: • You can get a pointer to anything with & (including to a pointer!)
2You’re very likely to get different numbers on yours. 3There is absolutely nothing in the spec that says this will always work this way, but it happens to work this way on my system. Chapter 23. Pointers III: Pointers to Pointers and More 152
• You can get the thing a pointer points to with * (including a pointer!) So you can think of & as being used to make pointers, and * being the inverse—it goes the opposite direction of &—to get to the thing pointed to. In terms of type, each time you &, that adds another pointer level to the type.
If you have Then you run The result type is int x &x int * int *x &x int ** int **x &x int *** int ***x &x int ****
And each time you use dereference (*), it does the opposite:
If you have Then you run The result type is int ****x *x int *** int ***x *x int ** int **x *x int * int *x *x int
Note that you can use multiple *s in a row to quickly dereference, just like we saw in the example code with **q, above. Each one strips away one level of indirection.
If you have Then you run The result type is int ****x ***x int * int ***x **x int * int **x **x int
In general, &*E == E4. The dereference “undoes” the address-of. But & doesn’t work the same way—you can only do those one at a time, and have to store the result in an intermediate variable: int x = 3490; // Type: int int *p = &x; // Type: int *, pointer to an int int **q = &p; // Type: int **, pointer to pointer to int int ***r = &q; // Type: int ***, pointer to pointer to pointer to int int ****s = &r; // Type: int ****, you get the idea int *****t = &s; // Type: int *****
23.1.1 Pointer Pointers and const If you recall, declaring a pointer like this: int *const p; means that you can’t modify p. Trying to p++ would give you a compile-time error. But how does that work with int ** or int ***? Where does the const go, and what does it mean? Let’s start with the simple bit. The const right next to the variable name refers to that variable. So if you want an int*** that you can’t change, you can do this:
4Even if E is NULL, it turns out, weirdly. Chapter 23. Pointers III: Pointers to Pointers and More 153
int ***const p;
p++; // Not allowed
But here’s where things get a little weird. What if we had this situation:
1 int main(void)
2 {
3 int x = 3490;
4 int *const p = &x;
5 int **q = &p;
6 }
When I build that, I get a warning: warning: initialization discards ‘const’ qualifier from pointer target type 7 | int **q = &p; | ^
What’s going on? The That is, we’re saying that q is type int **, and if you dereference that, the rightmost * in the type goes away. So after the dereference, we have type int *. And we’re assigning &p into it which is a pointer to an int *const, or, in other words, int *const *. But q is int **! A type with different constness on the first *! So we get a warning that the const in p’s int *const * is being ignored and thrown away. We can fix that by making sure q’s type is at least as const as p. int x = 3490; int *const p = &x; int *const *q = &p;
And now we’re happy. We could make q even more const. As it is, above, we’re saying, “q isn’t itself const, but the thing it points to is const.” But we could make them both const: int x = 3490; int *const p = &x; int *const *const q = &p; // More const!
And that works, too. Now we can’t modify q, or the pointer q points to.
23.2 Multibyte Values
We kinda hinted at this in a variety of places earlier, but clearly not every value can be stored in a single byte of memory. Things take up multiple bytes of memory (assuming they’re not chars). You can tell how many bytes by using sizeof. And you can tell which address in memory is the first byte of the object by using the standard & operator, which always returns the address of the first byte. And here’s another fun fact! If you iterate over the bytes of any object, you get its object representation. Two things with the same object representation in memory are equal. If you want to iterate over the object representation, you should do it with pointers to unsigned char. Let’s make our own version of memcpy() that does exactly this: void *my_memcpy(void *dest, const void *src, size_t n) { // Make local variables for src and dest, but of type unsigned char Chapter 23. Pointers III: Pointers to Pointers and More 154
const unsigned char *s = src; unsigned char *d = dest;
while (n-- > 0) // For the given number of bytes *d++ = *s++; // Copy source byte to dest byte
// Most copy functions return a pointer to the dest as a convenience // to the caller
return dest; }
(There are some good examples of post-increment and post-decrement in there for you to study, as well.) It’s important to note that the version, above, is probably less efficient than the one that comes with your system. But you can pass pointers to anything into it, and it’ll copy those objects. Could be int*, struct ani- mal*, or anything. Let’s do another example that prints out the object representation bytes of a struct so we can see if there’s any padding in there and what values it has5.
1 #include
2
3 struct foo {
4 char a;
5 int b;
6 };
7
8 int main(void)
9 {
10 struct foo x = {0x12, 0x12345678};
11 unsigned char *p = (unsigned char *)&x;
12
13 for (size_t i = 0; i < sizeof x; i++) {
14 printf("%02X\n", p[i]);
15 }
16 }
What we have there is a struct foo that’s built in such a way that should encourage a compiler to inject padding bytes (though it doesn’t have to). And then we get an unsigned char * to the first byte of the struct foo variable x. From there, all we need to know is the sizeof x and we can loop through that many bytes, printing out the values (in hex for ease). Running this gives a bunch of numbers as output. I’ve annotated it below to identify where the values were stored: 12 | x.a == 0x12
AB | BF | padding bytes with "random" value 26 |
78 | 56 | x.b == 0x12345678 34 | 12 |
5Your C compiler is not required to add padding bytes, and the values of any padding bytes that are added are indeterminate. Chapter 23. Pointers III: Pointers to Pointers and More 155
On all systems, sizeof(char) is 1, and we see that first byte at the top of the output holding the value 0x12 that we stored there. Then we have some padding bytes—for me, these varied from run to run. Finally, on my system, sizeof(int) is 4, and we can see those 4 bytes at the end. Notice how they’re the same bytes as are in the hex value 0x12345678, but strangely in reverse order6. So that’s a little peek under the hood at the bytes of a more complex entity in memory.
23.3 The NULL Pointer and Zero
These things can be used interchangeably: • NULL • 0 • '\0' • (void *)0 Personally, I always use NULL when I mean NULL, but you might see some other variants from time to time. Though '\0' (a byte with all bits set to zero) will also compare equal, it’s weird to compare it to a pointer; you should compare NULL against the pointer. (Of course, lots of times in string processing, you’re comparing the thing the pointer points to to '\0', and that’s right.) 0 is called the null pointer constant, and, when compared to or assigned into another pointer, it is converted to a null pointer of the same type.
23.4 Pointers as Integers
You can cast pointers to integers and vice-versa (since a pointer is just an index into memory), but you probably only ever need to do this if you’re doing some low-level hardware stuff. The results of such machinations are implementation-defined, so they aren’t portable. And weird things could happen. C does make one guarantee, though: you can convert a pointer to a uintptr_t type and you’ll be able to convert it back to a pointer without losing any data. uintptr_t is defined in
23.5 Pointer Differences
As you know from the section on pointer arithmetic, you can subtract one pointer from another8 to get the difference between them in count of array elements. Now the type of that difference is something that’s up to the implementation, so it could vary from system to system. To be more portable, you can store the result in a variable of type ptrdiff_t defined in
And you can print it by prefixing the integer format specifier with t:
6This will vary depending on the architecture, but my system is little endian, which means the least-significant byte of the number is stored first. Big endian systems will have the 12 first and the 78 last. But the spec doesn’t dictate anything about this representation. 7It’s an optional feature, so it might not be there—but it probably is. 8Assuming they point to the same array object. Chapter 23. Pointers III: Pointers to Pointers and More 156
printf("%td\n", d); // Print decimal: 40 printf("%tX\n", d); // Print hex: 28
23.6 Pointers to Functions
Functions are just collections of machine instructions in memory, so there’s no reason we can’t get a pointer to the first instruction of the function. And then call it. This can be useful for passing a pointer to a function into another function as an argument. Then the second one could call whatever was passed in. The tricky part with these, though, is that C needs to know the type of the variable that is the pointer to the function. And it would really like to know all the details. Like “this is a pointer to a function that takes two int arguments and returns void”. How do you write all that down so you can declare a variable? Well, it turns out it looks very much like a function prototype, except with some extra parentheses: // Declare p to be a pointer to a function. // This function returns a float, and takes two ints as arguments.
float (*p)(int, int);
Also notice that you don’t have to give the parameters names. But you can if you want; they’re just ignored. // Declare p to be a pointer to a function. // This function returns a float, and takes two ints as arguments.
float (*p)(int a, int b);
So now that we know how to declare a variable, how do we know what to assign into it? How do we get the address of a function? Turns out there’s a shortcut just like with getting a pointer to an array: you can just refer to the bare function name without parens. (You can put an & in front of this if you like, but it’s unnecessary and not idiomatic.) Once you have a pointer to a function, you can call it just by adding parens and an argument list. Let’s do a simple example where I effectively make an alias for a function by setting a pointer to it. Then we’ll call it. This code prints out 3490:
1 #include
2
3 void print_int(int n)
4 {
5 printf("%d\n", n);
6 }
7
8 int main(void)
9 {
10 // Assign p to point to print_int:
11
12 void (*p)(int) = print_int;
13 Chapter 23. Pointers III: Pointers to Pointers and More 157
14 p(3490); // Call print_int via the pointer
15 }
Notice how the type of p represents the return value and parameter types of print_int. It has to, or else C will complain about incompatible pointer types. One more example here shows how we might pass a pointer to a function as an argument to another function. We’ll write a function that takes a couple integer arguments, plus a pointer to a function that operates on those two arguments. Then it prints the result.
1 #include
2
3 int add(int a, int b)
4 {
5 return a + b;
6 }
7
8 int mult(int a, int b)
9 {
10 return a * b;
11 }
12
13 void print_math(int (*op)(int, int), int x, int y)
14 {
15 int result = op(x, y);
16
17 printf("%d\n", result);
18 }
19
20 int main(void)
21 {
22 print_math(add, 5, 7); // 12
23 print_math(mult, 5, 7); // 35
24 }
Take a moment to digest that. The idea here is that we’re going to pass a pointer to a function to print_math(), and it’s going to call that function to do some math. This way we can change the behavior of print_math() by passing another function into it. You can see we do that on lines 22-23 when we pass in pointers to functions add and mult, respectively. Now, on line 13, I think we can all agree the function signature of print_math() is a sight to behold. And, if you can believe it, this one is actually pretty straight-forward compared to some things you can construct9. But let’s digest it. Turns out there are only three parameters, but they’re a little hard to see: // op x y // |------| |---| |---| void print_math(int (*op)(int, int), int x, int y)
The first, op, is a pointer to a function that takes two ints as arguments and returns an int. This matches the signatures for both add() and mult(). The second and third, x and y, are just standard int parameters. Slowly and deliberately let your eyes play over the signature while you identify the working parts. One thing that always stands out for me is the sequence (*op)(, the parens and the asterisk. That’s the give- away it’s a pointer to a function.
9The Go Programming Language drew its type declaration syntax inspiration from the opposite of what C does. Chapter 23. Pointers III: Pointers to Pointers and More 158
Finally, jump back to the Pointers II chapter for a pointer-to-function example using the built-in qsort(). Chapter 24
Bitwise Operations
These numeric operations effectively allow you to manipulate individual bits in variables, fitting since C is such a low-level langauge1. If you’re not familiar with bitwise operations, Wikipedia has a good bitwise article2.
24.1 Bitwise AND, OR, XOR, and NOT
For each of these, the usual arithmetic conversions take place on the operands (which in this case must be an integer type), and then the appropriate bitwise operation is performed.
Operation Operator Example AND & a = b & c OR | a = b | c XOR ^ a = b ^ c NOT ~ a = ~c
Note how they’re similar to the Boolean operators && and ||. These have assignment shorthand variants similar to += and -=:
Operator Example Longhand equivalent &= a &= c a = a & c |= a |= c a = a | c ^= a ^= c a = a ^ c
24.2 Bitwise Shift
For these, the integer promotions are performed on each operand (which must be an integer type) and then a bitwise shift is executed. The type of the result is the type of the promoted left operand. New bits are filled with zeros, with a possible exception noted in the implementation-defined behavior, below.
1Not that other languages don’t do this—they do. It is interesting how many modern languages use the same operators for bitwise that C does. 2https://en.wikipedia.org/wiki/Bitwise_operation
159 Chapter 24. Bitwise Operations 160
Operation Operator Example Shift left << a = b << c Shift right >> a = b >> c
There’s also the same similar shorthand for shifting:
Operator Example Longhand equivalent >>= a >>= c a = a >> c <<= a <<= c a = a << c
Watch for undefined behavior: no negative shifts, and no shifts that are larger than the size of the promoted left operand. Also watch for implementation-defined behavior: if you right-shift a negative number, the results are implementation-defined. (It’s perfectly fine to right-shift a signed int, just make sure it’s positive.) Chapter 25
Variadic Functions
Variadic is a fancy word for functions that take arbitrary numbers of arguments. A regular function takes a specific number of arguments, for example: int add(int x, int y) { return x + y; }
You can only call that with exactly two arguments which correspond to parameters x and y. add(2, 3); add(5, 12);
But if you try it with more, the compiler won’t let you: add(2, 3, 4); // ERROR add(5); // ERROR
Variadic functions get around this limitation to a certain extent. We’ve already seen a famous example in printf()! You can pass all kinds of things to it. printf("Hello, world!\n"); printf("The number is %d\n", 2); printf("The number is %d and pi is %f\n", 2, 3.14159);
It seems to not care how many arguments you give it! Well, that’s not entirely true. Zero arguments will give you an error: printf(); // ERROR
This leads us to one of the limitations of variadic functions in C: they must have at least one argument. But aside from that, they’re pretty flexible, even allows arguments to have different types just like printf() does. Let’s see how they work!
25.1 Ellipses in Function Signatures
So how does it work, syntactically? What you do is put all the arguments that must be passed first (and remember there has to be at least one) and after that, you put .... Just like this: void func(int a, ...) // Literally 3 dots here
161 Chapter 25. Variadic Functions 162
Here’s some code to demo that: #include
void func(int a, ...) { printf("a is %d\n", a); // Prints "a is 2" }
int main(void) { func(2, 3, 4, 5, 6); }
So, great, we can get that first argument that’s in variable a, but what about the rest of the arguments? How do you get to them? Here’s where the fun begins!
25.2 Getting the Extra Arguments
You’re going to want to include
1 #include
2 #include
3
4 int add(int count, ...)
5 {
6 int total = 0;
7 va_list va;
8
9 va_start(va, count); // Start with arguments after "count"
10
11 for (int i = 0; i < count; i++) {
12 int n = va_arg(va, int); // Get the next int
13
14 total += n;
15 }
16
17 va_end(va); // All done
18
19 return total;
20 }
21
22 int main(void)
23 {
24 printf("%d\n", add(4, 6, 2, -4, 17)); // 6 + 2 - 4 + 17 = 21 Chapter 25. Variadic Functions 163
25 printf("%d\n", add(2, 22, 44)); // 22 + 44 = 66
26 }
When printf() is called, it uses the number of %ds (or whatever) in the format string to know how many more arguments there are! If the syntax of va_arg() is looking strange to you (because of that loose type name floating around in there), you’re not alone. These are implemented with preprocessor macros in order to get all the proper magic in there.
25.3 va_list Functionality
What is that va_list variable we’re using up there? It’s an opaque variable1 that holds information about which argument we’re going to get next with va_arg(). You see how we just call va_arg() over and over? The va_list variable is a placeholder that’s keeping track of progress so far. But we have to initialize that variable to some sensible value. That’s where va_start() comes into play. When we called va_start(va, count), above, we were saying, “Initialize the va variable to point to the variable argument immediately after count.” And that’s why we need to have at least one named variable in our argument list2. Once you have that pointer to the initial parameter, you can easily get subsequent argument values by calling va_arg() repeatedly. When you do, you have to pass in your va_list variable (so it can keep on keeping track of where you are), as well as the type of argument you’re about to copy off. It’s up to you as a programmer to figure out which type you’re going to pass to va_arg(). In the above example, we just did ints. But in the case of printf(), it uses the format specifier to determine which type to pull off next. And when you’re done, call va_end() to wrap it up. You must (the spec says) call this on a particular va_list variable before you decide to call either va_start() or va_copy() on it again. I know we haven’t talked about va_copy() yet. So the standard progression is: • va_start() to initialize your va_list variable • Repeatedly va_arg() to get the values • va_end() to deinitialize your va_list variable I also mentioned va_copy() up there; it makes a copy of your va_list variable in the exact same state. That is, if you haven’t started with va_arg() with the source variable, the new one won’t be started, either. If you’ve consumed 5 variables with va_arg() so far, the copy will also reflect that. va_copy() can be useful if you need to scan ahead through the arguments but need to also remember your current place.
25.4 Library Functions That Use va_lists
One of the other uses for these is pretty cool: writing your own custom printf() variant. It would be a pain to have to handle all those format specifiers right? All zillion of them? Luckily, there are printf() variants that accept a working va_list as an argument. You can use these to wrap up and make your own custom printf()s! These functions start with the letter v, such as vprintf(), vfprintf(), vsprintf(), and vsnprintf(). Basically all your printf() golden oldies except with a v in front.
1That is, us lowly developers aren’t supposed to know what’s in there or what it means. The spec doesn’t dictate what it is in detail. 2Honestly, it would be possible to remove that limitation from the language, but the idea is that the macros va_start(), va_arg(), and va_end() should be able to be written in C. And to make that happen, we need some way to initialize a pointer to the location of the first parameter. And to do that, we need the name of the first parameter. It would require a language extension to make this possible, and so far the committee hasn’t found a rationale for doing so. Chapter 25. Variadic Functions 164
Let’s make a function my_printf() that works just like printf() except it takes an extra argument up front.
1 #include
2 #include
3
4 int my_printf(int serial, const char *format, ...)
5 {
6 va_list va;
7
8 // Do my custom work
9 printf("The serial number is: %d\n", serial);
10
11 // Then pass the rest off to vprintf()
12 va_start(va, format);
13 int rv = vprintf(format, va);
14 va_end(va);
15
16 return rv;
17 }
18
19 int main(void)
20 {
21 int x = 10;
22 float y = 3.2;
23
24 my_printf(3490, "x is %d, y is %f\n", x, y);
25 }
See what we did there? On lines 12-14 we started a new va_list variable, and then just passed it right into vprintf(). And it knows just want to do with it, because it has all the printf() smarts built-in. We still have to call va_end() when we’re done, though, so don’t forget that! Chapter 26
Locale and Internationalization
Localization is the process of making your app ready to work well in different locales (or countries). As you might know, not everyone uses the same character for decimal points or for thousands separators… or for currency. These locales have names, and you can select one to use. For example, a US locale might write a number like: 100,000.00 Whereas in Brazil, the same might be written with the commas and decimal points swapped: 100.000,00 Makes it easier to write your code so it ports to other nationalities with ease! Well, sort of. Turns out C only has one built-in locale, and it’s limited. The spec really leaves a lot of ambiguity here; it’s hard to be completely portable. But we’ll do our best!
26.1 Setting the Localization, Quick and Dirty
For these calls, include
You’ll want to call that so that the program gets initialized with your current locale. Getting into more details, there is one more thing you can do and stay portable: set_locale(LC_ALL, "C"); // Use the default C locale but that’s called by default every time your program starts, so there’s not much need to do it yourself. In that second string, you can specify any locale supported by your system. This is completely system- dependent, so it will vary. On my system, I can specify this: setlocale(LC_ALL, "en_US.UTF-8"); // Non-portable!
And that’ll work. But it’s only portable to systems which have that exact same name for that exact same locale, and you can’t guarantee it. By passing in an empty string ("") for the second argument, you’re telling C, “Hey, figure out what the current locale on this system is so I don’t have to tell you.”
165 Chapter 26. Locale and Internationalization 166
26.2 Getting the Monetary Locale Settings
Because moving green pieces of paper around promises to be the key to happiness1, let’s talk about mon- etary locale. When you’re writing portable code, you have to know what to type for cash, right? Whether that’s “$”, “€”, “¥”, or “£”. How can you write that code without going insane? Luckily, once you call setlocale(LC_ALL, ""), you can just look these up with a call to localeconv(): struct lconv *x = localeconv();
This function returns a pointer to a statically-allocated struct lconv that has all that juicy information you’re looking for. Here are the fields of struct lconv and their meanings. First, some conventions. An _p_ means “positive”, and _n_ means “negative”, and int_ means “inter- national”. Though a lot of these are type char or char*, most (or the strings they point to) are actually treated as integers2. Before we go further, know that CHAR_MAX (from
Field Description char *mon_decimal_point Decimal pointer character for money, e.g. ".". char *mon_thousands_sep Thousands separator character for money, e.g. ",". char *mon_grouping Grouping description for money (see below). char *positive_sign Positive sign for money, e.g. "+" or "". char *negative_sign Negative sign for money, e.g. "-". char *currency_symbol Currency symbol, e.g. "$". char frac_digits When printing monetary amounts, how many digits to print past the decimal point, e.g. 2. char p_cs_precedes 1 if the currency_symbol comes before the value for a non-negative monetary amount, 0 if after. char n_cs_precedes 1 if the currency_symbol comes before the value for a negative monetary amount, 0 if after. char p_sep_by_space Determines the separation of the currency symbol from the value for non-negative amounts (see below). char n_sep_by_space Determines the separation of the currency symbol from the value for negative amounts (see below). char p_sign_posn Determines the positive_sign position for non-negative values. char p_sign_posn Determines the positive_sign position for negative values. char *int_curr_symbol International currency symbol, e.g. "USD ". char int_frac_digits International value for frac_digits. char int_p_cs_precedes International value for p_cs_precedes. char int_n_cs_precedes International value for n_cs_precedes. char int_p_sep_by_space International value for p_sep_by_space. char int_n_sep_by_space International value for n_sep_by_space. char int_p_sign_posn International value for p_sign_posn. char int_n_sign_posn International value for n_sign_posn.
1“This planet has—or rather had—a problem, which was this: most of the people living on it were unhappy for pretty much of the time. Many solutions were suggested for this problem, but most of these were largely concerned with the movement of small green pieces of paper, which was odd because on the whole it wasn’t the small green pieces of paper that were unhappy.” —The Hitchhiker’s Guide to the Galaxy, Douglas Adams 2Remember that char is just a byte-sized integer. Chapter 26. Locale and Internationalization 167
26.2.1 Monetary Digit Grouping
OK, this is a trippy one. mon_grouping is a char*, so you might be thinking it’s a string. But in this case, no, it’s really an array of chars. It should always end either with a 0 or CHAR_MAX. These values describe how to group sets of numbers in currency to the left of the decimal (the whole number part). For example, we might have: 2 1 0 ------$100,000,000.00
These are groups of three. Group 0 (just left of the decimal) has 3 digits. Group 1 (next group to the left) has 3 digits, and the last one also has 3. So we could describe these groups, from the right (the decimal) to the left with a bunch of integer values representing the group sizes: 3 3 3
And that would work for values up to $100,000,000. But what if we had more? We could keep adding 3s… 33333333333333333 but that’s crazy. Luckily, we can specify 0 to indicate that the previous group size repeats: 3 0
Which means to repeat every 3. That would handle $100, $1,000, $10,000, $10,000,000, $100,000,000,000, and so on. You can go legitimately crazy with these to indicate some weird groupings. For example: 4 3 2 1 0 would indicate: $1,0,0,0,0,00,000,0000.00
One more value that can occur is CHAR_MAX. This indicates that no more grouping should occur, and can appear anywhere in the array, including the first value. 3 2 CHAR_MAX would indicate: 100000000,00,000.00 for example. And simply having CHAR_MAX in the first array position would tell you there was to be no grouping at all.
26.2.2 Separators and Sign Position
All the sep_by_space variants deal with spacing around the currency sign. Valid values are:
Value Description 0 No space between currency symbol and value. 1 Separate the currency symbol (and sign, if any) from the value with a space. 2 Separate the sign symbol from the currency symbol (if adjacent) with a space, otherwise separate the sign symbol from the value with a space. Chapter 26. Locale and Internationalization 168
The sign_posn variants are determined by the following values:
Value Description 0 Put parens around the value and the currency symbol. 1 Put the sign string in front of the currency symbol and value. 2 Put the sign string after the currency symbol and value. 3 Put the sign string directly in front of the currency symbol. 4 Put the sign string directly behind the currency symbol.
26.2.3 Example Values When I get the values on my system, this is what I see (grouping string displayed as individual byte values): mon_decimal_point = "." mon_thousands_sep = "," mon_grouping = 3 3 0 positive_sign = "" negative_sign = "-" currency_symbol = "$" frac_digits = 2 p_cs_precedes = 1 n_cs_precedes = 1 p_sep_by_space = 0 n_sep_by_space = 0 p_sign_posn = 1 n_sign_posn = 1 int_curr_symbol = "USD " int_frac_digits = 2 int_p_cs_precedes = 1 int_n_cs_precedes = 1 int_p_sep_by_space = 1 int_n_sep_by_space = 1 int_p_sign_posn = 1 int_n_sign_posn = 1
26.3 Localization Specifics
Notice how we passed the macro LC_ALL to setlocale() earlier… this hints that there might be some variant that allows you to be more precise about which parts of the locale you’re setting. Let’s take a look at the values you can see for these:
Macro Description LC_ALL Set all of the following to the given locale. LC_COLLATE Controls the behavior of the strcoll() and strxfrm() functions. LC_CTYPE Controls the behavior of the character-handling functions3. LC_MONETARY Controls the values returned by localeconv(). LC_NUMERIC Controls the decimal point for the printf() family of functions. LC_TIME Controls time formatting of the strftime() and wcsftime() time and date printing functions.
It’s pretty common to see LC_ALL being set, but, hey, at least you have options.
3Except for isdigit() and isxdigit(). Chapter 26. Locale and Internationalization 169
Also I should point out that LC_CTYPE is one of the biggies because it ties into wide characters, a significant can of worms that we’ll talk about later. Chapter 27
Unicode, Wide Characters, and All That
Before we begin, note that this is an active area of language development in C as it works to get past some, erm, growing pains. When C2x comes out, updates here are probable. Most people are basically interested in the deceptively simple question, “How do I use such-and-such character set in C?” We’ll get to that. But as we’ll see, it might already work on your system. Or you might have to punt to a third-party library. We’re going to talk about a lot of things this chapter—some are platform agnostic, and some are C-specific. Let’s get an outline first of what we’re going to look at: • Unicode background • Character encoding background • Source and Execution character Sets • Using Unicode and UTF-8 • Using other character types like wchar_t, char16_t, and char32_t Let’s dive in!
27.1 What is Unicode?
Back in the day, it was popular in the US and much of the world to use a 7-bit or 8-bit encoding for characters in memory. This meant we could have 128 or 256 characters (including non-printable char- acters) total. That was fine for a US-centric world, but it turns out there are actually other alphabets out there—who knew? Chinese has over 50,000 characters, and that’s not fitting in a byte. So people came up with all kinds of alternate ways to represent their own custom character sets. And that was fine, but turned into a compatibility nightmare. To escape it, Unicode was invented. One character set to rule them all. It extends off into infinity (effec- tively) so we’ll never run out of space for new characters. It has Chinese, Latin, Greek, cuneiform, chess symbols, emojis… just about everything, really! And more is being added all the time!
27.2 Code Points
I want to talk about two concepts here. It’s confusing because they’re both numbers… different numbers for the same thing. But bear with me. Let’s loosely define code point to mean a numeric value representing a character. (Code points can also represent unprintable control characters, but just assume I mean something like the letter “B” or the char- acter “π”.)
170 Chapter 27. Unicode, Wide Characters, and All That 171
Each code point represents a unique character. And each character has a unique numeric code point associated with it. For example, in Unicode, the numeric value 66 represents “B”, and 960 represents “π”. Other character mappings that aren’t Unicode use different values, potentially, but let’s forget them and concentrate on Unicode, the future! So that’s one thing: there’s a number that represents each character. In Unicode, these numbers run from 0 to over 1 million. Got it? Because we’re about to flip the table a little.
27.3 Encoding
If you recall, an 8-bit byte can hold values from 0-255, inclusive. That’s great for “B” which is 66—that fits in a byte. But “π” is 960, and that doesn’t fit in a byte! We need another byte. How do we store all that in memory? Or what about bigger numbers, like 195,024? That’s going to need a number of bytes to hold. The Big Question: how are these numbers represented in memory? This is what we call the encoding of the characters. So we have two things: one is the code point which tells us effectively the serial number of a particular character. And we have the encoding which tells us how we’re going to represent that number in memory. There are plenty of encodings. You can make up your own right now, if you want1. But we’re going to look at some really common encodings that are in use with Unicode.
Encoding Description UTF-8 A byte-oriented encoding that uses a variable number of bytes per character. This is the one to use. UTF-16 A 16-bit per character2 encoding. UTF-32 A 32-bit per character encoding.
With UTF-16 and UTF-32, the byte order matters, so you might see UTF-16BE for big-endian and UTF- 16LE for little-endian. Same for UTF-32. Technically, if unspecified, you should assume big-endian. But since Windows uses UTF-16 extensively and is little-endian, sometimes that is assumed3. Let’s look at some examples. I’m going to write the values in hex because that’s exactly two digits per 8-bit byte, and it makes it easier to see how things are arranged in memory.
Character Code Point UTF-16BE UTF-32BE UTF-16LE UTF-32LE UTF-8 A 41 0041 00000041 4100 41000000 41 B 42 0042 00000042 4200 42000000 42 ~ 7E 007E 0000007E 7E00 7E000000 7E π 3C0 03C0 000003C0 C003 C0030000 CF80 € 20AC 20AC 000020AC AC20 AC200000 E282AC
Look in there for the patterns. Note that UTF-16BE and UTF-32BE are simply the code point represented directly as 16- and 32-bit values4.
1For example, we could store the code point in a big-endian 32-bit integer. Straightforward! We just invented an encoding! Actually not; that’s what UTF-32BE encoding is. Oh well—back to the grind! 2Ish. Technically, it’s variable width—there’s a way to represent code points higher than 216 by putting two UTF-16 characters together. 3There’s a special character called the Byte Order Mark (BOM), code point 0xFEFF, that can optionally precede the data stream and indicate the endianess. It is not required, however. 4Again, this is only true in UTF-16 for characters that fit in two bytes. Chapter 27. Unicode, Wide Characters, and All That 172
Little-endian is the same, except the bytes are in little-endian order. Then we have UTF-8 at the end. First you might notice that the single-byte code points are represented as a single byte. That’s nice. You might also notice that different code points take different number of bytes. This is a variable-width encoding. So as soon as we get above a certain value, UTF-8 starts using additional bytes to store the values. And they don’t appear to correlate with the code point value, either. The details of UTF-8 encoding5 are beyond the scope of this guide, but it’s enough to know that it has a variable number of bytes per code point, and those byte values don’t match up with the code point except for the first 128 code points. If you really want to learn more, Computerphile has a great UTF-8 video with Tom Scott6. That last bit is a neat thing about Unicode and UTF-8 from a North American perspective: it’s backward compatible with 7-bit ASCII encoding! So if you’re used to ASCII, UTF-8 is the same! Every ASCII- encoded document is also UTF-8 encoded! (But not the other way around, obviously.) It’s probably that last point more than any other that is driving UTF-8 to take over the world.
27.4 Source and Execution Character Sets
When programming in C, there are (at least) three character sets that are in play: • The one that your code exists on disk as. • The one the compiler translates that into just as compilation begins (the source character set). This might be the same as the one on disk, or it might not. • The one the compiler translates the source character set into for execution (the execution character set). This might be the same as the source character set, or it might not. Your compiler probably has options to select these character sets at build-time. The basic character set for both source and execution will contain the following characters: ABCDEFGHIJKLM NOPQRSTUVWXYZ abcdefghijklm nopqrstuvwxyz 0 1 2 3 4 5 6 7 8 9 !"#%&'()*+,-./: ;<=>?[\]^_{|}~ space tab vertical-tab form-feed end-of-line
Those are the characters you can use in your source and remain 100% portable. The execution character set will additionally have characters for alert (bell/flash), backspace, carriage return, and newline. But most people don’t go to that extreme and freely use their extended character sets in source and exe- cutable, especially now that Unicode and UTF-8 are getting more common. I mean, the basic character set doesn’t even allow for @, $, or `! Notably, it’s a pain (though possible with escape sequences) to enter Unicode characters using only the basic character set.
27.5 Unicode in C
Before I get into encoding in C, let’s talk about Unicode from a code point standpoint. There is a way in C to specify Unicode characters and these will get translated by the compiler into the execution character
5https://en.wikipedia.org/wiki/UTF-8 6https://www.youtube.com/watch?v=MijmeoH9LT4 Chapter 27. Unicode, Wide Characters, and All That 173 set7. So how do we do it? How about the euro symbol, code point 0x20AC. (I’ve written it in hex because both ways of representing it in C require hex.) How can we put that in our C code? Use the \u escape to put it in a string, e.g. "\u20AC" (case for the hex doesn’t matter). You must put exactly four hex digits after the \u, padding with leading zeros if necessary. Here’s an example: char *s = "\u20AC1.23"; printf("%s\n", s); // €1.23
So \u works for 16-bit Unicode code points, but what about ones bigger than 16 bits? For that, we need capitals: \U. For example: char *s = "\U0001D4D1"; printf("%s\n", s); // Prints a mathematical letter "B"
It’s the same as \u, just with 32 bits instead of 16. These are equivalent: \u03C0 \U000003C0
Again, these are translated into the execution character set during compilation. They represent Unicode code points, not any specific encoding. Furthermore, if a Unicode code point is not representable in the execution character set, the compiler can do whatever it wants with it. Now, you might wonder why you can’t just do this: char *s = "€1.23"; printf("%s\n", s); // €1.23
And you probably can, given a modern compiler. The source character set will be translated for you into the execution character set by the compiler. But compilers are free to puke out if they find any characters that aren’t included in their extended character set, and the € symbol certainly isn’t in the basic character set. Caveat from the spec: you can’t use \u or \U to encode any code points below 0xA0 except for 0x24 ($), 0x40 (@), and 0x60 (`)—yes, those are precisely the trio of common punctuation marks missing from the basic character set. Apparently this restriction is relaxed in the upcoming version of the spec. Finally, you can also use these in identifiers in your code, with some restrictions. But I don’t want to get into that here. We’re all about string handling in this chapter. And that’s about it for Unicode in C (except encoding).
27.6 A Quick Note on UTF-8 Before We Swerve into the Weeds
It could be that your source file on disk, the extended source characters, and the extended execution characters are all in UTF-8 format. And the libraries you use expect UTF-8. This is the glorious future of UTF-8 everywhere. If that’s the case, and you don’t mind being non-portable to systems that aren’t like that, then just run with it. Stick Unicode characters in your source and data at will. Use regular C strings and be happy.
7Presumably the compiler makes the best effort to translate the code point to whatever the output encoding is, but I can’t find any guarantees in the spec. Chapter 27. Unicode, Wide Characters, and All That 174
A lot of things will just work (albeit non-portably) because UTF-8 strings can safely be NUL-terminated just like any other C string. But maybe losing portability in exchange for easier character handling is a tradeoff that’s worth it to you. There are some caveats, however: • Things like strlen() report the number of bytes in a string, not the number of characters, neces- sarily. (Use mbstowcs() with a NULL first argument to get the number of characters in a multibyte string.) • The following won’t work properly with characters of more than one byte: strtok(), strchr() (use strstr() instead), strspn()-type functions, toupper(), tolower(), isalpha()-type functions, and probably more. Beware anything that operates on bytes. • printf() variants allow for a way to only print so many bytes of a string8. You want to make certain you print the correct number of bytes to end on a character boundary. • If you want to malloc() space for a string, or declare an array of chars for one, be aware that the maximum size could be more than you were expecting. Each character could take up to MB_LEN_MAX bytes (from
27.7 Different Character Types
I want to introduce more character types. We’re used to char, right? But that’s too easy. Let’s make things a lot more difficult! Yay!
27.7.1 Multibyte Characters
First of all, I want to potentially change your thinking about what a string (array of chars) is. These are multibyte strings made up of multibyte characters. That’s right—your run-of-the-mill string of characters is multibyte. Even if a particular character in the string is only a single byte, or if a string is made up of only single characters, it’s known as multibyte. For example: char c[128] = "Hello, world!"; // Multibyte string
What we’re saying here is that a particular character that’s not in the basic character set could be composed of multiple bytes. Up to MB_LEN_MAX of them (from
But here we’re getting into some weirdness, because check this out: char *s = "\u20AC1.23"; // €1.23 printf("%zu\n", strlen(s)); // 7!
The string length of "€1.23" is 7?! Yes! Well, on my system, yes! Remember that strlen() returns the number of bytes in the string, not the number of characters. (When we get to “wide characters”, coming up, we’ll see a way to get the number of characters in the string.)
8With a format specifier like "%s.12", for example. Chapter 27. Unicode, Wide Characters, and All That 175
Note that while C allows individual multibyte char constants, the behavior of these varies by implemen- tation and your compiler might warn on it. GCC, for example, warns of multi-character character constants for the following two lines (and, on my system, prints out the UTF-8 encoding): printf("%x\n", '€'); printf("%x\n", '\u20ac');
27.7.2 Wide Characters If you’re not a multibyte character, then you’re a wide character. A wide character is a single value that can uniquely represent any character in the current locale. It’s analogous to Unicode code points. But it might not be. Or it might be. Basically, where multibyte character strings are arrays of bytes, wide character strings are arrays of char- acters. So you can start thinking on a character-by-character basis rather than a byte-by-byte basis (the latter of which gets all messy when characters start taking up variable numbers of bytes). Wide characters can be represented by a number of types, but the big standout one is wchar_t. It’s the main one. You might be wondering if you can’t tell if it’s Unicode or not, how does that allow you much flexibility in terms of writing code? wchar_t opens some of those doors, as there are a rich set of function you can use to deal with wchar_t strings (like getting the length, etc.) without caring about the encoding.
27.8 Using Wide Characters and wchar_t
Time for a new type: wchar_t. This is the main wide character typre. Remember how a char is only one byte? And a byte’s not enough to represent all characters, potentially? Well, this one is enough. To use wchar_t, #include
Now—are those characters stored are Unicode code points, or not? Depends on the implementation. But you can test if they are with the macro __STDC_ISO_10646__. If this is defined, the answer is, “It’s Unicode!” More detailedly, the value in that macro is an integer in the form yyyymm that lets you know what Unicode standard you can rely on—whatever was in effect on that date. But how do you use them?
27.8.1 Multibyte to wchar_t Conversions So how do we get from the byte-oriented standard strings to the character-oriented wide strings and back? We can use a couple string conversion functions to make this happen. Chapter 27. Unicode, Wide Characters, and All That 176
First, some naming conventions you’ll see in these functions: • mb: multibyte • wc: wide character • mbs: multibyte string • wcs: wide character string So if we want to convert a multibyte string to a wide character string, we can call the mbstowcs(). And the other way around: wcstombs().
Conversion Function Description mbtowc() Convert a multibyte character to a wide character. wctomb() Convert a wide character to a multibyte character. mbstowcs() Convert a multibyte string to a wide string. wcstombs() Convert a wide string to a multibyte string.
Let’s do a quick demo where we convert a multibyte string to a wide character string, and compare the string lengths of the two using their respective functions.
1 #include
2 #include
3 #include
4 #include
5 #include
6
7 int main(void)
8 {
9 // Get out of the C locale to one that likely has the euro symbol
10 setlocale(LC_ALL, "");
11
12 // Original multibyte string with a euro symbol (Unicode point 20ac)
13 char *mb_string = "The cost is \u20ac1.23"; // €1.23
14 size_t mb_len = strlen(mb_string);
15
16 // Wide character array that will hold the converted string
17 wchar_t wc_string[128]; // Holds up to 128 wide characters
18
19 // Convert the MB string to WC; this returns the number of wide chars
20 size_t wc_len = mbstowcs(wc_string, mb_string, 128);
21
22 // Print result--note the %ls for wide char strings
23 printf("multibyte: \"%s\" (%zu bytes)\n", mb_string, mb_len);
24 printf("wide char: \"%ls\" (%zu characters)\n", wc_string, wc_len);
25 }
On my system, this outputs: multibyte: "The cost is €1.23" (19 bytes) wide char: "The cost is €1.23" (17 characters)
(Your system might vary on the number of bytes depending on your locale.) One interesting thing to note is that mbstowcs(), in addition to converting the multibyte string to wide, returns the length (in characters) of the wide character string. And, in fact, it has a special mode where it only returns the length-in-characters of a given multibyte string: you just pass NULL to the destination, and 0 to the maximum number of characters to convert (this value is ignored). (In the code below, I’m using my extended source character set—you might have to replace those with \u escapes.) Chapter 27. Unicode, Wide Characters, and All That 177 setlocale(LC_ALL, "");
// The following string has 7 characters size_t len_in_chars = mbstowcs(NULL, "§¶°±π€•", 0)); printf("%zu", len_in_chars); // 7
And, of course, if you want to convert the other way, it’s wcstombs().
27.9 Wide Character Functionality
Once we’re in wide character land, we have all kinds of functionality at our disposal. I’m just going to summarize a bunch of the functions here, but basically what we have here are the wide character versions of the multibyte string functions that we’re use to. (For example, we know strlen() for multibyte strings; there’s an wcslen() for wide character strings.)
27.9.1 wint_t
A lot of these functions use a wint_t to hold single characters, whether they are passed in or returned. It is related to wchar_t in nature. A wint_t is an integer that can represent all values in the extended character set, and also a special end-of-file character, WEOF. This is used by a number of single-character-oriented wide character functions.
27.9.2 I/O Stream Orientation
The tl;dr here is to not mix and match byte-oriented functions (like fprintf()) with wide-oriented func- tions (like fwprintf()). Decide if a stream will be byte-oriented or wide-oriented and stick with those types of I/O functions. In more detail: streams can be either byte-oriented or wide-oriented. When a stream is first created, it has no orientation, but the first read or write will set the orientation. If you first use a wide operation (like fwprintf()) it will orient the stream wide. If you first use a byte operation (like fprintf()) it will orient the stream by bytes. You can manually set an unoriented stream one way or the other with a call to fwide(). You can use that same function to get the orientation of a stream. If you need to change the orientation mid-flight, you can do it with freopen().
27.9.3 I/O Functions
Typically include
I/O Function Description wprintf() Formatted console output. wscanf() Formatted console input. getwchar() Character-based console input. putwchar() Character-based console output. fwprintf() Formatted file output. fwscanf() Formatted file input. fgetwc() Character-based file input. fputwc() Character-based file output. fgetws() String-based file input. fputws() String-based file output. swprintf() Formatted string output. swscanf() Formatted string input. Chapter 27. Unicode, Wide Characters, and All That 178
I/O Function Description vfwprintf() Variadic formatted file output. vfwscanf() Variadic formatted file input. vswprintf() Variadic formatted string output. vswscanf() Variadic formatted string input. vwprintf() Variadic formatted console output. vwscanf() Variadic formatted console input. ungetwc() Push a wide character back on an output stream. fwide() Get or set stream multibyte/wide orientation.
27.9.4 Type Conversion Functions
Typically include
Conversion Function Description wcstod() Convert string to double. wcstof() Convert string to float. wcstold() Convert string to long double. wcstol() Convert string to long. wcstoll() Convert string to long long. wcstoul() Convert string to unsigned long. wcstoull() Convert string to unsigned long long.
27.9.5 String and Memory Copying Functions
Typically include
Copying Function Description wcscpy() Copy string. wcsncpy() Copy string, length-limited. wmemcpy() Copy memory. wmemmove() Copy potentially-overlapping memory. wcscat() Concatenate strings. wcsncat() Concatenate strings, length-limited.
27.9.6 String and Memory Comparing Functions
Typically include
Comparing Function Description wcscmp() Compare strings lexicographically. wcsncmp() Compare strings lexicographically, length-limited. wcscoll() Compare strings in dictionary order by locale. wmemcmp() Compare memory lexicographically. wcsxfrm() Transform strings into versions such that wcscmp() behaves like wcscoll()9.
27.9.7 String Searching Functions
Typically include
9wcscoll() is the same as wcsxfrm() followed by wcscmp(). Chapter 27. Unicode, Wide Characters, and All That 179
Searching Function Description wcschr() Find a character in a string. wcsrchr() Find a character in a string from the back. wmemchr() Find a character in memory. wcsstr() Find a substring in a string. wcspbrk() Find any of a set of characters in a string. wcsspn() Find length of substring including any of a set of characters. wcscspn() Find length of substring before any of a set of characters. wcstok() Find tokens in a string.
27.9.8 Length/Miscellaneous Functions
Typically include
Length/Misc Function Description wcslen() Return the length of the string. wmemset() Set characters in memory. wcsftime() Formatted date and time output.
27.9.9 Character Classification Functions
Include
Length/Misc Function Description iswalnum() True if the character is alphanumeric. iswalpha() True if the character is alphabetic. iswblank() True if the character is blank (space-ish, but not a newline). iswcntrl() True if the character is a control character. iswdigit() True if the character is a digit. iswgraph() True if the character is printable (except space). iswlower() True if the character is lowercase. iswprint() True if the character is printable (including space). iswpunct() True if the character is punctuation. iswspace() True if the character is whitespace. iswupper() True if the character is uppercase. iswxdigit() True if the character is a hex digit. towlower() Convert character to lowercase. towupper() Convert character to uppercase.
27.10 Parse State, Restartable Functions
We’re going to get a little bit into the guts of multibyte conversion, but this is a good thing to understand, conceptually. Imagine how your program takes a sequence of multibyte characters and turns them into wide characters, or vice-versa. It might, at some point, be partway through parsing a character, or it might have to wait for more bytes before it makes the determination of the final value. This parse state is stored in an opaque variable of type mbstate_t and is used every time conversion is performed. That’s how the conversion functions keep track of where they are mid-work. And if you change to a different character sequence mid-stream, or try to seek to a different place in your input sequence, it could get confused over that. Now you might want to call me on this one: we just did some conversions, above, and I never mentioned any mbstate_t anywhere. Chapter 27. Unicode, Wide Characters, and All That 180
That’s because the conversion functions like mbstowcs(), wctomb(), etc. each have their own mb- state_t variable that they use. There’s only one per function, though, so if you’re writing multithreaded code, they’re not safe to use. Fortunately, C defines restartable versions of these functions where you can pass in your own mbstate_t on per-thread basis if you need to. If you’re doing multithreaded stuff, use these! Quick note on initializing an mbstate_t variable: just memset() it to zero. There is no built-in function to force it to be initialized. mbstate_t mbs;
// Set the state to the initial state memset(&mbs, 0, sizeof mbs);
Here is a list of the restartable conversion functions—note the naming convension of putting an “r” after the “from” type: • mbrtowc()—multibyte to wide character • wcrtomb()—wide character to multibyte • mbsrtowcs()—multibyte string to wide character string • wcsrtombs()—wide character string to multibyte string These are really similar to their non-restartable counterparts, except they require you pass in a pointer to your own mbstate_t variable. And also they modify the source string pointer (to help you out if invalid bytes are found), so it might be useful to save a copy of the original. Here’s the example from earlier in the chapter reworked to pass in our own mbstate_t.
1 #include
2 #include
3 #include
4 #include
5 #include
6 #include
7
8 int main(void)
9 {
10 // Get out of the C locale to one that likely has the euro symbol
11 setlocale(LC_ALL, "");
12
13 // Original multibyte string with a euro symbol (Unicode point 20ac)
14 char *mb_string = "The cost is \u20ac1.23"; // €1.23
15 size_t mb_len = strlen(mb_string);
16
17 // Wide character array that will hold the converted string
18 wchar_t wc_string[128]; // Holds up to 128 wide characters
19
20 // Set up the conversion state
21 mbstate_t mbs;
22 memset(&mbs, 0, sizeof mbs); // Initial state
23
24 // mbsrtowcs() modifies the input pointer to point at the first
25 // invalid character, or NULL if successful. Let's make a copy of
26 // the pointer for mbsrtowcs() to mess with so our original is
27 // unchanged.
28 //
29 // This example will probably be successful, but we check farther
30 // down to see.
31 const char *invalid = mb_string;
32
33 // Convert the MB string to WC; this returns the number of wide chars Chapter 27. Unicode, Wide Characters, and All That 181
34 size_t wc_len = mbsrtowcs(wc_string, &invalid, 128, &mbs);
35
36 if (invalid == NULL) {
37 printf("No invalid characters found\n");
38
39 // Print result--note the %ls for wide char strings
40 printf("multibyte: \"%s\" (%zu bytes)\n", mb_string, mb_len);
41 printf("wide char: \"%ls\" (%zu characters)\n", wc_string, wc_len);
42 } else {
43 ptrdiff_t offset = invalid - mb_string;
44 printf("Invalid character at offset %td\n", offset);
45 }
46 }
For the conversion functions that manage their own state, you can reset their internal state to the initial one by passing in NULL for their char* arguments, for example: mbstowcs(NULL, NULL, 0); // Reset the parse state for mbstowcs() mbstowcs(dest, src, 100); // Parse some stuff
For I/O, each wide stream manages its own mbstate_t and uses that for input and output conversions as it goes. And some of the byte-oriented I/O functions like printf() and scanf() keep their own internal state while doing their work. Finally, these restartable conversion functions do actually have their own internal state if you pass in NULL for the mbstate_t parameter. This makes them behave more like their non-restartable counterparts.
27.11 Unicode Encodings and C
In this section, we’ll see what C can (and can’t) do when it comes to three specific Unicode encodings: UTF-8, UTF-16, and UTF-32.
27.11.1 UTF-8 To refresh before this section, read the UTF-8 quick note, above. Aside from that, what are C’s UTF-8 capabilities? Well, not much, unfortunately. You can tell C that you specifically want a string literal to be UTF-8 encoded, and it’ll do it for you. You can prefix a string with u8: char *s = u8"Hello, world!";
printf("%s\n", s); // Hello, world!--if you can output UTF-8
Now, can you put Unicode characters in there? char *s = u8"€123";
Sure! If the extended source character set supports it. (gcc does.) What if it doesn’t? You can specify a Unicode code point with your friendly neighborhood \u and \U, as noted above. But that’s about it. There’s no portable way in the standard library to take arbirary input and turn it into UTF-8 unless your locale is UTF-8. Or to parse UTF-8 unless your locale is UTF-8. So if you want to do it, either be in a UTF-8 locale and: Chapter 27. Unicode, Wide Characters, and All That 182 setlocale(LC_ALL, ""); or figure out a UTF-8 locale name on your local machine and set it explicitly like so: setlocale(LC_ALL, "en_US.UTF-8"); // Non-portable name
Or use a third-party library.
27.11.2 UTF-16, UTF-32, char16_t, and char32_t char16_t and char32_t are a couple other potentially wide character types with sizes of 16 bits and 32 bits, respectively. Not necessarily wide, because if they can’t represent every character in the current locale, they lose their wide character nature. But the spec refers them as “wide character” types all over the place, so there we are. These are here to make things a little more Unicode-friendly, potentially. To use, include
Now—are values in these stored in UTF-16 or UTF-32? Depends on the implementation. But you can test to see if they are. If the macros __STDC_UTF_16__ or __STDC_UTF_32__ are defined (to 1) it means the types hold UTF-16 or UTF-32, respectively. If you’re curious, and I know you are, the values, if UTF-16 or UTF-32, are stored in the native endianess. That is, you should be able to compare them straight up to Unicode code point values: char16_t pi = u"\u03C0"; // pi symbol
#if __STDC_UTF_16__ pi == 0x3C0; // Always true #else pi == 0x3C0; // Probably not true #endif
27.11.3 Multibyte Conversions
You can convert from your multibyte encoding to char16_t or char32_t with a number of helper func- tions. (Like I said, though, the result might not be UTF-16 or UTF-32 unless the corresponding macro is set to 1.) All of these functions are restartable (i.e. you pass in your own mbstate_t), and all of them operate character by character10.
Conversion Function Description mbrtoc16() Convert a multibyte character to a char16_t character. mbrtoc32() Convert a multibyte character to a char32_t character. c16rtomb() Convert a char16_t character to a multibyte character. c32rtomb() Convert a char32_t character to a multibyte character.
10Ish—things get funky with multi-char16_t UTF-16 encodings. Chapter 27. Unicode, Wide Characters, and All That 183
27.11.4 Third-Party Libraries For heavy-duty conversion between different specific encodings, there are a couple mature libraries worth checking out. Note that I haven’t used either of these. • iconv11—Internationalization Conversion, a common POSIX-standard API available on the major platforms. • ICU12—International Components for Unicode. At least one blogger found this easy to use. If you have more noteworthy libraries, let me know.
11https://en.wikipedia.org/wiki/Iconv 12http://site.icu-project.org/ Chapter 28
Exiting a Program
Turns out there are a lot of ways to do this, and even ways to set up “hooks” so that a function runs when a program exits. In this chapter we’ll dive in and check them out. We already covered the meaning of the exit status code in the Exit Status section, so jump back there and review if you have to. All the functions in this section are in
28.1 Normal Exits
We’ll start with the regular ways to exit a program, and then jump to some of the rarer, more esoteric ones. When you exit a program normally, all open I/O streams are flushed and temporary files removed. Basi- cally it’s a nice exit where everything gets cleaned up and handled. It’s what you want to do almost all the time unless you have reasons to do otherwise.
28.1.1 Returning From main()
If you’ve noticed, main() has a return type of int… and yet I’ve rarely, if ever, been returning anything from main() at all. This is because for main() only (and I can’t stress enough this special case only applies to main() and no other functions anywhere) has an implicit return 0 if you fall off the end. You can explicitly return from main() any time you want, and some programmers feel it’s more Right to always have a return at the end of main(). But if you leave it off, C will put one there for you. So… here are the return rules for main(): • You can return an exit status from main() with a return statement. main() is the only function with this special behavior. Using return in any other function just returns from that function to the caller. • If you don’t explicitly return and just fall off the end of main(), it’s just as if you’d returned 0 or EXIT_SUCCESS.
28.1.2 exit()
This one has also made an appearance a few times. If you call exit() from anywhere in your program, it will exit at that point. The argument you pass to exit() is the exit status.
184 Chapter 28. Exiting a Program 185
28.1.3 Setting Up Exit Handlers with atexit()
You can register functions to be called when a program exits whether by returning from main() or calling the exit() function. A call to atexit() with the handler function name will get it done. You can register multiple exit handlers, and they’ll be called in the reverse order of registration. Here’s an example:
1 #include
2 #include
3
4 void on_exit_1(void)
5 {
6 printf("Exit handler 1 called!\n");
7 }
8
9 void on_exit_2(void)
10 {
11 printf("Exit handler 2 called!\n");
12 }
13
14 int main(void)
15 {
16 atexit(on_exit_1);
17 atexit(on_exit_2);
18
19 printf("About to exit...\n");
20 }
And the output is: About to exit... Exit handler 2 called! Exit handler 1 called!
28.2 Quicker Exits with quick_exit()
This is similar to a normal exit, except: • Open files might not be flushed. • Temporary files might not be removed. • atexit() handlers won’t be called. But there is a way to register exit handlers: call at_quick_exit() analogously to how you’d call atexit().
1 #include
2 #include
3
4 void on_quick_exit_1(void)
5 {
6 printf("Quick exit handler 1 called!\n");
7 }
8
9 void on_quick_exit_2(void)
10 {
11 printf("Quick exit handler 2 called!\n");
12 }
13
14 void on_exit(void) Chapter 28. Exiting a Program 186
15 {
16 printf("Normal exit--I won't be called!\n");
17 }
18
19 int main(void)
20 {
21 at_quick_exit(on_quick_exit_1);
22 at_quick_exit(on_quick_exit_2);
23
24 atexit(on_exit); // This won't be called
25
26 printf("About to quick exit...\n");
27
28 quick_exit(0);
29 }
Which gives this output: About to quick exit... Quick exit handler 2 called! Quick exit handler 1 called!
It works just like exit()/atexit(), except for the fact that file flushing and cleanup might not be done.
28.3 Nuke it from Orbit: _Exit()
Calling _Exit() exits immediately, period. No on-exit callback functions are executed. Files won’t be flushed. Temp files won’t be removed. Use this if you have to exit right fargin’ now.
28.4 Exiting Sometimes: assert()
The assert() statement is used to insist that something be true, or else the program will exit. Devs often use an assert to catch Should-Never-Happen type errors. #define PI 3.14159
assert(PI > 3); // Sure enough, it is, so carry on
versus: goats -= 100;
assert(goats >= 0); // Can't have negative goats
In that case, if I try to run it and goats falls under 0, this happens: goat_counter: goat_counter.c:8: main: Assertion `goats >= 0' failed. Aborted
and I’m dropped back to the command line. This isn’t very user-friendly, so it’s only used for things the user will never see. And often people write their own assert macros that can more easily be turned off.
28.5 Abnormal Exit: abort()
You can use this if something has gone horribly wrong and you want to indicate as much to the outside environment. This also won’t necessarily clean up any open files, etc. Chapter 28. Exiting a Program 187
I’ve rarely seen this used. Some foreshadowing about signals: this actually works by raising a SIGABRT which will end the process. What happens after that is up to the system, but on Unix-likes, it was common to dump core1 as the program terminated.
1https://en.wikipedia.org/wiki/Core_dump Chapter 29
Signal Handling
Before we start, I’m just going to advise you to generally ignore this entire chapter and use your OS’s (very likely) superior signal handling functions. Unix-likes have the sigaction() family of functions, and Windows has… whatever it does1. With that out of the way, what are signals?
29.1 What Are Signals?
A signal is raised on a variety of external events. Your program can be configured to be interrupted to handle the signal, and, optionally, continue where it left off once the signal has been handled. Think of it like a function that’s automatically called when one of these external events occurs. What are these events? On your system, there are probably a lot of them, but in the C spec there are just a few:
Signal Description SIGABRT Abnormal termination—what happens when abort() is called. SIGFPE Floating point exception. SIGILL Illegal instruction. SIGINT Interrupt—usually the result of CTRL-C being hit. SIGSEGV “Segmentation Violation”: invalid memory access. SIGTERM Termination requested.
You can set up your program to ignore, handle, or allow the default action for each of these by using the signal() function.
29.2 Handling Signals with signal()
The signal() call takes two parameters: the signal in question, and an action to take when that signal is raised. The action can be one of three things: • A pointer to a handler function. • SIG_IGN to ignore the signal. • SIG_DFL to restore the default handler for the signal. Let’s write a program that you can’t CTRL-C out of. (Don’t fret—in the following program, you can also hit RETURN and it’ll exit.)
1Apparently it doesn’t do Unix-style signals at all deep down, and they’re simulated for console apps.
188 Chapter 29. Signal Handling 189
1 #include
2 #include
3
4 int main(void)
5 {
6 char s[1024];
7
8 signal(SIGINT, SIG_IGN); // Ignore SIGINT, caused by ^C
9
10 printf("Try hitting ^C...\n");
11
12 // Wait for a line of input so the program doesn't just exit
13 fgets(s, sizeof s, stdin);
14 }
Check out line 8—we tell the program to ignore SIGINT, the interrupt signal that’s raised when CTRL-C is hit. No matter how much you hit it, the signal remains ignored. If you comment out line 8, you’ll see you can CTRL-C with impunity and quit the program on the spot.
29.3 Writing Signal Handlers
I mentioned you could also write a handler function that gets called with the signal is raised. These are pretty straightforward, are also very capability-limited when it comes to the spec. Before we start, let’s look at the function prototype for the signal() call: void (*signal(int sig, void (*func)(int)))(int);
Pretty easy to read, right? WRONG! :) Let’s take a moment to take it apart for practice. signal() takes two arguments: an integer sig representing the signal, and a pointer func to the handler (the handler returns void and takes an int as an argument), highlighted below: sig func |-----| |------| void (*signal(int sig, void (*func)(int)))(int);
Basically, we’re going to pass in the signal number we’re interesting in catching, and we’re going to pass a pointer to a function of the form: void f(int x);
that will do the actual catching. Now—what about the rest of that prototype? It’s basically all the return type. See, signal() will return whatever you passed as func on success… so that means it’s returning a pointer to a function that returns void and takes an int as an argument. returned function indicates we're and returns returning a that function void pointer to function takes an int |--| | |---| void (*signal(int sig, void (*func)(int)))(int);
Also, it can return SIG_ERR in case of an error. Let’s do an example where we make it so you have to hit CTRL-C twice to exit. Chapter 29. Signal Handling 190
I want to be clear that this program engages in undefined behavior in a couple ways. But it’ll probably work for you, and it’s hard to come up with portable non-trivial demos.
1 #include
2 #include
3 #include
4
5 int count = 0;
6
7 void sigint_handler(int signum)
8 {
9 // The compiler is allowed to run:
10 //
11 // signal(signum, SIG_DFL)
12 //
13 // when the handler is first called. So we reset the handler here:
14 signal(SIGINT, sigint_handler);
15
16 (void)signum; // Get rid of unused variable warning
17
18 count++; // Undefined behavior
19 printf("Count: %d\n", count); // Undefined behavior
20
21 if (count == 2) {
22 printf("Exiting!\n"); // Undefined behavior
23 exit(0);
24 }
25 }
26
27 int main(void)
28 {
29 signal(SIGINT, sigint_handler);
30
31 printf("Try hitting ^C...\n");
32
33 for(;;); // Wait here forever
34 }
One of the things you’ll notice is that on line 14 we reset the signal handler. This is because C has the option of resetting the signal handler to its SIG_DFL behavior before running your custom handler. In other words, it could be a one-off. So we reset it first thing so that we handle it again for the next one. We’re ignoring the return value from signal() in this case. If we’d set it to a different handler earlier, it would return a pointer to that handler, which we could get like this: // old_handler is type "pointer to function that takes a single // int parameter and returns void":
void (*old_handler)(int);
old_handler = signal(SIGINT, sigint_handler);
That said, I’m not sure of a common use case for this. But if you need the old handler for some reason, you can get it that way. Quick note on line 16—that’s just to tell the compiler to not warn that we’re not using this variable. It’s like saying, “I know I’m not using it; you don’t have to warn me.” And lastly you’ll see that I’ve marked undefined behavior in a couple places. More on that in the next section. Chapter 29. Signal Handling 191
29.4 What Can We Actually Do?
Turns out we’re pretty limited in what we can and can’t do in our signal handlers. This is one of the reasons why I say you shouldn’t even bother with this and instead use your OS’s signal handling instead (e.g. sigaction() for Unix-like systems). Wikipedia goes so far as to say the only really portable thing you can do is call signal() with SIG_IGN or SIG_DFL and that’s it. Here’s what we can’t portably do: • Call any standard library function. – Like printf(), for example. – I think it’s probably safe to call restartable/reentrant functions, but the spec doesn’t allow that liberty. • Get or set values from a local static, file scope, or thread-local variable. – Unless it’s a lock-free atomic object or… – You’re assigning into a variable of type volatile sig_atomic_t. That last bit–sig_atomic_t–is your ticket to getting data out of a signal handler. (Unless you want to use lock-free atomic objects, which is outside the scope of this section2.) It’s an integer type that might or might not be signed. And it’s bounded by what you can put in there. You can look at the minimum and maximum allowable values in the macros SIG_ATOMIC_MIN and SIG_ATOMIC_MAX3. Confusingly, the spec also says you can’t refer “to any object with static or thread storage duration that is not a lock-free atomic object other than by assigning a value to an object declared as volatile sig_atomic_t […]” My read on this is that you can’t read or write anything that’s not a lock-free atomic object. Also you can assign to an object that’s volatile sig_atomic_t. But can you read from it? I honestly don’t see why not, except that the spec is very pointed about men- tioning assigning into. But if you have to read it and make any kind of decision based on it, you might be opening up room for some kind of race conditions. With that in mind, we can rewrite our “hit CTRL-C twice to exit” code to be a little more portable, albeit less verbose on the output. Let’s change our SIGINT handler to do nothing except increment a value that’s of type volatile sig_atomic_t. So it’ll count the number of CTRL-Cs that have been hit. Then in our main loop, we’ll check to see if that counter is over 2, then bail out if it is.
1 #include
2 #include
3
4 volatile sig_atomic_t count = 0;
5
6 void sigint_handler(int signum)
7 {
8 (void)signum; // Unused variable warning
9
10 signal(SIGINT, sigint_handler); // Reset signal handler
11
12 count++; // Undefined behavior
13 }
14
15 int main(void)
16 {
17 signal(SIGINT, sigint_handler);
2Confusingly, sig_atomic_t predates the lock-free atomics and is not the same thing. 3If sig_action_t is signed, the range will be at least -127 to 127. If unsigned, at least 0 to 255. Chapter 29. Signal Handling 192
18
19 printf("Hit ^C twice to exit.\n");
20
21 while(count < 2);
22 }
Undefined behavior again? It’s my read that this is, because we have to read the value in order to increment and store it. We can do some ridiculous contortions so that we’re only assigning into the values and manage to avoid undefined behavior.
1 #include
2 #include
3
4 void sigint_handler_2(int signum)
5 {
6 (void)signum; // Unused variable warning
7 signal(SIGINT, SIG_DFL); // Reset signal handler
8 }
9
10 void sigint_handler_1(int signum)
11 {
12 (void)signum; // Unused variable warning
13 signal(SIGINT, sigint_handler_2); // Set to second handler
14 }
15
16 int main(void)
17 {
18 signal(SIGINT, sigint_handler_1);
19
20 printf("Hit ^C twice to exit.\n");
21
22 while(1);
23 }
That’s pretty ugly, all right. Later when we look at lock-free atomic variables, we’ll see a way to fix the count version (assuming lock-free atomic variables are available on your particular system), but we’re getting into zanyland here. This is why at the beginning, I was suggesting checking out your OS’s built-in signal system as a probably- superior alternative.
29.5 Friends Don’t Let Friends signal()
Again, use your OS’s built-in signal handling or the equivalent. It’s not in the spec, not as portable, but probably is far more capable. Plus your OS probably has a number of signals defined that aren’t in the C spec. And it’s difficult to write portable code using signal() anyway. Chapter 30
Variable-Length Arrays (VLAs)
C provides a way for you to declare an array whose size is determined at runtime. This gives you the benefits of dynamic runtime sizing like you get with malloc(), but without needing to worry about free()ing the memory after. Now, a lot of people don’t like VLAs. They’ve been banned from the Linux kernel, for example. We’ll dig into more of that rationale later. This is an optional feature of the language. The macro __STDC_NO_VLA__ is set to 1 if VLAs are not present. (They were mandatory in C99, and then became optional in C11.) #if __STDC_NO_VLA__ == 1 #error Sorry, need VLAs for this program! #endif
Let’s dive in first with an example, and then we’ll look for the devil in the details.
30.1 The Basics
A normal array is declared with a constant size, like this: int v[10];
But with VLAs, we can use a size determined at runtime to set the array, like this: int n = 10; int v[n];
Now, that looks like the same thing, and in many ways is, but this gives you the flexibility to compute the size you need, and then get an array of exactly that size. Let’s ask the user to input the size of the array, and then store the index-times-10 in each of those array elements:
1 #include
2
3 int main(void)
4 {
5 int n;
6
7 printf("Enter a number: "); fflush(stdout);
8 scanf(" %d", &n);
9
10 int v[n];
11
12 for (int i = 0; i < n; i++)
13 v[i] = i * 10;
193 Chapter 30. Variable-Length Arrays (VLAs) 194
14
15 for (int i = 0; i < n; i++)
16 printf("v[%d] = %d\n", i, v[i]);
17 }
(On line 7, I have an fflush() that should force the line to output even though I don’t have a newline at the end.) Line 10 is where we declare the VLA—once execution gets past that line, the size of the array is set to whatever n was at that moment. The array length can’t be changed later. You can put an expression in the brackets, as well: int v[x * 100];
Some restrictions: • You can’t declare a VLA at file scope, and you can’t make a static one in block scope1. • You can’t use an initializer list to initialize the array. Also, entering a negative value for the size of the array invokes undefined behavior—in this universe, anyway.
30.2 sizeof and VLAs
We’re used to sizeof giving us the size in bytes of any particular object, including arrays. And VLAs are no exception. The main difference is that sizeof on a VLA is executed at runtime, whereas on a non-variably-sized variable it is computed at compile time. But the usage is the same. You can even compute the number of elements in a VLA with the usual array trick: size_t num_elems = sizeof v / sizeof v[0];
There’s a subtle and correct implication from the above line: pointer arithmetic works just like you’d expect for a regular array. So go ahead and use it to your heart’s content:
1 #include
2
3 int main(void)
4 {
5 int n = 5;
6 int v[n];
7
8 int *p = v;
9
10 *(p+2) = 12;
11 printf("%d\n", v[2]); // 12
12
13 p[3] = 34;
14 printf("%d\n", v[3]); // 34
15 }
Like with regular arrays, you can use parentheses with sizeof() to get the size of a would-be VLA without actually declaring one: int x = 12;
printf("%zu\n", sizeof(int [x])); // Prints 48 on my system
1This is due to how VLAs are typically allocated on the stack, whereas static variables are on the heap. And the whole idea with VLAs is they’ll be automatically dellocated when the stack frame is popped at the end of the function. Chapter 30. Variable-Length Arrays (VLAs) 195
30.3 Multidimensional VLAs
You can go ahead and make all kinds of VLAs with one or more dimensions set to a variable int w = 10; int h = 20;
int x[h][w]; int y[5][w]; int z[10][w][20];
Again, you can navigate these just like you would a regular array.
30.4 Passing One-Dimensional VLAs to Functions
Passing single-dimensional VLAs into a function can be no different than passing a regular array in. You just go for it.
1 #include
2
3 int sum(int count, int *v)
4 {
5 int total = 0;
6
7 for (int i = 0; i < count; i++)
8 total += v[i];
9
10 return total;
11 }
12
13 int main(void)
14 {
15 int x[5]; // Standard array
16
17 int a = 5;
18 int y[a]; // VLA
19
20 for (int i = 0; i < a; i++)
21 x[i] = y[i] = i + 1;
22
23 printf("%d\n", sum(5, x));
24 printf("%d\n", sum(a, y));
25 }
But there’s a bit more to it than that. You can also let C know that the array is a specific VLA size by passing that in first and then giving that dimension in the parameter list: int sum(int count, int v[count]) { // ... }
Incidentally, there are a couple ways of listing a prototype for the above function; one of them involves an * if you don’t want to specifically name the value in the VLA. It just indicates that the type is a VLA as opposed to a regular pointer. VLA prototypes: void do_something(int count, int v[count]); // With names void do_something(int, int v[*]); // Without names Chapter 30. Variable-Length Arrays (VLAs) 196
Again, that * thing only works with the prototype—in the function itself, you’ll have to put the explicit size. Now—let’s get multidimensional! This is where the fun begins.
30.5 Passing Multi-Dimensional VLAs to Functions
Same thing as we did with the second form of one-dimensional VLAs, above, but this time we’re passing in two dimensions and using those. In the following example, we build a multiplication table matrix of a variable width and height, and then pass it to a function to print it out.
1 #include
2
3 void print_matrix(int h, int w, int m[h][w])
4 {
5 for (int row = 0; row < h; row++) {
6 for (int col = 0; col < w; col++)
7 printf("%2d ", m[row][col]);
8 printf("\n");
9 }
10 }
11
12 int main(void)
13 {
14 int rows = 4;
15 int cols = 7;
16
17 int matrix[rows][cols];
18
19 for (int row = 0; row < rows; row++)
20 for (int col = 0; col < cols; col++)
21 matrix[row][col] = row * col;
22
23 print_matrix(rows, cols, matrix);
24 }
30.5.1 Partial Multidimensional VLAs You can have some of the dimensions fixed and some variable. Let’s say we have a record length fixed at 5 elements, but we don’t know how many records there are.
1 #include
2
3 void print_records(int count, int record[count][5])
4 {
5 for (int i = 0; i < count; i++) {
6 for (int j = 0; j < 5; j++)
7 printf("%2d ", record[i][j]);
8 printf("\n");
9 }
10 }
11
12 int main(void)
13 {
14 int rec_count = 3;
15 int records[rec_count][5];
16 Chapter 30. Variable-Length Arrays (VLAs) 197
17 // Fill with some dummy data
18 for (int i = 0; i < rec_count; i++)
19 for (int j = 0; j < 5; j++)
20 records[i][j] = (i+1)*(j+2);
21
22 print_records(rec_count, records);
23 }
30.6 Compatibility with Regular Arrays
Because VLAs are just like regular arrays in memory, it’s perfectly permissible to pass them interchange- ably… as long as the dimensions match. For example, if we have a function that specifically wants a 3 × 5 array, we can still pass a VLA into it. int foo(int m[5][3]) {...}
\\ ...
int w = 3, h = 5; int matrix[h][w];
foo(matrix); // OK!
Likewise, if you have a VLA function, you can pass a regular array into it: int foo(int h, int w, int m[h][w]) {...}
\\ ...
int matrix[3][5];
foo(3, 5, matrix); // OK!
Beware, though: if your dimensions mismatch, you’re going to have some undefined behavior going on, likely.
30.7 typedef and VLAs
You can typedef a VLA, but the behavior might not be as you expect. Basically, typedef makes a new type with the values as they existed the moment the typedef was executed. So it’s not a typedef of a VLA so much as a new fixed size array type of the dimensions at the time.
1 #include
2
3 int main(void)
4 {
5 int w = 10;
6
7 typedef int goat[w];
8
9 // goat is an array of 10 ints
10 goat x;
11
12 // Init with squares of numbers
13 for (int i = 0; i < w; i++)
14 x[i] = i*i; Chapter 30. Variable-Length Arrays (VLAs) 198
15
16 // Print them
17 for (int i = 0; i < w; i++)
18 printf("%d\n", x[i]);
19
20 // Now let's change w...
21
22 w = 20;
23
24 // But goat is STILL an array of 10 ints, because that was the
25 // value of w when the typedef executed.
26 }
So it acts like an array of fixed size. But you still can’t use an initializer list on it.
30.8 Jumping Pitfalls
You have to watch out when using goto near VLAs because a lot of things aren’t legal. And when you’re using longjmp() there’s a case where you could leak memory with VLAs. But both of these things we’ll cover in their respective chapters.
30.9 General Issues
VLAs have been banned from the Linux kernel for a few reasons: • Lots of places they were used should have just been fixed-size. • The code behind VLAs is slower (to a degree that most people wouldn’t notice, but makes a differ- ence in an operating system). • VLAs are not supported to the same degree by all C compilers. • Stack size is limited, and VLAs go on the stack. If some code accidentally (or maliciously) passes a large value into a kernel function that allocates a VLA, Bad Things™ could happen. Other folks online point out that there’s no way to detect a VLA’s failure to allocate, and programs that suffered such problems would likely just crash. While fixed-size arrays also have the same issue, it’s far more likely that someone accidentally make a VLA Of Unusual Size than somehow accidentally declare a fixed-size, say, 30 megabyte array. Chapter 31
goto
The goto statement is universally revered and can be here presented without contest. Just kidding! Over the years, there has been a lot of back-and-forth over whether or not (often not) goto is considered harmful1. In this programmer’s opinion, you should use whichever constructs leads to the best code, factoring in maintainability and speed. And sometimes this might be goto! In this chapter, we’ll see how goto works in C, and then check out some of the common cases where it is used2.
31.1 A Simple Example
In this example, we’re going to use goto to skip a line of code and jump to a label. The label is the identifier that can be a goto target—it ends with a colon (:).
1 #include
2
3 int main(void)
4 {
5 printf("One\n");
6 printf("Two\n");
7
8 goto skip_3;
9
10 printf("Three\n");
11
12 skip_3:
13
14 printf("Five!\n");
15 }
The output is: One Two Five!
goto sends execution jumping to the specified label, skipping everything in between. You can jump forward or backward with goto.
1https://en.wikipedia.org/wiki/Goto#Criticism 2I’d like to point out that using goto in all these cases is avoidable. You can use variables and loops intead. It’s just that some people think goto produces the best code in those circumstances.
199 Chapter 31. goto 200 infinite_loop: print("Hello, world!\n"); goto infinite_loop;
Labels are skipped over during execution. The following will print all three numbers in order just as if the labels weren’t there: printf("Zero\n"); label_1: label_2: printf("One\n"); label 3: printf("Two\n"); label 4: printf("Three\n");
As you’ve noticed, it’s common convention to justify the labels all the way on the left. This increases readability because a reader can quickly scan to find the destination. Labels have function scope. That is, no matter how many levels deep in blocks they appear, you can still goto them from anywhere in the function. It also means you can only goto labels that are in the same function as the goto itself. Labels in other functions are out of scope from goto’s perspective. And it means you can use the same label name in two functions—just not the same label name in the same function.
31.2 Labeled continue
In some languages, you can actually specify a label for a continue statement. C doesn’t allow it, but you can easily use goto instead. To show the issue, check out continue in this nested loop: for (int i = 0; i < 3; i++) { for (int j = 0; j < 3; j++) { printf("%d, %d\n", i, j); continue; // Always goes to next j } }
As we see, that continue, like all continues, goes to the next iteration of the nearest enclosing loop. What if we want to continue in the next loop out, the loop with i? Well, we can break to get back to the outer loop, right? for (int i = 0; i < 3; i++) { for (int j = 0; j < 3; j++) { printf("%d, %d\n", i, j); break; // Gets us to the next iteration of i } }
That gets us two levels of nested loop. But then if we nest another loop, we’re out of options. What about this, where we don’t have any statement that will get us out to the next iteration of i? for (int i = 0; i < 3; i++) { for (int j = 0; j < 3; j++) { for (int k = 0; k < 3; k++) { printf("%d, %d, %d\n", i, j, k);
continue; // Gets us to the next iteration of k break; // Gets us to the next iteration of j Chapter 31. goto 201
????; // Gets us to the next iteration of i???
} } }
The goto statement offers us a way! for (int i = 0; i < 3; i++) { for (int j = 0; j < 3; j++) { for (int k = 0; k < 3; k++) { printf("%d, %d, %d\n", i, j, k);
goto continue_i; // Now continuing the i loop!! } } continue_i: ; }
We have a ; at the end there—that’s because you can’t have a label pointing to the plain end of a compound statement (or before a variable declaration).
31.3 Bailing Out
When you’re super nested in the middle of some code, you can use goto to get out of it in a manner that’s often cleaner than nesting more ifs and using flag variables. // Pseudocode
for(...) { for (...) { while (...) { do { if (some_error_condition) goto bail;
} while(...); } } } bail: // Cleanup here
Without goto, you’d have to check an error condition flag in all of the loops to get all the way out.
31.4 Labeled break
This is a very similar situation to how continue only continues the innermost loop. break also only breaks out of the innermost loop. for (int i = 0; i < 3; i++) { for (int j = 0; j < 3; j++) { printf("%d, %d\n", i, j); break; // Only breaks out of the j loop } }
printf("Done!\n"); Chapter 31. goto 202
But we can use goto to break farther: for (int i = 0; i < 3; i++) { for (int j = 0; j < 3; j++) { printf("%d, %d\n", i, j); goto break_i; // Now breaking out of the i loop! } } break_i:
printf("Done!\n");
31.5 Multi-level Cleanup
If you’re calling multiple functions to initialize multiple systems and one of them fails, you should only de-initialize the ones that you’ve gotten to so far. Let’s do a fake example where we start initializing systems and checking to see if any returns an error (we’ll use -1 to indicate an error). If one of them does, we have to shutdown only the systems we’ve initialized so far. if (init_system_1() == -1) goto shutdown;
if (init_system_2() == -1) goto shutdown_1;
if (init_system_3() == -1) goto shutdown_2;
if (init_system_4() == -1) goto shutdown_3;
do_main_thing(); // Run our program
shutdown_system4(); shutdown_3: shutdown_system3(); shutdown_2: shutdown_system2(); shutdown_1: shutdown_system1(); shutdown: print("All subsystems shut down.\n");
Note that we’re shutting down in the reverse order that we initialized the subsystems. So if subsystem 4 fails to start up, it will shut down 3, 2, then 1 in that order.
31.6 Restarting Interrupted System Calls
This is outside the spec, but commonly seen in Unix-like systems. Certain long-lived system calls might return an error if they’re interrupted by a signal, and errno will be set to EINTR to indicate the syscall was doing fine; it was just interrupted. Chapter 31. goto 203
In those cases, it’s really common for the programmer to want to restart the call and try it again. retry: byte_count = read(0, buf, sizeof(buf) - 1); // Unix read() syscall
if (byte_count == -1) { // An error occurred... if (errno == EINTR) { // But it was just interrupted printf("Restarting...\n"); goto retry; }
Many Unix-likes have an SA_RESTART flag you can pass to sigaction() to request the OS automatically restart any slow syscalls instead of failing with EINTR. Again, this is Unix-specific and is outside the C standard. That said, it’s possible to use a similar technique any time any function should be restarted.
31.7 goto and Variable Scope
We’ve already seen that labels have function scope, but weird things can happen if we jump past some variable initialization. Look at this example where we jump from a place where the variable x is out of scope into the middle of its scope (in the block). goto label;
{ int x = 12345; label: printf("%d\n", x); }
This will compile and run, but gives me a warning: warning: ‘x’ is used uninitialized in this function
And then it prints out 0 when I run it (your mileage may vary). Basically what has happened is that we jumped into x’s scope (so it was OK to reference it in the printf()) but we jumped over the line that actually initialized it to 12345. So the value was indeterminate. The fix is, of course, to get the initialization after the label one way or another. goto label;
{ int x; label: x = 12345; printf("%d\n", x); }
Let’s look at one more example. { int x = 10; label:
printf("%d\n", x); Chapter 31. goto 204
}
goto label;
What happens here? The first time through the block, we’re good. x is 10 and that’s what prints. But after the goto, we’re jumping into the scope of x, but past its initialization. Which means we can still print it, but the value is indeterminate (since it hasn’t been reinitialized). On my machine, it prints 10 again (to infinity), but that’s just luck. It could print any value after the goto since x is uninitialized.
31.8 goto and Variable-Length Arrays
When it comes to VLAs and goto, there’s one rule: you can’t jump from outside the scope of a VLA into the scope of that VLA. If I try to do this: int x = 10;
goto label;
{ int v[x]; label:
printf("Hi!\n"); }
I get an error: error: jump into scope of identifier with variably modified type
You can jump in ahead of the VLA declaration, like this: int x = 10;
goto label;
{ label: ; int v[x];
printf("Hi!\n"); }
Because that way the VLA gets allocated properly before its inevitable deallocation once it falls out of scope. Chapter 32
Types Part V: Compound Literals and Generic Selections
This is the final chapter for types! We’re going to talk about two things: • How to have “anonymous” unnamed objects and how that’s useful. • How to generate type-dependent code. They’re not particularly related, but don’t really each warrant their own chapters. So I crammed them in here like a rebel!
32.1 Compound Literals
This is a neat feature of the language that allows you to create an object of some type on the fly without ever assigning it to a variable. You can make simple types, arrays, structs, you name it. One of the main uses for this is passing complex arguments to functions when you don’t want to make a temporary variable to hold the value. The way you create a compound literal is to put the type name in parentheses, and then put an initializer list after. For example, an unnamed array of ints, might look like this: (int []){1,2,3,4}
Now, that line of code doesn’t do anything on its own. It creates an unnamed array of 4 ints, and then throws them away without using them. We could use a pointer to store a reference to the array… int *p = (int []){1 ,2 ,3 ,4}; printf("%d\n", p[1]); // 2
But that seems a little like a long-winded way to have an array. I mean, we could have just done this1: int p[] = {1, 2, 3, 4}; printf("%d\n", p[1]); // 2
So let’s take a look at a more useful example.
32.1.1 Passing Unnamed Objects to Functions
Let’s say we have a function to sum an array of ints:
1Which isn’t quite the same, since it’s an array, not a pointer to an int.
205 Chapter 32. Types Part V: Compound Literals and Generic Selections 206
int sum(int p[], int count) { int total = 0;
for (int i = 0; i < count; i++) total += p[i];
return total; }
If we wanted to call it, we’d normally have to do something like this, declaring an array and storing values in it to pass to the function: int a[] = {1, 2, 3, 4};
int s = sum(a, 4);
But unnamed objects give us a way to skip the variable by passing it directly in (parameter names listed above). Check it out—we’re going to replace the variable a with an unnamed array that we pass in as the second argument: // p[] count // |------| | int s = sum((int []){1, 2, 3, 4}, 4);
Pretty slick!
32.1.2 Unnamed structs
We can do something similar with structs. First, let’s do things without unnamed objects. We’ll define a struct to hold some x/y coordinates. Then we’ll define one, passing in values into its initializer. Finally, we’ll pass it to a function to print the values:
1 #include
2
3 struct coord {
4 int x, y;
5 };
6
7 void print_coord(struct coord c)
8 {
9 printf("%d, %d\n", c.x, c.y);
10 }
11
12 int main(void)
13 {
14 struct coord t = {.x=10, .y=20};
15
16 print_coord(t); // prints "10, 20"
17 }
Straightforward enough? Let’s modify it to use an unnamed object instead of the variable t we’re passing to print_coord(). We’ll just take t out of there and replace it with an unnamed struct:
7 //struct coord t = {.x=10, .y=20};
8
9 print_coord((struct coord){.x=10, .y=20}); // prints "10, 20"
Still works! Chapter 32. Types Part V: Compound Literals and Generic Selections 207
32.1.3 Pointers to Unnamed Objects
You might have noticed in the last example that even through we were using a struct, we were passing a copy of the struct to print_coord() as opposed to passing a pointer to the struct. Turns out, we can just take the address of an unnamed object with & like always. This is because, in general, if an operator would have worked on a variable of that type, you can use that operator on an unnamed object of that type. Let’s modify the above code so that we pass a pointer to an unnamed object
1 #include
2
3 struct coord {
4 int x, y;
5 };
6
7 void print_coord(struct coord *c)
8 {
9 printf("%d, %d\n", c->x, c->y);
10 }
11
12 int main(void)
13 {
14 // Note the &
15 // |
16 print_coord(&(struct coord){.x=10, .y=20}); // prints "10, 20"
17 }
Additionally, this can be a nice way to pass even pointers to simple objects: // Pass a pointer to an int with value 3490 foo(&(int){3490});
Easy as that.
32.1.4 Unnamed Objects and Scope The lifetime of an unnamed object ends at the end of its scope. The biggest way this could bite you is if you make a new unnamed object, get a pointer to it, and then leave the object’s scope. In that case, the pointer will refer to a dead object. So this is undefined behavior: int *p;
{ p = &(int){10}; }
printf("%d\n", *p); // INVALID: The (int){10} fell out of scope
Likewise, you can’t return a pointer to an unnamed object from a function. The object is deallocated when it falls out of scope:
1 #include
2
3 int *get3490(void)
4 {
5 // Don't do this
6 return &(int){3490};
7 } Chapter 32. Types Part V: Compound Literals and Generic Selections 208
8
9 int main(void)
10 {
11 printf("%d\n", *get3490()); // INVALID: (int){3490} fell out of scope
12 }
Just think of their scope like that of an ordinary local variable. You can’t return a pointer to a local variable, either.
32.1.5 Silly Unnamed Object Example You can put any type in there and make an unnamed object. For example, these are effectively equivalent: int x = 3490;
printf("%d\n", x); // 3490 (variable) printf("%d\n", 3490); // 3490 (constant) printf("%d\n", (int){3490}); // 3490 (unnamed object)
That last one is unnamed, but it’s silly. Might as well do the simple one on the line before. But hopefully that provides a little more clarity on the syntax.
32.2 Generic Selections
This is an expression that allows you to select different pieces of code depending on the type of the first argument to the expression. We’ll look at an example in just a second, but it’s important to know this is processed at compile time, not at runtime. There’s no runtime analysis going on here. The expression begins with _Generic, works kinda like a switch, and it takes at least two arguments. The first argument is an expression (or variable2) that has a type. All expressions have a type. The remaining arguments to _Generic are the cases of what to substitute in for the result of the expression if the first argument is that type. Wat? Let’s try it out and see.
1 #include
2
3 int main(void)
4 {
5 int i;
6 float f;
7 char c;
8
9 char *s = _Generic(i,
10 int: "that variable is an int",
11 float: "that variable is a float",
12 default: "that variable is some type"
13 );
14
15 printf("%s\n", s);
16 }
Check out the _Generic expression starting on line 9.
2A variable used here is an expression. Chapter 32. Types Part V: Compound Literals and Generic Selections 209
When the compiler sees it, it look at the type of the first argument. (In this example, the type of the variable i.) It then looks through the cases for something of that type. And then it substitutes the argument in place of the entire _Generic expression. In this case, i is an int, so it matches that case. Then the string is substituted in for the expression. So the line turns into this when the compiler sees it: char *s = "that variable is an int";
If the compiler can’t find a type match in the _Generic, it looks for the optional default case and uses that. If it can’t find a type match and there’s no default, you’ll get a compile error. The first expression must match one of the types or default. Because it’s inconvenient to write _Generic over and over, it’s often used to make the body of a macro that can be easily repeatedly reused. Let’s make a macro TYPESTR(x) that takes an argument and returns a string with the type of the argument. So TYPESTR(1) will return the string "int", for example. Here we go: #include
#define TYPESTR(x) _Generic((x), \ int: "int", \ long: "long", \ float: "float", \ double: "double", \ default: "something else") int main(void) { int i; long l; float f; double d; char c;
printf("i is type %s\n", TYPESTR(i)); printf("l is type %s\n", TYPESTR(l)); printf("f is type %s\n", TYPESTR(f)); printf("d is type %s\n", TYPESTR(d)); printf("c is type %s\n", TYPESTR(c)); }
This outputs: i is type int l is type long f is type float d is type double c is type something else
Which should be no surprise, because, like we said, that code in main() is replaced with the following when it is compiled: printf("i is type %s\n", "int"); printf("l is type %s\n", "long"); printf("f is type %s\n", "float"); printf("d is type %s\n", "double"); printf("c is type %s\n", "something else"); Chapter 32. Types Part V: Compound Literals and Generic Selections 210
And that’s exactly the output we see. Let’s do one more. I’ve included some macros here so that when you run: int i = 10; char *s = "Foo!";
PRINT_VAL(i); PRINT_VAL(s);
you get the output: i = 10 s = Foo!
We’ll have to make use of some macro magic to do that.
1 #include
2 #include
3
4 // Macro that gives back a format specifier for a type
5 #define FMTSPEC(x) _Generic((x), \
6 int: "%d", \
7 long: "%ld", \
8 float: "%f", \
9 double: "%f", \
10 char *: "%s")
11 // TODO: add more types
12
13 // Macro that prints a variable in the form "name = value"
14 #define PRINT_VAL(x) { \
15 char fmt[512]; \
16 snprintf(fmt, sizeof fmt, #x " = %s\n", FMTSPEC(x)); \
17 printf(fmt, (x)); \
18 }
19
20 int main(void)
21 {
22 int i = 10;
23 float f = 3.14159;
24 char *s = "Hello, world!";
25
26 PRINT_VAL(i);
27 PRINT_VAL(f);
28 PRINT_VAL(s);
29 }
for the output: i = 10 f = 3.141590 s = Hello, world!
We could have crammed that all in one big macro, but I broke it into two to prevent eye bleeding. Chapter 33
Arrays Part II
We’re going to go over a few extra misc things this chapter concerning arrays. • Type qualifiers with array parameters • The static keyword with array parameters • Partial multi-dimensional array initializers They’re not super-commonly seen, but we’ll peek at them since they’re part of the newer spec.
33.1 Type Qualifiers for Arrays in Parameter Lists
If you recall from earlier, these two things are equivalent in function parameter lists: int func(int *p) {...} int func(int p[]) {...}
And you might also recall that you can add type qualifiers to a pointer variable like so: int *const p; int *volatile p; int *const volatile p; // etc.
But how can we do that when we’re using array notation in your parameter list? Turns out it goes in the brackets. And you can put the optional count after. The two following lines are equivalent: int func(int *const volatile p) {...} int func(int p[const volatile]) {...} int func(int p[const volatile 10]) {...}
If you have a multidimensional array, you need to put the type qualifiers in the first set of brackets.
33.2 static for Arrays in Parameter Lists
Similarly, you can use the keyword static in the array in a parameter list. This is something I’ve never seen in the wild. It is always followed by a dimension: int func(int p[static 4]) {...}
What this means, in the above example, is the compiler is going to assume that any array you pass to the function will be at least 4 elements. Anything else is undefined behavior.
211 Chapter 33. Arrays Part II 212 int func(int p[static 4]) {...} int main(void) { int a[] = {11, 22, 33, 44}; int b[] = {11, 22, 33, 44, 55}; int c[] = {11, 22};
func(a); // OK! a is 4 elements, the minimum func(b); // OK! b is at least 4 elements func(c); // Undefined behavior! c is under 4 elements! }
This basically sets the minimum size array you can have. Important note: there is nothing in the compiler that prohibits you from passing in a smaller array. The compiler probably won’t warn you, and it won’t detect it at runtime. By putting static in there, you’re saying, “I double secret PROMISE that I will never pass in a smaller array than this.” And the compiler says, “Yeah, fine,” and trusts you to not do it. And then the compiler can make certain code optimizations, safe in the knowledge that you, the program- mer, will always do the right thing.
33.3 Equivalent Initializers
C is a little bit, shall we say, flexible when it comes to array initializers. We’ve already seen some of this, where any missing values are replaced with zero. For example, we can initialize a 5 element array to 1,2,0,0,0 with this: int a[5] = {1, 2};
Or set an array entirely to zero with: int a[5] = {0};
But things get interesting when initializing multidimensional arrays. Let’s make an array of 3 rows and 2 columns: int a[3][2];
Let’s write some code to initialize it and print the result: #include
for (int row = 0; row < 3; row++) { for (int col = 0; col < 2; col++) printf("%d ", a[row][col]); printf("\n"); } }
And when we run it, we get the expected: Chapter 33. Arrays Part II 213
1 2 3 4 5 6
Let’s leave off some of the initializer elements and see they get set to zero: int a[3][2] = { {1, 2}, {3}, // Left off the 4! {5, 6} }; which produces: 1 2 3 0 5 6
Now let’s leave off the entire last middle element: int a[3][2] = { {1, 2}, // {3, 4}, // Just cut this whole thing out {5, 6} };
And now we get this, which might not be what you expect: 1 2 5 6 0 0
But if you stop to think about it, we only provided enough initializers for two rows, so they got used for the first two rows. And the remaining elements were initialized to zero. So far so good. Generally, if we leave off parts of the initializer, the compiler sets the corresponding elements to 0. But let’s get crazy. int a[3][2] = { 1, 2, 3, 4, 5, 6 };
What—? That’s a 2D array, but it only has a 1D initializer! Turns out that’s legal (though GCC will warn about it with the proper warnings turned on). Basically, what it does is starts filling in elements in row 0, then row 1, then row 2 from left to right. So when we print, it prints in order: 1 2 3 4 5 6
If we leave some off: int a[3][2] = { 1, 2, 3 }; they fill with 0: 1 2 3 0 0 0
So if you want to fill the whole array with 0, then go ahead and: int a[3][2] = {0}; Chapter 33. Arrays Part II 214
But my recommendation is if you have a 2D array, use a 2D initializer. It just makes the code more readable. (Except for initializing the whole array with 0, in which case it’s idiomatic to use {0} no matter the dimension of the array.) Chapter 34
Long Jumps with setjmp, longjmp
We’ve already seen goto, which jumps in function scope. But longjmp() allows you to jump back to an earlier point in execution, back to a function that called this one. There are a lot of limitations and caveats, but this can be a useful function for bailing out from deep in the call stack back up to an earlier state. In my experience, this is very rarely-used functionality.
34.1 Using setjmp and longjmp
The dance we’re going to do here is to basically put a bookmark in execution with setjmp(). Later on, we’ll call longjmp() and it’ll jump back to the earlier point in execution where we set the bookmark with setjmp(). And it can do this even if you’ve called subfunctions. Here’s a quick demo where we call into functions a couple levels deep and then bail out of it. We’re going to use a file scope variable env to keep the state of things when we call setjmp() so we can restore them when we call longjmp() later. This is the variable in which we remember our “place”. The variable env is of type jmp_buf, an opaque type declared in
1 #include
2 #include
3
4 jmp_buf env;
5
6 void depth2(void)
7 {
8 printf("Entering depth 2\n");
9 longjmp(env, 3490); // Bail out
10 printf("Leaving depth 2\n"); // This won't happen
11 }
12
13 void depth1(void)
14 {
15 printf("Entering depth 1\n");
16 depth2();
17 printf("Leaving depth 1\n"); // This won't happen
18 }
19
20 int main(void)
21 {
215 Chapter 34. Long Jumps with setjmp, longjmp 216
22 switch (setjmp(env)) {
23 case 0:
24 printf("Calling into functions, setjmp() returned 0\n");
25 depth1();
26 printf("Returned from functions\n"); // This won't happen
27 break;
28
29 case 3490:
30 printf("Bailed back to main, setjmp() returned 3490\n");
31 break;
32 }
33 }
When run, this outputs: Calling into functions, setjmp() returned 0 Entering depth 1 Entering depth 2 Bailed back to main, setjmp() returned 3490
If you try to take that output and match it up with the code, it’s clear there’s some really funky stuff going on. One of the most notable things is that setjmp() returns twice. What the actual frank? What is this sorcery?! So here’s the deal: if setjmp() returns 0, it means that you’ve successfully set the “bookmark” at that point. If it returns non-zero, it means you’ve just returned to the “bookmark” set earlier. (And the value returned is the one you pass to longjmp().) This way you can tell the difference between setting the bookmark and returning to it later. So when the code, above, calls setjmp() the first time, setjmp() stores the state in the env variable and returns 0. Later when we call longjmp() with that same env, it restores the state and setjmp() returns the value longjmp() was passed.
34.2 Pitfalls
Under the hood, this is pretty straightforward. Typically the stack pointer keeps track of the locations in memory that local variables are stored, and the program counter keeps track of the address of the currently-executing instruction1. So if we want to jump back to an earlier function, it’s basically only a matter of restoring the stack pointer and program counter to the values kept in the jmp_buf variable, and making sure the return value is set correctly. And then execution will resume there. But a variety of factors confound this, making a significant number of undefined behavior traps.
34.2.1 The Values of Local Variables
If you want the values of automatic (non-static and non-extern) local variables to persist in the function that called setjmp() after a longjmp() happens, you must declare those variables to be volatile. Technically, they only have to be volatile if they change between the time setjmp() is called and longjmp() is called2.
1Both “stack pointer” and “program counter” are related to the underlying architecture and C implementation, and are not part of the spec. 2The rationale here is that the program might store a value temporarily in a CPU register while it’s doing work on it. In that timeframe, the register holds the correct value, and the value on the stack might be out of date. Then later the register values would get overwritten and the changes to the variable lost. Chapter 34. Long Jumps with setjmp, longjmp 217
For example, if we run this code: int x = 20; if (setjmp(env) == 0) { x = 30; } and then later longjmp() back, the value of x will be indeterminate. If we want to fix this, x must be volatile: volatile int x = 20; if (setjmp(env) == 0) { x = 30; }
Now the value will be the correct 30 after a longjmp() returns us to this point.
34.2.2 How Much State is Saved?
When you longjmp(), execution resumes at the point of the corresponding setjmp(). And that’s it. The spec points out that it’s just as if you’d jumped back into the function at that point with local variables set to whatever values they had when the longjmp() call was made. Things that aren’t restored include, paraphrasing the spec: • Floating point status flags • Open files • Any other component of the abstract machine
34.2.3 You Can’t Name Anything setjmp
You can’t have any extern identifiers with the name setjmp. Or, if setjmp is a macro, you can’t undefine it. Both are undefined behavior.
34.2.4 You Can’t setjmp() in a Larger Expression That is, you can’t do something like this: if (x == 12 && setjmp(env) == 0) { ... }
That’s too complex to be allowed by the spec due to the machinations that must occur when unrolling the stack and all that. We can’t longjmp() back into some complex expression that’s only been partially executed. So there are limits on the complexity of that expression. • It can be the entire controlling expression of the conditional. if (setjmp(env)) {...}
switch (setjmp(env)) {...}
• It can be part of a relational or equality expression, as long as the other operand is an integer constant. And the whole thing is the controlling expression of the conditional. if (setjmp(env) == 0) {...}
• The operand to a logical NOT (!) operation, being the entire controlling expression. if (!setjmp(env)) {...} Chapter 34. Long Jumps with setjmp, longjmp 218
• A standalone expression, possibly cast to void. setjmp(env);
(void)setjmp(env);
34.2.5 When Can’t You longjmp()? It’s undefined behavior if: • You didn’t call setjmp() earlier • You called setjmp() from another thread • You called setjmp() in the scope of a variable length array (VLA), and execution left the scope of that VLA before longjmp() was called. • The function containing the setjmp() exited before longjmp() was called. On that last one, “exited” includes normal returns from the function, as well as the case if another longjmp() jumped back to “earlier” in the call stack than the function in question.
34.2.6 You Can’t Pass 0 to longjmp()
If you try to pass the value 0 to longjmp(), it will silently change that value to 1. Since setjmp() ultimately returns this value, and having setjmp() return 0 has special meaning, return- ing 0 is prohibited.
34.2.7 longjmp() and Variable Length Arrays
If you are in scope of a VLA and longjmp() out there, the memory allocated to the VLA could leak3. Same thing happens if you longjmp() back over any earlier functions that had VLAs still in scope. This is one thing that really bugged me able VLAs—that you could write perfectly legitimate C code that squandered memory. But, hey—I’m not in charge of the spec.
3That is, remain allocated until the program ends with no way to free it. Chapter 35
Incomplete Types
It might surprise you to learn that this builds without error:
1 extern int a[];
2
3 int main(void)
4 {
5 struct foo *x;
6 union bar *y;
7 enum baz *z;
8 }
We never gave a size for a. And we have pointers to structs foo, bar, and baz that never seem to be declared anywhere. And the only warnings I get are that x, y, and z are unused. These are examples of incomplete types. An incomplete type is a type the size (i.e. the size you’d get back from sizeof) for which is not known. Another way to think of it is a type that you haven’t finished declaring. You can have a pointer to an incomplete type, but you can’t dereference it or use pointer arithmetic on it. And you can’t sizeof it. So what can you do with it?
35.1 Use Case: Self-Referential Structures
I only know of one real use case: forward references to structs or unions with self-referential or co- dependent structures. (I’m going to use struct for the rest of these examples, but they all apply equally to unions, as well.) Let’s do the classic example first. But before I do, know this! As you declare a struct, the struct is incomplete until the closing brace is reached! struct antelope { // struct antelope is incomplete here int leg_count; // Still incomplete float stomach_fullness; // Still incomplete float top_speed; // Still incomplete char *nickname; // Still incomplete }; // NOW it's complete.
So what? Seems sane enough.
219 Chapter 35. Incomplete Types 220
But what if we’re doing a linked list? Each linked list node needs to have a reference to another node. But how can we create a reference to another node if we haven’t finished even declaring the node yet? C’s allowance for incomplete types makes it possible. We can’t declare a node, but we can declare a pointer to one, even if it’s incomplete!
1 struct node {
2 int val;
3 struct node *next; // struct node is incomplete, but that's OK!
4 };
Even though the struct node is incomplete on line 3, we can still declare a pointer to one1. We can do the same thing if we have two different structs that refer to each other: struct a { struct b *x; // Refers to a `struct b` };
struct b { struct a *x; // Refers to a `struct a` };
We’d never be able to make that pair of structures without the relaxed rules for incomplete types.
35.2 Incomplete Type Error Messages
Are you getting errors like these? invalid application of ‘sizeof’ to incomplete type
invalid use of undefined type
dereferencing pointer to incomplete type
Most likely culprit: you probably forgot to #include the header file that declares the type.
35.3 Other Incomplete Types
Declaring a struct or union with no body makes an incomplete type, e.g. struct foo;. enums are incomplete until the closing brace. void is an incomplete type. Arrays declared extern with no size are incomplete, e.g.: extern int a[];
If it’s a non-extern array with no size followed by an initializer, it’s incomplete until the closing brace of the initializer.
35.4 Use Case: Arrays in Header Files
It can be useful to declare incomplete array types in header files. In those cases, the actual storage (where the complete array is declared) should be in a single .c file. If you put it in the .h file, it will be duplicated every time the header file is included. So what you can do is make a header file with an incomplete type that refers to the array, like so:
1This works because in C, pointers are the same size regardless of the type of data they point to. So the compiler doesn’t need to know the size of the struct node at this point; it just needs to know the size of a pointer. Chapter 35. Incomplete Types 221
1 // File: bar.h
2
3 #ifndef BAR_H
4 #define BAR_H
5
6 extern int my_array[]; // Incomplete type
7
8 #endif
And the in the .c file, actually define the array:
1 // File: bar.c
2
3 int my_array[1024]; // Complete type!
Then you can include the header from as many places as you’d like, and every one of those places will refer to the same underlying my_array.
1 // File: foo.c
2
3 #include
4 #include "bar.h" // includes the incomplete type for my_array
5
6 int main(void)
7 {
8 my_array[0] = 10;
9
10 printf("%d\n", my_array[0]);
11 }
When compiling multiple files, remember to specific all the .c files to the compiler, but not the .h files, e.g.: gcc -o foo foo.c bar.c
35.5 Completing Incomplete Types
If you have an incomplete type, you can complete it by defining the complete struct, union, enum, or array in the same scope. struct foo; // incomplete type
struct foo *p; // pointer, no problem
// struct foo f; // Error: incomplete type!
struct foo { int x, y, z; }; // Now the struct foo is complete!
struct foo f; // Success!
Note that though void is an incomplete type, there’s no way to complete it. Not that anyone ever thinks of doing that weird thing. But it does explain why you can do this: void *p; // OK: pointer to incomplete type
and not either of these: void v; // Error: declare variable of incomplete type Chapter 35. Incomplete Types 222 printf("%d\n", *p); // Error: dereference incomplete type
The more you know… Chapter 36
Complex Numbers
A tiny primer on Complex numbers1 stolen directly from Wikipedia: A complex number is a number that can be expressed in the form 푎 + 푏푖, where 푎 and 푏 are real numbers [i.e. floating point types in C], and 푖 represents the imaginary unit, satisfying the equation 푖2 = −1. Because no real number satisfies this equation, 푖 is called an imagi- nary number. For the complex number 푎 + 푏푖, 푎 is called the real part, and 푏 is called the imaginary part. But that’s as far as I’m going to go. We’ll assume that if you’re reading this chapter, you know what a complex number is and what you want to do with them. And all we need to cover is C’s faculties for doing so. Turns out, though, that complex number support in a compiler is an optional feature. Not all compliant compilers can do it. And the ones that do, might do it to various degrees of completeness. You can test if your system supports complex numbers with: #ifdef __STDC_NO_COMPLEX__ #error Complex numbers not supported! #endif
Furthermore, there is a macro that indicates adherence to the ISO 60559 (IEEE 754) standard for floating point math with complex numbers, as well as the presence of the _Imaginary type. #if __STDC_IEC_559_COMPLEX__ != 1 #error Need IEC 60559 complex support! #endif
More details on that are spelled out in Annex G in the C11 spec.
36.1 Complex Types
To use complex numbers, #include
Those both mean the same thing, so you might as well use the prettier complex. You also get some types for imaginary numbers if you implementation is IEC 60559-compliant: _Imaginary imaginary
1https://en.wikipedia.org/wiki/Complex_number
223 Chapter 36. Complex Numbers 224
These also both mean the same thing, so you might as well use the prettier imaginary. You also get values for the imaginary number 푖, itself: I _Complex_I _Imaginary_I
The macro I is set to _Imaginary_I (if available), or _Complex_I. So just use I for the imaginary number. One aside: I’ve said that if a compiler has __STDC_IEC_559_COMPLEX__ set to 1, it must support _Imag- inary types to be compliant. That’s my read of the spec. However, I don’t know of a single compiler that actually supports _Imaginary even though they have __STDC_IEC_559_COMPLEX__ set. So I’m going to write some code with that type in here I have no way of testing. Sorry! OK, so now we know there’s a complex type, how can we use it?
36.2 Assigning Complex Numbers
Since the complex number has a real and imaginary part, but both of them rely on floating point numbers to store values, we need to also tell C what precision to use for those parts of the complex number. We do that by just pinning a float, double, or long double to the complex, either before or after it. Let’s define a complex number that uses float for its components: float complex c; // Spec prefers this way complex float c; // Same thing--order doesn't matter
So that’s great for declarations, but how do we initialize them or assign to them? Turns out we get to use some pretty natural notation. Example! double complex x = 5 + 2*I; double complex y = 10 + 3*I;
For 5 + 2푖 and 10 + 3푖, respectively.
36.3 Constructing, Deconstructing, and Printing
We’re getting there… We’ve already seen one way to write a complex number: double complex x = 5 + 2*I;
There’s also no problem using other floating point numbers to build it: double a = 5; double b = 2; double complex x = a + b*I;
There is also a set of macros to help build these. The above code could be written using the CMPLX() macro, like so: double complex x = CMPLX(5, 2);
As far as I can tell in my research, these are almost equivalent: double complex x = 5 + 2*I; double complex x = CMPLX(5, 2); Chapter 36. Complex Numbers 225
But the CMPLX() macro will handle negative zeros in the imaginary part correctly every time, whereas the other way might convert them to positive zeros. I think2 This seems to imply that if there’s a chance the imaginary part will be zero, you should use the macro… but someone should correct me on this if I’m mistaken! The CMPLX() macro works on double types. There are two other macros for float and long double: CMPLXF() and CMPLXL(). (These “f” and “l” suffixes appear in virtually all the complex-number-related functions.) Now let’s try the reverse: if we have a complex number, how do we break it apart into its real and imaginary parts? Here we have a couple functions that will extract the real and imaginary parts from the number: creal() and cimag(): double complex x = 5 + 2*I; double complex y = 10 + 3*I;
printf("x = %f + %fi\n", creal(x), cimag(x)); printf("y = %f + %fi\n", creal(y), cimag(y));
for the output: x = 5.000000 + 2.000000i y = 10.000000 + 3.000000i
Note that the i I have in the printf() format string is a literal i that gets printed—it’s not part of the format specifier. Both return values from creal() and cimag() are double. And as usual, there are float and long double variants of these functions: crealf(), cimagf(), creall(), and cimagl().
36.4 Complex Arithmetic and Comparisons
Arithmetic can be performed on complex numbers, though how this works mathematically is beyond the scope of the guide.
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2*I;
7 double complex y = 3 + 4*I;
8 double complex z;
9
10 z = x + y;
11 printf("x + y = %f + %fi\n", creal(z), cimag(z));
12
13 z = x - y;
14 printf("x - y = %f + %fi\n", creal(z), cimag(z));
15
16 z = x * y;
17 printf("x * y = %f + %fi\n", creal(z), cimag(z));
18
19 z = x / y;
2This was a harder one to research, and I’ll take any more information anyone can give me. I could be defined as _Complex_I or _Imaginary_I, if the latter exists. _Imaginary_I will handle signed zeros, but _Complex_I might not. This has implications with branch cuts and other complex-numbery-mathy things. Maybe. Can you tell I’m really getting out of my element here? In any case, the CMPLX() macros behave as if I were defined as _Imaginary_I, with signed zeros, even if _Imaginary_I doesn’t exist on the system. Chapter 36. Complex Numbers 226
20 printf("x / y = %f + %fi\n", creal(z), cimag(z));
21 }
for a result of: x + y = 4.000000 + 6.000000i x - y = -2.000000 + -2.000000i x * y = -5.000000 + 10.000000i x / y = 0.440000 + 0.080000i
You can also compare two complex numbers for equality (or inequality):
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2*I;
7 double complex y = 3 + 4*I;
8
9 printf("x == y = %d\n", x == y); // 0
10 printf("x != y = %d\n", x != y); // 1
11 }
with the output: x == y = 0 x != y = 1
They are equal if both components test equal. Note that as with all floating point, they could be equal if they’re close enough due to rounding error3.
36.5 Complex Math
But wait! There’s more than just simple complex arithmetic! Here’s a summary table of all the math functions available to you with complex numbers. I’m only going to list the double version of each function, but for all of them there is a float version that you can get by appending f to the function name, and a long double version that you can get by appending l. For example, the cabs() function for computing the absolute value of a complex number also has cabsf() and cabsl() variants. I’m omitting them for brevity.
36.5.1 Trigonometry Functions
Function Description ccos() Cosine csin() Sine ctan() Tangent cacos() Arc cosine casin() Arc sine catan() Play Settlers of Catan ccosh() Hyperbolic cosine csinh() Hyperbolic sine ctanh() Hyperbolic tangent cacosh() Arc hyperbolic cosine
3The simplicity of this statement doesn’t do justice to the incredible amount of work that goes into simply understanding how floating point actually functions. https://randomascii.wordpress.com/2012/02/25/comparing-floating-point-numbers-2012-edition/ Chapter 36. Complex Numbers 227
Function Description casinh() Arc hyperbolic sine catanh() Arc hyperbolic tangent
36.5.2 Exponential and Logarithmic Functions
Function Description cexp() Base-푒 exponential clog() Natural (base-푒) logarithm
36.5.3 Power and Absolute Value Functions
Function Description cabs() Absolute value cpow() Power csqrt() Square root
36.5.4 Manipulation Functions
Function Description creal() Return real part cimag() Return imaginary part CMPLX() Construct a complex number carg() Argument/phase angle conj() Conjugate4 cproj() Projection on Riemann sphere
4This is the only one that doesn’t begin with an extra leading c, strangely. Chapter 37
Fixed Width Integer Types
C has all those small, bigger, and biggest integer types like int and long and all that. And you can look in the section on limits to see what the largest int is with INT_MAX and so on. How big are those types? That is, how many bytes do they take up? We could use sizeof to get that answer. But what if I wanted to go the other way? What if I needed a type that was exactly 32 bits (4 bytes) or at least 16 bits or somesuch? How can we declare a type that’s a certain size? The header
37.1 The Bit-Sized Types
For both signed and unsigned integers, we can specify a type that is a certain number of bits, with some caveats, of course. And there are three main classes of these types (in these examples, the N would be replaced by a certain number of bits): • Integers of exactly a certain size (intN_t) • Integers that are at least a certain size (int_leastN_t) • Integers that are at least a certain size and are as fast as possible (int_fastN_t)1 How much faster is fast? Definitely maybe some amount faster. Probably. The spec doesn’t say how much faster, just that they’ll be the fastest on this architecture. Most C compilers are pretty good, though, so you’ll probably only see this used in places where the most possible speed needs to be guaranteed (rather than just hoping the compiler is producing pretty-dang-fast code, which it is). Finally, these unsigned number types have a leading u to differentiate them. For example, these types have the corresponding listed meaning: int32_t w; // x is exactly 32 bits, signed uint16_t x; // y is exactly 16 bits, unsigned int_least8_t y; // y is at least 8 bits, signed uint_fast64_t z; // z is the fastest representation at least 64 bits, unsigned
The following types are guaranteed to be defined:
1Some architectures have different sized data that the CPU and RAM can operate with at a faster rate than others. In those cases, if you need the fastest 8-bit number, it might give you have a 16- or 32-bit type instead because that’s just faster. So with this, you won’t know how big the type is, but it will be least as big as you say.
228 Chapter 37. Fixed Width Integer Types 229 int_least8_t uint_least8_t int_least16_t uint_least16_t int_least32_t uint_least32_t int_least64_t uint_least64_t int_fast8_t uint_fast8_t int_fast16_t uint_fast16_t int_fast32_t uint_fast32_t int_fast64_t uint_fast64_t
There might be others of different widths, as well, but those are optional. Hey! Where are the fixed types like int16_t? Turns out those are entirely optional…unless certain con- ditions are met2. And if you have an average run-of-the-mill modern computer system, those conditions probably are met. And if they are, you’ll have these types: int8_t uint8_t int16_t uint16_t int32_t uint32_t int64_t uint64_t
Other variants with different widths might be defined, but they’re optional.
37.2 Maximum Integer Size Type
There’s a type you can use that holds the largest representable integers available on the system, both signed and unsigned: intmax_t uintmax_t
Use these types when you want to go as big as possible. Obviously values from any other integer types of the same sign will fit in this type, necessarily.
37.3 Using Fixed Size Constants
If you have a constant that you want to have fit in a certain number of bits, you can use these macros to automatically append the proper suffix onto the number (e.g. 22L or 3490ULL). INT8_C(x) UINT8_C(x) INT16_C(x) UINT16_C(x) INT32_C(x) UINT32_C(x) INT64_C(x) UINT64_C(x) INTMAX_C(x) UINTMAX_C(x)
Again, these work only with constant integer values. For example, we can use one of these to assign constant values like so: uint16_t x = UINT16_C(12); intmax_t y = INTMAX_C(3490);
37.4 Limits of Fixed Size Integers
We also have some limits defined so you can get the maximum and minimum values for these types:
2Namely, the system has 8, 16, 32, or 64 bit integers with no padding that use two’s complement representation, in which case the intN_t variant for that particular number of bits must be defined. Chapter 37. Fixed Width Integer Types 230
INT8_MAX INT8_MIN UINT8_MAX INT16_MAX INT16_MIN UINT16_MAX INT32_MAX INT32_MIN UINT32_MAX INT64_MAX INT64_MIN UINT64_MAX
INT_LEAST8_MAX INT_LEAST8_MIN UINT_LEAST8_MAX INT_LEAST16_MAX INT_LEAST16_MIN UINT_LEAST16_MAX INT_LEAST32_MAX INT_LEAST32_MIN UINT_LEAST32_MAX INT_LEAST64_MAX INT_LEAST64_MIN UINT_LEAST64_MAX
INT_FAST8_MAX INT_FAST8_MIN UINT_FAST8_MAX INT_FAST16_MAX INT_FAST16_MIN UINT_FAST16_MAX INT_FAST32_MAX INT_FAST32_MIN UINT_FAST32_MAX INT_FAST64_MAX INT_FAST64_MIN UINT_FAST64_MAX
INTMAX_MAX INTMAX_MIN UINTMAX_MAX
Note the MIN for all the unsigned types is 0, so, as such, there’s no macro for it.
37.5 Format Specifiers
In order to print these types, you need to send the right format specifier to printf(). (And the same issue for getting input with scanf().) But how are you going to know what size the types are under the hood? Luckily, once again, C provides some macros to help with this. All this can be found in
Look for the patterns there. You can see there are variants for the fixed, least, fast, and max types. And you also have a lowercase d and a lowercase i. Those correspond to the printf() format specifiers %d and %i. So if I have something of type: int_least16_t x = 3490;
I can print that with the equivalent format specifier for %d by using PRId16. But how? How do we use that macro? First of all, that macro specifies a string containing the letter or letters printf() needs to use to print that type. Like, for example, it could be "d" or "ld". So all we need to do is embed that in our format string to the printf() call. To do this, we can take advantage of a fact about C that you might have forgotten: adjacent string literals are automatically concatenated to a single string. E.g.: printf("Hello, " "world!\n"); // Prints "Hello, world!"
And since these macros are string literals, we can use them like so:
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 int_least16_t x = 3490;
8
9 printf("The value is %" PRIdLEAST16 "!\n", x);
10 }
We also have a pile of macros for printing unsigned types: PRIon PRIoLEASTn PRIoFASTn PRIoMAX PRIun PRIuLEASTn PRIuFASTn PRIuMAX PRIxn PRIxLEASTn PRIxFASTn PRIxMAX PRIXn PRIXLEASTn PRIXFASTn PRIXMAX
In this case, o, u, x, and X correspond to the documented format specifiers in printf(). And, as before, the lowercase n should be substituted with 8, 16, 32, or 64. But just when you think you had enough of the macros, it turns out we have a complete complementary set of them for scanf()! SCNdn SCNdLEASTn SCNdFASTn SCNdMAX SCNin SCNiLEASTn SCNiFASTn SCNiMAX SCNon SCNoLEASTn SCNoFASTn SCNoMAX SCNun SCNuLEASTn SCNuFASTn SCNuMAX SCNxn SCNxLEASTn SCNxFASTn SCNxMAX
Remember: when you want to print out a fixed size integer type with printf() or scanf(), grab the correct corresponding format specifer from
Date and Time Functionality
“Time is an illusion. Lunchtime doubly so.” —Ford Prefect, The Hitchhikers Guide to the Galaxy This isn’t too complex, but it can be a little intimidating at first, both with the different types available and the way we can convert between them. Mix in GMT (UTC) and local time and we have all the Usual Fun™ one gets with times and dates. And of course never forget the golden rule of dates and times: Never attempt to write your own date and time functionality. Only use what the library gives you. Time is too complex for mere mortal programmers to handle correctly. Seriously, we all owe a point to everyone who worked on any date and time library, so put that in your budget.
38.1 Quick Terminology and Information
Just a couple quick terms in case you don’t have them down. • UTC: Coordinated Universal Time is a universally1 agreed upon, absolute time. Everyone on the planet thinks it’s the same time right now in UTC… even though they have different local times. • GMT: Greenwich Mean Time, effectively the same as UTC2. You probably want to say UTC, or “universal time”. If you’re talking specifically about the GMT timezone, say GMT. Confusingly, many of C’s UTC functions predate UTC and still refer to Greenwich Mean Time. When you see that, know the C means UTC. • Local time: what time it is where the computer running the program is located. This is described as an offset from UTC. Although there are many timezones in the world, most computers do work in either local time or UTC. If you are describing an event that happens one time, like a log entry, or a rocket launch, or when pointers finally clicked for you, use UTC. On the other hand, if it’s something that happens the same time in every timezone, like New Year’s Eve or dinner time, use local time. Since a lot of languages are only good at converting between UTC and local time, you can cause yourself a lot of pain by choosing to store your dates in the wrong form. (Ask me how I know.)
38.2 Date Types
There are two3 main types in C when it comes to dates: time_t and struct tm.
1On Earth, anyway. Who know what crazy systems they use out there… 2OK, don’t murder me! GMT is technically a timezone while UTC is a global time system. Also some countries might adjust GMT for daylight saving time, whereas UTC is never adjusted for daylight saving time. 3Admittedly, there are more than two.
232 Chapter 38. Date and Time Functionality 233
The spec doesn’t actually say much about them: • time_t: a real type capable of holding a time. So by the spec, this could be a floating type or integer type. In POSIX (Unix-likes), it’s an integer. This holds calendar time. Which you can think of as UTC time. • struct tm: holds the components of a calendar time. This is a broken-down time, i.e. the compo- nents of the time, like hour, minute, second, day, month, year, etc. On a lot of systems, time_t represents the number of seconds since Epoch4. Epoch is in some ways the start of time from the computer’s perspective, which is commonly January 1, 1970 UTC. time_t can go negative to represent times before Epoch. Windows behaves the same way as Unix from what I can tell. And what’s in a struct tm? The following fields: struct tm { int tm_sec; // seconds after the minute -- [0, 60] int tm_min; // minutes after the hour -- [0, 59] int tm_hour; // hours since midnight -- [0, 23] int tm_mday; // day of the month -- [1, 31] int tm_mon; // months since January -- [0, 11] int tm_year; // years since 1900 int tm_wday; // days since Sunday -- [0, 6] int tm_yday; // days since January 1 -- [0, 365] int tm_isdst; // Daylight Saving Time flag };
Note that everything is zero-based except the day of the month. It’s important to know that you can put any values in these types you want. There are functions to help get the time now, but the types hold a time, not the time. So the question becomes: “How do you initialize data of these types, and how do you convert between them?”
38.3 Initialization and Conversion Between Types
First, you can get the current time and store it in a time_t with the time() function. time_t now; // Variable to hold the time now now = time(NULL); // You can get it like this... time(&now); // ...or this. Same as the previous line.
Great! You have a variable that gets you the time now. Amusingly, there’s only one portable way to print out what’s in a time_t, and that’s the rarely-used ctime() function that prints the value in local time: now = time(NULL); printf("%s", ctime(&now));
This returns a string with a very specific form that includes a newline at the end: Sun Feb 28 18:47:25 2021
So that’s kind of inflexible. If you want more control, you should convert that time_t into a struct tm.
38.3.1 Converting time_t to struct tm There are two amazing ways to do this conversion:
4https://en.wikipedia.org/wiki/Unix_time Chapter 38. Date and Time Functionality 234
• localtime(): this function converts a time_t to a struct tm in local time. • gmtime(): this function converts a time_t to a struct tm in UTC. (See ye olde GMT creeping into that function name?) Let’s see what time it is now by printing out a struct tm with the asctime() function: printf("Local: %s", asctime(localtime(&now))); printf(" UTC: %s", asctime(gmtime(&now)));
Output (I’m in Pacific Standard Time, out of daylight ): Local: Sun Feb 28 20:15:27 2021 UTC: Mon Mar 1 04:15:27 2021
Once you have your time_t in a struct tm, it opens all kinds of doors. You can print out the time in a variety of ways, figure out which day of the week a date is, and so on. Or convert it back into a time_t. More on that soon!
38.3.2 Converting struct tm to time_t
If you want to go the other way, you can use mktime() to get that information. mktime() sets the values of tm_wday and tm_yday for you, so don’t bother filling them out because they’ll just be overwritten. Also, you can set tm_isdst to -1 to have it make the determination for you. Or you can manually set it to true or false. struct tm some_time = { .tm_year=82, // years since 1900 .tm_mon=3, // months since January -- [0, 11] .tm_mday=12, // day of the month -- [1, 31] .tm_hour=12, // hours since midnight -- [0, 23] .tm_min=00, // minutes after the hour -- [0, 59] .tm_sec=04, // seconds after the minute -- [0, 60] .tm_isdst=-1, // Daylight Saving Time flag }; time_t some_time_epoch; some_time_epoch = mktime(&some_time); printf("%s", ctime(&some_time_epoch)); printf("Is DST: %d\n", some_time.tm_isdst);
Output: Mon Apr 12 12:00:04 1982 Is DST: 0
When you manually load a struct tm like that, it should be in local time. mktime() will convert that local time into a time_t calendar time. Weirdly, however, the standard doesn’t give us a way to load up a struct tm with a UTC time and convert that to a time_t. If you want to do that with Unix-likes, try the non-standard timegm(). On Windows, _mkgmtime().
38.4 Formatted Date Output
We’ve already seen a couple ways to print formatted date output to the screen. With time_t we can use ctime(), and with struct tm we can use asctime(). Chapter 38. Date and Time Functionality 235
time_t now = time(NULL); struct tm *local = localtime(&now); struct tm *utc = gmtime(&now);
printf("Local time: %s", ctime(&now)); // Local time with time_t printf("Local time: %s", asctime(local)); // Local time with struct tm printf("UTC : %s", asctime(utc)); // UTC with a struct tm
But what if I told you, dear reader, that there’s a way to have much more control over how the date was printed? Sure, we could fish individual fields out of the struct tm, but there’s a great function called strftime() that will do a lot of the hard work for you. It’s like printf(), except for dates! Let’s see some examples. In each of these, we pass in a destination buffer, a maximum number of char- acters to write, and then a format string (in the style of—but not the same as—printf()) which tells strftime() which components of a struct tm to print and how. You can add other constant characters to include in the output in the format string, as well, just like with printf(). We get a struct tm in this case from localtime(), but any source works fine.
1 #include
2 #include
3
4 int main(void)
5 {
6 char s[128];
7 time_t now = time(NULL);
8
9 // %c: print date as per current locale
10 strftime(s, sizeof s, "%c", localtime(&now));
11 puts(s); // Sun Feb 28 22:29:00 2021
12
13 // %A: full weekday name
14 // %B: full month name
15 // %d: day of the month
16 strftime(s, sizeof s, "%A, %B %d", localtime(&now));
17 puts(s); // Sunday, February 28
18
19 // %I: hour (12 hour clock)
20 // %M: minute
21 // %S: second
22 // %p: AM or PM
23 strftime(s, sizeof s, "It's %I:%M:%S %p", localtime(&now));
24 puts(s); // It's 10:29:00 PM
25
26 // %F: ISO 8601 yyyy-mm-dd
27 // %T: ISO 8601 hh:mm:ss
28 // %z: ISO 8601 timezone offset
29 strftime(s, sizeof s, "ISO 8601: %FT%T%z", localtime(&now));
30 puts(s); // ISO 8601: 2021-02-28T22:29:00-0800
31 }
There are a ton of date printing format specifiers for strftime(), so be sure to check them out in the strftime() reference page. Chapter 38. Date and Time Functionality 236
38.5 More Resolution with timespec_get()
You can get the number of seconds and nanoseconds since Epoch with timespec_get(). Maybe. Implementations might not have nanosecond resolution (that’s one billionth of a second) so who knows how many significant places you’ll get, but give it a shot and see. timespec_get() takes two arguments. One is a pointer to a struct timespec to hold the time informa- tion. And the other is the base, which the spec lets you set to TIME_UTC indicating that you’re interested in seconds since Epoch. (Other implementations might give you more options for the base.) And the structure itself has two fields: struct timespec { time_t tv_sec; // Seconds long tv_nsec; // Nanoseconds (billionths of a second) };
Here’s an example where we get the time and print it out both as integer values and also a floating value: struct timespec ts;
timespec_get(&ts, TIME_UTC);
printf("%ld s, %ld ns\n", ts.tv_sec, ts.tv_nsec);
double float_time = ts.tv_sec + ts.tv_nsec/1000000000.0; printf("%f seconds since epoch\n", float_time);
Example output: 1614581530 s, 806325800 ns 1614581530.806326 seconds since epoch
struct timespec also makes an appearance in a number of the threading functions that need to be able to specify time with that resolution.
38.6 Differences Between Times
One quick note about getting the difference between two time_ts: since the spec doesn’t dictate how that type represents a time, you might not be able to simply subtract two time_ts and get anything sensible5. Luckily you can use difftime() to compute the difference in seconds between two dates. In the following example, we have two events that occur some time apart, and we use difftime() to compute the difference.
1 #include
2 #include
3
4 int main(void)
5 {
6 struct tm time_a = {
7 .tm_year=82, // years since 1900
8 .tm_mon=3, // months since January -- [0, 11]
9 .tm_mday=12, // day of the month -- [1, 31]
10 .tm_hour=4, // hours since midnight -- [0, 23]
11 .tm_min=00, // minutes after the hour -- [0, 59]
12 .tm_sec=04, // seconds after the minute -- [0, 60]
13 .tm_isdst=-1, // Daylight Saving Time flag
5You will on POSIX, where time_t is definitely an integer. Unfortunately the entire world isn’t POSIX, so there we are. Chapter 38. Date and Time Functionality 237
14 };
15
16 struct tm time_b = {
17 .tm_year=120, // years since 1900
18 .tm_mon=10, // months since January -- [0, 11]
19 .tm_mday=15, // day of the month -- [1, 31]
20 .tm_hour=16, // hours since midnight -- [0, 23]
21 .tm_min=27, // minutes after the hour -- [0, 59]
22 .tm_sec=00, // seconds after the minute -- [0, 60]
23 .tm_isdst=-1, // Daylight Saving Time flag
24 };
25
26 time_t cal_a = mktime(&time_a);
27 time_t cal_b = mktime(&time_b);
28
29 double diff = difftime(cal_b, cal_a);
30
31 double years = diff / 60 / 60 / 24 / 365.2425; // close enough
32
33 printf("%f seconds (%f years) between events\n", diff, years);
34 }
Output: 1217996816.000000 seconds (38.596783 years) between events
And there you have it! Remember to use difftime() to take the time difference. Even though you can just subtract on a POSIX system, might as well stay portable. Chapter 39
Multithreading
C11 introduced, formally, multithreading to the C language. It’s very eerily similar to POSIX threads1, if you’ve ever used those. And if you’re not, no worries. We’ll talk it through. Do note, however, that I’m not intending this to be a full-blown classic multithreading how-to2; you’ll have to pick up a different very thick book for that, specifically. Sorry! Threading is an optional feature. If a C11+ compiler defines __STDC_NO_THREADS__, threads will not be present in the library. Why they decided to go with a negative sense in that macro is beyond me, but there we are. You can test for it like this: #ifdef __STDC_NO_THREADS__ #error I need threads to build this program! #endif
Also, you might need to specify certain linker options when building. In the case of Unix-likes, try ap- pending a -lpthreads to the end of the command line to link the pthreads library3: gcc -std=c11 -o foo foo.c -lpthreads
If you’re getting linker errors on your system, it could be because the appropriate library wasn’t included.
39.1 Background
Threads are a way to have all those shiny CPU cores you paid for do work for you in the same program. Normally, a C program just runs on a single CPU core. But if you know how to split up the work, you can give pieces of it to a number of threads and have them do the work simultaneously. Though the spec doesn’t say it, on your system it’s very likely that C (or the OS at its behest) will attempt to balance the threads over all your CPU cores. And if you have more threads than cores, that’s OK. You just won’t realize all those gains if they’re all trying to compete for CPU time.
39.2 Things You Can Do
You can create a thread. It will begin running the function you specify. The parent thread that spawned it will also continue to run. 1https://en.wikipedia.org/wiki/POSIX_Threads 2I’m more a fan of shared-nothing, myself, and my skills with classic multithreading constructs are rusty, to say the least. 3Yes, pthreads with a “p”. It’s short for POSIX threads, a library that C11 borrowed liberally from for its threads implementa- tion.
238 Chapter 39. Multithreading 239
And you can wait for the thread to complete. This is called joining. Or if you don’t care when the thread completes and don’t want to wait, you can detach it. A thread can explicitly exit, or it can implicitly call it quits by returning from its main function. A thread can also sleep for a period of time, doing nothing while other threads run. The main() program is a thread, as well. Additionally, we have thread local storage, mutexes, and conditional variables. But more on those later. Let’s just look at the basics for now.
39.3 Data Races and the Standard Library
Some of the functions in the standard library (e.g. asctime() and strtok()) return or use static data elements that aren’t threadsafe. But in general unless it’s said otherwise, the standard library makes an effort to be so4. But keep an eye out. If a standard library function is maintaining state between calls in a variable you don’t own, or if a function is returning a pointer to a thing that you didn’t pass in, it’s not threadsafe.
39.4 Creating and Waiting for Threads
Let’s hack something up! We’ll make some threads (create) and wait for them to complete (join). We have a tiny bit to understand first, though. Every single thread is identified by an opaque variable of type thrd_t. It’s a unique identifier per thread in your program. When you create a thread, it’s given a new ID. Also when you make the thread, you have to give it a pointer to a function to run, and a pointer to an argument to pass to it (or NULL if you don’t have anything to pass). The thread will begin execution on the function you specify. When you want to wait for a thread to complete, you have to specify its thread ID so C knows which one to wait for. So the basic idea is: 1. Write a function to act as the thread’s “main”. It’s not main()-proper, but analogous to it. The thread will start running there. 2. From the main thread, launch a new thread with thrd_create(), and pass it a pointer to the func- tion to run. 3. In that function, have the thread do whatever it has to do. 4. Meantimes, the main thread can continue doing whatever it has to do. 5. When the main thread decides to, it can wait for the child thread to complete by calling thrd_join(). Generally you must thrd_join() the thread to clean up after it or else you’ll leak memory5 thrd_create() takes a pointer to the function to run, and it’s of type thrd_start_t, which is int (*)(void *). That’s Greek for “a pointer to a function that takes an void* as an argument, and returns an int.” Let’s make a thread! We’ll launch it from the main thread with thrd_create() to run a function, do some other things, then wait for it to complete with thrd_join(). I’ve named the thread’s main function run(), but you can name it anything as long as the types match thrd_start_t.
4Per §7.1.4¶5. 5Unless you thrd_detach(). More on this later. Chapter 39. Multithreading 240
1 #include
2 #include
3
4 // This is the function the thread will run. It can be called anything.
5 //
6 // arg is the argument pointer passed to `thrd_create()`.
7 //
8 // The parent thread will get the return value back from `thrd_join()`'
9 // later.
10
11 int run(void *arg)
12 {
13 int *a = arg; // We'll pass in an int* from thrd_create()
14
15 printf("THREAD: Running thread with arg %d\n", *a);
16
17 return 12; // Value to be picked up by thrd_join() (chose 12 at random)
18 }
19
20 int main(void)
21 {
22 thrd_t t; // t will hold the thread ID
23 int arg = 3490;
24
25 printf("Launching a thread\n");
26
27 // Launch a thread to the run() function, passing a pointer to 3490
28 // as an argument. Also stored the thread ID in t:
29
30 thrd_create(&t, run, &arg);
31
32 printf("Doing other things while the thread runs\n");
33
34 printf("Waiting for thread to complete...\n");
35
36 int res; // Holds return value from the thread exit
37
38 // Wait here for the thread to complete; store the return value
39 // in res:
40
41 thrd_join(t, &res);
42
43 printf("Thread exited with return value %d\n", res);
44 }
See how we did the thrd_create() there to call the run() function? Then we did other things in main() and then stopped and waited for the thread to complete with thrd_join(). Sample output (yours might vary): Launching a thread Doing other things while the thread runs Waiting for thread to complete... THREAD: Running thread with arg 3490 Thread exited with return value 12
The arg that you pass to the function has to have a lifetime long enough so that the thread can pick it up before it goes away. Also, it needs to not be overwritten by the main thread before the new thread can use it. Chapter 39. Multithreading 241
Let’s look at an example that launches 5 threads. One thing to note here is how we use an array of thrd_ts to keep track of all the thread IDs.
1 #include
2 #include
3
4 int run(void *arg)
5 {
6 int i = *(int*)arg;
7
8 printf("THREAD %d: running!\n", i);
9
10 return i;
11 }
12
13 #define THREAD_COUNT 5
14
15 int main(void)
16 {
17 thrd_t t[THREAD_COUNT];
18
19 int i;
20
21 printf("Launching threads...\n");
22 for (i = 0; i < THREAD_COUNT; i++)
23
24 // NOTE! In the following line, we pass a pointer to i,
25 // but each thread sees the same pointer. So they'll
26 // print out weird things as i changes value here in
27 // the main thread! (More in the text, below.)
28
29 thrd_create(t + i, run, &i);
30
31 printf("Doing other things while the thread runs...\n");
32 printf("Waiting for thread to complete...\n");
33
34 for (int i = 0; i < THREAD_COUNT; i++) {
35 int res;
36 thrd_join(t[i], &res);
37
38 printf("Thread %d complete!\n", res);
39 }
40
41 printf("All threads complete!\n");
42 }
When I run the threads, I count i up from 0 to 4. And pass a pointer to it to thrd_create(). This pointer ends up in the run() routine where we make a copy of it. Simple enough? Here’s the output: Launching threads... THREAD 2: running! THREAD 3: running! THREAD 4: running! THREAD 2: running! Doing other things while the thread runs... Waiting for thread to complete... Thread 2 complete! Thread 2 complete! Chapter 39. Multithreading 242
THREAD 5: running! Thread 3 complete! Thread 4 complete! Thread 5 complete! All threads complete!
Whaaa—? Where’s THREAD 0? And why do we have a THREAD 5 when clearly i is never more than 4 when we call thrd_create()? And two THREAD 2s? Madness! This is getting into the fun land of race conditions. The main thread is modifying i before the thread has a chance to copy it. Indeed, i makes it all the way to 5 and ends the loop before the last thread gets a chance to copy it. We’ve got to have a per-thread variable that we can refer to so we can pass it in as the arg. We could have a big array of them. Or we could malloc() space (and free it somewhere—maybe in the thread itself.) Let’s give that a shot:
1 #include
2 #include
3 #include
4
5 int run(void *arg)
6 {
7 int i = *(int*)arg; // Copy the arg
8
9 free(arg); // Done with this
10
11 printf("THREAD %d: running!\n", i);
12
13 return i;
14 }
15
16 #define THREAD_COUNT 5
17
18 int main(void)
19 {
20 thrd_t t[THREAD_COUNT];
21
22 int i;
23
24 printf("Launching threads...\n");
25 for (i = 0; i < THREAD_COUNT; i++) {
26
27 // Get some space for a per-thread argument:
28
29 int *arg = malloc(sizeof *arg);
30 *arg = i;
31
32 thrd_create(t + i, run, arg);
33 }
34
35 // ...
Notice on lines 27-30 we malloc() space for an int and copy the value of i into it. Each new thread gets its own freshly-malloc()d variable and we pass a pointer to that to the run() function. Once run() makes its own copy of the arg on line 7, it free()s the malloc()d int. And now that it has its own copy, it can do with it what it pleases. Chapter 39. Multithreading 243
And a run shows the result: Launching threads... THREAD 0: running! THREAD 1: running! THREAD 2: running! THREAD 3: running! Doing other things while the thread runs... Waiting for thread to complete... Thread 0 complete! Thread 1 complete! Thread 2 complete! Thread 3 complete! THREAD 4: running! Thread 4 complete! All threads complete!
There we go! Threads 0-4 all in effect! Your run might vary—how the threads get scheduled to run is beyond the C spec. We see in the above example that thread 4 didn’t even begin until threads 0-1 had completed. Indeed, if I run this again, I likely get different output. We cannot guarantee a thread execution order.
39.5 Detaching Threads
If you want to fire-and-forget a thread (i.e. so you don’t have to thrd_join() it later), you can do that with thrd_detach(). This removes the parent thread’s ability to get the return value from the child thread, but if you don’t care about that and just want threads to clean up nicely on their own, this is the way to go. Basically we’re going to do this: thrd_create(&t, run, NULL); thrd_detach(t);
where the thrd_detach() call is the parent thread saying, “Hey, I’m not going to wait for this child thread to complete with thrd_join(). So go ahead and clean it up on your own when it completes.”
1 #include
2 #include
3
4 int run(void *arg)
5 {
6 (void)arg;
7
8 //printf("Thread running! %lu\n", thrd_current()); // non-portable!
9 printf("Thread running!\n");
10
11 return 0;
12 }
13
14 #define THREAD_COUNT 10
15
16 int main(void)
17 {
18 thrd_t t;
19
20 for (int i = 0; i < THREAD_COUNT; i++) {
21 thrd_create(&t, run, NULL);
22 thrd_detach(t); // <-- DETACH! Chapter 39. Multithreading 244
23 }
24
25 // Sleep for a second to let all the threads finish
26 thrd_sleep(&(struct timespec){.tv_sec=1}, NULL);
27 }
Note that in this code, we put the main thread to sleep for 1 second with thrd_sleep()—more on that later. Also in the run() function, I have a commented-out line in there that prints out the thread ID as an unsigned long. This is non-portable, because the spec doesn’t say what type a thrd_t is under the hood—it could be a struct for all we know. But that line works on my system. Something interesting I saw when I ran the code, above, and printed out the thread IDs was that some threads had duplicate IDs! This seems like it should be impossible, but C is allowed to reuse thread IDs after the corresponding thread has exited. So what I was seeing was that some threads completed their run before other threads were launched.
39.6 Thread Local Data
Threads are interesting because they don’t have their own memory beyond local variables. If you want a static variable or file scope variable, all threads will see that same variable. This can lead to race conditions, where you get Weird Things™ happening. Check out this example. We have a static variable foo in block scope in run(). This variable will be visible to all threads that pass through the run() function. And the various threads can effectively step on each other’s toes. Each thread copies foo into a local variable x (which is not shared between threads—all the threads have their own call stacks). So they should be the same, right? And the first time we print them, they are6. But then right after that, we check to make sure they’re still the same. And they usually are. But not always!
1 #include
2 #include
3 #include
4
5 int run(void *arg)
6 {
7 int n = *(int*)arg; // Thread number for humans to differentiate
8
9 free(arg);
10
11 static int foo = 10; // Static value shared between threads
12
13 int x = foo; // Automatic local variable--each thread has its own
14
15 // We just assigned x from foo, so they'd better be equal here.
16 // (In all my test runs, they were, but even this isn't guaranteed!)
17
18 printf("Thread %d: x = %d, foo = %d\n", n, x, foo);
19
20 // And they should be equal here, but they're not always!
21 // (Sometimes they were, sometimes they weren't!)
22
6Though I don’t think they have to be. It’s just that the threads don’t seem to get rescheduled until some system call like might happen with a printf()… which is why I have the printf() in there. Chapter 39. Multithreading 245
23 // What happens is another thread gets in and increments foo
24 // right now, but this thread's x remains what it was before!
25
26 if (x != foo) {
27 printf("Thread %d: Craziness! x != foo! %d != %d\n", n, x, foo);
28 }
29
30 foo++; // Increment shared value
31
32 return 0;
33 }
34
35 #define THREAD_COUNT 5
36
37 int main(void)
38 {
39 thrd_t t[THREAD_COUNT];
40
41 for (int i = 0; i < THREAD_COUNT; i++) {
42 int *n = malloc(sizeof *n); // Holds a thread serial number
43 *n = i;
44 thrd_create(t + i, run, n);
45 }
46
47 for (int i = 0; i < THREAD_COUNT; i++) {
48 thrd_join(t[i], NULL);
49 }
50 }
Here’s an example output (though this varies from run to run): Thread 0: x = 10, foo = 10 Thread 1: x = 10, foo = 10 Thread 1: Craziness! x != foo! 10 != 11 Thread 2: x = 12, foo = 12 Thread 4: x = 13, foo = 13 Thread 3: x = 14, foo = 14
In thread 1, between the two printf()s, the value of foo somehow changed from 10 to 11, even though clearly there’s no increment between the printf()s! It was another thread that got in there (probably thread 0, from the look of it) and incremented the value of foo behind thread 1’s back! Let’s solve this problem two different ways. (If you want all the threads to share the variable and not step on each other’s toes, you’ll have to read on to the mutex section.)
39.6.1 _Thread_local Storage-Class
First things first, let’s just look at the easy way around this: the _Thread_local storage-class. Basically we’re just going to slap this on the front of our block scope static variable and things will work! It tells C that every thread should have its own version of this variable, so none of them step on each other’s toes. The
5 int run(void *arg)
6 {
7 int n = *(int*)arg; // Thread number for humans to differentiate
8
9 free(arg);
10
11 thread_local static int foo = 10; // <-- No longer shared!!
And running we get: Thread 0: x = 10, foo = 10 Thread 1: x = 10, foo = 10 Thread 2: x = 10, foo = 10 Thread 4: x = 10, foo = 10 Thread 3: x = 10, foo = 10
No more weird problems! One thing: if a thread_local variable is block scope, it must be static. Them’s the rules. (But this is OK because non-static variables are per-thread already since each thread has it’s own non-static variables.) A bit of a lie there: block scope thread_local variables can also be extern.
39.6.2 Another Option: Thread-Specific Storage Thread-specific storage (TSS) is another way of getting per-thread data. One additional feature is that these functions allow you to specify a destructor that will be called on the data when the TSS variable is deleted. Commonly this destructor is free() to automatically clean up malloc()d per-thread data. Or NULL if you don’t need to destroy anything. The destructor is type tss_dtor_t which is a pointer to a function that returns void and takes a void* as an argument (the void* points to the data stored in the variable). In other words, it’s a void (*)(void*), if that clears it up. Which I admit it probably doesn’t. Check out the example, below. Generally, thread_local is probably your go-to, but if you like the destructor idea, then you can make use of that. The usage is a bit weird in that we need a variable of type tss_t to be alive to represent the value on a per thread basis. Then we initialize it with tss_create(). Eventually we get rid of it with tss_delete(). Note that calling tss_delete() doesn’t run all the destructors—it’s thrd_exit() (or returning from the run function) that does that. tss_delete() just releases any memory allocated by tss_create(). In the middle, threads can call tss_set() and tss_get() to set and get the value. In the following code, we set up the TSS variable before creating the threads, then clean up after the threads. In the run() function, the threads malloc() some space for a string and store that pointer in the TSS variable. When the thread exits, the destructor function (free() in this case) is called for all the threads.
1 #include
2 #include
3 #include
4
5 tss_t str;
6
7 void some_function(void)
8 {
9 // Retrieve the per-thread value of this string
10 char *tss_string = tss_get(str); Chapter 39. Multithreading 247
11
12 // And print it
13 printf("TSS string: %s\n", tss_string);
14 }
15
16 int run(void *arg)
17 {
18 int serial = *(int*)arg; // Get this thread's serial number
19 free(arg);
20
21 // malloc() space to hold the data for this thread
22 char *s = malloc(64);
23 sprintf(s, "thread %d! :)", serial); // Happy little string
24
25 // Set this TSS variable to point at the string
26 tss_set(str, s);
27
28 // Call a function that will get the variable
29 some_function();
30
31 return 0; // Equivalent to thrd_exit(0)
32 }
33
34 #define THREAD_COUNT 15
35
36 int main(void)
37 {
38 thrd_t t[THREAD_COUNT];
39
40 // Make a new TSS variable, the free() function is the destructor
41 tss_create(&str, free);
42
43 for (int i = 0; i < THREAD_COUNT; i++) {
44 int *n = malloc(sizeof *n); // Holds a thread serial number
45 *n = i;
46 thrd_create(t + i, run, n);
47 }
48
49 for (int i = 0; i < THREAD_COUNT; i++) {
50 thrd_join(t[i], NULL);
51 }
52
53 // All threads are done, so we're done with this
54 tss_delete(str);
55 }
Again, this is kind of a painful way of doing things compared to thread_local, so unless you really need that destructor functionality, I’d use that instead.
39.7 Mutexes
If you want to only allow a single thread into a critical section of code at a time, you can protect that section with a mutex7. For example, if we had a static variable and we wanted to be able to get and set it in two operations without another thread jumping in the middle and corrupting it, we could use a mutex for that.
7Short for “mutual exclusion”, AKA a “lock” on a section of code that only one thread is permitted to execute. Chapter 39. Multithreading 248
You can acquire a mutex or release it. If you attempt to acquire the mutex and succeed, you may continue execution. If you attempt and fail (because someone else holds it), you will block8 until the mutex is released. If multiple threads are blocked waiting for a mutex to be released, one of them will be chosen to run (at random, from our perspective), and the others will continue to sleep. The gameplan is that first we’ll initialize a mutex variable to make it ready to use with mtx_init(). Then subsequent threads can call mtx_lock() and mtx_unlock() to get and release the mutex. When we’re completely done with the mutex, we can destroy it with mtx_destroy(), the logical opposite of mtx_init(). First, let’s look at some code that does not use a mutex, and endeavors to print out a shared (static) serial number and then increment it. Because we’re not using a mutex over the getting of the value (to print it) and the setting (to increment it), threads might get in each other’s way in that critical section.
1 #include
2 #include
3
4 int run(void *arg)
5 {
6 (void)arg;
7
8 static int serial = 0; // Shared static variable!
9
10 printf("Thread running! %d\n", serial);
11
12 serial++;
13
14 return 0;
15 }
16
17 #define THREAD_COUNT 10
18
19 int main(void)
20 {
21 thrd_t t[THREAD_COUNT];
22
23 for (int i = 0; i < THREAD_COUNT; i++) {
24 thrd_create(t + i, run, NULL);
25 }
26
27 for (int i = 0; i < THREAD_COUNT; i++) {
28 thrd_join(t[i], NULL);
29 }
30 }
When I run this, I get something that looks like this: Thread running! 0 Thread running! 0 Thread running! 0 Thread running! 3 Thread running! 4 Thread running! 5 Thread running! 6 Thread running! 7 Thread running! 8
8That is, your process will go to sleep. Chapter 39. Multithreading 249
Thread running! 9
Clearly multiple threads are getting in there and running the printf() before anyone gets a change to update the serial variable. What we want to do is wrap the getting of the variable and setting of it into a single mutex-protected stretch of code. We’ll add a new variable to represent the mutex of type mtx_t in file scope, initialize it, and then the threads can lock and unlock it in the run() function.
1 #include
2 #include
3
4 mtx_t serial_mtx; // <-- MUTEX VARIABLE
5
6 int run(void *arg)
7 {
8 (void)arg;
9
10 static int serial = 0; // Shared static variable!
11
12 // Acquire the mutex--all threads will block on this call until
13 // they get the lock:
14
15 mtx_lock(&serial_mtx); // <-- ACQUIRE MUTEX
16
17 printf("Thread running! %d\n", serial);
18
19 serial++;
20
21 // Done getting and setting the data, so free the lock. This will
22 // unblock threads on the mtx_lock() call:
23
24 mtx_unlock(&serial_mtx); // <-- RELEASE MUTEX
25
26 return 0;
27 }
28
29 #define THREAD_COUNT 10
30
31 int main(void)
32 {
33 thrd_t t[THREAD_COUNT];
34
35 // Initialize the mutex variable, indicating this is a normal
36 // no-frills, mutex:
37
38 mtx_init(&serial_mtx, mtx_plain); // <-- CREATE MUTEX
39
40 for (int i = 0; i < THREAD_COUNT; i++) {
41 thrd_create(t + i, run, NULL);
42 }
43
44 for (int i = 0; i < THREAD_COUNT; i++) {
45 thrd_join(t[i], NULL);
46 }
47
48 // Done with the mutex, destroy it: Chapter 39. Multithreading 250
49
50 mtx_destroy(&serial_mtx); // <-- DESTROY MUTEX
51 }
See how on lines 38 and 50 of main() we initialize and destroy the mutex. But each individual thread acquires the mutex on line 15 and releases it on line 24. In between the mtx_lock() and mtx_unlock() is the critical section, the area of code where we don’t want multiple threads mucking about at the same time. And now we get proper output! Thread running! 0 Thread running! 1 Thread running! 2 Thread running! 3 Thread running! 4 Thread running! 5 Thread running! 6 Thread running! 7 Thread running! 8 Thread running! 9
If you need multiple mutexes, no problem: just have multiple mutex variables. And always remember the Number One Rule of Multiple Mutexes: Unlock mutexes in the opposite order in which you lock them!
39.7.1 Different Mutex Types
As hinted earlier, we have a few mutex types that you can create with mtx_init(). (Some of these types are the result of a bitwise-OR operation, as noted in the table.)
Type Description mtx_plain Regular ol’ mutex mtx_timed Mutex that supports timeouts mtx_plain|mtx_recursive Recursive mutex mtx_timed|mtx_recursive Recursive mutex that supports timeouts
“Recursive” means that the holder of a lock can call mtx_lock() multiple times on the same lock. (They have to unlock it an equal number of times before anyone else can take the mutex.) This might ease coding from time to time, especially if you call a function that needs to lock the mutex when you already hold the mutex. And the timeout gives a thread a chance to try to get the lock for a while, but then bail out if it can’t get it in that timeframe. For a timeout mutex, be sure to create it with mtx_timed: mtx_init(&serial_mtx, mtx_timed);
And then when you wait for it, you have to specify a time in UTC when it will unlock9. The function timespec_get() from
9You might have expected it to be “time from now”, but you’d just like to think that, wouldn’t you! Chapter 39. Multithreading 251
Then call mtx_timedlock() instead of mtx_lock(). If it returns the value thrd_timedout, it timed out. struct timespec timeout; timespec_get(&timeout, TIME_UTC); // Get current time timeout.tv_sec += 1; // Timeout 1 second after now int result = mtx_timedlock(&serial_mtx, &timeout)); if (result == thrd_timedout) { printf("Mutex lock timed out!\n"); }
Other than that, timed locks are the same as regular locks.
39.8 Condition Variables
Condition Variables are the last piece of the puzzle we need to make performant multithreaded applications and to compose more complex multithreaded structures. A condition variable provides a way for threads to go to sleep until some event on another thread occurs. In other words, we might have a number of threads that are rearing to go, but they have to wait until some event is true before they continue. Basically they’re being told “wait for it!” until they get notified. And this works hand-in-hand with mutexes since what we’re going to wait on generally depends on the value of some data, and that data generally needs to be protected by a mutex. It’s important to note that the condition variable itself isn’t the holder of any particular data from our perspective. It’s merely the variable by which C keeps track of the waiting/not-waiting status of a particular thread or group of threads. Let’s write a contrived program that reads in groups of 5 numbers from the main thread one at a time. Then, when 5 numbers have been entered, the child thread wakes up, sums up those 5 numbers, and prints the result. The numbers will be stored in a global, shared array, as will the index into the array of the about-to-be- entered number. Since these are shared values, we at least have to hide them behind a mutex for both the main and child threads. (The main will be writing data to them and the child will be reading data from them.) But that’s not enough. The child thread needs to block (“sleep”) until 5 numbers have been read into the array. And then the parent thread needs to wake up the child thread so it can do its work. And when it wakes up, it needs to be holding that mutex. And it will! When a thread waits on a condition variable, it also acquires a mutex when it wakes up. How’s that work? Let’s look at the outline of what the child thread will do: 1. Lock the mutex with mtx_lock() 2. If we haven’t entered all the numbers, wait on the condition variable with cnd_wait() 3. Do the work that needs doing 4. Unlock the mutex with mtx_unlock() Meanwhile the main thread will be doing this: 1. Lock the mutex with mtx_lock() 2. Store the recently-read number into the array 3. If the array is full, signal the child to wake up with cnd_signal() 4. Unlock the mutex with mtx_unlock() Chapter 39. Multithreading 252
If you didn’t skim that too hard (it’s OK—I’m not offended), you might notice something weird: how can the main thread hold the mutex lock and signal the child, if the child has to hold the mutex lock to wait for the signal? They can’t both hold the lock! And indeed they don’t! There’s some behind-the-scenes magic with condition variables: when you cnd_wait(), it releases the mutex that you specify and the thread goes to sleep. And when someone signals that thread to wake up, it reacquires the lock as if nothing had happened. It’s a little different on the cnd_signal() side of things. This doesn’t do anything with the mutex. The signalling thread still must manually release the mutex before the waiting threads can wake up. One more thing on the cnd_wait(). You’ll probably be calling cnd_wait() if some condition10 is not yet met (e.g. in this case, if not all the numbers have yet been entered). Here’s the deal: this condition should be in a while loop, not an if statement. Why? It’s because of a mysterious phenomenon called a spurious wakeup. Sometimes, in some implementations, a thread can be woken up out of a cnd_wait() sleep for seemingly no reason. [X-Files music]11. And so we have to check to see that the condition we need is still actually met when we wake up. And if it’s not, back to sleep with us! So let’s do this thing! Starting with the main thread: • The main thread will set up the mutex and condition variable, and will launch the child thread. • Then it will, in an infinite loop, get numbers as input from the console. • It will also acquire the mutex to store the inputted number into a global array. • When the array has 5 numbers in it, the main thread will signal the child thread that it’s time to wake up and do its work. • Then the main thread will unlock the mutex and go back to reading the next number from the console. Meanwhile, the child thread has been up to its own shenanigans: • The child thread grabs the mutex • While the condition is not met (i.e. while the shared array doesn’t yet have 5 numbers in it), the child thread sleeps by waiting on the condition variable. When it waits, it unlocks the mutex. • Once the main thread signals the child thread to wake up, it wakes up to do the work and gets the mutex lock back. • The child thread sums the numbers and resets the variable that is the index into the array. • It then releases the mutex and runs again in an infinite loop. And here’s the code! Give it some study so you can see where all the above pieces are being handled:
1 #include
2 #include
3
4 #define VALUE_COUNT_MAX 5
5
6 int value[VALUE_COUNT_MAX]; // Shared global
7 int value_count = 0; // Shared global, too
8
9 mtx_t value_mtx; // Mutex around value
10 cnd_t value_cnd; // Condition variable on value
11
12 int run(void *arg)
13 {
14 (void)arg;
10And that’s why they’re called condition variables! 11I’m not saying it’s aliens… but it’s aliens. OK, really more likely another thread might have been woken up and gotten to the work first. Chapter 39. Multithreading 253
15
16 for (;;) {
17 mtx_lock(&value_mtx); // <-- GRAB THE MUTEX
18
19 while (value_count < VALUE_COUNT_MAX) {
20 printf("Thread: is waiting\n");
21 cnd_wait(&value_cnd, &value_mtx); // <-- CONDITION WAIT
22 }
23
24 printf("Thread: is awake!\n");
25
26 int t = 0;
27
28 // Add everything up
29 for (int i = 0; i < VALUE_COUNT_MAX; i++)
30 t += value[i];
31
32 printf("Thread: total is %d\n", t);
33
34 // Reset input index for main thread
35 value_count = 0;
36
37 mtx_unlock(&value_mtx); // <-- MUTEX UNLOCK
38 }
39
40 return 0;
41 }
42
43 int main(void)
44 {
45 thrd_t t;
46
47 // Spawn a new thread
48
49 thrd_create(&t, run, NULL);
50 thrd_detach(t);
51
52 // Set up the mutex and condition variable
53
54 mtx_init(&value_mtx, mtx_plain);
55 cnd_init(&value_cnd);
56
57 for (;;) {
58 int n;
59
60 scanf("%d", &n);
61
62 mtx_lock(&value_mtx); // <-- LOCK MUTEX
63
64 value[value_count++] = n;
65
66 if (value_count == VALUE_COUNT_MAX) {
67 printf("Main: signaling thread\n");
68 cnd_signal(&value_cnd); // <-- SIGNAL CONDITION
69 }
70
71 mtx_unlock(&value_mtx); // <-- UNLOCK MUTEX
72 } Chapter 39. Multithreading 254
73
74 // Clean up (I know that's an infinite loop above here, but I
75 // want to at least pretend to be proper):
76
77 mtx_destroy(&value_mtx);
78 cnd_destroy(&value_cnd);
79 }
And here’s some sample output (individual numbers on lines are my input): Thread: is waiting 1 1 1 1 1 Main: signaling thread Thread: is awake! Thread: total is 5 Thread: is waiting 2 8 5 9 0 Main: signaling thread Thread: is awake! Thread: total is 24 Thread: is waiting
It’s a common use of condition variables in producer-consumer situations like this. If we didn’t have a way to put the child thread to sleep while it waited for some condition to be met, it would be force to poll which is a big waste of CPU.
39.8.1 Timed Condition Wait
There’s a variant of cnd_wait() that allows you to specify a timeout so you can stop waiting. Since the child thread must relock the mutex, this doesn’t necessarily mean that you’ll be popping back to life the instant the timeout occurs; you still must wait for any other threads to release the mutex. But it does mean that you won’t be waiting until the cnd_signal() happens. To make this work, call cnd_timedwait() instead of cnd_wait(). If it returns the value thrd_timedout, it timed out. The timestamp is an absolute time in UTC, not a time-from-now. Thankfully the timespec_get() func- tion in
timespec_get(&timeout, TIME_UTC); // Get current time timeout.tv_sec += 1; // Timeout 1 second after now
int result = cnd_timedwait(&condition, &mutex, &timeout));
if (result == thrd_timedout) { printf("Condition variable timed out!\n"); } Chapter 39. Multithreading 255
39.8.2 Broadcast: Wake Up All Waiting Threads cnd_signal() only wakes up one thread to continue working. Depending on how you have your logic done, it might make sense to wake up more than one thread to continue once the condition is met. Of course only one of them can grab the mutex, but if you have a situation where: • The newly-awoken thread is responsible for waking up the next one, and— • There’s a chance the spurious-wakeup loop condition will prevent it from doing so, then— you’ll want to broadcast the wake up so that you’re sure to get at least one of the threads out of that loop to launch the next one. How, you ask? Simply use cnd_broadcast() instead of cnd_signal(). Exact same usage, except cnd_broadcast() wakes up all the sleeping threads that were waiting on that condition variable.
39.9 Running a Function One Time
Let’s say you have a function that could be run by many threads, but you don’t know when, and it’s not work trying to write all that logic. There’s a way around it: use call_once(). Tons of threads could try to run the function, but only the first one counts12 To work with this, you need a special flag variable you declare to keep track of whether or not the thing’s been run. And you need a function to run, which takes no parameters and returns no value. once_flag of = ONCE_FLAG_INIT; // Initialize it like this void run_once_function(void) { printf("I'll only run once!\n"); } int run(void *arg) { (void)arg;
call_once(&of, run_once_function);
// ...
In this example, no matter how many threads get to the run() function, the run_once_function() will only be called a single time.
12Survival of the fittest! Right? I admit it’s actually nothing like that. Chapter 40
Function Description clearerr() Clear the feof and ferror status flags fclose() Close an open file feof() Return the file end-of-file status ferror() Return the file error status fflush() Flush all buffered output to a file fgetc() Read a character in a file fgetpos() Get the file I/O position fgets() Read a line from a file fopen() Open a file fprintf() Print formatted output to a file fputc() Print a character to a file fputs() Print a string to a file fread() Read binary data from a file freopen() Change file associated with a stream fscanf() Read formatted input from a file fseek() Set the file I/O position fsetpos() Set the file I/O position ftell() Get the file I/O position fwrite() Write binary data to a file getc() Get a character from stdin getchar() Get a character from stdin gets() Get a string from stdin (removed in C11) perror() Print a human-formatted error message printf() Print formatted output to stdout putc() Print a character to stdout putchar() Print a character to stdout puts() Print a string to stdout remove() Delete a file from disk rename() Rename or move a file on disk rewind() Set the I/O position to the beginning of a file scanf() Read formatted input from stdin setbuf() Configure buffering for I/O operations setvbuf() Configure buffering for I/O operations snprintf() Print length-limited formatted output to a string sprintf() Print formatted output to a string sscanf() Read formatted input from a string tmpfile() Create a temporary file tmpnam() Generate a unique name for a temporary file ungetc() Push a character back on the input stream vfprintf() Variadic print formatted output to a file
256 Chapter 40.
Function Description vfscanf() Variadic read formatted input from a file vprintf() Variadic print formatted output to stdout vscanf() Variadic read formatted input from stdin vsnprintf() Variadic length-limited print formatted output to a string vsprintf() Variadic print formatted output to a string vsscanf() Variadic read formatted input to a string
The most basic of all libraries in the whole of the standard C library is the standard I/O library. It’s used for reading from and writing to files. I can see you’re very excited about this. So I’ll continue. It’s also used for reading and writing to the console, as we’ve already often seen with the printf() function. (A little secret here—many many things in various operating systems are secretly files deep down, and the console is no exception. “Everything in Unix is a file!” :-)) You’ll probably want some prototypes of the functions you can use, right? To get your grubby little mittens on those, you’ll want to include stdio.h. Anyway, so we can do all kinds of cool stuff in terms of file I/O. LIE DETECTED. Ok, ok. We can do all kinds of stuff in terms of file I/O. Basically, the strategy is this: 1. Use fopen() to get a pointer to a file structure of type FILE*. This pointer is what you’ll be passing to many of the other file I/O calls. 2. Use some of the other file calls, like fscanf(), fgets(), fprintf(), or etc. using the FILE* returned from fopen(). 3. When done, call fclose() with the FILE*. This let’s the operating system know that you’re truly done with the file, no take-backs. What’s in the FILE*? Well, as you might guess, it points to a struct that contains all kinds of information about the current read and write position in the file, how the file was opened, and other stuff like that. But, honestly, who cares. No one, that’s who. The FILE structure is opaque to you as a programmer; that is, you don’t need to know what’s in it, and you don’t even want to know what’s in it. You just pass it to the other standard I/O functions and they know what to do. This is actually pretty important: try to not muck around in the FILE structure. It’s not even the same from system to system, and you’ll end up writing some really non-portable code. One more thing to mention about the standard I/O library: a lot of the functions that operate on files use an “f” prefix on the function name. The same function that is operating on the console will leave the “f” off. For instance, if you want to print to the console, you use printf(), but if you want to print to a file, use fprintf(), see? Wait a moment! If writing to the console is, deep down, just like writing to a file, since everything in Unix is a file, why are there two functions? Answer: it’s more convenient. But, more importantly, is there a FILE* associated with the console that you can use? Answer: YES! There are, in fact, three (count ’em!) special FILE*s you have at your disposal merely for just including stdio.h. There is one for input, and two for output. That hardly seems fair—why does output get two files, and input only get one? That’s jumping the gun a bit—let’s just look at them:
Stream Description stdin Input from the console. stdout Output to the console. stderr Output to the console on the error file stream. Chapter 40.
So standard input (stdin) is by default just what you type at the keyboard. You can use that in fscanf() if you want, just like this: /* this line: */ scanf("%d", &x);
/* is just like this line: */ fscanf(stdin, "%d", &x);
And stdout works the same way: printf("Hello, world!\n"); fprintf(stdout, "Hello, world!\n"); /* same as previous line! */
So what is this stderr thing? What happens when you output to that? Well, generally it goes to the console just like stdout, but people use it for error messages, specifically. Why? On many systems you can redirect the output from the program into a file from the command line…and sometimes you’re interested in getting just the error output. So if the program is good and writes all its errors to stderr, a user can redirect just stderr into a file, and just see that. It’s just a nice thing you, as a programmer, can do. Finally, a lot of these functions return int where you might expect char. This is because the function can return a character or end-of-file (EOF), and EOF is potentially an integer. If you don’t get EOF as a return value, you can safely store the result in a char.
40.1 remove()
Delete a file
Synopsis
#include
int remove(const char *filename);
Description Removes the specified file from the filesystem. It just deletes it. Nothing magical. Simply call this function and sacrifice a small chicken and the requested file will be deleted.
Return Value
Returns zero on success, and -1 on error, setting errno.
Example
1 char *filename = "/home/beej/evidence.txt";
2
3 remove(filename);
4 remove("/disks/d/Windows/system.ini");
See Also
rename() Chapter 40.
40.2 rename()
Renames a file and optionally moves it to a new location
Synopsis
#include
int rename(const char *old, const char *new);
Description
Renames the file old to name new. Use this function if you’re tired of the old name of the file, and you are ready for a change. Sometimes simply renaming your files makes them feel new again, and could save you money over just getting all new files! One other cool thing you can do with this function is actually move a file from one directory to another by specifying a different path for the new name.
Return Value
Returns zero on success, and -1 on error, setting errno.
Example
1 rename("foo", "bar"); // changes the name of the file "foo" to "bar"
2
3 // the following moves the file "evidence.txt" from "/tmp" to
4 // "/home/beej", and also renames it to "nothing.txt":
5 rename("/tmp/evidence.txt", "/home/beej/nothing.txt");
See Also
remove()
40.3 tmpfile()
Create a temporary file
Synopsis
#include
FILE *tmpfile(void);
Description
This is a nifty little function that will create and open a temporary file for you, and will return a FILE* to it that you can use. The file is opened with mode “r+b”, so it’s suitable for reading, writing, and binary data. By using a little magic, the temp file is automatically deleted when it is close()’d or when your program exits. (Specifically, in Unix terms, tmpfile() unlinks1 the file right after it opens it. This means that
1https://man.archlinux.org/man/unlinkat.2.en#DESCRIPTION Chapter 40.
it’s primed to be deleted from disk, but still exists because your process still has it open. As soon as your process exits, all open files are closed, and the temp file vanishes into the ether.)
Return Value
This function returns an open FILE* on success, or NULL on failure.
Example
1 #include
2
3 int main(void)
4 {
5 FILE *temp;
6 char s[128];
7
8 temp = tmpfile();
9
10 fprintf(temp, "What is the frequency, Alexander?\n");
11
12 rewind(temp); // back to the beginning
13
14 fscanf(temp, "%s", s); // read it back out
15
16 fclose(temp); // close (and magically delete)
17 }
See Also
fopen(), fclose(), tmpnam()
40.4 tmpnam()
Generate a unique name for a temporary file
Synopsis
#include
char *tmpnam(char *s);
Description This function takes a good hard look at the existing files on your system, and comes up with a unique name for a new file that is suitable for temporary file usage. Let’s say you have a program that needs to store off some data for a short time so you create a temporary file for the data, to be deleted when the program is done running. Now imagine that you called this file foo.txt. This is all well and good, except what if a user already has a file called foo.txt in the directory that you ran your program from? You’d overwrite their file, and they’d be unhappy and stalk you forever. And you wouldn’t want that, now would you? Ok, so you get wise, and you decide to put the file in /tmp so that it won’t overwrite any important content. But wait! What if some other user is running your program at the same time and they both want to use that filename? Or what if some other program has already created that file? Chapter 40.
See, all of these scary problems can be completely avoided if you just use tmpnam() to get a safe-ready- to-use filename. So how do you use it? There are two amazing ways. One, you can declare an array (or malloc() it— whatever) that is big enough to hold the temporary file name. How big is that? Fortunately there has been a macro defined for you, L_tmpnam, which is how big the array must be. And the second way: just pass NULL for the filename. tmpnam() will store the temporary name in a static array and return a pointer to that. Subsequent calls with a NULL argument will overwrite the static array, so be sure you’re done using it before you call tmpnam() again. Again, this function just makes a file name for you. It’s up to you to later fopen() the file and use it. One more note: some compilers warn against using tmpnam() since some systems have better functions (like the Unix function mkstemp().) You might want to check your local documentation to see if there’s a better option. Linux documentation goes so far as to say, “Never use this function. Use mkstemp() instead.” I, however, am going to be a jerk and not talk about mkstemp() because it’s not in the standard I’m writing about. Nyaah. The macro TMP_MAX holds the number of unique filenames that can be generated by tmpnam(). Ironically, it is the minimum number of such filenames.
Return Value Returns a pointer to the temporary file name. This is either a pointer to the string you passed in, or a pointer to internal static storage if you passed in NULL. On error (like it can’t find any temporary name that is unique), tmpnam() returns NULL.
Example
1 char filename[L_tmpnam];
2 char *another_filename;
3
4 if (tmpnam(filename) != NULL)
5 printf("We got a temp file named: \"%s\"\n", filename);
6 else
7 printf("Something went wrong, and we got nothing!\n");
8
9 another_filename = tmpnam(NULL);
10 printf("We got another temp file named: \"%s\"\n", another_filename);
11 printf("And we didn't error check it because we're too lazy!\n");
On my Linux system, this generates the following output:
We got a temp file named: "/tmp/filew9PMuZ" We got another temp file named: "/tmp/fileOwrgPO" And we didn't error check it because we're too lazy!
See Also
fopen(), tmpfile()
40.5 fclose()
The opposite of fopen()—closes a file when you’re done with it so that it frees system resources Chapter 40.
Synopsis
#include
int fclose(FILE *stream);
Description When you open a file, the system sets aside some resources to maintain information about that open file. Usually it can only open so many files at once. In any case, the Right Thing to do is to close your files when you’re done using them so that the system resources are freed. Also, you might not find that all the information that you’ve written to the file has actually been written to disk until the file is closed. (You can force this with a call to fflush().) When your program exits normally, it closes all open files for you. Lots of times, though, you’ll have a long-running program, and it’d be better to close the files before then. In any case, not closing a file you’ve opened makes you look bad. So, remember to fclose() your file when you’re done with it!
Return Value
On success, 0 is returned. Typically no one checks for this. On error EOF is returned. Typically no one checks for this, either.
Example
1 FILE *fp;
2
3 fp = fopen("spoonDB.dat", r"); // (you should error-check this)
4 sort_spoon_database(fp);
5 fclose(fp); // pretty simple, huh.
See Also
fopen()
40.6 fflush()
Process all buffered I/O for a stream right now
Synopsis
#include
int fflush(FILE *stream);
Description
When you do standard I/O, as mentioned in the section on the setvbuf() function, it is usually stored in a buffer until a line has been entered or the buffer is full or the file is closed. Sometimes, though, you really want the output to happen right this second, and not wait around in the buffer. You can force this to happen by calling fflush(). The advantage to buffering is that the OS doesn’t need to hit the disk every time you call fprintf(). The disadvantage is that if you look at the file on the disk after the fprintf() call, it might not have actually been written to yet. (“I called fputs(), but the file is still zero bytes long! Why?!”) In virtually Chapter 40.
all circumstances, the advantages of buffering outweigh the disadvantages; for those other circumstances, however, use fflush(). Note that fflush() is only designed to work on output streams according to the spec. What will happen if you try it on an input stream? Use your spooky voice: who knooooows!
Return Value
On success, fflush() returns zero. If there’s an error, it returns EOF and sets the error condition for the stream (see ferror().)
Example
In this example, we’re going to use the carriage return, which is '\r'. This is like newline ('\n'), except that it doesn’t move to the next line. It just returns to the front of the current line. What we’re going to do is a little text-based status bar like so many command line programs implement. It’ll do a countdown from 10 to 0 printing over itself on the same line. What is the catch and what does this have to do with fflush()? The catch is that the terminal is most likely “line buffered” (see the section on setvbuf() for more info), meaning that it won’t actually display anything until it prints a newline. But we’re not printing newlines; we’re just printing carriage returns, so we need a way to force the output to occur even though we’re on the same line. Yes, it’s fflush()!
1 #include
2 #include
3
4 int main(void)
5 {
6 int count;
7
8 for(count = 10; count >= 0; count--) {
9 printf("\rSeconds until launch: "); // lead with a CR
10 if (count > 0)
11 printf("%2d", count);
12 else
13 printf("blastoff!\n");
14
15 // force output now!!
16 fflush(stdout);
17
18 // the sleep() function is non-standard, but virtually every
19 // system implements it--it simply delays for the specificed
20 // number of seconds:
21 sleep(1);
22 }
23 }
See Also
setbuf(), setvbuf()
40.7 fopen()
Opens a file for reading or writing Chapter 40.
Synopsis
#include
FILE *fopen(const char *path, const char *mode);
Description
The fopen() opens a file for reading or writing. Parameter path can be a relative or fully-qualified path and file name to the file in question. Parameter mode tells fopen() how to open the file (reading, writing, or both), and whether or not it’s a binary file. Possible modes are:
Mode Description r Open the file for reading (read-only). w Open the file for writing (write-only). The file is created if it doesn’t exist. r+ Open the file for reading and writing. The file has to already exist. w+ Open the file for writing and reading. The file is created if it doesn’t already exist. a Open the file for append. This is just like opening a file for writing, but it positions the file pointer at the end of the file, so the next write appends to the end. The file is created if it doesn’t exist. a+ Open the file for reading and appending. The file is created if it doesn’t exist.
Any of the modes can have the letter “b” appended to the end, as is “wb” (“write binary”), to signify that the file in question is a binary file. (“Binary” in this case generally means that the file contains non-alphanumeric characters that look like garbage to human eyes.) Many systems (like Unix) don’t differentiate between binary and non-binary files, so the “b” is extraneous. But if your data is binary, it doesn’t hurt to throw the “b” in there, and it might help someone who is trying to port your code to another system. The macro FOPEN_MAX tells you how many streams (at least) you can have open at once. The macro FILENAME_MAX tells you what the longest valid filename can be. Don’t go crazy, now.
Return Value
fopen() returns a FILE* that can be used in subsequent file-related calls. If something goes wrong (e.g. you tried to open a file for read that didn’t exist), fopen() will return NULL.
Example
1 int main(void)
2 {
3 FILE *fp;
4
5 if ((fp = fopen("datafile.dat", "r")) == NULL) {
6 printf("Couldn't open datafile.dat for reading\n");
7 exit(1);
8 }
9 Chapter 40.
10 // fp is now initialized and can be read from it
11 }
See Also
fclose(), freopen()
40.8 freopen()
Reopen an existing FILE*, associating it with a new path
Synopsis
#include
FILE *freopen(const char *filename, const char *mode, FILE *stream);
Description
Let’s say you have an existing FILE* stream that’s already open, but you want it to suddenly use a different file than the one it’s using. You can use freopen() to “re-open” the stream with a new file. Why on Earth would you ever want to do that? Well, the most common reason would be if you had a program that normally would read from stdin, but instead you wanted it to read from a file. Instead of changing all your scanf()s to fscanf()s, you could simply reopen stdin on the file you wanted to read from. Another usage that is allowed on some systems is that you can pass NULL for filename, and specify a new mode for stream. So you could change a file from “r+” (read and write) to just “r” (read), for instance. It’s implementation dependent which modes can be changed. When you call freopen(), the old stream is closed. Otherwise, the function behaves just like the stan- dard fopen().
Return Value
freopen() returns stream if all goes well. If something goes wrong (e.g. you tried to open a file for read that didn’t exist), freopen() will return NULL.
Example
1 #include
2
3 int main(void)
4 {
5 int i, i2;
6
7 scanf("%d", &i); // read i from stdin
8
9 // now change stdin to refer to a file instead of the keyboard
10 freopen("someints.txt", "r", stdin);
11
12 scanf("%d", &i2); // now this reads from the file "someints.txt"
13 Chapter 40.
14 printf("Hello, world!\n"); // print to the screen
15
16 // change stdout to go to a file instead of the terminal:
17 freopen("output.txt", "w", stdout);
18
19 printf("This goes to the file \"output.txt\"\n");
20
21 // this is allowed on some systems--you can change the mode of a file:
22 freopen(NULL, "wb", stdout); // change to "wb" instead of "w"
23 }
See Also
fclose(), fopen()
40.9 setbuf(), setvbuf()
Configure buffering for standard I/O operations
Synopsis
#include
void setbuf(FILE *stream, char *buf);
int setvbuf(FILE *stream, char *buf, int mode, size_t size);
Description
Now brace yourself because this might come as a bit of a surprise to you: when you printf() or fprintf() or use any I/O functions like that, it does not normally work immediately. For the sake of efficiency, and to irritate you, the I/O on a FILE* stream is buffered away safely until certain conditions are met, and only then is the actual I/O performed. The functions setbuf() and setvbuf() allow you to change those conditions and the buffering behavior. So what are the different buffering behaviors? The biggest is called “full buffering”, wherein all I/O is stored in a big buffer until it is full, and then it is dumped out to disk (or whatever the file is). The next biggest is called “line buffering”; with line buffering, I/O is stored up a line at a time (until a newline ('\n') character is encountered) and then that line is processed. Finally, we have “unbuffered”, which means I/O is processed immediately with every standard I/O call. You might have seen and wondered why you could call putchar() time and time again and not see any output until you called putchar('\n'); that’s right—stdout is line-buffered! Since setbuf() is just a simplified version of setvbuf(), we’ll talk about setvbuf() first. The stream is the FILE* you wish to modify. The standard says you must make your call to setvbuf() before any I/O operation is performed on the stream, or else by then it might be too late. The next argument, buf allows you to make your own buffer space (using malloc() or just a char array) to use for buffering. If you don’t care to do this, just set buf to NULL. Now we get to the real meat of the function: mode allows you to choose what kind of buffering you want to use on this stream. Set it to one of the following:
Mode Description _IOFBF stream will be fully buffered. Chapter 40.
Mode Description _IOLBF stream will be line buffered. _IONBF stream will be unbuffered.
Finally, the size argument is the size of the array you passed in for buf…unless you passed NULL for buf, in which case it will resize the existing buffer to the size you specify. Now what about this lesser function setbuf()? It’s just like calling setvbuf() with some specific parameters, except setbuf() doesn’t return a value. The following example shows the equivalency: // these are the same: setbuf(stream, buf); setvbuf(stream, buf, _IOFBF, BUFSIZ); // fully buffered
// and these are the same: setbuf(stream, NULL); setvbuf(stream, NULL, _IONBF, BUFSIZ); // unbuffered
Return Value
setvbuf() returns zero on success, and nonzero on failure. setbuf() has no return value.
Example
1 FILE *fp;
2 char lineBuf[1024];
3
4 fp = fopen("somefile.txt", "r");
5 setvbuf(fp, lineBuf, _IOLBF, 1024); // set to line buffering
6 // ...
7 fclose(fp);
8
9 fp = fopen("another.dat", "rb");
10 setbuf(fp, NULL); // set to unbuffered
11 // ...
12 fclose(fp);
See Also
fflush()
40.10 printf(), fprintf(), sprintf(), snprintf()
Print a formatted string to the console or to a file
Synopsis
#include
int printf(const char *format, ...);
int fprintf(FILE *stream, const char *format, ...);
int sprintf(char * restrict s, const char * restrict format, ...); Chapter 40.
int snprintf(char * restrict s, size_t n, const char * restrict format, ...);
Description These functions print formatted output to a variety of destinations.
Function Output Destination printf() Print to console (screen by default, typically). fprintf() Print to a file. sprintf() Print to a string. snprintf() Print to a string (safely).
The only differences between these is are the leading parameters that you pass to them before the format string.
Function What you pass before format printf() Nothing comes before format. fprintf() Pass a FILE*. sprintf() Pass a char* to a buffer to print into. snprintf() Pass a char* to the buffer and a maximum buffer length.
The printf() function is legendary as being one of the most flexible outputting systems ever devised. It can also get a bit freaky here or there, most notably in the format string. We’ll take it a step at a time here. The easiest way to look at the format string is that it will print everything in the string as-is, unless a character has a percent sign (%) in front of it. That’s when the magic happens: the next argument in the printf() argument list is printed in the way described by the percent code. These percent codes are called format specifiers. Here are the most common format specifiers.
Specifier Description %d Print the next argument as a signed decimal number, like 3490. The argument printed this way should be an int, or something that gets promoted to int. %f Print the next argument as a signed floating point number, like 3.14159. The argument printed this way should be a double, or something that gets promoted to a double. %c Print the next argument as a character, like 'B'. The argument printed this way should be a char variant. %s Print the next argument as a string, like "Did you remember your mittens?". The argument printed this way should be a char* or char[]. %% No arguments are converted, and a plain old run-of-the-mill percent sign is printed. This is how you print a ‘%’ using printf().
So those are the basics. I’ll give you some more of the format specifiers in a bit, but let’s get some more breadth before then. There’s actually a lot more that you can specify in there after the percent sign. For one thing, you can put a field width in there—this is a number that tells printf() how many spaces to put on one side or the other of the value you’re printing. That helps you line things up in nice columns. If the number is negative, the result becomes left-justified instead of right-justified. Example: printf("%10d", x); /* prints X on the right side of the 10-space field */ printf("%-10d", x); /* prints X on the left side of the 10-space field */
If you don’t know the field width in advance, you can use a little kung-foo to get it from the argument list Chapter 40.
You can also put a “0” in front of the number if you want it to be padded with zeros: int x = 17; printf("%05d", x); /* "00017" */
When it comes to floating point, you can also specify how many decimal places to print by making a field width of the form “x.y” where x is the field width (you can leave this off if you want it to be just wide enough) and y is the number of digits past the decimal point to print: float f = 3.1415926535; printf("%.2f", f); /* "3.14" */ printf("%7.3f", f); /* " 3.141" <-- 7 spaces across */
Ok, those above are definitely the most common uses of printf(), but let’s get total coverage.
40.10.0.1 Format Specifier Layout Technically, the layout of the format specifier is these things in this order: 1. %, followed by… 2. Optional: zero or more flags, left justify, leading zeros, etc. 3. Optional: Field width, how wide the output field should be. 4. Optional: Precision, or how many decimal places to print. 5. Optional: Length modifier, for printing things bigger than int or double. 6. Conversion specifier, like d, f, etc. In short, the whole format specifier is laid out like this:
%[flags][fieldwidth][.precision][lengthmodifier]conversionspecifier What could be easier?
40.10.0.2 Conversion Specifiers Let’s talk conversion specifiers first. Each of the following specifies what type it can print, but it can also print anything that gets promoted to that type. For example, %d can print int, short, and char.
Conversion Specifier Description d Print an int argument as a decimal number. i Identical to d. o Print an unsigned int in octal (base 8). u Print an unsigned int in decimal. x Print an unsigned int in hexadecimal with lowercase letters. X Print an unsigned int in hexadecimal with uppercase letters. f Print a double in decimal notation. Infinity is printed as infinity or inf, and NaN is printed as nan, any of which could have a leading minus sign. F Same as f, except it prints out INFINITY, INF, or NAN in all caps. e Print a number in scientific notation, e.g. 1.234e56. Does infinity and NaN like f. E Just like e, except prints the exponent E (and infinity and NaN) in uppercase. g Print small numbers like f and large numbers like e. See note below. G Print small numbers like F and large numbers like E. See note below. Chapter 40.
Conversion Specifier Description a Print a double in hexadecimal form 0xh.hhhhpd where h is a lowercase hex digit and d is a decimal exponent of 2. Infinity and NaN in the form of f. More below. A Like a except everything’s uppercase. c Convert int argument to unsigned char and print as a character. s Print a string starting at the given char*. p Print a void* out as a number, probably the numeric address, possibly in hex. n Store the number of characters written so far in the given int*. Doesn’t print anything. See below. % Print a literal percent sign.
40.10.0.2.1 Note on %a and %A When printing floating point numbers in hex form, there is one number before the decimal point, and the rest of are out to the precision. double pi = 3.14159265358979; printf("%.3a\n", pi); // 0x1.922p+1
C can choose the leading number in such a way to ensure subsequent digits align to 4-bit boundaries. If the precision is left out and the macro FLT_RADIX is a power of 2, enough precision is used to represent the number exactly. If FLT_RADIX is not a power of two, enough precision is used to be able to tell any two floating values apart. If the precision is 0 and the # flag isn’t specified, the decimal point is omitted.
40.10.0.2.2 Note on %g and %G The gist of this is to use scientific notation when the number gets too “extreme”, and regular decimal notation otherwise. The exact behavior for whether these print as %f or %e depends on a number of factors: If the number’s exponent is greater than or equal to -4 and the precision is greater than the exponent, we use %f. In this case, the precision is converted according to 푝 = 푝 − (푥 + 1), where 푝 is the specified precision and 푥 is the exponent. Otherwise we use %e, and the precision becomes 푝 − 1. Trailing zeros in the decimal portion are removed. And if there are none left, the decimal point is removed, too. All this unless the # flag is specified.
40.10.0.2.3 Note on %n This specifier is cool and different, and rarely needed. It doesn’t actually print anything, but stores the number of characters printed so far in the next pointer argument in the list. int numChars; float a = 3.14159; int b = 3490; printf("%f %d%n\n", a, b, &numChars); printf("The above line contains %d characters.\n", numChars);
The above example will print out the values of a and b, and then store the number of characters printed so far into the variable numChars. The next call to printf() prints out that result. 3.141590 3490 The above line contains 13 characters
40.10.0.3 Length Modifiers You can stick a length modifier in front of each of the conversion specifiers, if you want. most of those format specifiers work on int or double types, but what if you want larger or smaller types? That’s what these are good for. Chapter 40.
For example, you could print out a long long int with the ll modifier: long long int x = 3490; printf("%lld\n", x); // 3490
Length Modifier Conversion Specifier Description hh d, i, o, u, x, X Convert argument to char (signed or unsigned as appropriate) before printing. h d, i, o, u, x, X Convert argument to short int (signed or unsigned as appropriate) before printing. l d, i, o, u, x, X Argument is a long int (signed or unsigned as appropriate). ll d, i, o, u, x, X Argument is a long long int (signed or unsigned as appropriate). j d, i, o, u, x, X Argument is a intmax_t or uintmax_t (as appropriate). z d, i, o, u, x, X Argument is a size_t. t d, i, o, u, x, X Argument is a ptrdiff_t. L a, A, e, E, f, F, g, G Argument is a long double. l c Argument is in a wint_t, a wide character. l s Argument is in a wchar_t*, a wide character string. hh n Store result in signed char* argument. h n Store result in short int* argument. l n Store result in long int* argument. ll n Store result in long long int* argument. j n Store result in intmax_t* argument. z n Store result in size_t* argument. t n Store result in ptrdiff_t* argument.
40.10.0.4 Precision In front of the length modifier, you can put a precision, which generally means how many decimal places you want on your floating point numbers. To do this, you put a decimal point (.) and the decimal places afterward. For example, we could print π rounded to two decimal places like this: double pi = 3.14159265358979; printf("%.2f\n", pi); // 3.14
Conversion Specifier Precision Value Meaning d, i, o, u, x, X For integer types, minimum number of digits (will pad with leading zeros) a, e, f, A, E, F For floating types, the precision is the number of digits past the decimal. g, G For floating types, the precision is the number of significant digits printed. s The maximum number of bytes (not multibyte characters!) to be written.
If no number is specified in the precision after the decimal point, the precision is zero. If an * is specified after the decimal, something amazing happens! It means the int argument to printf() before the number to be printed holds the precision. You can use this if you don’t know the precision at compile time. Chapter 40.
Which gives: Enter precision: 4 3.1416
40.10.0.5 Field Width In front of the optional precision, you can indicate a field width. This is a decimal number that indicates how wide the region should be in which the argument is printed. The region is padding with leading (or trailing) spaces to make sure it’s wide enough. If the field width specified is too small to hold the output, it is ignored. As a preview, you can give a negative field width to justify the item the other direction. So let’s print a number in a field of width 10. We’ll put some angle brackets around it so we can see the padding spaces in the output. printf("<<%10d>>\n", 3490); // right justified printf("<<%-10d>>\n", 3490); // left justified
<< 3490>> <<3490 >>
Like with the precision, you can use an asterisk (*) as the field width int field_width; int val = 3490; printf("Enter field_width: "); fflush(stdout); scanf("%d", &field_width); printf("<<%*d>>\n", field_width, val);
40.10.0.6 Flags Before the field width, you can put some optional flags that further control the output of the subsequent fields. We just saw that the - flag can be used to left- or right-justify fields. But there are plenty more!
Flag Description - For a field width, left justify in the field (right is default). + If the number is signed, always prefix a + or - on the front. [SPACE] If the number is signed, prefix a space for positive, or a - for negative. 0 Pad the right-justified field with leading zeros instead of leading spaces. # Print using an alternate form. See below.
For example, we could pad a hexadecimal number with leading zeros to a field width of 8 with: printf("%08x\n", 0x1234); // 00001234
The # “alternate form” result depends on the conversion specifier. Chapter 40.
Conversion Specifier Alternate Form (#) Meaning o Increase precision of a non-zero number just enough to get one leading 0 on the octal number. x Prefix a non-zero number with 0x. X Same as x, except capital 0X. a, e, f Always print a decimal point, even if nothing follows it. A, E, F Identical to a, e, f. g, G Always print a decimal point, even if nothing follows it, and keep trailing zeros.
40.10.0.7 sprintf() and snprintf() Details
Both sprintf() and snprintf() have the quality that if you pass in NULL as the buffer, nothing is written—but you can still check the return value to see how many characters would have been written. snprintf() always terminates the string with a NUL character. So if you try to write out more than the maximum specified characters, the universe ends. Just kidding. If you do, snprintf() will write 푛 − 1 characters so that it has enough room to write the terminator at the end.
Return Value Returns the number of characters outputted, or a negative number on error.
Example
1 int a = 100;
2 float b = 2.717;
3 char *c = "beej!";
4 char d = 'X';
5 int e = 5;
6
7 printf("%d", a); /* "100" */
8 printf("%f", b); /* "2.717000" */
9 printf("%s", c); /* "beej!" */
10 printf("%c", d); /* "X" */
11 printf("110%%"); /* "110%" */
12
13 printf("%10d\n", a); /* " 100" */
14 printf("%-10d\n", a); /* "100 " */
15 printf("%*d\n", e, a); /* " 100" */
16 printf("%.2f\n", b); /* "2.71" */
17
18 printf("%hhd\n", c); /* "88" <-- ASCII code for 'X' */
19
20 printf("%5d %5.2f %c\n", a, b, d); /* " 100 2.71 X" */
See Also
sprintf(), vprintf()
40.11 scanf(), fscanf(), sscanf()
Read formatted string, character, or numeric data from the console or from a file Chapter 40.
Synopsis
#include
Description These functions read formatted output from a variety of sources.
Function Input Source scanf() Read from the console (keyboard by default, typically). fscanf() Read from a file. sscanf() Read from a string.
The only differences between these is are the leading parameters that you pass to them before the format string.
Function What you pass before format scanf() Nothing comes before format. fscanf() Pass a FILE*. sscanf() Pass a char* to a buffer to read from.
The scanf() family of functions reads data from the console or from a FILE stream, parses it, and stores the results away in variables you provide in the argument list. The format string is very similar to that in printf() in that you can tell it to read a "%d", for instance for an int. But it also has additional capabilities, most notably that it can eat up other characters in the input that you specify in the format string. But let’s start simple, and look at the most basic usage first before plunging into the depths of the function. We’ll start by reading an int from the keyboard: int a; scanf("%d", &a); scanf() obviously needs a pointer to the variable if it is going to change the variable itself, so we use the address-of operator to get the pointer. In this case, scanf() walks down the format string, finds a “%d”, and then knows it needs to read an integer and store it in the next variable in the argument list, a. Here are some of the other format specifiers you can put in the format string:
Format Specifier Description %d Reads an integer to be stored in an int. This integer can be signed. %u Reads an integer to be stored in an unsigned int. %f Reads a floating point number, to be stored in a float. %s Reads a string up to the first whitespace character. %c Reads a char. Chapter 40.
And that’s the end of the story! Ha! Just kidding. If you’ve just arrived from the printf() page, you know there’s a near-infinite amount of additional material.
40.11.0.1 Consuming Other Characters scanf() will move along the format string matching any characters you include. For example, you could read a hyphenated date like so: scanf("%u-%u-%u", &yyyy, &mm, &dd);
In that case, scanf() will attempt to consume an unsigned decimal number, then a hyphen, then another unsigned number, then another hypen, then another unsigned number. If it fails to match at any point (e.g. the user entered “foo”), scanf() will bail without consuming the offending characters. And it will return the number of variables successfully converted. In the example above, if the user entered a valid string, scanf() would return 3, one for each variable successfully read.
40.11.0.2 Problems with scanf()
I (and the C FAQ and a lot of people) recommend against using scanf() to read directly from the key- board. It’s too easy for it to stop consuming characters when the user enters some bad data. If you have data in a file and you’re confident it’s in good shape, fscanf() can be really useful. But in the case of the keyboard or file, you can always use fgets() to read a complete line into a buffer, and then use sscanf() to scan things out of the buffer. This gives you the best of both worlds.
40.11.0.3 Problems with sscanf() A while back, a third-party programmer rose to fame for figuring out how to cut GTA Online load times by 70%2. What they’d discovered was that the implementation of sscanf() first effectively calls strlen()… so even if you’re just using sscanf() to peel the first few characters off the string, it still runs all the way out to the end of the string first. On small strings, no big deal, but on large strings with repeated calls (which is what was happening in GTA) it got sloooooooooowwwww… So if you’re just converting a string to a number, consider atoi(), atof(), or the strtol() and str- tod() families of functions, instead. (The programmer collected a $10,000 bug bounty for the effort.)
40.11.0.4 The Deep Details
Let’s check out what a scanf() And here are some more codes, except these don’t tend to be used as often. You, of course, may use them as often as you wish! First, the format string. Like we mentioned, it can hold ordinary characters as well as % format specifiers. And whitespace characters. Whitespace characters have a special role: a whitespace character will cause scanf() to consume as many whitespace characters as it can up to the next non-whitespace character. You can use this to ignore all leading or trailing whitespace. Also, all format specifiers except for s, c, and [ automatically consume leading whitespace.
2https://nee.lv/2021/02/28/How-I-cut-GTA-Online-loading-times-by-70/ Chapter 40.
But I know what you’re thinking: the meat of this function is in the format specifiers. What do those look like? These consist of the following, in sequence: 1. A % sign 2. Optional: an * to suppress assignment—more later 3. Optional: a field width—max characters to read 4. Optional: length modifier, for specifying longer or shorter types 5. A conversion specifier, like d or f indicating the type to read
40.11.0.5 The Conversion Specifier Let’s start with the best and last: the conversion specifier. This is the part of the format specifier that tells us what type of variable scanf() should be reading into, like %d or %f.
Conversion Specifier Description d Matches a decimal int. Can have a leading sign. i Like d, except will handle it if you put a leading 0x (hex) or 0 (octal) on the number. o Matches an octal (base 8) unsigned int. Leading zeros are ignored. u Matches a decimal unsigned int. x Matches a hex (base 16) unsigned int. f Match a floating point number (or scientific notation, or anything strtod() can handle). c Match a char, or mutiple chars if a field width is given. s Match a sequence of non-whitespace chars. [ Match a sequence of characters from a set. The set ends with ]. More below. p Match a pointer, the opposite of %p for printf(). n Store the number of characters written so far in the given int*. Doesn’t consume anything. % Match a literal percent sign.
All of the following are equivalent to the f specifier: a, e, g, A, E, F, G. And capital X is equivalent to lowercase x.
40.11.0.5.1 The Scanset %[] Conversion Specifier This is about the weirdest format specifier there is. It allows you to specify a set of characters (the scanset) to be stored away (likely in an array of chars). Conversion stops when a character that is not in the set is matched. For example, %[0-9] means “match all numbers zero through nine.” And %[AD-G34] means “match A, D through G, 3, or 4”. Now, to convolute matters, you can tell scanf() to match characters that are not in the set by putting a caret (^) directly after the %[ and following it with the set, like this: %[^A-C], which means “match all characters that are not A through C.” To match a close square bracket, make it the first character in the set, like this: %[]A-C] or %[^]A-C]. (I added the “A-C” just so it was clear that the “]” was first in the set.) To match a hyphen, make it the last character in the set, e.g. to match A-through-C or hyphen: %[A-C-]. So if we wanted to match all letters except “%”, “^”, “]”, “B”, “C”, “D”, “E”, and “-”, we could use this format string: %[^]%^B-E-]. Got it? Now we can go onto the next func—no wait! There’s more! Yes, still more to know about scanf(). Does it never end? Try to imagine how I feel writing about it! Chapter 40.
40.11.0.6 The Length Modifier
So you know that “%d” stores into an int. But how do you store into a long, short, or double? Well, like in printf(), you can add a modifier before the type specifier to tell scanf() that you have a longer or shorter type. The following is a table of the possible modifiers:
Length Modifier Conversion Specifier Description hh d, i, o, u, x, X Convert input to char (signed or unsigned as appropriate) before printing. h d, i, o, u, x, X Convert input to short int (signed or unsigned as appropriate) before printing. l d, i, o, u, x, X Convert input to long int (signed or unsigned as appropriate). ll d, i, o, u, x, X Convert input to long long int (signed or unsigned as appropriate). j d, i, o, u, x, X Convert input to intmax_t or uintmax_t (as appropriate). z d, i, o, u, x, X Convert input to size_t. t d, i, o, u, x, X Convert input to ptrdiff_t. L a, A, e, E, f, F, g, G Convert input to long double. l c,s,[ Convert input to wchar_t, a wide character. l s Argument is in a wchar_t*, a wide character string. hh n Store result in signed char* argument. h n Store result in short int* argument. l n Store result in long int* argument. ll n Store result in long long int* argument. j n Store result in intmax_t* argument. z n Store result in size_t* argument. t n Store result in ptrdiff_t* argument.
40.11.0.7 Field Widths The field width generally allows you to specify a maximum number of characters to consume. If the thing you’re trying to match is shorter than the field width, that input will stop being processed before the field width is reached. So a string will stop being consumed when whitespace is found, even if fewer than the field width char- acters are matched. And a float will stop being consumed at the end of the number, even if fewer characters than the field width are matched. But %c is an interesting one—it doesn’t stop consuming characters on anything. So it’ll go exactly to the field width. (Or 1 character if no field width is given.)
40.11.0.8 Skip Input with *
If you put an * in the format specifier, it tells scanf() do to the conversion specified, but not store it anywhere. It simply discards the data as it reads it. This is what you use if you want scanf() to eat some data but you don’t want to store it anywhere; you don’t give scanf() an argument for this conversion. // Read 3 ints, but discard the middle one scanf("%d %*d %d", &int1, &int3);
Return Value scanf() returns the number of items assigned into variables. Since assignment into variables stops when given invalid input for a certain format specifier, this can tell you if you’ve input all your data correctly. Also, scanf() returns EOF on end-of-file. Chapter 40.
Example
1 int a;
2 long int b;
3 unsigned int c;
4 float d;
5 double e;
6 long double f;
7 char s[100];
8
9 scanf("%d", &a); // store an int
10 scanf(" %d", &a); // eat any whitespace, then store an int
11 scanf("%s", s); // store a string
12 scanf("%Lf", &f); // store a long double
13
14 // store an unsigned, read all whitespace, then store a long int:
15 scanf("%u %ld", &c, &b);
16
17 // store an int, read whitespace, read "blendo", read whitespace,
18 // and store a float:
19 scanf("%d blendo %f", &a, &d);
20
21 // read all whitespace, then store all characters up to a newline
22 scanf(" %[^\n]", s);
23
24 // store a float, read (and ignore) an int, then store a double:
25 scanf("%f %*d %lf", &d, &e);
26
27 // store 10 characters:
28 scanf("%10c", s);
See Also
sscanf(), vscanf(), vsscanf(), vfscanf()
40.12 vprintf(), vfprintf(), vsprintf(), vsnprintf()
printf() variants using variable argument lists (va_list)
Synopsis
#include
int vprintf(const char * restrict format, va_list arg);
int vfprintf(FILE * restrict stream, const char * restrict format, va_list arg);
int vsprintf(char * restrict s, const char * restrict format, va_list arg);
int vsnprintf(char * restrict s, size_t n, const char * restrict format, va_list arg); Chapter 40.
Description
These are just like the printf() variants except instead of taking an actual variable number of arguments, they take a fixed number—the last of which is a va_list that refers to the variable arguments. Like with printf(), the different variants send output different places.
Function Output Destination vprintf() Print to console (screen by default, typically). vfprintf() Print to a file. vsprintf() Print to a string. vsnprintf() Print to a string (safely).
Both vsprintf() and vsnprintf() have the quality that if you pass in NULL as the buffer, nothing is written—but you can still check the return value to see how many characters would have been written. If you try to write out more than the maximum number of characters, vsnprintf() will graciously write only 푛 − 1 characters so that it has enough room to write the terminator at the end. As for why in the heck would you ever want to do this, the most common reason is to create your own specialized versions of printf()-type functions, piggybacking on all that printf() functionality good- ness. See the example for an example, predictably.
Return Value
vprintf() and vfprintf() return the number of characters printed, or a negative value on error. vsprintf() returns the number of characters printed to the buffer, not counting the NUL terminator, or a negative value if an error occurred. vnsprintf() returns the number of characters printed to the buffer. Or the number that would have been printed if the buffer had been large enough.
Example
In this example, we make our own version of printf() called logger() that timestamps output. Notice how the calls to logger() have all the bells and whistles of printf().
1 #include
2 #include
3 #include
4
5 int logger(char *format, ...)
6 {
7 va_list va;
8 time_t now_secs = time(NULL);
9 struct tm *now = gmtime(&now_secs);
10
11 // Output timestamp in format "YYYY-MM-DD hh:mm:ss : "
12 printf("%04d-%02d-%02d %02d:%02d:%02d : ",
13 now->tm_year + 1900, now->tm_mon + 1, now->tm_mday,
14 now->tm_hour, now->tm_min, now->tm_sec);
15
16 va_start(va, format);
17 int result = vprintf(format, va);
18 va_end(va);
19
20 printf("\n"); Chapter 40.
21
22 return result;
23 }
24
25 int main(void)
26 {
27 int x = 12;
28 float y = 3.2;
29
30 logger("Hello!");
31 logger("x = %d and y = %.2f", x, y);
32 }
Output: 2021-03-30 04:25:49 : Hello! 2021-03-30 04:25:49 : x = 12 and y = 3.20
See Also
printf()
40.13 vscanf(), vfscanf(), vsscanf()
scanf() variants using variable argument lists (va_list)
Synopsis
#include
int vscanf(const char * restrict format, va_list arg);
int vfscanf(FILE * restrict stream, const char * restrict format, va_list arg);
int vsscanf(const char * restrict s, const char * restrict format, va_list arg);
Description
These are just like the scanf() variants except instead of taking an actual variable number of arguments, they take a fixed number—the last of which is a va_list that refers to the variable arguments.
Function Input Source vscanf() Read from the console (keyboard by default, typically). vfscanf() Read from a file. vsscanf() Read from a string.
Like with the vprintf() functions, this would be a good way to add additional functionality that took advantage of the power scanf() has to offer. Chapter 40.
Return Value
Returns the number of items successfully scanned, or EOF on end-of-file or error.
Example I have to admit I was wracking my brain to think of when you’d ever want to use this. The best example I could find was one on Stack Overflow3 that error-checks the return value from scanf() against the expected. A variant of that is shown below.
1 #include
2 #include
3 #include
4
5 int error_check_scanf(int expected_count, char *format, ...)
6 {
7 va_list va;
8
9 va_start(va, format);
10 int count = vscanf(format, va);
11 va_end(va);
12
13 // This line will crash the program if the condition is false:
14 assert(count == expected_count);
15
16 return count;
17 }
18
19 int main(void)
20 {
21 int a, b;
22 float c;
23
24 error_check_scanf(3, "%d, %d/%f", &a, &b, &c);
25 error_check_scanf(2, "%d", &a);
26 }
See Also
scanf()
40.14 getc(), fgetc(), getchar()
Get a single character from the console or from a file
Synopsis
#include
int getc(FILE *stream);
int fgetc(FILE *stream);
int getchar(void);
3https://stackoverflow.com/questions/17017331/c99-vscanf-for-dummies/17018046#17018046 Chapter 40.
Description
All of these functions in one way or another, read a single character from the console or from a FILE. The differences are fairly minor, and here are the descriptions: getc() returns a character from the specified FILE. From a usage standpoint, it’s equivalent to the same fgetc() call, and fgetc() is a little more common to see. Only the implementation of the two functions differs. fgetc() returns a character from the specified FILE. From a usage standpoint, it’s equivalent to the same getc() call, except that fgetc() is a little more common to see. Only the implementation of the two functions differs. Yes, I cheated and used cut-n-paste to do that last paragraph. getchar() returns a character from stdin. In fact, it’s the same as calling getc(stdin).
Return Value
All three functions return the unsigned char that they read, except it’s cast to an int. If end-of-file or an error is encountered, all three functions return EOF.
Example
1 // read all characters from a file, outputting only the letter 'b's
2 // it finds in the file
3
4 #include
5
6 int main(void)
7 {
8 FILE *fp;
9 int c;
10
11 fp = fopen("datafile.txt", "r"); // error check this!
12
13 // this while-statement assigns into c, and then checks against EOF:
14
15 while((c = fgetc(fp)) != EOF) {
16 if (c == 'b') {
17 putchar(c);
18 }
19 }
20
21 fclose(fp);
22 }
See Also
40.15 gets(), fgets()
Read a string from console or file
Synopsis Chapter 40.
#include
Description These are functions that will retrieve a newline-terminated string from the console or a file. In other normal words, it reads a line of text. The behavior is slightly different, and, as such, so is the usage. For instance, here is the usage of gets(): Don’t use gets(). In fact, as of C11, it ceases to exist! This is one of the rare cases of a function being removed from the standard. Admittedly, rationale would be useful, yes? For one thing, gets() doesn’t allow you to specify the length of the buffer to store the string in. This would allow people to keep entering data past the end of your buffer, and believe me, this would be Bad News. And that’s what the size parameter in fgets() is for. fgets() will read at most size-1 characters and then stick a NUL terminator on after that. I was going to add another reason, but that’s basically the primary and only reason not to use gets(). As you might suspect, fgets() allows you to specify a maximum string length. One difference here between the two functions: gets() will devour and throw away the newline at the end of the line, while fgets() will store it at the end of your string (space permitting). Here’s an example of using fgets() from the console, making it behave more like gets() (with the exception of the newline inclusion): char s[100]; gets(s); // don't use this--read a line (from stdin) fgets(s, sizeof(s), stdin); // read a line from stdin
In this case, the sizeof() operator gives us the total size of the array in bytes, and since a char is a byte, it conveniently gives us the total size of the array. Of course, like I keep saying, the string returned from fgets() probably has a newline at the end that you might not want. You can write a short function to chop the newline off—in fact, let’s just roll that into our own version of gets() #include
if (rv == NULL) return NULL;
char *p = strchr(s, '\n'); // Find a newline
if (p != NULL) // if there's a newline *p = '\0'; // truncate the string there
return s; }
So, in summary, use fgets() to read a line of text from the keyboard or a file, and don’t use gets(). Chapter 40.
Return Value
Both gets() and fgets() return a pointer to the string passed. On error or end-of-file, the functions return NULL.
Example
1 char s[100];
2
3 gets(s); // read from standard input (don't use this--use fgets()!)
4
5 fgets(s, sizeof(s), stdin); // read 100 bytes from standard input
6
7 fp = fopen("datafile.dat", "r"); // (you should error-check this)
8 fgets(s, 100, fp); // read 100 bytes from the file datafile.dat
9 fclose(fp);
10
11 fgets(s, 20, stdin); // read a maximum of 20 bytes from stdin
See Also
getc(), fgetc(), getchar(), puts(), fputs(), ungetc()
40.16 putc(), fputc(), putchar()
Write a single character to the console or to a file
Synopsis
#include
int putc(int c, FILE *stream);
int fputc(int c, FILE *stream);
int putchar(int c);
Description
All three functions output a single character, either to the console or to a FILE. putc() takes a character argument, and outputs it to the specified FILE. fputc() does exactly the same thing, and differs from putc() in implementation only. Most people use fputc(). putchar() writes the character to the console, and is the same as calling putc(c, stdout).
Return Value
All three functions return the character written on success, or EOF on error.
Example
1 // print the alphabet
2
3 #include
4
5 int main(void)
6 {
7 char i;
8
9 for(i = 'A'; i <= 'Z'; i++)
10 putchar(i);
11
12 putchar('\n'); // put a newline at the end to make it pretty
13 }
See Also
40.17 puts(), fputs()
Write a string to the console or to a file
Synopsis
#include
int puts(const char *s);
int fputs(const char *s, FILE *stream);
Description
Both these functions output a NUL-terminated string. puts() outputs to the console, while fputs() allows you to specify the file for output.
Return Value
Both functions return non-negative on success, or EOF on error.
Example
1 // read strings from the console and save them in a file
2
3 #include
4
5 int main(void)
6 {
7 FILE *fp;
8 char s[100];
9
10 fp = fopen("datafile.txt", "w"); // error check this!
11
12 while(fgets(s, sizeof(s), stdin) != NULL) { // read a string
13 fputs(s, fp); // write it to the file we opened
14 }
15
16 fclose(fp);
17 } Chapter 40.
See Also
40.18 ungetc()
Pushes a character back into the input stream
Synopsis
#include
int ungetc(int c, FILE *stream);
Description
You know how getc() reads the next character from a file stream? Well, this is the opposite of that—it pushes a character back into the file stream so that it will show up again on the very next read from the stream, as if you’d never gotten it from getc() in the first place. Why, in the name of all that is holy would you want to do that? Perhaps you have a stream of data that you’re reading a character at a time, and you won’t know to stop reading until you get a certain character, but you want to be able to read that character again later. You can read the character, see that it’s what you’re supposed to stop on, and then ungetc() it so it’ll show up on the next read. Yeah, that doesn’t happen very often, but there we are. Here’s the catch: the standard only guarantees that you’ll be able to push back one character. Some implementations might allow you to push back more, but there’s really no way to tell and still be portable.
Return Value
On success, ungetc() returns the character you passed to it. On failure, it returns EOF.
Example This example reads a piece of punctuation, then everything after it up to the next piece of punctuation. It returns the leading punctuation, and stores the rest in a string.
1 #include
2 #include
3
4 int read_punctstring(FILE *fp, char *s)
5 {
6 int origpunct, c;
7
8 origpunct = fgetc(fp);
9
10 if (origpunct == EOF) // return EOF on end-of-file
11 return EOF;
12
13 while (c = fgetc(fp), !ispunct(c) && c != EOF)
14 *s++ = c; // save it in the string
15
16 *s = '\0'; // nul-terminate the string
17
18 // if we read punctuation last, ungetc it so we can fgetc it next
19 // time:
20 if (ispunct(c)) Chapter 40.
21 ungetc(c, fp);
22
23 return origpunct;
24 }
25
26 int main(void)
27 {
28 char s[128];
29 char c;
30
31 while((c = read_punctstring(stdin, s)) != EOF) {
32 printf("%c: %s\n", c, s);
33 }
34 }
Sample Input: !foo#bar*baz
Sample output: !: foo #: bar *: baz
See Also
fgetc()
40.19 fread()
Read binary data from a file
Synopsis
#include
size_t fread(void *p, size_t size, size_t nmemb, FILE *stream);
Description
You might remember that you can call fopen() with the “b” flag in the open mode string to open the file in “binary” mode. Files open in not-binary (ASCII or text mode) can be read using standard character- oriented calls like fgetc() or fgets(). Files open in binary mode are typically read using the fread() function. All this function does is says, “Hey, read this many things where each thing is a certain number of bytes, and store the whole mess of them in memory starting at this pointer.” This can be very useful, believe me, when you want to do something like store 20 ints in a file. But wait—can’t you use fprintf() with the “%d” format specifier to save the ints to a text file and store them that way? Yes, sure. That has the advantage that a human can open the file and read the numbers. It has the disadvantage that it’s slower to convert the numbers from ints to text and that the numbers are likely to take more space in the file. (Remember, an int is likely 4 bytes, but the string “12345678” is 8 bytes.) So storing the binary data can certainly be more compact and faster to read. Chapter 40.
Return Value This function returns the number of items successfully read. If all requested items are read, the return value will be equal to that of the parameter nmemb. If EOF occurs, the return value will be zero. To make you confused, it will also return zero if there’s an error. You can use the functions feof() or ferror() to tell which one really happened.
Example
1 // read 10 numbers from a file and store them in an array
2
3 int main(void)
4 {
5 int i;
6 int n[10]
7 FILE *fp;
8
9 fp = fopen("binarynumbers.dat", "rb");
10 fread(n, sizeof(int), 10, fp); // read 10 ints
11 fclose(fp);
12
13 // print them out:
14 for(i = 0; i < 10; i++)
15 printf("n[%d] == %d\n", i, n[i]);
16 }
See Also
fopen(), fwrite(), feof(), ferror()
40.20 fwrite()
Write binary data to a file
Synopsis
#include
size_t fwrite(const void *p, size_t size, size_t nmemb, FILE *stream);
Description
This is the counterpart to the fread() function. It writes blocks of binary data to disk. For a description of what this means, see the entry for fread().
Return Value
fwrite() returns the number of items successfully written, which should hopefully be nmemb that you passed in. It’ll return zero on error.
Example
1 // save 10 random numbers to a file
2 Chapter 40.
3 int main(void)
4 {
5 int i;
6 int r[10];
7 FILE *fp;
8
9 // populate the array with random numbers:
10 for(i = 0; i < 10; i++) {
11 r[i] = rand();
12 }
13
14 // save the random numbers (10 ints) to the file
15 fp = fopen("binaryfile.dat", "wb");
16 fwrite(r, sizeof(int), 10, fp); // write 10 ints
17 fclose(fp);
18 }
See Also
fopen(), fread()
40.21 fgetpos(), fsetpos()
Get the current position in a file, or set the current position in a file. Just like ftell() and fseek() for most systems
Synopsis
#include
int fgetpos(FILE *stream, fpos_t *pos);
int fsetpos(FILE *stream, fpos_t *pos);
Description
These functions are just like ftell() and fseek(), except instead of counting in bytes, they use an opaque data structure to hold positional information about the file. (Opaque, in this case, means you’re not supposed to know what the data type is made up of.) On virtually every system (and certainly every system that I know of), people don’t use these functions, using ftell() and fseek() instead. These functions exist just in case your system can’t remember file positions as a simple byte offset. Since the pos variable is opaque, you have to assign to it using the fgetpos() call itself. Then you save the value for later and use it to reset the position using fsetpos().
Return Value
Both functions return zero on success, and -1 on error.
Example
1 char s[100];
2 fpos_t pos; Chapter 40.
3
4 fgets(s, sizeof(s), fp); // read a line from the file
5
6 fgetpos(fp, &pos); // save the position
7
8 fgets(s, sizeof(s), fp); // read another line from the file
9
10 fsetpos(fp, &pos); // now restore the position to where we saved
See Also
fseek(), ftell(), rewind()
40.22 fseek(), rewind()
Position the file pointer in anticipition of the next read or write
Synopsis
#include
int fseek(FILE *stream, long offset, int whence);
void rewind(FILE *stream);
Description When doing reads and writes to a file, the OS keeps track of where you are in the file using a counter generically known as the file pointer. You can reposition the file pointer to a different point in the file using the fseek() call. Think of it as a way to randomly access you file. The first argument is the file in question, obviously. offset argument is the position that you want to seek to, and whence is what that offset is relative to. Of course, you probably like to think of the offset as being from the beginning of the file. I mean, “Seek to position 3490, that should be 3490 bytes from the beginning of the file.” Well, it can be, but it doesn’t have to be. Imagine the power you’re wielding here. Try to command your enthusiasm. You can set the value of whence to one of three things:
whence Description SEEK_SET offset is relative to the beginning of the file. This is probably what you had in mind anyway, and is the most commonly used value for whence. SEEK_CUR offset is relative to the current file pointer position. So, in effect, you can say, “Move to my current position plus 30 bytes,” or, “move to my current position minus 20 bytes.” SEEK_END offset is relative to the end of the file. Just like SEEK_SET except from the other end of the file. Be sure to use negative values for offset if you want to back up from the end of the file, instead of going past the end into oblivion.
Speaking of seeking off the end of the file, can you do it? Sure thing. In fact, you can seek way off the end and then write a character; the file will be expanded to a size big enough to hold a bunch of zeros way out to that character. Now that the complicated function is out of the way, what’s this rewind() that I briefly mentioned? It Chapter 40.
repositions the file pointer at the beginning of the file: fseek(fp, 0, SEEK_SET); // same as rewind() rewind(fp); // same as fseek(fp, 0, SEEK_SET)
Return Value
For fseek(), on success zero is returned; -1 is returned on failure. The call to rewind() never fails.
Example
1 fseek(fp, 100, SEEK_SET); // seek to the 100th byte of the file
2 fseek(fp, -30, SEEK_CUR); // seek backward 30 bytes from the current pos
3 fseek(fp, -10, SEEK_END); // seek to the 10th byte before the end of file
4
5 fseek(fp, 0, SEEK_SET); // seek to the beginning of the file
6 rewind(fp); // seek to the beginning of the file
See Also
ftell(), fgetpos(), fsetpos()
40.23 ftell()
Tells you where a particular file is about to read from or write to
Synopsis
#include
long ftell(FILE *stream);
Description
This function is the opposite of fseek(). It tells you where in the file the next file operation will occur relative to the beginning of the file. It’s useful if you want to remember where you are in the file, fseek() somewhere else, and then come back later. You can take the return value from ftell() and feed it back into fseek() (with whence parameter set to SEEK_SET) when you want to return to your previous position.
Return Value
Returns the current offset in the file, or -1 on error.
Example
1 long pos;
2
3 // store the current position in variable "pos":
4 pos = ftell(fp);
5
6 // seek ahead 10 bytes:
7 fseek(fp, 10, SEEK_CUR); Chapter 40.
8
9 // do some mysterious writes to the file
10 do_mysterious_writes_to_file(fp);
11
12 // and return to the starting position, stored in "pos":
13 fseek(fp, pos, SEEK_SET);
See Also
fseek(), rewind(), fgetpos(), fsetpos()
40.24 feof(), ferror(), clearerr()
Determine if a file has reached end-of-file or if an error has occurred
Synopsis
#include
int feof(FILE *stream);
int ferror(FILE *stream);
void clearerr(FILE *stream);
Description
Each FILE* that you use to read and write data from and to a file contains flags that the system sets when certain events occur. If you get an error, it sets the error flag; if you reach the end of the file during a read, it sets the EOF flag. Pretty simple really. The functions feof() and ferror() give you a simple way to test these flags: they’ll return non-zero (true) if they’re set. Once the flags are set for a particular stream, they stay that way until you call clearerr() to clear them.
Return Value
feof() and ferror() return non-zero (true) if the file has reached EOF or there has been an error, re- spectively.
Example
1 // read binary data, checking for eof or error
2 int main(void)
3 {
4 int a;
5 FILE *fp;
6
7 fp = fopen("binaryints.dat", "rb");
8
9 // read single ints at a time, stopping on EOF or error:
10
11 while(fread(&a, sizeof(int), 1, fp), !feof(fp) && !ferror(fp)) {
12 printf("I read %d\n", a); Chapter 40.
13 }
14
15 if (feof(fp))
16 printf("End of file was reached.\n");
17
18 if (ferror(fp))
19 printf("An error occurred.\n");
20
21 fclose(fp);
22 }
See Also
fopen(), fread()
40.25 perror()
Print the last error message to stderr
Synopsis
#include
void perror(const char *s);
Description Many functions, when they encounter an error condition for whatever reason, will set a global variable called errno (in
errno Description EDOM Math operation outside domain. EILSEQ Invalid sequence in multibyte to wide character encoding. ERANGE Result of operation doesn’t fit in specified type.
The catch is that different systems define different values for errno, so it’s not very portable beyond the above 3. The good news is that at least the values are largely portable between Unix-like systems, at least.
Return Value Returns nothing at all! Sorry! Chapter 40.
Example
fseek() returns -1 on error, and sets errno, so let’s use it. Seeking on stdin makes no sense, so it should generate an error:
1 #include
2 #include
3
4 int main(void)
5 {
6 if (fseek(stdin, 10L, SEEK_SET) < 0)
7 perror("fseek");
8
9 fclose(stdin); // stop using this stream
10
11 if (fseek(stdin, 20L, SEEK_CUR) < 0) {
12
13 // specifically check errno to see what kind of
14 // error happened...this works on Linux, but your
15 // mileage may vary on other systems!
16
17 if (errno == EBADF) {
18 perror("fseek again, EBADF");
19 } else {
20 perror("fseek again");
21 }
22 }
23 }
And the output is: fseek: Illegal seek fseek again, EBADF: Bad file descriptor
See Also
feof(), ferror(), strerror() Chapter 41
Function Description memchr() Find the first occurrence of a character in memory. memcmp() Compare two regions of memory. memcpy() Copy a region of memory to another. memmove() Move a (potentially overlapping) region of memory. memset() Set a region of memory to a value. strcat() Concatenate (join) two strings together. strchr() Find the first occurrence of a character in a string. strcmp() Compare two strings. strcoll() Compare two strings accounting for locale. strcpy() Copy a string. strcspn() Find length of a string not consisting of a set of characters. strerror() Return a human-readable error message for a given code. strlen() Return the length of a string. strncat() Concatenate (join) two strings, length-limited. strncmp() Compare two strings, length-limited. strncpy() Copy two strings, length-limited. strpbrk() Search a string for one of a set of character. strrchr() Find the last occurrence of a character in a string. strspn() Find length of a string consisting of a set of characters. strstr() Find a substring in a string. strtok() Tokenize a string. strxfrm() Prepare a string for comparison as if by strcoll().
As has been mentioned earlier in the guide, a string in C is a sequence of bytes in memory, terminated by a NUL character (‘\0’). The NUL at the end is important, since it lets all these string functions (and printf() and puts() and everything else that deals with a string) know where the end of the string actually is. Fortunately, when you operate on a string using one of these many functions available to you, they add the NUL terminator on for you, so you actually rarely have to keep track of it yourself. (Sometimes you do, especially if you’re building a string from scratch a character at a time or something.) In this section you’ll find functions for pulling substrings out of strings, concatenating strings together, getting the length of a string, and so forth and so on.
41.1 memcpy(), memmove()
Copy bytes of memory from one location to another
295 Chapter 41.
Synopsis
#include
void *memcpy(void * restrict s1, const void * restrict s2, size_t n);
void *memmove(void *s1, const void *s2, size_t n);
Description These functions copy memory—as many bytes as you want! From source to destination! The main difference between the two is that memcpy() cannot safely copy overlapping memory regions, whereas memmove() can. On the one hand, I’m not sure why you’d want to ever use memcpy() instead of memmove(), but I’ll bet it’s possibly more performant. The parameters are in a particular order: destination first, then source. I remember this order because it behaves like an “=” assignment: the destination is on the left.
Return Value
Both functions return whatever you passed in for parameter s1 for your convenience.
Example
1 char s[100] = "Goats";
2 char t[100];
3
4 memcpy(t, s, 6); // Copy non-overlapping memory
5
6 memmove(s + 2, s, 6); // Copy overlapping memory
See Also
strcpy(), strncpy()
41.2 strcpy(), strncpy()
Copy a string
Synopsis
#include
char *strcpy(char *dest, char *src);
char *strncpy(char *dest, char *src, size_t n);
Description These functions copy a string from one address to another, stopping at the NUL terminator on the srcstring. Chapter 41.
strncpy() is just like strcpy(), except only the first n characters are actually copied. Beware that if you hit the limit, n before you get a NUL terminator on the src string, your dest string won’t be NUL-terminated. Beware! BEWARE! (If the src string has fewer than n characters, it works just like strcpy().) You can terminate the string yourself by sticking the '\0' in there yourself: char s[10]; char foo = "My hovercraft is full of eels."; // more than 10 chars
strncpy(s, foo, 9); // only copy 9 chars into positions 0-8 s[9] = '\0'; // position 9 gets the terminator
Return Value
Both functions return dest for your convenience, at no extra charge.
Example
1 char *src = "hockey hockey hockey hockey hockey hockey hockey hockey";
2 char dest[20];
3
4 int len;
5
6 strcpy(dest, "I like "); // dest is now "I like "
7
8 len = strlen(dest);
9
10 // tricky, but let's use some pointer arithmetic and math to append
11 // as much of src as possible onto the end of dest, -1 on the length to
12 // leave room for the terminator:
13 strncpy(dest+len, src, sizeof(dest)-len-1);
14
15 // remember that sizeof() returns the size of the array in bytes
16 // and a char is a byte:
17 dest[sizeof(dest)-1] = '\0'; // terminate
18
19 // dest is now: v null terminator
20 // I like hockey hocke
21 // 01234567890123456789012345
See Also
memcpy(), strcat(), strncat()
41.3 strcat(), strncat()
Concatenate two strings into a single string
Synopsis
#include
int strcat(const char *dest, const char *src); Chapter 41.
int strncat(const char *dest, const char *src, size_t n);
Description “Concatenate”, for those not in the know, means to “stick together”. These functions take two strings, and stick them together, storing the result in the first string. These functions don’t take the size of the first string into account when it does the concatenation. What this means in practical terms is that you can try to stick a 2 megabyte string into a 10 byte space. This will lead to unintended consequences, unless you intended to lead to unintended consequences, in which case it will lead to intended unintended consequences. Technical banter aside, your boss and/or professor will be irate. If you want to make sure you don’t overrun the first string, be sure to check the lengths of the strings first and use some highly technical subtraction to make sure things fit. You can actually only concatenate the first n characters of the second string by using strncat() and specifying the maximum number of characters to copy.
Return Value Both functions return a pointer to the destination string, like most of the string-oriented functions.
Example
1 char dest[30] = "Hello";
2 char *src = ", World!";
3 char numbers[] = "12345678";
4
5 printf("dest before strcat: \"%s\"\n", dest); // "Hello"
6
7 strcat(dest, src);
8 printf("dest after strcat: \"%s\"\n", dest); // "Hello, world!"
9
10 strncat(dest, numbers, 3); // strcat first 3 chars of numbers
11 printf("dest after strncat: \"%s\"\n", dest); // "Hello, world!123"
Notice I mixed and matched pointer and array notation there with src and numbers; this is just fine with string functions.
See Also
strlen()
41.4 strcmp(), strncmp(), memcmp()
Compare two strings or memory regions and return a difference
Synopsis
#include
int strcmp(const char *s1, const char *s2);
int strncmp(const char *s1, const char *s2, size_t n); Chapter 41.
int memcmp(const void *s1, const void *s2, size_t n);
Description All these functions compare chunks of bytes in memory. strcmp() and strncmp() operate on NUL-terminated strings, whereas memcmp() will compare the num- ber of bytes you specify, brazenly ignoring any NUL characters it finds along the way. strcmp() compares the entire string down to the end, while strncmp() only compares the first n char- acters of the strings. It’s a little funky what they return. Basically it’s a difference of the strings, so if the strings are the same, it’ll return zero (since the difference is zero). It’ll return non-zero if the strings differ; basically it will find the first mismatched character and return less-than zero if that character in s1 is less than the corresponding character in s2. It’ll return greater-than zero if that character in s1 is greater than that in s2. So if they return 0, the comparison was equal (i.e. the difference was 0.) These functions can be used as comparison functions for qsort() if you have an array of char*s you want to sort.
Return Value
Returns zero if the strings or memory are the same, less-than zero if the first different character in s1 is less than that in s2, or greater-than zero if the first difference character in s1 is greater than than in s2.
Example
1 char *s1 = "Muffin";
2 char *s2 = "Muffin Sandwich";
3 char *s3 = "Muffin";
4
5 strcmp("Biscuits", "Kittens"); // returns < 0 since 'B' < 'K'
6 strcmp("Kittens", "Biscuits"); // returns > 0 since 'K' > 'B'
7
8 if (strcmp(s1, s2) == 0)
9 printf("This won't get printed because the strings differ");
10
11 if (strcmp(s1, s3) == 0)
12 printf("This will print because s1 and s3 are the same");
13
14 // this is a little weird...but if the strings are the same, it'll
15 // return zero, which can also be thought of as "false". Not-false
16 // is "true", so (!strcmp()) will be true if the strings are the
17 // same. yes, it's odd, but you see this all the time in the wild
18 // so you might as well get used to it:
19
20 if (!strcmp(s1, s3))
21 printf("The strings are the same!");
22
23 if (!strncmp(s1, s2, 6))
24 printf("The first 6 characters of s1 and s2 are the same");
See Also
memcmp(), qsort() Chapter 41.
41.5 strcoll()
Compare two strings accounting for locale
Synopsis
#include
int strcoll(const char *s1, const char *s2);
Description
This is basically strcmp(), except that it handles accented characters better depending on the locale. For example, my strcmp() reports that the character “é” (with accent) is greater than “f”. But that’s hardly useful for alphabetizing. By setting the LC_COLLATE locale value (either by name or via LC_ALL), you can have strcoll() sort in a way that’s more meaningful by the current locale. For example, by having “é” appear sanely before “f”. It’s also a lot slower than strcmp() so use it only if you have to. See strxfrm() for a potential speedup.
Return Value
Like the other string comparison functions, strcoll() returns a negative value if s1 is less than s2, or a positive value if s1 is greater than s2. Or 0 if they are equal.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 setlocale(LC_ALL, "");
8
9 // If your source character set doesn't support "é" in a string
10 // you can replace it with `\u00e9`, the Unicode code point
11 // for "é".
12
13 printf("%d\n", strcmp("é", "f")); // Reports é > f, yuck.
14 printf("%d\n", strcoll("é", "f")); // Reports é < f, yay!
15 }
See Also
strcmp()
41.6 strxfrm()
Transform a string for comparing based on locale Chapter 41.
Synopsis
#include
size_t strxfrm(char * restrict s1, const char * restrict s2, size_t n);
Description This is a strange little function, so bear with me. Firstly, if you haven’t done so, get familiar with strcoll() because this is closely related to that. OK! Now that you’re back, you can think of strxfrm() as the first part of the strcoll() internals. Basically, strcoll() has to transform a string into a form that can be compared with strcmp(). And it does this with strxfrm() for both strings every time you call it. strxform() takes string s2 and transforms it (readies it for strcmp()) storing the result in s1. It writes no more than n bytes, protecting us from terrible buffer overflows. But hang on—there’s another mode! If you pass NULL for s1 and 0 for n, it will return the number of bytes that the transformed string would have used1. This is useful if you need to allocate some space to hold the transformed string before you strcmp() it against another. What I’m getting at, not to be too blunt, is that strcoll() is slow compared to strcmp(). It does a lot of extra work running strxfrm() on all its strings. In fact, we can see how it works by writing our own like this:
1 int my_strcoll(char *s1, char *s2)
2 {
3 // Use n = 0 to just get the lengths of the transformed strings
4 int len1 = strxfrm(NULL, s1, 0) + 1;
5 int len2 = strxfrm(NULL, s2, 0) + 1;
6
7 // Allocate enough room for each
8 char *d1 = malloc(len1);
9 char *d2 = malloc(len2);
10
11 // Transform the strings for comparison
12 strxfrm(d1, s1, len1);
13 strxfrm(d2, s2, len2);
14
15 // Compare the transformed strings
16 int result = strcmp(d1, d2);
17
18 // Free up the transformed strings
19 free(d2);
20 free(d1);
21
22 return result;
23 }
You see on lines 12, 13, and 16, above how we transform the two input strings and then call strcmp() on the result. So why do we have this function? Can’t we just call strcoll() and be done with it? The idea is that if you have one string that you’re going to be comparing against a whole lot of other ones, maybe you just want to transform that string one time, then use the faster strcmp() saving yourself a bunch of the work we had to do in the function, above.
1It always returns the number of bytes the transformed string took, but in this case because s1 was NULL, it doesn’t actually write a transformed string. Chapter 41.
We’ll do that in the example.
Return Value
Returns the number of bytes in the transformed sequence. If the value is greater than n, the results in s1 are meaningless.
Example
1 #include
2 #include
3 #include
4 #include
5
6 // Transform a string for comparison, returning a malloc'd
7 // result
8 char *get_xfrm_str(char *s)
9 {
10 int len = strxfrm(NULL, s, 0) + 1;
11 char *d = malloc(len);
12
13 strxfrm(d, s, len);
14
15 return d;
16 }
17
18 // Does half the work of a regular strcoll() because the second
19 // string arrives already transformed.
20 int half_strcoll(char *s1, char *s2_transformed)
21 {
22 char *s1_transformed = get_xfrm_str(s1);
23
24 int result = strcmp(s1_transformed, s2_transformed);
25
26 free(s1_transformed);
27
28 return result;
29 }
30
31 int main(void)
32 {
33 setlocale(LC_ALL, "");
34
35 // Pre-transform the string to compare against
36 char *s = get_xfrm_str("éfg");
37
38 // Repeatedly compare against "éfg"
39 printf("%d\n", half_strcoll("fgh", s)); // "fgh" > "éfg"
40 printf("%d\n", half_strcoll("àbc", s)); // "àbc" < "éfg"
41 printf("%d\n", half_strcoll("ĥij", s)); // "ĥij" > "éfg"
42
43 free(s);
44 }
See Also
strcoll() Chapter 41.
41.7 strchr(), strrchr(), memchr()
Find a character in a string
Synopsis
#include
char *strchr(char *str, int c);
char *strrchr(char *str, int c);
void *memchr(const void *s, int c, size_t n);
Description
The functions strchr() and strrchr find the first or last occurrence of a letter in a string, respectively. (The extra “r” in strrchr() stands for “reverse”–it looks starting at the end of the string and working backward.) Each function returns a pointer to the char in question, or NULL if the letter isn’t found in the string. memchr() is similar, except that instead of stopping on the first NUL character, it continues searching for however many bytes you specify. Quite straightforward. One thing you can do if you want to find the next occurrence of the letter after finding the first, is call the function again with the previous return value plus one. (Remember pointer arithmetic?) Or minus one if you’re looking in reverse. Don’t accidentally go off the end of the string!
Return Value
Returns a pointer to the occurrence of the letter in the string, or NULL if the letter is not found.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // "Hello, world!"
7 // ^ ^ ^
8 // ABC
9
10 char *str = "Hello, world!";
11 char *p;
12
13 p = strchr(str, ','); // p now points at position A
14 p = strrchr(str, 'o'); // p now points at position B
15
16 p = memchr(str, '!', 13); // p now points at position C
17
18 // repeatedly find all occurrences of the letter 'B'
19 str = "A BIG BROWN BAT BIT BEEJ";
20 Chapter 41.
21 for(p = strchr(str, 'B'); p != NULL; p = strchr(p + 1, 'B')) {
22 printf("Found a 'B' here: %s\n", p);
23 }
24
25 // output is:
26 //
27 // Found a 'B' here: BIG BROWN BAT BIT BEEJ
28 // Found a 'B' here: BROWN BAT BIT BEEJ
29 // Found a 'B' here: BAT BIT BEEJ
30 // Found a 'B' here: BIT BEEJ
31 // Found a 'B' here: BEEJ
32 }
41.8 strspn(), strcspn()
Return the length of a string consisting entirely of a set of characters, or of not a set of characters
Synopsis
#include
size_t strspn(char *str, const char *accept);
size_t strcspn(char *str, const char *reject);
Description
strspn() will tell you the length of a string consisting entirely of the set of characters in accept. That is, it starts walking down str until it finds a character that is not in the set (that is, a character that is not to be accepted), and returns the length of the string so far. strcspn() works much the same way, except that it walks down str until it finds a character in the reject set (that is, a character that is to be rejected.) It then returns the length of the string so far.
Return Value
The length of the string consisting of all characters in accept (for strspn()), or the length of the string consisting of all characters except reject (for strcspn()).
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 char str1[] = "a banana";
7 char str2[] = "the bolivian navy on maenuvers in the south pacific";
8 int n;
9
10 // how many letters in str1 until we reach something that's not a vowel?
11 n = strspn(str1, "aeiou");
12 printf("%d\n", n); // n == 1, just "a"
13
14 // how many letters in str1 until we reach something that's not a, b, Chapter 41.
15 // or space?
16 n = strspn(str1, "ab ");
17 printf("%d\n", n); // n == 4, "a ba"
18
19 // how many letters in str2 before we get a "y"?
20 n = strcspn(str2, "y");
21 printf("%d\n", n); // n = 16, "the bolivian nav"
22 }
See Also
strchr(), strrchr()
41.9 strpbrk()
Search a string for one of a set of characters
Synopsis
#include
char *strpbrk(const char *s1, const char *s2);
Description
This function searches string s1 for any of the characters that are found in string s2. It’s just like how strchr() searches for a specific character in a string, except it will match any of the characters found in s2. Think of the power!
Return Value
Returns a pointer to the first character matched in s1, or NULL if the string isn’t found.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // p points here after strpbrk
7 // v
8 char *s1 = "Hello, world!";
9 char *s2 = "dow!"; // Match any of these chars
10
11 char *p = strpbrk(s1, s2); // p points to the o
12
13 printf("%s\n", p); // "o, world!"
14 } Chapter 41.
See Also
strchr(), memchr()
41.10 strstr()
Find a string in another string
Synopsis
#include
char *strstr(const char *str, const char *substr);
Description Let’s say you have a big long string, and you want to find a word, or whatever substring strikes your fancy, inside the first string. Then strstr() is for you! It’ll return a pointer to the substr within the str!
Return Value
You get back a pointer to the occurrence of the substr inside the str, or NULL if the substring can’t be found.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 char *str = "The quick brown fox jumped over the lazy dogs.";
7 char *p;
8
9 p = strstr(str, "lazy");
10 printf("%s\n", p == NULL? "null": p); // "lazy dogs."
11
12 // p is NULL after this, since the string "wombat" isn't in str:
13 p = strstr(str, "wombat");
14 printf("%s\n", p == NULL? "null": p); // "null"
15 }
See Also
strchr(), strrchr(), strspn(), strcspn()
41.11 strtok()
Tokenize a string Chapter 41.
Synopsis
#include
char *strtok(char *str, const char *delim);
Description If you have a string that has a bunch of separators in it, and you want to break that string up into individual pieces, this function can do it for you. The usage is a little bit weird, but at least whenever you see the function in the wild, it’s consistently weird. Basically, the first time you call it, you pass the string, str that you want to break up in as the first argument. For each subsequent call to get more tokens out of the string, you pass NULL. This is a little weird, but strtok() remembers the string you originally passed in, and continues to strip tokens off for you. Note that it does this by actually putting a NUL terminator after the token, and then returning a pointer to the start of the token. So the original string you pass in is destroyed, as it were. If you need to preserve the string, be sure to pass a copy of it to strtok() so the original isn’t destroyed.
Return Value
A pointer to the next token. If you’re out of tokens, NULL is returned.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // break up the string into a series of space or
7 // punctuation-separated words
8 char str[] = "Where is my bacon, dude?";
9 char *token;
10
11 // Note that the following if-do-while construct is very very
12 // very very very common to see when using strtok().
13
14 // grab the first token (making sure there is a first token!)
15 if ((token = strtok(str, ".,?! ")) != NULL) {
16 do {
17 printf("Word: \"%s\"\n", token);
18
19 // now, the while continuation condition grabs the
20 // next token (by passing NULL as the first param)
21 // and continues if the token's not NULL:
22 } while ((token = strtok(NULL, ".,?! ")) != NULL);
23 }
24
25 // output is:
26 //
27 // Word: "Where"
28 // Word: "is"
29 // Word: "my"
30 // Word: "bacon"
31 // Word: "dude" Chapter 41.
32 //
33 }
See Also
strchr(), strrchr(), strspn(), strcspn()
41.12 memset()
Set a region of memory to a certain value
Synopsis
#include
void *memset(void *s, int c, size_t n);
Description
This function is what you use to set a region of memory to a particular value, namely c converted into unsigned char. The most common usage is to zero out an array or struct.
Return Value
memset() returns whatever you passed in as s for happy convenience.
Example
1 struct banana {
2 float ripeness;
3 char *peel_color;
4 int grams;
5 };
6
7 struct banana b;
8
9 memset(&b, 0, sizeof b);
10
11 b.ripeness == 0.0; // True
12 b.peel_color == NULL; // True
13 b.grams == 0; // True
See Also
memcpy(), memmove()
41.13 strerror()
Get a string version of an error number Chapter 41.
Synopsis
#include
char *strerror(int errnum);
Description
This function ties closely into perror() (which prints a human-readable error message corresponding to errno). But instead of printing, strerror() returns a pointer to the locale-specific error message string. So if you ever need that string back for some reason (e.g. you’re going to fprintf() it to a file or something), this function will give it to you. All you need to do is pass in errno as an argument. (Recall that errno gets set as an error status by a variety of functions.) You can actually pass in any integer for errnum you want. The function will return some message, even if the number doesn’t correspond to any known value for errno. The values of errno and the strings returned by strerror() are system-dependent.
Return Value A string error message corresponding to the given error number. You are not allowed to modify the returned string.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 FILE *fp = fopen("NONEXISTENT_FILE.TXT", "r");
8
9 if (fp == NULL) {
10 char *errmsg = strerror(errno);
11 printf("Error %d opening file: %s\n", errno, errmsg);
12 }
13 }
Output: Error 2 opening file: No such file or directory
See Also
perror()
41.14 strlen()
Returns the length of a string
Synopsis Chapter 41.
#include
size_t strlen(const char *s);
Description This function returns the length of the passed null-terminated string (not counting the NUL character at the end). It does this by walking down the string and counting the bytes until the NUL character, so it’s a little time consuming. If you have to get the length of the same string repeatedly, save it off in a variable somewhere.
Return Value Returns the number of bytes in the string. Note that this might be different than the number of characters in a multibyte string.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 char *s = "Hello, world!"; // 13 characters
7
8 // prints "The string is 13 characters long.":
9
10 printf("The string is %zu characters long.\n", strlen(s));
11 }
See Also Chapter 42
Many of the following functions have float and long double variants as described below (e.g. pow(), powf(), powl()). The float and long double variants are omitted from the following table to keep your eyeballs from melting out.
Function Description acos() Calculate the arc cosine of a number. acosh() Compute arc hyperbolic cosine. asin() Calculate the arc sine of a number. asinh() Compute arc hyperbolic sine. atan(), atan2() Calculate the arc tangent of a number. atanh() Compute the arc hyperbolic tangent. cbrt() Compute the cube root. ceil() Ceiling—return the next whole number not smaller than the given number. copysign() Copy the sign of one value into another. cos() Calculate the cosine of a number. cosh() Compute the hyperbolic cosine. erf() Compute the error function of the given value. erfc() Compute the complementary error function of a value. exp() Compute 푒 raised to a power. exp2() Compute 2 to a power. expm1() Compute 푒푥 − 1. fabs() Compute the absolute value. fdim() Return the positive difference between two numbers clamped at 0. floor() Compute the largest whole number not larger than the given value. fma() Floating (AKA “Fast”) multiply and add. fmax(), fmin() Return the maximum or minimum of two numbers. fmod() Compute the floating point remainder. fpclassify() Return the classification of a given floating point number. frexp() Break a number into its fraction part and exponent (as a power of 2). hypot() Compute the length of the hypotenuse of a triangle. ilogb() Return the exponent of a floating point number. isfinite() True if the number is not infinite or NaN. isgreater() True if one argument is greater than another. isgreatereequal() True if one argument is greater than or equal to another. isinf() True if the number is infinite. isless() True if one argument is less than another. islesseequal() True if one argument is less than or equal to another. islessgreater() Test if a floating point number is less than or greater than another. isnan() True if the number is Not-a-Number. isnormal() True if the number is normal. isunordered() Macro returns true if either floating point argument is NaN.
311 Chapter 42.
Function Description ldexp() Multiply a number by an integral power of 2. lgamma() Compute the natural logarithm of the absolute value of Γ(푥). log() Compute the natural logarithm. log10() Compute the log-base-10 of a number. log2() Compute the base-2 logarithm of a number. logb() Extract the exponent of a number given FLT_RADIX. logp1() Compute the natural logarithm of a number plus 1. lrint() Returns x rounded in the current rounding direction as an integer. lround(), llround() Round a number in the good old-fashioned way, returning an integer. modf() Extract the integral and fractional parts of a number. nan() Return NAN. nearbyint() Rounds a value in the current rounding direction. nextafter() Get the next (or previous) representable floating point value. nexttoward() Get the next (or previous) representable floating point value. pow() Compute a value raised to a power. remainder() Compute the remainder IEC 60559-style. remquo() Compute the remainder and (some of the) quotient. rint() Rounds a value in the current rounding direction. round() Round a number in the good old-fashioned way. scalbn(), scalbln() Efficiently compute 푥 × 푟푛, where 푟 is FLT_RADIX. signbit() Return the sign of a number. sin() Calculate the sine of a number. sqrt() Calculate the square root of a number. tan() Calculate the tangent of a number. tanh() Compute the hyperbolic tangent. tgamma() Compute the gamma function, Γ(푥). trunc() Truncate the fractional part off a floating point value.
It’s your favorite subject: Mathematics! Hello, I’m Doctor Math, and I’ll be making math FUN and EASY! [vomiting sounds] Ok, I know math isn’t the grandest thing for some of you out there, but these are merely functions that quickly and easily do math you either know, want, or just don’t care about. That pretty much covers it.
42.1 Math Function Idioms
Many of these math functions exist in three forms, each corresponding to the argument and/or return types the function uses, float, double, or long double. The alternate form for float is made by appending f to the end of the function name. The alternate form for long double is made by appending l to the end of the function name. For example, the pow() function, which computes 푥푦, exists in these forms: double pow(double x, double y); // double float powf(float x, float y); // float long double powl(long double x, long double y); // long double
Remember that parameters are given values as if you assigned into them. So if you pass a double to powf(), it’ll choose the closest float it can to hold the double. If the double doesn’t fit, undefined behavior happens.
42.2 Math Types
We have two exciting new types in
• float_t • double_t The float_t type is at least as accurate as a float, and the double_t type is at least as accurate as a double. The idea with these types is they can represent the most efficient way of storing numbers for maximum speed. Their actual types vary by implementation, but can be determined by the value of the FLT_EVAL_METHOD macro.
FLT_EVAL_METHOD float_t type double_t type 0 float double 1 double double 2 long double long double Other Implementation-defined Implementation-defined
For all defined values of FLT_EVAL_METHOD, float_t is the least-precise type used for all floating cal- culations.
42.3 Math Macros
There are actually a number of these defined, but we’ll cover most of them in their relevant reference sections, below. But here are a couple: NAN represents Not-A-Number. Defined in
42.4 Math Errors
As we know, nothing can ever go wrong with math… except everything! So there are just a couple errors that might occur when using some of these functions. • Range errors mean that some result is beyond what can be stored in the result type. • Domain errors mean that you’ve passed in an argument that doesn’t have a defined result for this function. • Pole errors mean that the limit of the function as 푥 approaches the given argument is infinite. • Overflow errors are when the result is really large, but can’t be stored without incurring large roundoff error. • Underflow errors are like overflow errors, except with very small numbers. Now, the C math library can do a couple things when these errors occur: • Set errno to some value, or… • Raise a floating point exception. Your system might vary on what happens. You can check it by looking at the value of the variable math_errhandling. It will be equivalent to one of the following1:
1Though the system defines MATH_ERRNO as 1 and MATH_ERREXCEPT as 2, it’s best to always use their symbolic names. Just in case. Chapter 42.
math_errhandling Description MATH_ERRNO The system uses errno for math errors. MATH_ERREXCEPT The system uses exceptions for math errors. MATH_ERRNO | MATH_ERREXCEPT The system does both! (That’s a bitwise-OR!)
You are not allowed to change math_errhandling. For a fuller description on how exceptions work and their meanings, see the
42.5 Math Pragmas
In case you don’t remember, you can brush up on pragmas back in the C Preprocessor section. But in a nutshell, they offer various ways to control the compiler’s behavior. In this case, we have a pragma FP_CONTRACT that can be turned off and on. What does it mean? First of all, keep in mind that any operation in an expression can cause rounding error. So each step of the expression can introduce more rounding error. But what if the compiler knows a double secret way of taking the expression you wrote and converting it to a single instruction that reduced the number of steps such that the intermediate rounding error didn’t occur? Could it use it? I mean, the results would be different than if you let the rounding error settle each step of the way… Because the results would be different, you can tell the compiler if you want to allow it to do this or not. If you want to allow it: #pragma STDC FP_CONTRACT ON and to disallow it: #pragma STDC FP_CONTRACT OFF
If you do this at global scope, it stays at whatever state you set it to until you change it. If you do it at block scope, it reverts to the value outside the block when the block ends. The initial value of the FP_CONTRACT pragma varies from system to system.
42.6 fpclassify()
Return the classification of a given floating point number.
Synopsis
#include
Description What kind of entity does this floating point number represent? What are the options? We’re used to floating point numbers being regular old things like 3.14 or 3490.0001. Chapter 42.
But floating point numbers can also represent things like infinity. Or Not-A-Number (NAN). This function will let you know which type of floating point number the argument is. This is a macro, so you can use it with float, double, long double or anything similar.
Return Value Returns one of these macros depending on the argument’s classification:
Classification Description FP_INFINITE Number is infinite. FP_NAN Number is Not-A-Number (NAN). FP_NORMAL Just a regular number. FP_SUBNORMAL Number is a sub-normal number. FP_ZERO Number is zero.
A discussion of subnormal numbers is beyond the scope of the guide, and is something that most devs go their whole lives without dealing with. In a nutshell, it’s a way to represent really small numbers that might normally round down to zero. If you want to know more, see the Wikipedia page on denormal numbers2.
Example Print various number classifications.
1 #include
2 #include
3
4 const char *get_classification(double n)
5 {
6 switch (fpclassify(n)) {
7 case FP_INFINITE: return "infinity";
8 case FP_NAN: return "not a number";
9 case FP_NORMAL: return "normal";
10 case FP_SUBNORMAL: return "subnormal";
11 case FP_ZERO: return "zero";
12 }
13
14 return "unknown";
15 }
16
17 int main(void)
18 {
19 printf(" 1.23: %s\n", get_classification(1.23));
20 printf(" 0.0: %s\n", get_classification(0.0));
21 printf("sqrt(-1): %s\n", get_classification(sqrt(-1)));
22 printf("1/tan(0): %s\n", get_classification(1/tan(0)));
23 printf(" 1e-310: %s\n", get_classification(1e-310)); // very small!
24 }
Output3: 1.23: normal 0.0: zero sqrt(-1): not a number
2https://en.wikipedia.org/wiki/Denormal_number 3This is on my system. Some systems will have different points at which numbers become subnormal, or they might not support subnormal values at all. Chapter 42.
1/tan(0): infinity 1e-310: subnormal
See Also
isfinite(), isinf(), isnan(), isnormal(), signbit()
42.7 isfinite(), isinf(), isnan(), isnormal()
Return true if a number matches a classification.
Synopsis
#include
int isfinite(any_floating_type x);
int isinf(any_floating_type x);
int isnan(any_floating_type x);
int isnormal(any_floating_type x);
Description
These are helper macros to fpclassify(). Bring macros, they work on any floating point type.
Macro Description isfinite() True if the number is not infinite or NaN. isinf() True if the number is infinite. isnan() True if the number is Not-a-Number. isnormal() True if the number is normal.
For more superficial discussion on normal and subnormal numbers, see fpclassify().
Return Value Returns non-zero for true, and zero for false.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 printf(" isfinite(1.23): %d\n", isfinite(1.23)); // 1
7 printf(" isinf(1/tan(0)): %d\n", isinf(1/tan(0))); // 1
8 printf(" isnan(sqrt(-1)): %d\n", isnan(sqrt(-1))); // 1
9 printf("isnormal(1e-310): %d\n", isnormal(1e-310)); // 0
10 } Chapter 42.
See Also
fpclassify(), signbit(),
42.8 signbit()
Return the sign of a number.
Synopsis
#include
int signbit(any_floating_type x);
Description This macro takes any floating point number and returns a value indicating the sign of the number, positive or negative.
Return Value
Returns 1 if the sign is negative, otherwise 0.
Example
1 printf("%d\n", signbit(3490.0)); // 0
2 printf("%d\n", signbit(-37.0)); // 1
See Also
fpclassify(), isfinite(), isinf(), isnan(), isnormal(), copysign()
42.9 acos(), acosf(), acosl()
Calculate the arc cosine of a number.
Synopsis
#include
double acos(double x); float acosf(float x); long double acosl(long double x);
Description
Calculates the arc cosine of a number in radians. (That is, the value whose cosine is x.) The number must be in the range -1.0 to 1.0. For those of you who don’t remember, radians are another way of measuring an angle, just like degrees. To convert from degrees to radians or the other way around, use the following code: Chapter 42.
pi = 3.14159265358979; degrees = radians * 180 / pi; radians = degrees * pi / 180;
Return Value
Returns the arc cosine of x, unless x is out of range. In that case, errno will be set to EDOM and the return value will be NaN. The variants return different types.
Example
1 double acosx;
2 long double ldacosx;
3
4 acosx = acos(0.2);
5 ldacosx = acosl((long double)0.3);
See Also
asin(), atan(), atan2(), cos()
42.10 asin(), asinf(), asinl()
Calculate the arc sine of a number.
Synopsis
#include
double asin(double x); float asinf(float x); long double asinl(long double x);
Description
Calculates the arc sine of a number in radians. (That is, the value whose sine is x.) The number must be in the range -1.0 to 1.0. For those of you who don’t remember, radians are another way of measuring an angle, just like degrees. To convert from degrees to radians or the other way around, use the following code: pi = 3.14159265358979; degrees = radians * 180 / pi; radians = degrees * pi / 180;
Return Value
Returns the arc sine of x, unless x is out of range. In that case, errno will be set to EDOM and the return value will be NaN. The variants return different types.
Example
1 double asinx;
2 long double ldasinx; Chapter 42.
3
4 asinx = asin(0.2);
5 ldasinx = asinl((long double)0.3);
See Also
acos(), atan(), atan2(), sin()
42.11 atan(), atanf(), atanl(), atan2(), atan2f(), atan2l()
Calculate the arc tangent of a number.
Synopsis
#include
double atan(double x); float atanf(float x); long double atanl(long double x);
double atan2(double y, double x); float atan2f(float y, float x); long double atan2l(long double y, long double x);
Description
Calculates the arc tangent of a number in radians. (That is, the value whose tangent is x.) The atan2() variants are pretty much the same as using atan() with y/x as the argument…except that atan2() will use those values to determine the correct quadrant of the result. For those of you who don’t remember, radians are another way of measuring an angle, just like degrees. To convert from degrees to radians or the other way around, use the following code: pi = 3.14159265358979; degrees = radians * 180 / pi; radians = degrees * pi / 180;
Return Value
The atan() functions return the arc tangent of x, which will be between PI/2 and -PI/2. The atan2() functions return an angle between PI and -PI.
Example
1 double atanx;
2 long double ldatanx;
3
4 atanx = atan(0.2);
5 ldatanx = atanl((long double)0.3);
6
7 atanx = atan2(0.2);
8 ldatanx = atan2l((long double)0.3); Chapter 42.
See Also
tan(), asin(), atan()
42.12 cos(), cosf(), cosl()
Calculate the cosine of a number.
Synopsis
#include
double cos(double x) float cosf(float x) long double cosl(long double x)
Description
Calculates the cosine of the value x, where x is in radians. For those of you who don’t remember, radians are another way of measuring an angle, just like degrees. To convert from degrees to radians or the other way around, use the following code: pi = 3.14159265358979; degrees = radians * 180 / pi; radians = degrees * pi / 180;
Return Value
Returns the cosine of x. The variants return different types.
Example
1 double sinx;
2 long double ldsinx;
3
4 sinx = sin(3490.0); // round and round we go!
5 ldsinx = sinl((long double)3.490);
See Also
sin(), tan(), acos()
42.13 sin(), sinf(), sinl()
Calculate the sine of a number.
Synopsis
#include
double sin(double x); Chapter 42.
float sinf(float x); long double sinl(long double x);
Description
Calculates the sine of the value x, where x is in radians. For those of you who don’t remember, radians are another way of measuring an angle, just like degrees. To convert from degrees to radians or the other way around, use the following code: pi = 3.14159265358979; degrees = radians * 180 / pi; radians = degrees * pi / 180;
Return Value
Returns the sine of x. The variants return different types.
Example
1 double sinx;
2 long double ldsinx;
3
4 sinx = sin(3490.0); // round and round we go!
5 ldsinx = sinl((long double)3.490);
See Also
cos(), tan(), asin()
42.14 tan(), tanf(), tanl()
Calculate the tangent of a number.
Synopsis
#include
double tan(double x) float tanf(float x) long double tanl(long double x)
Description
Calculates the tangent of the value x, where x is in radians. For those of you who don’t remember, radians are another way of measuring an angle, just like degrees. To convert from degrees to radians or the other way around, use the following code: pi = 3.14159265358979; degrees = radians * 180 / pi; radians = degrees * pi / 180;
Return Value
Returns the tangent of x. The variants return different types. Chapter 42.
Example
1 double tanx;
2 long double ldtanx;
3
4 tanx = tan(3490.0); // round and round we go!
5 ldtanx = tanl((long double)3.490);
See Also
sin(), cos(), atan(), atan2()
42.15 acosh(), acoshf(), acoshl()
Compute arc hyperbolic cosine.
Synopsis
#include
double acosh(double x);
float acoshf(float x);
long double acoshl(long double x);
Description Trig lovers can rejoice! C has arc hyperbolic cosine! These functions return the nonnegative acosh of x, which must be greater than or equal to 1.
Return Value Returns the arc hyperbolic cosince in the range [0, +∞].
Example
1 printf("acosh 1.8 = %f\n", acosh(1.8)); // 1.192911
See Also
asinh()
42.16 asinh(), asinhf(), asinhl()
Compute arc hyperbolic sine.
Synopsis Chapter 42.
#include
double asinh(double x);
float asinhf(float x);
long double asinhl(long double x);
Description Trig lovers can rejoice! C has arc hyperbolic sine! These functions return the asinh of x.
Return Value Returns the arc hyperbolic sine.
Example
1 printf("acosh 1.8 = %f\n", asinh(1.8)); // 1.350441
See Also
acosh()
42.17 atanh(), atanhf(), atanhl()
Compute the arc hyperbolic tangent.
Synopsis
#include
double atanh(double x);
float atanhf(float x);
long double atanhl(long double x);
Description
These functions compute the arc hyperbolic tangent of x, which must be in the range [−1, +1]. Passing exactly −1 or +1 might result in a pole error.
Return Value
Returns the arc hyperbolic tangent of x.
Example
1 printf("atanh 0.5 = %f\n", atanh(0.5)); // 0.549306 Chapter 42.
See Also
acosh(), asinh()
42.18 cosh(), coshf(), coshl()
Compute the hyperbolic cosine.
Synopsis
#include
double cosh(double x);
float coshf(float x);
long double coshl(long double x);
Description
These functions predictably compute the hyperbolic cosine of x. A range error might occur if x is too large.
Return Value
Returns the hyperbolic cosine of x.
Example
1 printf("cosh 0.5 = %f\n", cosh(0.5)); // 1.127626
See Also
sinh(), tanh()
42.19 sinh(), sinhf(), sinhl()
Compute the hyperbolic sine.
Synopsis
#include
double sinh(double x);
float sinhf(float x);
long double sinhl(long double x);
Description
These functions predictably compute the hyperbolic sine of x. A range error might occur if x is too large. Chapter 42.
Return Value
Returns the hyperbolic sine of x.
Example
1 printf("sinh 0.5 = %f\n", sinh(0.5)); // 0.521095
See Also
sinh(), tanh()
42.20 tanh(), tanhf(), tanhl()
Compute the hyperbolic tangent.
Synopsis
#include
double tanh(double x);
float tanhf(float x);
long double tanhl(long double x);
Description
These functions predictably compute the hyperbolic tangent of x. Mercifully, this is the last trig-related man page I’m going to write.
Return Value
Returns the hyperbolic tangent of x.
Example
1 printf("tanh 0.5 = %f\n", tanh(0.5)); // 0.462117
See Also
cosh(), sinh()
42.21 exp(), expf(), expl()
Compute 푒 raised to a power.
Synopsis Chapter 42.
#include
double exp(double x);
float expf(float x);
long double expl(long double x);
Description Compute 푒푥 where 푒 is Euler’s number4. The number 푒 is named after Leonard Euler, born April 15, 1707, who is responsible, among other things, for making this reference page longer than it needed to be.
Return Value Returns 푒푥.
Example
1 printf("exp(1) = %f\n", exp(1)); // 2.718282
2 printf("exp(2) = %f\n", exp(2)); // 7.389056
See Also
exp2(), expm1(), pow(), log()
42.22 exp2(), exp2f(), exp2l()
Compute 2 to a power.
Synopsis
#include
double exp2(double x);
float exp2f(float x);
long double exp2l(long double x);
Description These functions raise 2 to a power. Very exciting, since computers are all about twos-to-powers! These are likely to be faster than using pow() to do the same thing. They support fractional exponents, as well. A range error occurs if x is too large.
Return Value
exp2() returns 2푥.
4https://en.wikipedia.org/wiki/E_(mathematical_constant) Chapter 42.
Example
1 printf("2^3 = %f\n", exp2(3)); // 2^3 = 8.000000
2 printf("2^8 = %f\n", exp2(8)); // 2^8 = 256.000000
3 printf("2^0.5 = %f\n", exp2(0.5)); // 2^0.5 = 1.414214
See Also
exp(), pow()
42.23 expm1(), expm1f(), expm1l()
Compute 푒푥 − 1.
Synopsis
#include
double expm1(double x);
float expm1f(float x);
long double expm1l(long double x);
Description
This is just like exp() except—plot twist!–it computes that result minus one. For more discussion about what 푒 is, see the exp() man page. If x is giant, a range error might occur. For small values of x near zero, expm1(x) might be more accurate than computing exp(x)-1.
Return Value Returns 푒푥 − 1.
Example
1 printf("%f\n", expm1(2.34)); // 9.381237
See Also
exp()
42.24 frexp(), frexpf(), frexpl()
Break a number into its fraction part and exponent (as a power of 2).
Synopsis Chapter 42.
#include
double frexp(double value, int *exp);
float frexpf(float value, int *exp);
long double frexpl(long double value, int *exp);
Description If you have a floating point number, you can break it into its fractional part and exponent part (as a power of 2). For example, if you have the number 1234.56, this can be represented as a multiple of a power of 2 like so: 1234.56 = 0.6028125 × 211 And you can use this function to get the 0.6028125 and 11 parts of that equation. As for why, I have a simple answer: I don’t know. I can’t find a use. K&R2 and everyone else I can find just says how to use it, but not why you might want to. The C99 Rationale document says: The functions frexp, ldexp, and modf are primitives used by the remainder of the library. There was some sentiment for dropping them for the same reasons that ecvt, fcvt, and gcvt were dropped, but their adherents rescued them for general use. Their use is problematic: on non-binary architectures, ldexp may lose precision and frexp may be inefficient. So there you have it. If you need it.
Return Value
frexp() returns the fractional part of value in the range 0.5 (inclusive) to 1 (exclusive), or 0. And it stores the exponent power-of-2 in the variable pointed to by exp. If you pass in zero, the return value and the variable exp points to are both zero.
Example
1 double frac;
2 int expt;
3
4 frac = frexp(1234.56, &expt);
5 printf("1234.56 = %.7f x 2^%d\n", frac, expt);
Output: 1234.56 = 0.6028125 x 2^11
See Also
ldexp(), ilogb(), modf()
42.25 ilogb(), ilogbf(), ilogbl()
Return the exponent of a floating point number. Chapter 42.
Synopsis
#include
int ilogb(double x);
int ilogbf(float x);
int ilogbl(long double x);
Description This gives you the exponent of the given number… it’s a little weird, because the exponent depends on the value of FLT_RADIX. Now, this is very often 2—but no guarantees!
It actually returns log푟 |푥| where 푟 is FLT_RADIX. Domain or range errors might occur for invalid values of x, or for return values that are outside the range of the return type.
Return Value
The exponent of the absolute value of the given number, depending on FLT_RADIX.
Specifically log푟 |푥| where 푟 is FLT_RADIX. If you pass in 0, it’ll return FP_ILOGB0. If you pass in infinity, it’ll return INT_MAX. If you pass in NaN, it’ll return FP_ILOGBNAN. The spec goes on to say that the value of FP_ILOGB0 will be either INT_MIN or -INT_MAX. And the value of FP_ILOGBNAN shall be either INT_MAX or INT_MIN, if that’s useful in any way.
Example
1 printf("%d\n", ilogb(257)); // 8
2 printf("%d\n", ilogb(256)); // 8
3 printf("%d\n", ilogb(255)); // 7
See Also
frexp(), logb()
42.26 ldexp(), ldexpf(), ldexpl()
Multiply a number by an integral power of 2.
Synopsis
#include
double ldexp(double x, int exp);
float ldexpf(float x, int exp);
long double ldexpl(long double x, int exp); Chapter 42.
Description
These functions multiply the given number x by 2 raised to the exp power.
Return Value Returns 푥 × 2푒푥푝.
Example
1 printf("1 x 2^10 = %f\n", ldexp(1, 10));
2 printf("5.67 x 2^7 = %f\n", ldexp(5.67, 7));
Output: 1 x 2^10 = 1024.000000 5.67 x 2^7 = 725.760000
See Also
exp()
42.27 log(), logf(), logl()
Compute the natural logarithm.
Synopsis
#include
double log(double x);
float logf(float x);
long double logl(long double x);
Description Natural logarithms! And there was much rejoycing.
These compute the base-푒 logarithm of a number, log푒 푥, ln 푥. In other words, for a given 푥, solves 푥 = 푒푦 for 푦.
Return Value
The base-푒 logarithm of the given value, log푒 푥, ln 푥.
Example
1 const double e = 2.718281828459045;
2
3 printf("%f\n", log(3490.2)); // 8.157714
4 printf("%f\n", log(e)); // 1.000000 Chapter 42.
See Also
exp(), log10(), logp1()
42.28 log10(), log10f(), log10l()
Compute the log-base-10 of a number.
Synopsis
#include
double log10(double x);
float log10f(float x);
long double log10l(long double x);
Description Just when you thought you might have to use Laws of Logarithms to compute this, here’s a function coming out of the blue to save you.
These compute the base-10 logarithm of a number, log10 푥. In other words, for a given 푥, solves 푥 = 10푦 for 푦.
Return Value
Returns the log base-10 of x, log10 푥.
Example
1 printf("%f\n", log10(3490.2)); // 3.542850
2 printf("%f\n", log10(10)); // 1.000000
See Also
pow(), log()
42.29 logp1(), logp1f(), logp1l()
Compute the natural logarithm of a number plus 1.
Synopsis
#include
double log1p(double x);
float log1pf(float x);
long double log1pl(long double x); Chapter 42.
Description
This computes log푒(1 + 푥), ln(1 + 푥). This works just like calling: log(1 + x)
except it could be more accurate for small values of x. So if your x is small magnitude, use this.
Return Value
Returns log푒(1 + 푥), ln(1 + 푥).
Example
Compute some big and small logarithm values to see the difference between log1p() and log():
1 printf("Big log1p() : %.*Lf\n", LDBL_DECIMAL_DIG-1, log1pl(9));
2 printf("Big log() : %.*Lf\n", LDBL_DECIMAL_DIG-1, logl(1 + 9));
3
4 printf("Small log1p(): %.*Lf\n", LDBL_DECIMAL_DIG-1, log1pl(0.01));
5 printf("Small log() : %.*Lf\n", LDBL_DECIMAL_DIG-1, logl(1 + 0.01));
Output on my system: Big log1p() : 2.30258509299404568403 Big log() : 2.30258509299404568403 Small log1p(): 0.00995033085316808305 Small log() : 0.00995033085316809164
See Also
log()
42.30 log2(), log2f(), log2l()
Compute the base-2 logarithm of a number.
Synopsis
#include
double log2(double x);
float log2f(float x);
long double log2l(long double x);
Description Wow! Were you thinking we were done with the logarithm functions? We’re only getting started! 푦 This one computes log2 푥. That is, computes 푦 that satisfies 푥 = 2 . Love me those powers of 2! Chapter 42.
Return Value
Returns the base-2 logarithm of the given value, log2 푥.
Example
1 printf("%f\n", log2(3490.2)); // 11.769094
2 printf("%f\n", log2(256)); // 8.000000
See Also
log()
42.31 logb(), logbf(), logbl()
Extract the exponent of a number given FLT_RADIX.
Synopsis
#include
double logb(double x);
float logbf(float x);
long double logbl(long double x);
Description
This function returns the whole number portion of the exponent of the number with radix FLT_RADIX,
namely the whole number portion log푟 |푥| where 푟 is FLT_RADIX. Fractional numbers are truncated. If the number is subnormal5, logb() treats it as if it were normalized. If x is 0, there could be a domain error or pole error.
Return Value
This function returns the whole number portion of log푟 |푥| where 푟 is FLT_RADIX.
Example
1 printf("FLT_RADIX = %d\n", FLT_RADIX);
2 printf("%f\n", logb(3490.2));
3 printf("%f\n", logb(256));
Output: FLT_RADIX = 2 11.000000 8.000000
5https://en.wikipedia.org/wiki/Denormal_number Chapter 42.
See Also
ilogb()
42.32 modf(), modff(), modfl()
Extract the integral and fractional parts of a number.
Synopsis
#include
double modf(double value, double *iptr);
float modff(float value, float *iptr);
long double modfl(long double value, long double *iptr);
Description
If you have a floating point number, like 123.456, this function will extract the integral part (123.0) and the fractional part (0.456). It’s total coincidence that this is exactly the plot for the latest Jason Statham action spectacular. Both the integral part and fractional parts keep the sign of the passed in value. The integral part is stored in the address pointed to by iptr. See the note in frexp() regarding why this is in the library.
Return Value These functions return the fractional part of the number. The integral part is stored in the address pointed to by iptr. Both the integral and fractional parts preserve the sign of the passed-in value.
Example
1 #include
2 #include
3
4 void print_parts(double x)
5 {
6 double i, f;
7
8 f = modf(x, &i);
9
10 printf("Entire number : %f\n", x);
11 printf("Integral part : %f\n", i);
12 printf("Fractional part: %f\n\n", f);
13 }
14
15 int main(void)
16 {
17 print_parts(123.456);
18 print_parts(-123.456);
19 } Chapter 42.
Output: Entire number : 123.456000 Integral part : 123.000000 Fractional part: 0.456000
Entire number : -123.456000 Integral part : -123.000000 Fractional part: -0.456000
See Also
frexp()
42.33 scalbn(), scalbnf(), scalbnl() scalbln(), scalblnf(), scalblnl()
Efficiently compute 푥 × 푟푛, where 푟 is FLT_RADIX.
Synopsis
#include
double scalbn(double x, int n);
float scalbnf(float x, int n);
long double scalbnl(long double x, int n);
double scalbln(double x, long int n);
float scalblnf(float x, long int n);
long double scalblnl(long double x, long int n);
Description
These functions efficiently compute 푥 × 푟푛, where 푟 is FLT_RADIX. If FLT_RADIX happens to be 2 (no guarantees!), then this works like exp2(). The name of this function should have an obvious meaning to you. Clearly they all start with the prefix “scalb” which means… …OK, I confess! I have no idea what it means. My searches are futile! But let’s look at the suffixes:
Suffix Meaning n scalbn()—exponent n is an int nf scalbnf()—float version of scalbn() nl scalbnl()—long double version of scalbn() ln scalbln()—exponent n is a long int lnf scalblnf()—float version of scalbln() lnl scalblnl()—long double version of scalbln() Chapter 42.
So while I’m still in the dark about “scalb”, at least I have that part down. A range error might occur for large values.
Return Value
Returns 푥 × 푟푛, where 푟 is FLT_RADIX.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 printf("FLT_RADIX = %d\n\n", FLT_RADIX);
8 printf("scalbn(3, 8) = %f\n", scalbn(2, 8));
9 printf("scalbn(10.2, 20.7) = %f\n", scalbn(10.2, 20.7));
10 }
Output on my system: FLT_RADIX = 2
scalbn(3, 8) = 512.000000 scalbn(10.2, 20.7) = 10695475.200000
See Also
exp2(), pow()
42.34 cbrt(), cbrtf(), cbrtl()
Compute the cube root.
Synopsis
#include
double cbrt(double x);
float cbrtf(float x);
long double cbrtl(long double x);
Description √ Computes the cube root of x, 푥1/3, 3 푥.
Return Value √ Returns the cube root of x, 푥1/3, 3 푥. Chapter 42.
Example
1 printf("cbrt(1729.03) = %f\n", cbrt(1729.03));
Output: cbrt(1729.03) = 12.002384
See Also
sqrt(), pow()
42.35 fabs(), fabsf(), fabsl()
Compute the absolute value.
Synopsis
#include
double fabs(double x);
float fabsf(float x);
long double fabsl(long double x);
Description
These functions straightforwardly return the absolute value of x, that is |푥|. If you’re rusty on your absolute values, all it means is that the result will be positive, even if x is negative. It’s just strips negative signs off.
Return Value
Returns the absolute value of x, |푥|.
Example
1 printf("fabs(3490.0) = %f\n", fabs(3490.0)); // 3490.000000
2 printf("fabs(-3490.0) = %f\n", fabs(3490.0)); // 3490.000000
See Also
abs(), copysign(), imaxabs()
42.36 hypot(), hypotf(), hypotl()
Compute the length of the hypotenuse of a triangle.
Synopsis Chapter 42.
#include
double hypot(double x, double y);
float hypotf(float x, float y);
long double hypotl(long double x, long double y);
Description Pythagorean Theorem6 fans rejoice! This is the function you’ve been waiting for! If you know the lengths of the two sides of a right triangle, x and y, you can compute the length of the hypotenuse (the longest, diagonal side) with this function. In particular, it computes the square root of the sum of the squares of the sides: √푥2 + 푦2.
Return Value
Returns the lenght of the hypotenuse of a right triangle with side lengths x and y: √푥2 + 푦2.
Example
1 printf("%f\n", hypot(3, 4)); // 5.000000
See Also
sqrt()
42.37 pow(), powf(), powl()
Compute a value raised to a power.
Synopsis
#include
double pow(double x, double y);
float powf(float x, float y);
long double powl(long double x, long double y);
Description
Computes x raised to the yth power: 푥푦. These arguments can be fractional.
Return Value
Returns x raised to the yth power: 푥푦. A domain error can occur if: 6https://en.wikipedia.org/wiki/Pythagorean_theorem Chapter 42.
• x is a finite negative number and y is a finite non-integer • x is zero and y is zero. A domain error or pole error can occur if x is zero and y is negative. A range error can occur for large values.
Example
1 printf("%f\n", pow(3, 4)); // 3^4 = 81.000000
2 printf("%f\n", pow(2, 0.5)); // sqrt 2 = 1.414214
See Also
exp(), exp2(), sqrt(), cbrt()
42.38 sqrt()
Calculate the square root of a number.
Synopsis
#include
double sqrt(double x);
float sqrtf(float x);
long double sqrtl(long double x);
Description √ Computes the square root of a number: 푥. To those of you who don’t know what a square root is, I’m not going to explain. Suffice it to say, the square root of a number delivers a value that when squared (multiplied by itself) results in the original number. Ok, fine—I did explain it after all, but only because I wanted to show off. It’s not like I’m giving you examples or anything, such as the square root of nine is three, because when you multiply three by three you get nine, or anything like that. No examples. I hate examples! And I suppose you wanted some actual practical information here as well. You can see the usual trio of functions here—they all compute square root, but they take different types as arguments. Pretty straight- forward, really. A domain error occurs if x is negative.
Return Value √ Returns (and I know this must be something of a surprise to you) the square root of x: 푥.
Example
1 // example usage of sqrt()
2
3 float something = 10;
4
5 double x1 = 8.2, y1 = -5.4; Chapter 42.
6 double x2 = 3.8, y2 = 34.9;
7 double dx, dy;
8
9 printf("square root of 10 is %.2f\n", sqrtf(something));
10
11 dx = x2 - x1;
12 dy = y2 - y1;
13 printf("distance between points (x1, y1) and (x2, y2): %.2f\n",
14 sqrt(dx*dx + dy*dy));
And the output is: square root of 10 is 3.16 distance between points (x1, y1) and (x2, y2): 40.54
See Also
hypot(), pow()
42.39 erf(), erff(), erfl()
Compute the error function of the given value.
Synopsis
#include
double erfc(double x);
float erfcf(float x);
long double erfcl(long double x);
Description These functions compute the error function7 of a value.
Return Value
Returns the error function of x: 푥 2 2 √ ∫ 푒−푡 푑푡 휋 0
Example
1 for (float i = -2; i <= 2; i += 0.5)
2 printf("% .1f: %f\n", i, erf(i));
Output: -2.0: -0.995322 -1.5: -0.966105 -1.0: -0.842701 -0.5: -0.520500
7https://en.wikipedia.org/wiki/Error_function Chapter 42.
0.0: 0.000000 0.5: 0.520500 1.0: 0.842701 1.5: 0.966105 2.0: 0.995322
See Also
erfc()
42.40 erfc(), erfcf(), erfcl()
Compute the complementary error function of a value.
Synopsis
#include
double erfc(double x);
float erfcf(float x);
long double erfcl(long double x);
Description These functions compute the complementary error function8 of a value. This is the same as: 1 - erf(x)
A range error can occur if x is too large.
Return Value
Returns 1 - erf(x), namely: ∞ 2 2 √ ∫ 푒−푡 푑푡 휋 푥
Example
1 for (float i = -2; i <= 2; i += 0.5)
2 printf("% .1f: %f\n", i, erfc(i));
Output: -2.0: 1.995322 -1.5: 1.966105 -1.0: 1.842701 -0.5: 1.520500 0.0: 1.000000 0.5: 0.479500 1.0: 0.157299
8https://en.wikipedia.org/wiki/Error_function Chapter 42.
1.5: 0.033895 2.0: 0.004678
See Also
erf()
42.41 lgamma(), lgammaf(), lgammal()
Compute the natural logarithm of the absolute value of Γ(푥).
Synopsis
#include
double lgamma(double x);
float lgammaf(float x);
long double lgammal(long double x);
Description 9 Compute the natural log of the absolute value of gamma x, log푒 |Γ(푥)|. A range error can occur if x is too large. A pole error can occur is x is non-positive.
Return Value
Returns log푒 |Γ(푥)|.
Example
1 for (float i = 0.5; i <= 4; i += 0.5)
2 printf("%.1f: %f\n", i, lgamma(i));
Output: 0.5: 0.572365 1.0: 0.000000 1.5: -0.120782 2.0: 0.000000 2.5: 0.284683 3.0: 0.693147 3.5: 1.200974 4.0: 1.791759
See Also
tgamma()
9https://en.wikipedia.org/wiki/Gamma_function Chapter 42.
42.42 tgamma(), tgammaf(), tgammal()
Compute the gamma function, Γ(푥).
Synopsis
#include
double tgamma(double x);
float tgammaf(float x);
long double tgammal(long double x);
Description
Computes the gamma function10 of x, Γ(푥). A domain or pole error might occur if x is non-positive. A range error might occur if x is too large or too small.
Return Value
Returns the gamma function of x, Γ(푥).
Example
1 for (float i = 0.5; i <= 4; i += 0.5)
2 printf("%.1f: %f\n", i, tgamma(i));
Output: 0.5: 1.772454 1.0: 1.000000 1.5: 0.886227 2.0: 1.000000 2.5: 1.329340 3.0: 2.000000 3.5: 3.323351 4.0: 6.000000
See Also
lgamma()
42.43 ceil(), ceilf(), ceill()
Ceiling—return the next whole number not smaller than the given number.
Synopsis
#include
double ceil(double x);
10https://en.wikipedia.org/wiki/Gamma_function Chapter 42.
float ceilf(float x);
long double ceill(long double x);
Description
Returns the ceiling of the x: ⌈푥⌉. This is the next whole number not smaller than x. Beware this minor dragon: it’s not just “rounding up”. Well, it is for positive numbers, but negative numbers effectively round toward zero. (Because the ceiling function is headed for the next largest whole number and −4 is larger than −5.)
Return Value
Returns the next largest whole number larger than x.
Example Notice for the negative numbers it heads toward zero, i.e. toward the next largest whole number—just like the positives head toward the next largest whole number.
1 printf("%f\n", ceil(4.0)); // 4.000000
2 printf("%f\n", ceil(4.1)); // 5.000000
3 printf("%f\n", ceil(-2.0)); // -2.000000
4 printf("%f\n", ceil(-2.1)); // -2.000000
5 printf("%f\n", ceil(-3.1)); // -3.000000
See Also
floor(), round()
42.44 floor(), floorf(), floorl()
Compute the largest whole number not larger than the given value.
Synopsis
#include
Description
Returns the floor of the value: ⌊푥⌋. This is the opposite of ceil(). This is the largest whole number that is not greater than x. For positive numbers, this is like rounding down: 4.5 becomes 4.0. For negative numbers, it’s like rounding up: -3.6 becomes -4.0. In both cases, those results are the largest whole number not bigger than the given number. Chapter 42.
Return Value
Returns the largest whole number not greater than x: ⌊푥⌋.
Example Note how the negative numbers effectively round away from zero, unlike the positives.
1 printf("%f\n", floor(4.0)); // 4.000000
2 printf("%f\n", floor(4.1)); // 4.000000
3 printf("%f\n", floor(-2.0)); // -2.000000
4 printf("%f\n", floor(-2.1)); // -3.000000
5 printf("%f\n", floor(-3.1)); // -4.000000
See Also
ceil(), round()
42.45 nearbyint(), nearbyintf(), nearbyintl()
Rounds a value in the current rounding direction.
Synopsis
#include
double nearbyint(double x);
float nearbyintf(float x);
long double nearbyintl(long double x);
Description
This function rounds x to the nearest integer in the current rounding direction. The rounding direction can be set with fesetround() in
Return Value
Returns x rounded in the current rounding direction.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 #pragma STDC FENV_ACCESS ON // If supported
8
9 fesetround(FE_TONEAREST); // round to nearest
10
11 printf("%f\n", nearbyint(3.14)); // 3.000000 Chapter 42.
12 printf("%f\n", nearbyint(3.74)); // 4.000000
13
14 fesetround(FE_TOWARDZERO); // round toward zero
15
16 printf("%f\n", nearbyint(1.99)); // 1.000000
17 printf("%f\n", nearbyint(-1.99)); // -1.000000
18 }
See Also
rint(), lrint(), round(), fesetround(), fegetround()
42.46 rint(), rintf(), rintl()
Rounds a value in the current rounding direction.
Synopsis
#include
double rint(double x);
float rintf(float x);
long double rintl(long double x);
Description
This works just like nearbyint() except that is can raise the “inexact” floating point exception.
Return Value
Returns x rounded in the current rounding direction.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 #pragma STDC FENV_ACCESS ON
8
9 fesetround(FE_TONEAREST);
10
11 printf("%f\n", rint(3.14)); // 3.000000
12 printf("%f\n", rint(3.74)); // 4.000000
13
14 fesetround(FE_TOWARDZERO);
15
16 printf("%f\n", rint(1.99)); // 1.000000
17 printf("%f\n", rint(-1.99)); // -1.000000
18 } Chapter 42.
See Also
nearbyint(), lrint(), round(), fesetround(), fegetround()
42.47 lrint(), lrintf(), lrintl(), llrint(), llrintf(), llrintl()
Returns x rounded in the current rounding direction as an integer.
Synopsis
#include
long int lrint(double x); long int lrintf(float x); long int lrintl(long double x);
long long int llrint(double x); long long int llrintf(float x); long long int llrintl(long double x);
Description Round a floating point number in the current rounding direction, but this time return an integer intead of a float. You know, just to mix it up. These come in two variants: • lrint()—returns long int • llrint()—returns long long int If the result doesn’t fit in the return type, a domain or range error might occur.
Return Value
The value of x rounded to an integer in the current rounding direction.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 #pragma STDC FENV_ACCESS ON
8
9 fesetround(FE_TONEAREST);
10
11 printf("%ld\n", lrint(3.14)); // 3
12 printf("%ld\n", lrint(3.74)); // 4
13
14 fesetround(FE_TOWARDZERO);
15
16 printf("%ld\n", lrint(1.99)); // 1 Chapter 42.
17 printf("%ld\n", lrint(-1.99)); // -1
18 }
See Also
nearbyint(), rint(), round(), fesetround(), fegetround()
42.48 round(), roundf(), roundl()
Round a number in the good old-fashioned way.
Synopsis
#include
double round(double x);
float roundf(float x);
long double roundl(long double x);
Description Rounds a number to the nearest whole value. In case of halfsies, rounds away from zero (i.e. “round up” in magnitude). The current rounding direction’s Jedi mind tricks don’t work on this function.
Return Value
The rounded value of x.
Example
1 printf("%f\n", round(3.14)); // 3.000000
2 printf("%f\n", round(3.5)); // 4.000000
3
4 printf("%f\n", round(-1.5)); // -2.000000
5 printf("%f\n", round(-1.14)); // -1.000000
See Also
lround(), nearbyint(), rint(), lrint(), trunc()
42.49 lround(), lroundf(), lroundl() llround(), llroundf(), llroundl()
Round a number in the good old-fashioned way, returning an integer. Chapter 42.
Synopsis
#include
long int lround(double x); long int lroundf(float x); long int lroundl(long double x);
long long int llround(double x); long long int llroundf(float x); long long int llroundl(long double x);
Description
These are just like round() except they return integers. Halfway values round away from zero, e.g. 1.5 rounds to 2 and −1.5 rounds to −2. The functions are grouped by return type: • lround()—returns a long int • llround()—returns a long long int If the rounded value can’t find in the return time, a domain or range error can occur.
Return Value
Returns the rounded value of x as an integer.
Example
1 printf("%ld\n", lround(3.14)); // 3
2 printf("%ld\n", lround(3.5)); // 4
3
4 printf("%ld\n", lround(-1.5)); // -2
5 printf("%ld\n", lround(-1.14)); // -1
See Also
round(), nearbyint(), rint(), lrint(), trunc()
42.50 trunc(), truncf(), truncl()
Truncate the fractional part off a floating point value.
Synopsis
#include
double trunc(double x);
float truncf(float x);
long double truncl(long double x); Chapter 42.
Description These functions just drop the fractional part of a floating point number. Boom. In other words, they always round toward zero.
Return Value Returns the truncated floating point number.
Example
1 printf("%f\n", trunc(3.14)); // 3.000000
2 printf("%f\n", trunc(3.8)); // 3.000000
3
4 printf("%f\n", trunc(-1.5)); // -1.000000
5 printf("%f\n", trunc(-1.14)); // -1.000000
See Also
round(), lround(), nearbyint(), rint(), lrint()
42.51 fmod(), fmodf(), fmodl()
Compute the floating point remainder.
Synopsis
#include
double fmod(double x, double y);
float fmodf(float x, float y);
long double fmodl(long double x, long double y);
Description 푥 Returns the remainder of 푦 . The result will have the same sign as x. Under the hood, the computation performed is: x - trunc(x / y) * y
But it might be easier just to think of the remainder.
Return Value 푥 Returns the remainder of 푦 with the same sign as x.
Example
1 printf("%f\n", fmod(-9.2, 5.1)); // -4.100000
2 printf("%f\n", fmod(9.2, 5.1)); // 4.100000 Chapter 42.
See Also
remainder()
42.52 remainder(), remainderf(), remainderl()
Compute the remainder IEC 60559-style.
Synopsis
#include
double remainder(double x, double y);
float remainderf(float x, float y);
long double remainderl(long double x, long double y);
Description
This is similar to fmod(), but not quite the same. fmod() is probably what you’re after if you’re expecting remainders to wrap around like an odometer. The C spec quotes IEC 60559 on how this works: When 푦 ≠ 0, the remainder 푟 = 푥 REM 푦 is defined regardless of the rounding mode by the mathematical relation 푟 = 푥 − 푛푦, where 푛 is the integer nearest the exact value of 푥/푦; whenever |푛 − 푥/푦| = 1/2, then 푛 is even. If 푟 = 0, its sign shall be that of 푥. Hope that clears it up! OK, maybe not. Here’s the upshot: You know how if you fmod() something by, say 2.0 you get a result that is somewhere between 0.0 and 2.0? And how if you just increase the number that you’re modding by 2.0, you can see the result climb up to 2.0 and then wrap around to 0.0 like your car’s odometer? remainder() works just like that, except if y is 2.0, it wraps from -1.0 to 1.0 instead of from 0.0 to 2.0. In other words, the range of the function runs from -y/2 to y/2. Contrasted to fmod() that runs from 0.0 to y, remainder()’s output is just shifted down half a y. And zero-remainder-anything is 0. Except if y is zero, the function might return zero or a domain error might occur.
Return Value
The IEC 60559 result of x-remainder-y.
Example
1 printf("%f\n", remainder(3.7, 4)); // -0.300000
2 printf("%f\n", remainder(4.3, 4)); // 0.300000 Chapter 42.
See Also fmod(), remquo()
42.53 remquo(), remquof(), remquol()
Compute the remainder and (some of the) quotient.
Synopsis
#include
Description This is a funky little thing. First of all, the return value is the remainder, the same as the remainder() function, so check that out. And the quotient comes back in the quo pointer. Or at least some of it does. You’ll get at least 3 bits worth of the quotient. But why? So a couple things. One is that the quotient of some very large floating point numbers can easily be far too gigantic to fit in even a long long unsigned int. So some of it might very well need to be lopped off, anyway. But at 3 bits? How’s that even useful? That only gets you from 0 to 7! The C99 Rationale document states: The remquo functions are intended for implementing argument reductions which can exploit a few low-order bits of the quotient. Note that 푥 may be so large in magnitude relative to 푦 that an exact representation of the quotient is not practical. So… implementing argument reductions… which can exploit a few low-order bits… Ooookay. CPPReference has this to say11 on the matter, which is spoken so well, I will quote wholesale: This function is useful when implementing periodic functions with the period exactly repre- sentable as a floating-point value: when calculating sin(휋푥) for a very large x, calling sin directly may result in a large error, but if the function argument is first reduced with remquo, the low-order bits of the quotient may be used to determine the sign and the octant of the result within the period, while the remainder may be used to calculate the value with high precision. And there you have it. If you have another example that works for you… congratulations! :)
Return Value
Returns the same as remainder: The IEC 60559 result of x-remainder-y. In addition, at least the lowest 3 bits of the quotient will be stored in quo with the same sign as x/y.
11https://en.cppreference.com/w/c/numeric/math/remquo Chapter 42.
Example
There’s a great cos() example at CPPReference12 that covers a genuine use case. But instead of stealing it, I’ll just post a simple example here and you can visit their site for a real one.
1 int quo;
2 double rem;
3
4 rem = remquo(12.75, 2.25, &quo);
5
6 printf("%d remainder %f\n", quo, rem); // 6 remainder -0.750000
See Also
remainder(), imaxdiv()
42.54 copysign(), copysignf(), copysignl()
Copy the sign of one value into another.
Synopsis
#include
double copysign(double x, double y);
float copysignf(float x, float y);
long double copysignl(long double x, long double y);
Description
These functions return a number that has the magnitude of x and the sign of y. You can use them to coerce the sign to that of another value. Neither x nor y are modified, of course. The return value holds the result.
Return Value
Returns a value with the magnitude of x and the sign of y.
Example
1 double x = 34.9;
2 double y = -999.9;
3 double z = 123.4;
4
5 printf("%f\n", copysign(x, y)); // -34.900000
6 printf("%f\n", copysign(x, z)); // 34.900000
See Also
signbit()
12https://en.cppreference.com/w/c/numeric/math/remquo Chapter 42.
42.55 nan(), nanf(), nanl()
Return NAN.
Synopsis
#include
double nan(const char *tagp);
float nanf(const char *tagp);
long double nanl(const char *tagp);
Description
These functions return a quiet NaN13. It is produced as if calling strtod() with "NAN" (or a variant thereof) as an argument. tagp points to a string which could be several things, including empty. The contents of the string deter- mine which variant of NaN might get returned depending on the implementation. Which version of NaN? Did you even know it was possible to get this far into the weeds with something that wasn’t a number? Case 1 in which you pass in an empty string, in which case these are the same: nan("");
strtod("NAN()", NULL);
Case 2 in which the string contains only digits 0-9, letters a-z, letters A-Z, and/or underscore: nan("goats");
strtod("NAN(goats)", NULL);
And Case 3, in which the string contains anything else and is ignored: nan("!");
strtod("NAN", NULL);
As for what strtod() does with those values in parens, see the [strtod()] reference page. Spoiler: it’s implementation-defined.
Return Value Returns the requested quiet NaN, or 0 if such things aren’t supported by your system.
Example
1 printf("%f\n", nan("")); // nan
2 printf("%f\n", nan("goats")); // nan
3 printf("%f\n", nan("!")); // nan
See Also
strtod()
13A quiet NaN is one that doesn’t raise any exceptions. Chapter 42.
42.56 nextafter(), nextafterf(), nextafterl()
Get the next (or previous) representable floating point value.
Synopsis
#include
double nextafter(double x, double y);
float nextafterf(float x, float y);
long double nextafterl(long double x, long double y);
Description As you probably know, floating point numbers can’t represent every possible real number. There are limits. And, as such, there exists a “next” and “previous” number after or before any floating point number. These functions return the next (or previous) representable number. That is, no floating point numbers exist between the given number and the next one. The way it figures it out is it works from x in the direction of y, answering the question of “what is the next representable number from x as we head toward y.
Return Value
Returns the next representable floating point value from x in the direction of y. If x equals y, returns y. And also x, I suppose.
Example
1 printf("%.*f\n", DBL_DECIMAL_DIG, nextafter(0.5, 1.0));
2 printf("%.*f\n", DBL_DECIMAL_DIG, nextafter(0.349, 0.0));
Output on my system: 0.50000000000000011 0.34899999999999992
See Also
nexttoward()
42.57 nexttoward(), nexttowardf(), nexttowardl()
Get the next (or previous) representable floating point value.
Synopsis Chapter 42.
include
double nexttoward(double x, long double y);
float nexttowardf(float x, long double y);
long double nexttowardl(long double x, long double y);
Description
These functions are the same as nextafter() except the second parameter is always long double.
Return Value
Returns the same as nextafter() except if x equals y, returns y cast to the function’s return type.
Example
1 printf("%.*f\n", DBL_DECIMAL_DIG, nexttoward(0.5, 1.0));
2 printf("%.*f\n", DBL_DECIMAL_DIG, nexttoward(0.349, 0.0));
Output on my system: 0.50000000000000011 0.34899999999999992
See Also
nextafter()
42.58 fdim(), fdimf(), fdiml()
Return the positive difference between two numbers clamped at 0.
Synopsis
#include
double fdim(double x, double y);
float fdimf(float x, float y);
long double fdiml(long double x, long double y);
Description
The positive difference between x and y is the difference… except if the difference is less than 0, it’s clamped to 0. These functions might throw a range error.
Return Value
Returns the difference of x-y if the difference is greater than 0. Otherwise it returns 0. Chapter 42.
Example
1 printf("%f\n", fdim(10.0, 3.0)); // 7.000000
2 printf("%f\n", fdim(3.0, 10.0)); // 0.000000, clamped
42.59 fmax(), fmax(), fmax(), fmin(), fmin(), fmin()
Return the maximum or minimum of two numbers.
Synopsis
#include
double fmax(double x, double y);
float fmaxf(float x, float y);
long double fmaxl(long double x, long double y);
double fmin(double x, double y);
float fminf(float x, float y);
long double fminl(long double x, long double y);
Description Straightforwardly, these functions return the minimum or maximum of two given numbers. If one of the numbers is NaN, the functions return the non-NaN number. If both arguments are NaN, the functions return NaN.
Return Value Returns the minimum or maximum values, with NaN handled as mentioned above.
Example
1 printf("%f\n", fmin(10.0, 3.0)); // 3.000000
2 printf("%f\n", fmax(3.0, 10.0)); // 10.000000
42.60 fma(), fmaf(), fmal()
Floating (AKA “Fast”) multiply and add.
Synopsis
#include
double fma(double x, double y, double z);
float fmaf(float x, float y, float z); Chapter 42.
long double fmal(long double x, long double y, long double z);
Description This performs the operation (푥 × 푦) + 푧, but does so in a nifty way. It does the computation as if it had infinite precision, and then rounds the final result to the final data type according to the current rounding mode. Contrast to if you’d do the math yourself, where it would have rounded each step of the way, potentially. Also some architectures have a CPU instruction to do exactly this calculation, so it can do it super quick. (If it doesn’t, it’s considerably slower.) You can tell if your CPU supports the fast version by checking that the macro FP_FAST_FMA is set to 1. (The float and long variants of fma() can be tested with FP_FAST_FMAF and FP_FAST_FMAL, respec- tively.) These functions might cause a range error to occur.
Return Value
Returns (x * y) + z.
Example
1 printf("%f\n", fma(1.0, 2.0, 3.0)); // 5.000000
42.61 isgreater(), isgreaterequal(), isless(), islesse- qual()
Floating point comparison macros.
Synopsis
#include
int isgreater(any_floating_type x, any_floating_type y);
int isgreaterequal(any_floating_type x, any_floating_type y);
int isless(any_floating_type x, any_floating_type y);
int islessequal(any_floating_type x, any_floating_type y);
Description These macros compare floating point numbers. Being macros, we can pass in any floating point type. You might think you can already do that with just regular comparison operators—and you’d be right! One one exception: the comparison operators raise the “invalid” floating exception if one or more of the operands is NaN. These macros do not. Note that you must only pass floating point types into these functions. Passing an integer or any other type is undefined behavior. Chapter 42.
Return Value
isgreater() returns the result of x > y. isgreaterequal() returns the result of x >= y. isless() returns the result of x < y. islessequal() returns the result of x <= y.
Example
1 printf("%d\n", isgreater(10.0, 3.0)); // 1
2 printf("%d\n", isgreaterequal(10.0, 10.0)); // 1
3 printf("%d\n", isless(10.0, 3.0)); // 0
4 printf("%d\n", islessequal(10.0, 3.0)); // 0
See Also
islessgreater(), isunordered()
42.62 islessgreater()
Test if a floating point number is less than or greater than another.
Synopsis
#include
int islessgreater(any_floating_type x, any_floating_type y);
Description
This macro is similar to isgreater() and all those, except it made the section name too long if I included it up there. So it gets its own spot. This returns true if 푥 < 푦 or 푥 > 푦. Even though it’s a macro, we can rest assured that x and y are only evaluated once. And even if x or y are NaN, this will not throw an “invalid” exception, unlike the normal comparison operators. If you pass in a non-floating type, the behavior is undefined.
Return Value
Returns (x < y) || (x > y).
Example
1 printf("%d\n", islessgreater(10.0, 3.0)); // 1
2 printf("%d\n", islessgreater(10.0, 30.0)); // 1
3 printf("%d\n", islessgreater(10.0, 10.0)); // 0 Chapter 42.
See Also
isgreater(), isgreaterequal(), isless(), islessequal(), isunordered()
42.63 isunordered()
Macro returns true if either floating point argument is NaN.
Synopsis
#include
int isunordered(any_floating_type x, any_floating_type y);
Description The spec writes: The isunordered macro determines whether its arguments are unordered. See? Told you C was easy! It does also elaborate that the arguments are unordered if one or both of them are NaN.
Return Value This macro returns true if one or both of the arguments are NaN.
Example
1 printf("%d\n", isunordered(1.0, 2.0)); // 0
2 printf("%d\n", isunordered(1.0, sqrt(-1))); // 1
3 printf("%d\n", isunordered(NAN, 30.0)); // 1
4 printf("%d\n", isunordered(NAN, NAN)); // 1
5 }
See Also
isgreater(), isgreaterequal(), isless(), islessequal(), islessgreater() Chapter 43
Some of the following functions have variants that handle different types: atoi(), strtod(), strtol(), abs(), and div(). Only a single one is listed here for brevity.
Function Description _Exit() Exit the currently-running program and don’t look back abort() Abruptly end program execution abs() Compute the absolute value of an integer aligned_alloc() Allocate specifically-aligned memory at_quick_exit() Set up handlers to run when the program quickly exits atexit() Set up handlers to run when the program exits atof() Convert a string to a floating point value atoi() Convert an integer in a string into a integer type bsearch() Binary Search (maybe) an array of objects calloc() Allocate and clear memory for arbitrary use div() Compute the quotient and remainder of two numbers exit() Exit the currently-running program free() Free a memory region getenv() Get the value of an environment variable malloc() Allocate memory for arbitrary use mblen() Return the number of bytes in a multibyte character mbstowcs() Convert a multibyte string to a wide character string mbtowc() Convert a multibyte character to a wide character qsort() Quicksort (maybe) some data quick_exit() Exit the currently-running program quickly rand() Return a pseudorandom number realloc() Resize a previously allocated stretch of memory srand() Seed the built-in pseudorandom number generator strtod() Convert a string to a floating point number strtol() Convert a string to an integer system() Run an external program wcstombs() Convert a wide character string to a multibyte string wctomb() Convert a wide character to a multibyte character
The
361 Chapter 43.
• Pseudorandom number generation • Dynamic memory allocation • Various ways to exit the program • Ability to run external programs • Binary search (or some fast search) • Quicksort (or some fast sort) • Integer arithmetic functions • Multibyte and wide character and string conversions So, you know… a little of everything.
43.1
A couple new types and macros are introduced, though some of these might also be defined elsewhere:
Type Description size_t Returned from sizeof and used elsewhere wchar_t For wide character operations div_t For the div() function ldiv_t For the ldiv() function lldiv_t for the lldiv() function
And some macros:
Type Description NULL Our good pointer friend EXIT_SUCCESS Good exit status when things go well EXIT_FAILURE Good exit status when things go poorly RAND_MAX The maximum value that can be returned by the rand() function MB_CUR_MAX Maximum number of bytes in a multibyte character in the current locale
And there you have it. Just a lot of fun, useful functions in here. Let’s check ’em out!
43.2 atof()
Convert a string to a floating point value
Synopsis
#include
Description This stood for “ASCII-To-Floating” back in the day1, but no one would dare to use such coarse language now. 1http://man.cat-v.org/unix-1st/3/atof Chapter 43.
But the gist is the same: we’re going to convert a string with numbers and (optionally) a decimal point into a floating point value. Leading whitespace is ignored, and translation stops at the first invalid character. If the result doesn’t fit in a double, behavior is undefined. It generally works as if you’d called strtod(): strtod(nptr, NULL)
So check out that reference page for more info. In fact, strtod() is just better and you should probably use that.
Return Value
Returns the string converted to a double.
Example
1 double x = atof("3.141593");
2 printf("%f\n", x); // 3.141593
See Also
atoi(), strtod()
43.3 atoi(), atol(), atoll()
Convert an integer in a string into a integer type
Synopsis
#include
int atoi(const char *nptr);
long int atol(const char *nptr);
long long int atoll(const char *nptr);
Description
Back in the day, atoi() stood for “ASCII-To_Integer”2 but now the spec makes no mention of that. These functions take a string with a number in them and convert it to an integer of the specified return type. Leading whitespace is ignored. Translation stops at the first invalid character. If the result doesn’t fit in the return type, behavior is undefined. It generally works as if you’d called strtol() family of functions: atoi(nptr) // is basically the same as... (int)strtol(nptr, NULL, 10)
atol(nptr) // is basically the same as... strtol(nptr, NULL, 10)
2http://man.cat-v.org/unix-1st/3/atoi Chapter 43.
atoll(nptr) // is basically the same as... strtoll(nptr, NULL, 10)
Again, the strtol() functions are generally better, so I recommend them instead of these.
Return Value Returns an integer result corresponding to the return type.
Example
1 int x = atof("3490");
2 printf("%d\n", x); // 3490
See Also
atof(), strtol()
43.4 strtod(), strtof(), strtold()
Convert a string to a floating point number
Synopsis
#include
double strtod(const char * restrict nptr, char ** restrict endptr);
float strtof(const char * restrict nptr, char ** restrict endptr);
long double strtold(const char * restrict nptr, char ** restrict endptr);
Description These are some neat functions that convert strings to floating point numbers (or even NaN or Infinity) and provide some error checking, besides. Firstly, leading whitespace is skipped. Then the functions attempt to convert characters into the floating point result. Finally, when an invalid character (or NUL character) is reached, they set endptr to point to the invalid character. Set endptr to NULL if you don’t care about where the first invalid character is. If you didn’t set endptr to NULL, it will point to a NUL character if the translation didn’t find any bad characters. That is: if (*endptr == '\0') { printf("What a perfectly-formed number!\n"); } else { printf("I found badness in your number: \"%s\"\n", endptr); }
But guess what! You can also translate strings into special values, like NaN and Infinity! If nptr points to a string containing INF or INFINITY (upper or lowercase), the value for Infinity will be returned. Chapter 43.
If nptr points to a string containing NAN, then (a quiet, non-signalling) NaN will be returned. You can tag the NAN with a sequence of characters from the set 0-9, a-z, A-Z, and _ by enclosing them in parens: NAN(foobar_3490)
What your compiler does with this is implementation-defined, but it can be used to specify different kinds of NaN. You can also specify a number in hexadecimal with a power-of-two exponent (2푥) if you lead with 0x (or 0X). For the exponent, use a p followed by a base 10 exponent. (You can’t use e because that’s a valid hex digit!) Example: 0xabc.123p15
Which computes to 0푥푎푏푐.123 × 215. You can put in FLT_DECIMAL_DIG, DBL_DECIMAL_DIG, or LDBL_DECIMAL_DIG digits and get a correctly- rounded result for the type.
Return Value
Returns the converted number. If there was no number, returns 0. endptr is set to point to the first invalid character, or the NUL terminator if all characters were consumed. If there’s an overflow, HUGE_VAL, HUGE_VALF, or HUGE_VALL is returned, signed like the input, and errno is set to ERANGE. If there’s an underflow, it returns the smallest number closest to zero with the input sign. errno may be set to ERANGE.
Example
1 char *inp = " 123.4567beej";
2 char *badchar;
3
4 double val = strtod(inp, &badchar);
5
6 printf("Converted string to %f\n", val);
7 printf("Encountered bad characters: %s\n", badchar);
8
9 val = strtod("987.654321beej", NULL);
10 printf("Ignoring bad chars: %f\n", val);
11
12 val = strtod("11.2233", &badchar);
13
14 if (*badchar == '\0')
15 printf("No bad chars: %f\n", val);
16 else
17 printf("Found bad chars: %f, %s\n", val, badchar);
Output: Converted string to 123.456700 Encountered bad characters: beej Ignoring bad chars: 987.654321 No bad chars: 11.223300
See Also
atof(), strtol() Chapter 43.
43.5 strtol(), strtoll(), strtoul(), strtoull()
Convert a string to an integer
Synopsis
#include
Description
These convert a string to an integer like atoi(), but they have a few more bells and whistles. Most notable, they can tell you where conversion started going wrong, i.e. where invalid characters, if any, appear. Leading spaces are ignored. A + or - sign may precede the number. The basic idea is that if things go well, these functions will return the integer values contained in the strings. And if you pass in the char** typed endptr, it’ll set it to point at the NUL at the end of the string. If things don’t go well, they’ll set endptr to point at the first character where things have gone awry. That is, if you’re converting a value 103z2! in base 10, they’ll send endptr to point at the z because that’s the first non-numeric character. You can pass in NULL for endptr if you don’t care to do any of that kind of error checking. Wait—did I just say we could set the number base for the conversion? Yes! Yes, I did. Now number bases3 are out of scope for this document, but certainly some of the more well-known are binary (base 2), octal (base 8), decimal (base 10), and hexadecimal (base 16). You can specify the number base for the conversion as the third parameter. Bases from 2 to 36 are sup- ported, with case-insensitive digits running from 0 to Z. If you specify a base of 0, the function will make an effort to determine it. It’ll default to base 10 except for a couple cases: • If the number has a leading 0, it will be octal (base 8) • If the number has a leading 0x or 0X, it will be hex (base 16) The locale might affect the behavior of these functions.
Return Value Returns the converted value. endptr, if not NULL is set to the first invalid character, or to the beginning of the string if no conversion was performed, or to the string terminal NUL if all characters were valid.
3https://en.wikipedia.org/wiki/Radix Chapter 43.
If there’s overflow, one of these values will be returned: LONG_MIN, LONG_MAX, LLONG_MIN, LLONG_MAX, ULONG_MAX, ULLONG_MAX. And errno is set to ERANGE.
Example
1 // All output in decimal (base 10)
2
3 printf("%ld\n", strtol("123", NULL, 0)); // 123
4 printf("%ld\n", strtol("123", NULL, 10)); // 123
5 printf("%ld\n", strtol("101010", NULL, 2)); // binary, 42
6 printf("%ld\n", strtol("123", NULL, 8)); // octal, 83
7 printf("%ld\n", strtol("123", NULL, 16)); // hex, 291
8
9 printf("%ld\n", strtol("0123", NULL, 0)); // octal, 83
10 printf("%ld\n", strtol("0x123", NULL, 0)); // hex, 291
11
12 char *badchar;
13 long int x = strtol(" 1234beej", &badchar, 0);
14
15 printf("Value is %ld\n", x); // Value is 1234
16 printf("Bad chars at \"%s\"\n", badchar); // Bad chars at "beej"
Output: 123 123 42 83 291 83 291 Value is 1234 Bad chars at "beej"
See Also
atoi(), strtod(), setlocale(), strtoimax(), strtoumax()
43.6 rand()
Return a pseudorandom number
Synopsis
#include
int rand(void);
Description
This gives us back a pseudorandom number in the range 0 to RAND_MAX, inclusive. (RAND_MAX will be at least 32767.) If you want to force this to a certain range, the classic way to do this is to force it with the modulo operator Chapter 43.
%, although this introduces biases4 if RAND_MAX+1 is not a multiple of the number you’re modding by. Dealing with this is out of scope for this guide. If you want to to make a floating point number between 0 and 1 inclusive, you can divide the result by RAND_MAX. Or RAND_MAX+1 if you don’t want to include 1. But of course, there are out-of-scope problems with this, as well5. In short, rand() is a great way to get potentially poor random numbers with ease. Probably good enough for the game you’re writing. The spec elaborates: There are no guarantees as to the quality of the random sequence produced and some imple- mentations are known to produce sequences with distressingly non-random low-order bits. Applications with particular requirements should use a generator that is known to be suffi- cient for their needs. Your system probably has a good random number generator on it if you need a stronger source. Linux users have getrandom(), for example, and Windows has CryptGenRandom(). For other more demanding random number work, you might find a library like the GNU Scientific Library6 of use. With most implementations, the numbers produced by rand() will be the same from run to run. To get around this, you need to start it off in a different place by passing a seed into the random number generator. You can do this with srand().
Return Value
Returns a random number in the range 0 to RAND_MAX, inclusive.
Example Note that all of these examples don’t produce perfectly uniform distributions. But good enough for the untrained eye, and really common in general use when mediocre random number quality is acceptable.
1 #include
2 #include
3
4 int main(void)
5 {
6 printf("RAND_MAX = %d\n", RAND_MAX);
7
8 printf("0 to 9: %d\n", rand() % 10);
9
10 printf("10 to 44: %d\n", rand() % 35 + 10);
11 printf("0 to 0.99999: %f\n", rand() / ((float)RAND_MAX + 1));
12 printf("10.5 to 15.7: %f\n", 10.5 + 5.2 * rand() / (float)RAND_MAX);
13 }
Output on my system: RAND_MAX = 2147483647 0 to 9: 3 10 to 44: 21 0 to 0.99999: 0.783099 10.5 to 15.7: 14.651888
Example of seeding the RNG with the time:
4https://stackoverflow.com/questions/10984974/why-do-people-say-there-is-modulo-bias-when-using-a-random-number- generator 5https://mumble.net/~campbell/2014/04/28/uniform-random-float 6https://www.gnu.org/software/gsl/doc/html/rng.html Chapter 43.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 // time(NULL) very likely returns the number of seconds since
8 // January 1, 1970:
9
10 srand(time(NULL));
11
12 for (int i = 0; i < 5; i++)
13 printf("%d\n", rand());
14 }
See Also
srand()
43.7 srand()
Seed the built-in pseudorandom number generator
Synopsis
#include
void srand(unsigned int seed);
Description The dirty little secret of pseudorandom number generation is that they’re completely deterministic. There’s nothing random about them. They just look random. If you use rand() and run your program several times, you might notice something fishy: they produce the same random numbers over and over again. To mix it up, we need to give the pseudorandom number generator a new “starting point”, if you will. We call that the seed. It’s just a number, but it is used as the basic for subsequent number generation. Give a different seed, and you’ll get a different sequence of random numbers. Give the same seed, and you’ll get the same sequence of random numbers corresponding to it7. So if you call srand(3490) before you start generating numbers with rand(), you’ll get the same se- quence every time. srand(37) would also give you the same sequence every time, but it would be a different sequence than the one you got with srand(3490). But if you can’t hardcode the seed (because that would give you the same sequence every time), how are you supposed to do this? It’s really common to use the number of seconds since January 1, 1970 (this date is known as the Unix epoch8) to seed the generator. This sounds pretty arbitrary except for the fact that it’s exactly the value most implementations return from the library call time(NULL)9.
7Minecraft enthusiasts might recall that when generating a new world, they were given the option to enter a random number seed. That single value is used to generate that entire random world. And if your friend starts a world with the same seed you did, they’ll get the same world you did. 8https://en.wikipedia.org/wiki/Unix_time 9The C spec doesn’t say exactly what time(NULL) will return, but the POSIX spec does! And virtually everyone returns exactly that: the number of seconds since epoch. Chapter 43.
We’ll do that in the example. If you don’t call srand(), it’s as if you called srand(1).
Return Value Returns nothing!
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 srand(time(NULL));
8
9 for (int i = 0; i < 5; i++)
10 printf("%d\n", rand() % 32);
11 }
Output: 4 20 22 14 9
Output from a subsequent run: 19 0 31 31 24
See Also
rand(), time()
43.8 aligned_alloc()
Allocate specifically-aligned memory
Synopsis
#include
void *aligned_alloc(size_t alignment, size_t size);
Description
Maybe you wanted malloc() or calloc() instead of this. But if you’re sure you don’t, read on! Chapter 43.
Normally you don’t have to think about this, since malloc() and realloc() both provide memory regions that are suitably aligned10 for use with any data type. But if you need a more specific alignment, you can specify it with this function. When you’re done using the memory region, be sure to free it with a call to free(). Don’t pass in 0 for the size. It probably won’t do anything you want. In case you’re wondering, all dynamically-allocated memory is automatically freed by the system when the program ends. That said, it’s considered to be Good Form to explicitly free() everything you allocate. This way other programmers don’t think you were being sloppy.
Return Value
Returns a pointer to the newly-allocated memory, aligned as specified. Returns NULL if something goes wrong.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 int *p = aligned_alloc(256, 10 * sizeof(int));
8
9 // Just for fun, let's convert to intptr_t and mod with 256
10 // to make sure we're actually aligned on a 256-byte boundary.
11 //
12 // This is probably some kind of implementation-defined
13 // behavior, but I'll bet it works.
14
15 intptr_t ip = (intptr_t)p;
16
17 printf("%ld\n", ip % 256); // 0!
18
19 // Free it up
20 free(p);
21 }
See Also
malloc(), calloc(), free()
43.9 calloc(), malloc()
Allocate memory for arbitrary use
Synopsis
#include
void *calloc(size_t nmemb, size_t size);
10https://en.wikipedia.org/wiki/Data_structure_alignment Chapter 43.
void *malloc(size_t size);
Description Both of these functions allocate memory for general-purpose use. It will be aligned such that it’s useable for storing any data type. malloc() allocates exactly the specified number of bytes of memory in a contiguous block. The memory might be full of garbage data. (You can clear it with memset(), if you wish.) calloc() is different in that it allocates space for nmemb objects of size bytes each. (You can do the same with malloc(), but you have to do the multiplication yourself.) calloc() has an additional feature: it clears all the memory to 0. So if you’re planning to zero the memory anyway, calloc() is probably the way to go. If you’re not, you can avoid that overhead by calling malloc(). When you’re done using the memory region, free it with a call to free(). Don’t pass in 0 for the size. It probably won’t do anything you want. In case you’re wondering, all dynamically-allocated memory is automatically freed by the system when the program ends. That said, it’s considered to be Good Form to explicitly free() everything you allocate. This way other programmers don’t think you were being sloppy.
Return Value
Both functions return a pointer to the shiny, newly-allocated memory. Or NULL if something’s gone awry.
Example
Comparison of malloc() and calloc() for allocating 5 ints:
1 // Allocate space for 5 ints
2 int *p = malloc(5 * sizeof(int));
3
4 p[0] = 12;
5 p[1] = 30;
6
7 // Allocate space for 5 ints
8 // (Also clear that memory to 0)
9 int *q = calloc(5, sizeof(int));
10
11 q[0] = 12;
12 q[1] = 30;
13
14 // All done
15 free(p);
16 free(q);
See Also
aligned_alloc(), free()
43.10 free()
Free a memory region Chapter 43.
Synopsis
#include
void free(void *ptr);
Description
You know that pointer you got back from malloc(), calloc(), or aligned_alloc()? You pass that pointer to free() to free the memory associated with it. If you don’t do this, the memory will stay allocated FOREVER AND EVER! (Well, until your program exits, anyway.) Fun fact: free(NULL) does nothing. You can safely call that. Sometimes it’s convenient. Don’t free() a pointer that’s already been free()d. Don’t free() a pointer that you didn’t get back from one of the allocation functions. It would be Bad11.
Return Value Returns nothing!
Example
1 // Allocate space for 5 ints
2 int *p = malloc(5 * sizeof(int));
3
4 p[0] = 12;
5 p[1] = 30;
6
7 // Free that space
8 free(p);
See Also
malloc(), calloc(), aligned_alloc()
43.11 realloc()
Resize a previously allocated stretch of memory
Synopsis
#include
void *realloc(void *ptr, size_t size);
Description
This takes a pointer to some memory previously allocated with malloc() or calloc() and resizes it to the new size. If the new size is smaller than the old size, any data larger than the new size is discarded.
11“Try to imagine all life as you know it stopping instantaneously and every molecule in your body exploding at the speed of light.” —Egon Spengler Chapter 43.
If the new size is larger than the old size, the new larger part is uninitialized. (You can clear it with memset().) Important note: the memory might move! If you resize, the system might need to relocate the memory to a larger continguous chunk. If this happens, realloc() will copy the old data to the new location for you. Because of this, it’s important to save the returned value to your pointer to update it to the new location if things move. (Also, be sure to error-check so that you don’t overwrite your old pointer with NULL, leaking the memory.) You can also relloc() memory allocated with aligned_alloc(), but it will potentially lose its align- ment if the block is moved.
Return Value
Returns a pointer to the resized memory region. This might be equivalent to the ptr passed in, or it might be some other location.
Example
1 // Allocate space for 5 ints
2 int *p = malloc(5 * sizeof(int));
3
4 p[0] = 12;
5 p[1] = 30;
6
7 // Reallocate for 10 bytes
8 int *new_p = realloc(p, 10 * sizeof(int));
9
10 if (new_p == NULL) {
11 printf("Error reallocing\n");
12 } else {
13 p = new_p; // It's good; let's keep it
14 p[7] = 99;
15 }
16
17 // All done
18 free(p);
See Also
malloc(), calloc()
43.12 abort()
Abruptly end program execution
Synopsis
#include
_Noreturn void abort(void); Chapter 43.
Description This ends program execution abnormally and immediately. Use this in rare, unexpected circumstances. Open streams might not be flushed. Temporary files created might not be removed. Exit handlers are not called. A non-zero exit status is returned to the environment. On some systems, abort() might dump core12, but this is outside the scope of the spec. You can cause the equivalent of an abort() by calling raise(SIGABRT), but I don’t know why you’d do that. The only portable way to stop an abort() call midway is to use signal() to catch SIGABRT and then exit() in the signal handler.
Return Value This function never returns.
Example
1 if (bad_thing) {
2 printf("This should never have happened!\n");
3 fflush(stdout); // Make sure the message goes out
4 abort();
5 }
On my system, this outputs: This should never have happened! zsh: abort (core dumped) ./foo
See Also
signal()
43.13 atexit(), at_quick_exit()
Set up handlers to run when the program exits
Synopsis
#include
int atexit(void (*func)(void));
int at_quick_exit(void (*func)(void));
Description
When the program does a normal exit with exit() or returns from main(), it looks for previously- registered handlers to call on the way out. These handlers are registered with the atexit() call. Think of it like, “Hey, when you’re about to exit, do these extra things.” For the quick_exit() call, you can use the at_quick_exit() function to register handlers for that13.
12https://en.wikipedia.org/wiki/Core_dump 13quick_exit() differs from exit() in that open files might not be flushed and temporary files might not be removed. Chapter 43.
There’s no crossover in handlers from exit() to quick_exit(), i.e. for a call to one, none of the other’s handlers will fire. You can register multiple handlers to fire—at least 32 handlers are supported by both exit() and quick_exit(). The argument func to the functions looks a little weird—it’s a pointer to a function to call. Basically just put the function name to call in there (without parentheses after). See the example, below. If you call atexit() from inside your atexit() handler (or equivalent in your at_quick_exit() han- dler), it’s unspecified if it will get called. So get them all registered before you exit. When exiting, the functions will be called in the reverse order they were registered.
Return Value
These functions return 0 on success, or nonzero on failure.
Example
atexit():
1 #include
2 #include
3
4 void exit_handler_1(void)
5 {
6 printf("Exit handler 1 called!\n");
7 }
8
9 void exit_handler_2(void)
10 {
11 printf("Exit handler 2 called!\n");
12 }
13
14 int main(void)
15 {
16 atexit(exit_handler_1);
17 atexit(exit_handler_2);
18
19 exit(0);
20 }
For the output: Exit handler 2 called! Exit handler 1 called!
And a similar example with quick_exit():
1 #include
2 #include
3
4 void exit_handler_1(void)
5 {
6 printf("Exit handler 1 called!\n");
7 }
8
9 void exit_handler_2(void)
10 {
11 printf("Exit handler 2 called!\n");
12 } Chapter 43.
13
14 int main(void)
15 {
16 at_quick_exit(exit_handler_1);
17 at_quick_exit(exit_handler_2);
18
19 quick_exit(0);
20 }
See Also
exit(), quick_exit()
43.14 exit(), quick_exit(), _Exit()
Exit the currently-running program
Synopsis
#include
_Noreturn void exit(int status);
_Noreturn void quick_exit(int status);
_Noreturn void _Exit(int status);
Description All these functions cause the program to exit, with various levels of cleanup performed. exit() does the most cleanup and is the most normal exit. quick_exit() is the second most. _Exit() unceremoniously drops everything and ragequits on the spot. Calling either of exit() or quick_exit() causes their respective atexit() or at_quick_exit() han- dlers to be called in the reverse order in which they were registered. exit() will flush all streams and delete all temporary files. quick_exit() or _Exit() might not perform that nicety. _Exit() doesn’t call any of the at-exit handlers, either. For all functions, the exit status is returned to the environment. Defined exit statuses are:
Status Description EXIT_SUCCESS Typically returned when good things happen 0 Same as EXIT_SUCCESS EXIT_FAILURE Oh noes! Definitely failure! Any positive value Generally indicates another failure of some kind Chapter 43.
Return Value None of these functions ever return.
Example
1 exit(EXIT_SUCCESS);
1 quick_exit(EXIT_FAILURE);
1 _Exit(2);
See Also
atexit(), at_quick_exit()
43.15 getenv()
Get the value of an environment variable
Synopsis
#include
char *getenv(const char *name);
Description The environment often provides variables that are set before the program run that you can access at run- time. Of course the exact details are system dependent, but these variables are key/value pairs, and you can get the value by passing the key to getenv() as the name parameter. You’re not allowed to overwrite the string that’s returned. This is pretty limited in the standard, but your OS often provides better functionality. See the Environment Variables section for more details.
Return Value
Returns a pointer to the environment variable value, or NULL if the variable doesn’t exist.
Example
1 printf("PATH is %s\n", getenv("PATH"));
Output (truncated in my case): PATH is /usr/bin:/usr/local/bin:/usr/sbin:/home/beej/.cargo/bin [...]
43.16 system()
Run an external program Chapter 43.
Synopsis
#include
int system(const char *string);
Description This will run an external program and then return to the caller. The manner in which it runs the program is system-defined, but typically you can pass something to it just like you’d run on the command line, searching the PATH, etc. Not all systems have this capability, but you can test for it by passing NULL to system() and seeing if it returns 0 (no command processor is available) or non-zero (a command processor is available! Yay!) If you’re getting user input and passing it to the system() call, be extremely careful to escape all special shell characters (everything that’s not alphanumeric) with a backslash to keep a villain from running something you don’t want them to.
Return Value
If NULL is passed, returns nonzero if a command processor is available (i.e. system() will work at all). Otherwise returns an implementation-defined value.
Example
1 printf("Here's a directory listing:\n\n");
2
3 system("ls -l"); // Run this command and return
4
5 printf("\nAll done!\n");
Output: Here's a directory listing:
total 92 drwxr-xr-x 3 beej beej 4096 Oct 14 21:38 bin drwxr-xr-x 2 beej beej 4096 Dec 20 20:07 examples -rwxr-xr-x 1 beej beej 16656 Feb 23 21:49 foo -rw-rw-rw- 1 beej beej 155 Feb 23 21:49 foo.c -rw-r--r-- 1 beej beej 1350 Jan 27 22:11 Makefile -rw-r--r-- 1 beej beej 4644 Jan 18 09:12 README.md drwxr-xr-x 3 beej beej 4096 Feb 23 20:21 src drwxr-xr-x 6 beej beej 4096 Feb 21 20:24 stage drwxr-xr-x 2 beej beej 4096 Sep 27 20:54 translations drwxr-xr-x 2 beej beej 4096 Sep 27 20:54 website
All done!
43.17 bsearch()
Binary Search (maybe) an array of objects Chapter 43.
Synopsis
#include
void *bsearch(const void *key, const void *base, size_t nmemb, size_t size, int (*compar)(const void *, const void *));
Description This crazy-looking function searches an array for a value. It probably is a binary search or some fast, efficient search. But the spec doesn’t really say. However, the array must be sorted! So binary search seems likely. • key is a pointer to the value to find. • base is a pointer to the start of the array—the array must be sorted! • nmemb is the number of elements in the array. • size is the sizeof each element in the array. • compar is a pointer to a function that will compare the key against other values. The comparison function takes the key as the first argument and the value to compare against as the second. It should return a negative number if the key is less than the value, 0 if the key equals the value, and a positive number if the key is greater than the value. This is commonly computed by taking the difference between the key and the value to be compared. If subtraction is supported. The return value from the strcmp() function can be used for comparing strings. Again, the array must be sorted according to the order of the comparison function before running bsearch(). Luckily for you, you can just call qsort() with the same comparison function to get this done. It’s a general-purpose function—it’ll search any type of array for anything. The catch is you have to write the comparison function. And that’s not as scary as it looks. Jump down to the example
Return Value
The function returns a pointer to the found value, or NULL if it can’t be found.
Example
1 #include
2 #include
3
4 int compar(const void *key, const void *value)
5 {
6 const int *k = key, *v = value; // Need ints, not voids
7
8 return *k - *v;
9 }
10
11 int main(void)
12 {
13 int a[9] = {2, 6, 9, 12, 13, 18, 20, 32, 47};
14
15 int *r, key;
16 Chapter 43.
17 key = 12; // 12 is in there
18 r = bsearch(&key, a, 9, sizeof(int), compar);
19 printf("Found %d\n", *r);
20
21 key = 30; // Won't find a 30
22 r = bsearch(&key, a, 9, sizeof(int), compar);
23 if (r == NULL)
24 printf("Didn't find 30\n");
25
26 // Searching with an unnamed key, pointer to 32
27 r = bsearch(&(int){32}, a, 9, sizeof(int), compar);
28 printf("Found %d\n", *r); // Found it
29 }
Output: Found 12 Didn't find 30 Found 32
See Also
strcmp(), qsort()
43.18 qsort()
Quicksort (maybe) some data
Synopsis
#include
void qsort(void *base, size_t nmemb, size_t size, int (*compar)(const void *, const void *));
Description This function will quicksort (or some other sort, probably speedy) an array of data in-place14. Like bsearch(), it’s data-agnostic. Any data for which you can define a relative ordering can be sorted, whether ints, structs, or anything else. Also like bsearch(), you have to give a comparison function to do the actual compare. • base is a pointer to the start of the array to be sorted. • nmemb is the number of elements in the array. • size is the sizeof each element. • compar is a pointer to the comparison function. The comparison function takes pointers to two elements of the array as arguments and compares them. It should return a negative number if the first argument is less than the second, 0 if they are equal, and a positive number if the first argument is greater than the second. This is commonly computed by taking the difference between the first argument and the second. If sub- traction is supported. The return value from the strcmp() function can provide sort order for strings.
14“In-place” meaning that the original array will hold the results; no new array is allocated. Chapter 43.
If you have to sort a struct, just subtract the specific field you want to sort by. This comparison function can be used by bsearch() to do searches after the list is sorted. To reverse the sort, subtract the second argument from the first, i.e. negate the return value from compar().
Return Value Returns nothing!
Example
1 #include
2 #include
3
4 int compar(const void *elem0, const void *elem1)
5 {
6 const int *x = elem0, *y = elem1; // Need ints, not voids
7
8 return *x - *y;
9 }
10
11 int main(void)
12 {
13 int a[9] = {14, 2, 3, 17, 10, 8, 6, 1, 13};
14
15 // Sort the list
16
17 qsort(a, 9, sizeof(int), compar);
18
19 // Print sorted list
20
21 for (int i = 0; i < 9; i++)
22 printf("%d ", a[i]);
23
24 putchar('\n');
25
26 // Use the same compar() function to binary search
27 // for 17 (passed in as an unnamed object)
28
29 int *r = bsearch(&(int){17}, a, 9, sizeof(int), compar);
30 printf("Found %d!\n", *r);
31 }
Output: 1 2 3 6 8 10 13 14 17 Found 17!
See Also
strcmp(), bsearch()
43.19 abs(), labs(), llabs()
Compute the absolute value of an integer Chapter 43.
Synopsis
#include
int abs(int j);
long int labs(long int j);
long long int llabs(long long int j);
Description
Compute the absolute value of j. If you don’t remember, that’s how far from zero j is. In other words, if j is negative, return it as a positive. If it’s positive, return it as a positive. Always be positive. Enjoy life. If the result cannot be represented, the behavior is undefined. Be especially aware of the upper half of unsigned numbers.
Return Value
Returns the absolute value of j, |푗|.
Example
1 printf("|-2| = %d\n", abs(-2));
2 printf("|4| = %d\n", abs(4));
Output: |-2| = 2 |4| = 4
See Also
fabs()
43.20 div(), ldiv(), lldiv()
Compute the quotient and remainder of two numbers
Synopsis
#include
div_t div(int numer, int denom);
ldiv_t ldiv(long int numer, long int denom);
lldiv_t lldiv(long long int numer, long long int denom);
Description These functions get you the quotient and remainder of a pair of numbers in one go. Chapter 43.
They return a structure that has two fields, quot, and rem, the types of which match types of numer and denom. Note how each function returns a different variant of div_t. These div_t variants are equivalent to the following: typedef struct { int quot, rem; } div_t;
typedef struct { long int quot, rem; } ldiv_t;
typedef struct { long long int quot, rem; } lldiv_t;
Why use these instead of the division operator? The C99 Rationale says: Because C89 had implementation-defined semantics for division of signed integers when negative operands were involved, div and ldiv, and lldiv in C99, were invented to provide well-specified semantics for signed integer division and remainder operations. The semantics were adopted to be the same as in Fortran. Since these functions return both the quotient and the remainder, they also serve as a convenient way of efficiently modeling underlying hardware that computes both results as part of the same operation. Table 7.2 summarizes the semantics of these functions. Indeed, K&R2 (C89) says: The direction of truncation for / and the sign of the result for % are machine-dependent for negative operands […] The Rationale then goes on to spell out what the signs of the quotient and remainder will be given the signs of a numerator and denominator when using the div() functions:
numer denom quot rem + + + + − + − − + − − + − − + −
Return Value
A div_t, ldiv_t, or lldiv_t structure with the quot and rem fields loaded with the quotient and re- mainder of the operation of numer/denom.
Example
1 div_t d = div(64, -7);
2
3 printf("64 / -7 = %d\n", d.quot);
4 printf("64 %% -7 = %d\n", d.rem);
Output: 64 / -7 = -9 64 % -7 = 1 Chapter 43.
See Also
fmod(), remainder()
43.21 mblen()
Return the number of bytes in a multibyte character
Synopsis
#include
int mblen(const char *s, size_t n);
Description If you have a multibyte character in a string, this will tell you how many bytes long it is. n is the maximum number of bytes mblen() will scan before giving up. If s is a NULL pointer, tests if this encoding has state dependency, as noted in the return value, below. It also resets the state, if there is one. The behavior of this function is influenced by the locale.
Return Value
Returns the number of bytes used to encode this character, or -1 if there is no valid multibyte character in the next n bytes. Or, if s is NULL, returns true if this encoding has state dependency.
Example For the example, I used my extended character set to put Unicode characters in the source. If this doesn’t work for you, use the \uXXXX escape.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 setlocale(LC_ALL, "");
8
9 printf("State dependency: %d\n", mblen(NULL, 0));
10 printf("Bytes for €: %d\n", mblen("€", 5));
11 printf("Bytes for \u00e9: %d\n", mblen("\u00e9", 5)); // \u00e9 == é
12 printf("Bytes for &: %d\n", mblen("&", 5));
13 }
Output (in my case, the encoding is UTF-8, but your mileage may vary): State dependency: 0 Bytes for €: 3 Bytes for é: 2 Bytes for &: 1 Chapter 43.
See Also
mbtowc(), mbstowcs()), setlocale()
43.22 mbtowc()
Convert a multibyte character to a wide character
Synopsis
#include
int mbtowc(wchar_t * restrict pwc, const char * restrict s, size_t n);
Description If you have a multibyte character, this function will convert it to a wide character and stored at the address pointed to by pwc. Up to n bytes of the multibyte character will be analyzed. If pwc is NULL, the resulting character will not be stored. (Useful for just getting the return value.) If s is a NULL pointer, tests if this encoding has state dependency, as noted in the return value, below. It also resets the state, if there is one. The behavior of this function is influenced by the locale.
Return Value
Returns the number of bytes used in the encoded wide character, or -1 if there is no valid multibyte character in the next n bytes. Returns 0 if s points to the NUL character. Or, if s is NULL, returns true if this encoding has state dependency.
Example
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 setlocale(LC_ALL, "");
9
10 printf("State dependency: %d\n", mbtowc(NULL, NULL, 0));
11
12 wchar_t wc;
13 int bytes;
14
15 bytes = mbtowc(&wc, "€", 5);
16
17 printf("L'%lc' takes %d bytes as multibyte char '€'\n", wc, bytes);
18 }
Output on my system: Chapter 43.
State dependency: 0 L'€' takes 3 bytes as multibyte char '€'
See Also
mblen(), mbstowcs(), wcstombs(), setlocale()
43.23 wctomb()
Convert a wide character to a multibyte character
Synopsis
#include
int wctomb(char *s, wchar_t wc);
Description If you have your hands on a wide character, you can use this to make it multibyte. The wide character wc is stored as a multibyte character in the string pointed to by s. The buffer s points to should be at least MB_CUR_MAX characters long. Note that MB_CUR_MAX changes with locale. If wc is a NUL wide character, a NUL is stored in s after the bytes needed to reset the shift state (if any). If s is a NULL pointer, tests if this encoding has state dependency, as noted in the return value, below. It also resets the state, if there is one. The behavior of this function is influenced by the locale.
Return Value
Returns the number of bytes used in the encoded multibyte character, or -1 if wc does not correspond to any valid multibyte character. Or, if s is NULL, returns true if this encoding has state dependency.
Example
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 setlocale(LC_ALL, "");
9
10 printf("State dependency: %d\n", mbtowc(NULL, NULL, 0));
11
12 int bytes;
13 char mb[MB_CUR_MAX + 1];
14
15 bytes = wctomb(mb, L'€');
16 mb[bytes] = '\0';
17 Chapter 43.
18 printf("L'€' takes %d bytes as multibyte char '%s'\n", bytes, mb);
19 }
Output on my system: State dependency: 0 L'€' takes 3 bytes as multibyte char '€'
See Also
mbtowc(), mbstowcs(), wcstombs(), setlocale()
43.24 mbstowcs()
Convert a multibyte string to a wide character string
Synopsis
#include
size_t mbstowcs(wchar_t * restrict pwcs, const char * restrict s, size_t n);
Description If you have a multibyte string (AKA a regular string), you can convert it wto a wide character string with this function. At most n wide characters are written to the destination pwcs from the source s. A NUL character is stored as a wide NUL character. Non-portable POSIX extension: if you’re using a POSIX-complaint library, this function allows pwcs to be NULL if you’re only interested in the return value. Most notably, this will give you the number of characters in a multibyte string (as opposed to strlen() which counts the bytes.)
Return Value
Returns the number of wide characters written to the destination pwcs. If an invalid multibyte character was found, returns (size_t)(-1). If the return value is n, it means the result was not NUL-terminated.
Example This source uses an extended character set. If your compiler doesn’t support it, you’ll have to replace them with \u escapes.
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 setlocale(LC_ALL, "");
9
10 wchar_t wcs[128]; Chapter 43.
11 char *s = "€200 for this spoon?"; // 20 characters
12
13 size_t char_count, byte_count;
14
15 char_count = mbstowcs(wcs, s, 128);
16 byte_count = strlen(s);
17
18 printf("Wide string: L\"%ls\"\n", wcs);
19 printf("Char count : %zu\n", char_count); // 20
20 printf("Byte count : %zu\n\n", byte_count); // 22 on my system
21
22 // POSIX Extension that allows you to pass NULL for
23 // the destination so you can just use the return
24 // value (which is the character count of the string,
25 // if no errors have occurred)
26
27 s = "§¶°±π€•"; // 7 characters
28
29 char_count = mbstowcs(NULL, s, 0); // POSIX-only, nonportable
30 byte_count = strlen(s);
31
32 printf("Multibyte str: \"%s\"\n", s);
33 printf("Char count : %zu\n", char_count); // 7
34 printf("Byte count : %zu\n", byte_count); // 16 on my system
35 }
Output on my system (byte count will depend on your encoding): Wide string: L"€200 for this spoon?" Char count : 20 Byte count : 22
Multibyte str: "§¶°±π€•" Char count : 7 Byte count : 16
See Also
mblen(), mbtowc(), wcstombs(), setlocale()
43.25 wcstombs()
Convert a wide character string to a multibyte string
Synopsis
#include
size_t wcstombs(char * restrict s, const wchar_t * restrict pwcs, size_t n);
Description If you have a wide character string and you want it as multibyte string, this is the function for you! It’ll take the wide characters pointed to by pwcs and convert them to multibyte characters stored in s. No more than n bytes will be written to s. Chapter 43.
Non-portable POSIX extension: if you’re using a POSIX-complaint library, this function allows s to be NULL if you’re only interested in the return value. Most notably, this will give you the number of bytes needed to encode the wide characters in a multibyte string.
Return Value
Returns the number of bytes written to s, or (size_t)(-1) if one of the characters can’t be encoded into a multibyte string. If the return value is n, it means the result was not NUL-terminated.
Example This source uses an extended character set. If your compiler doesn’t support it, you’ll have to replace them with \u escapes.
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 setlocale(LC_ALL, "");
9
10 char mbs[128];
11 wchar_t *wcs = L"€200 for this spoon?"; // 20 characters
12
13 size_t byte_count;
14
15 byte_count = wcstombs(mbs, wcs, 128);
16
17 printf("Wide string: L\"%ls\"\n", wcs);
18 printf("Multibyte : \"%s\"\n", mbs);
19 printf("Byte count : %zu\n\n", byte_count); // 22 on my system
20
21 // POSIX Extension that allows you to pass NULL for
22 // the destination so you can just use the return
23 // value (which is the character count of the string,
24 // if no errors have occurred)
25
26 wcs = L"§¶°±π€•"; // 7 characters
27
28 byte_count = wcstombs(NULL, wcs, 0); // POSIX-only, nonportable
29
30 printf("Wide string: L\"%ls\"\n", wcs);
31 printf("Byte count : %zu\n", byte_count); // 16 on my system
32 }
Output on my system (byte count will depend on your encoding): Wide string: L"€200 for this spoon?" Multibyte : "€200 for this spoon?" Byte count : 22
Wide string: L"§¶°±π€•" Byte count : 16 Chapter 43.
See Also mblen(), wctomb(), mbstowcs(), setlocale() Chapter 44
Function Description clock() How much processor time has been used by this process difftime() Compute the difference between two times mktime() Convert a struct tm into a time_t time() Get the current calendar time timespec_get() Get a higher resolution time, probably now asctime() Return a human-readable version of a struct tm ctime() Return a human-readable version of a time_t gmtime() Convert a calendar time into a UTC broken-down time localtime() Convert a calendar time into a broken-down local time strftime() Formatted date and time output
When it comes to time and C, there are two main types to look for: • time_t holds a calendar time. This is an potentially opaque numeric type that represents an absolute time that can be converted to UTC1 or local time. • struct tm holds a broken-down time. This has things like the day of the week, the day of the month, the hour, the minute, the second, etc. On POSIX systems and Windows, time_t is an integer and represents the number of seconds that have elapsed since January 1, 1970 at 00:00 UTC. A struct tm contains the following fields: struct tm { int tm_sec; // seconds after the minute -- [0, 60] int tm_min; // minutes after the hour -- [0, 59] int tm_hour; // hours since midnight -- [0, 23] int tm_mday; // day of the month -- [1, 31] int tm_mon; // months since January -- [0, 11] int tm_year; // years since 1900 int tm_wday; // days since Sunday -- [0, 6] int tm_yday; // days since January 1 -- [0, 365] int tm_isdst; // Daylight Saving Time flag };
You can convert between the two with mktime(), gmtime(), and localtime(). You can print time information to strings with ctime(), asctime(), and strftime().
1When you say GMT, unless you’re talking specifically about the timezone and not the time, you probably mean “UTC”.
392 Chapter 44.
44.1 Thread Safety Warning
asctime(), ctime(): These two functions return a pointer to a static memory region. They both might return the same pointer. If you need thread safety, you’ll need a mutex across them. If you need both results at once, strcpy() one of them out. All these problems with asctime() and ctime() can be avoided by using the more flexible and thread- safe strftime() function instead. localtime(), gmtime(): These other two functions also return a pointer to a static memory region. They both might return the same pointer. If you need thread safety, you’ll need a mutex across them. If you need both results at once, copy the struct to another.
44.2 clock()
How much processor time has been used by this process
Synopsis
#include
clock_t clock(void);
Description Your processor is juggling a lot of things right now. Just because a process has been alive for 20 minutes doesn’t mean that it used 20 minutes of “CPU time”. Most of the time your average process spends asleep, and that doesn’t count toward the CPU time spent. This function returns an opaque type representing the number of “clock ticks”2 the process has spent in operation. You can get the number of seconds out of that by dividing by the macro CLOCKS_PER_SEC. This is an integer, so you will have to cast part of the expression to a floating type to get a fractional time. Note that this is not the “wall clock time” of the program. If you want to get that loosely use time() and difftime() (which might only offer 1-second resolution) or timespec_get() (which might only also offer low resolution, but at least it might go to nanosecond level).
Return Value Returns the amount of CPU time spent by this process. This comes back in a form that can be divided by CLOCKS_PER_SEC to determine the time in seconds.
Example
1 #include
2 #include
3
4 // Deliberately naive Fibonacci
5 long long int fib(long long int n) {
6 if (n <= 1) return n;
7
8 return fib(n-1) + fib(n-2);
9 }
2The spec doesn’t actually say “clock ticks”, but I… am. Chapter 44.
10
11 int main(void)
12 {
13 printf("The 42nd Fibonacci Number is %lld\n", fib(42));
14
15 printf("CPU time: %f\n", clock() / (double)CLOCKS_PER_SEC);
16 }
Output on my system: The 42nd Fibonacci Number is 267914296 CPU time: 1.863078
See Also
time(), difftime(), timespec_get()
44.3 difftime()
Compute the difference between two times
Synopsis
#include
double difftime(time_t time1, time_t time0);
Description
Since the time_t type is technically opaque, you can’t just straight-up subtract to get the difference between two of them3. Use this function to do it. There is no guarantee as to the resolution of this difference, but it’s probably to the second.
Return Value
Returns the difference between two time_ts in seconds.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // April 12, 1982 and change
7 struct tm time_a = { .tm_year=82, .tm_mon=3, .tm_mday=12,
8 .tm_hour=4, .tm_min=00, .tm_sec=04, .tm_isdst=-1,
9 };
10
11 // November 15, 2020 and change
12 struct tm time_b = { .tm_year=120, .tm_mon=10, .tm_mday=15,
13 .tm_hour=16, .tm_min=27, .tm_sec=00, .tm_isdst=-1,
14 };
3Unless you’re on a POSIX system where time_t is definitely an integer, in which case you can subtract. But you should still use difftime() for maximum portability. Chapter 44.
15
16 time_t cal_a = mktime(&time_a);
17 time_t cal_b = mktime(&time_b);
18
19 double diff = difftime(cal_b, cal_a);
20
21 double years = diff / 60 / 60 / 24 / 365.2425; // close enough
22
23 printf("%f seconds (%f years) between events\n", diff, years);
24 }
Output: 1217996816.000000 seconds (38.596783 years) between events
See Also
time(), mktime()
44.4 mktime()
Convert a struct tm into a time_t
Synopsis
#include
time_t mktime(struct tm *timeptr);
Description
If you have a local date and time and want it converted to a time_t (so that you can difftime() it or whatever), you can convert it with this function. Basically you fill out the fields in your struct tm in local time and mktime() will convert those to the UTC time_t equivalent. A couple notes: • Don’t bother filling out tm_wday or tm_yday. mktime() will fill these out for you. • You can set tm_isdst to 0 to indicate your time isn’t Daylight Saving Time (DST), 1 to indicate it is, and -1 to have mktime() fill it in according to your locale’s preference. If you need input in UTC, see the non-standard functions timegm()4 for Unix-likes and _mkgmtime()5 for Windows.
Return Value
Returns the local time in the struct tm as a time_t calendar time. Returns (time_t)(-1) on error.
4https://man.archlinux.org/man/timegm.3.en 5https://docs.microsoft.com/en-us/cpp/c-runtime-library/reference/mkgmtime-mkgmtime32-mkgmtime64?view=msvc-160 Chapter 44.
Example
In the following example, we have mktime() tell us if that time was DST or not.
1 #include
2 #include
3
4 int main(void)
5 {
6 struct tm broken_down_time = {
7 .tm_year=82, // years since 1900
8 .tm_mon=3, // months since January -- [0, 11]
9 .tm_mday=12, // day of the month -- [1, 31]
10 .tm_hour=4, // hours since midnight -- [0, 23]
11 .tm_min=00, // minutes after the hour -- [0, 59]
12 .tm_sec=04, // seconds after the minute -- [0, 60]
13 .tm_isdst=-1, // Daylight Saving Time flag
14 };
15
16 time_t calendar_time = mktime(&broken_down_time);
17
18 char *days[] = {"Sunday", "Monday", "Tuesday",
19 "Wednesday", "Furzeday", "Friday", "Saturday"};
20
21 // This will print what was in broken_down_time
22 printf("Local time : %s", asctime(localtime(&calendar_time)));
23 printf("Is DST : %d\n", broken_down_time.tm_isdst);
24 printf("Day of week: %s\n\n", days[broken_down_time.tm_wday]);
25
26 // This will print UTC for the local time, above
27 printf("UTC : %s", asctime(gmtime(&calendar_time)));
28 }
Output (for me in Pacific Time—UTC is 8 hours ahead): Local time : Mon Apr 12 04:00:04 1982 Is DST : 0 Day of week: Monday
UTC : Mon Apr 12 12:00:04 1982
See Also
localtime(), gmtime()
44.5 time()
Get the current calendar time
Synopsis
#include
time_t time(time_t *timer); Chapter 44.
Description Returns the current calendar time right now. I mean, now. No, now! If timer is not NULL, it gets loaded with the current time, as well. This can be converted into a struct tm with localtime() or gmtime(), or printed directly with ctime().
Return Value
Returns the current calendar time. Also loads timer with the current time if it’s not NULL. Or returns (time_t)(-1) if the time isn’t available because you’ve fallen out of the space-time continuum and/or the system doesn’t support times.
Example
1 time_t now = time(NULL);
2
3 printf("The local time is %s", ctime(&now));
Example output: The local time is Mon Mar 1 18:45:14 2021
See Also
localtime(), gmtime(), ctime()
44.6 timespec_get()
Get a higher resolution time, probably now
Synopsis
#include
int timespec_get(struct timespec *ts, int base);
Description
This function loads the current time UTC (unless directed otherwise) into the given struct timespec, ts. That structure has two fields: struct timespec { time_t tv_sec; // Whole seconds long tv_nsec; // Nanoseconds, 0-999999999 }
Nanoseconds are billionths of a second. You can divide by 1000000000.0 to convert to seconds. The base parameter has only one defined value, by the spec: TIME_UTC. So portably make it that. This will load ts with the current time in seconds since a system-defined Epoch6, often January 1, 1970 at 00:00 UTC. Your implementation might define other values for base.
6https://en.wikipedia.org/wiki/Unix_time Chapter 44.
Return Value
When base is TIME_UTC, loads ts with the current UTC time. On success, returns base, valid values for which will always be non-zero. On error, returns 0.
Example
1 struct timespec ts;
2
3 timespec_get(&ts, TIME_UTC);
4
5 printf("%ld s, %ld ns\n", ts.tv_sec, ts.tv_nsec);
6
7 double float_time = ts.tv_sec + ts.tv_nsec/1000000000.0;
8 printf("%f seconds since epoch\n", float_time);
Example output: 1614654187 s, 825540756 ns 1614654187.825541 seconds since epoch
Here’s a helper function to add values to a struct timespec that handles negative values and nanosecond overflow. #include
// Add delta seconds and delta nanoseconds to ts. // Negative values are allowed. Each component is added individually. // // Subtract 1.5 seconds from the current value: // // timespec_add(&ts, -1, -500000000L);
struct timespec *timespec_add(struct timespec *ts, long dsec, long dnsec) { long sec = (long)ts->tv_sec + dsec; long nsec = ts->tv_nsec + dnsec;
ldiv_t qr = ldiv(nsec, 1000000000L);
if (qr.rem < 0) { nsec = 1000000000L + qr.rem; sec += qr.quot - 1; } else { nsec = qr.rem; sec += qr.quot; }
ts->tv_sec = sec; ts->tv_nsec = nsec;
return ts; }
And here are some functions to convert from long double to struct timespec and back, just in case you like thinking in decimals. This is more limited in significant figures than using the integer values. #include
// Convert a struct timespec into a long double Chapter 44.
// Convert a long double to a struct timespec struct timespec ld_to_timespec(long double t) { long double f; struct timespec ts; ts.tv_nsec = modfl(t, &f) * 1000000000L; ts.tv_sec = f;
return ts; }
See Also time(), mtx_timedlock(), cnd_timedwait()
44.7 asctime()
Return a human-readable version of a struct tm
Synopsis
#include
Description
This takes a time in a struct tm and returns a string with that date in the form: Sun Sep 16 01:03:52 1973 with a newline included at the end, rather unhelpfully. (strftime() will give you more flexibility.) It’s just like ctime(), except it takes a struct tm instead of a time_t. WARNING: This function returns a pointer to a static char* region that isn’t thread-safe and might be shared with the ctime() function. If you need thread safety, use strftime() or use a mutex that covers ctime() and asctime(). Behavior is undefined for: • Years less than 1000 • Years greater than 9999 • Any members of timeptr are out of range
Return Value Returns a pointer to the human-readable date string.
Example Chapter 44.
1 time_t now = time(NULL);
2
3 printf("Local: %s", asctime(localtime(&now)));
4 printf("UTC : %s", asctime(gmtime(&now)));
Sample output: Local: Mon Mar 1 21:17:34 2021 UTC : Tue Mar 2 05:17:34 2021
See Also
ctime(), localtime(), gmtime()
44.8 ctime()
Return a human-readable version of a time_t
Synopsis
#include
char *ctime(const time_t *timer);
Description
This takes a time in a time_t and returns a string with the local time and date in the form: Sun Sep 16 01:03:52 1973
with a newline included at the end, rather unhelpfully. (strftime() will give you more flexibility.) It’s just like asctime(), except it takes a time_t instead of a struct tm. WARNING: This function returns a pointer to a static char* region that isn’t thread-safe and might be shared with the asctime() function. If you need thread safety, use strftime() or use a mutex that covers ctime() and asctime(). Behavior is undefined for: • Years less than 1000 • Years greater than 9999 • Any members of timeptr are out of range
Return Value A pointer to the human-readable local time and data string.
Example
1 time_t now = time(NULL);
2
3 printf("Local: %s", ctime(&now));
Sample output: Local: Mon Mar 1 21:32:23 2021 Chapter 44.
See Also
asctime()
44.9 gmtime()
Convert a calendar time into a UTC broken-down time
Synopsis
#include
struct tm *gmtime(const time_t *timer);
Description
If you have a time_t, you can run it through this function to get a struct tm back full of the correspond- ing broken-down UTC time information. This is just like localtime(), except it does UTC instead of local time. Once you have that struct tm, you can feed it to strftime() to print it out. WARNING: This function returns a pointer to a static struct tm* region that isn’t thread-safe and might be shared with the localtime() function. If you need thread safety use a mutex that covers gmtime() and localtime().
Return Value
Returns a pointer to the broken-down UTC time, or NULL if it can’t be obtained.
Example
1 time_t now = time(NULL);
2
3 printf("UTC : %s", asctime(gmtime(&now)));
4 printf("Local: %s", asctime(localtime(&now)));
Sample output: UTC : Tue Mar 2 05:40:05 2021 Local: Mon Mar 1 21:40:05 2021
See Also
localtime(), asctime(), strftime()
44.10 localtime()
Convert a calendar time into a broken-down local time
Synopsis Chapter 44.
#include
struct tm *localtime(const time_t *timer);
Description
If you have a time_t, you can run it through this function to get a struct tm back full of the correspond- ing broken-down local time information. This is just like gmtime(), except it does local time instead of UTC. Once you have that struct tm, you can feed it to strftime() to print it out. WARNING: This function returns a pointer to a static struct tm* region that isn’t thread-safe and might be shared with the gmtime() function. If you need thread safety use a mutex that covers gmtime() and localtime().
Return Value
Returns a pointer to the broken-down local time, or NULL if it can’t be obtained.
Example
1 time_t now = time(NULL);
2
3 printf("Local: %s", asctime(localtime(&now)));
4 printf("UTC : %s", asctime(gmtime(&now)));
Sample output: Local: Mon Mar 1 21:40:05 2021 UTC : Tue Mar 2 05:40:05 2021
See Also
gmtime(), asctime(), strftime()
44.11 strftime()
Formatted date and time output
Synopsis
#include
size_t strftime(char * restrict s, size_t maxsize, const char * restrict format, const struct tm * restrict timeptr);
Description
This is the sprintf() of date and time functions. It’ll take a struct tm and produce a string in just about whatever form you desire, for example: 2021-03-01 Monday, March 1 at 9:54 PM It's Monday! Chapter 44.
It’s a super flexible version of asctime(). And thread-safe, besides, since it doesn’t rely on a static buffer to hold the results. Basically what you do is give it a destination, s, and its max size in bytes in maxsize. Also, provide a format string that’s analogous to printf()’s format string, but with different format specifiers. And lastly, a struct tm with the broken-down time information to use for printing. The format string works like this, for example: "It's %A, %B %d!"
Which produces: It's Monday, March 1!
The %A is the full day-of-week name, the %B is the full month name, and the %d is the day of the month. strftime() substitutes the right thing to produce the result. Brilliant! So what are all the format specifiers? Glad you asked! I’m going to be lazy and just drop this table in right from the spec.
Specifier Description %a Locale’s abbreviated weekday name. [tm_wday] %A Locale’s full weekday name. [tm_wday] %b Locale’s abbreviated month name. [tm_mon] %B Locale’s full month name. [tm_mon] %c Locale’s appropriate date and time representation. %C Year divided by 100 and truncated to an integer, as a decimal number (00–99). [tm_year] %d Day of the month as a decimal number (01–31). [tm_mday] %D Equivalent to "%m/%d/%y".[tm_mon, tm_mday, tm_year] %e Day of the month as a decimal number (1–31); a single digit is preceded by a space. [tm_mday] %F Equivalent to “%Y-%m-%d” (the ISO 8601 date format). [tm_year, tm_mon, tm_mday] %g Last 2 digits of the week-based year (see below) as a decimal number (00–99). [tm_year, tm_wday, tm_yday] %G Week-based year (see below) as a decimal number (e.g., 1997). [tm_year, tm_wday, tm_yday] %h Equivalent to “%b”. [tm_mon] %H Hour (24-hour clock) as a decimal number (00–23). [tm_hour] %I Hour (12-hour clock) as a decimal number (01–12). [tm_hour] %j Day of the year as a decimal number (001–366). [tm_yday] %m Month as a decimal number (01–12). %M Minute as a decimal number (00–59). [tm_min] %n A new-line character. %p Locale’s equivalent of the AM/PM designations associated with a 12-hour clock. [tm_hour] %r Locale’s 12-hour clock time. [tm_hour, tm_min, tm_sec] %R Equivalent to "%H:%M".[tm_hour, tm_min] %S Second as a decimal number (00–60). [tm_sec] %t A horizontal-tab character. %T Equivalent to "%H:%M:%S" (the ISO 8601 time format). [tm_hour, tm_min, tm_sec] %u ISO 8601 weekday as a decimal number (1–7), where Monday is 1. [tm_wday] %U Week number of the year (the first Sunday as the first day of week 1) as a decimal number (00–53). [tm_year, tm_wday, tm_yday] %V ISO 8601 week number (see below) as a decimal number (01–53). [tm_year, tm_wday, tm_yday] %w Weekday as a decimal number (0–6), where Sunday is 0. %W Week number of the year (the first Monday as the first day of week 1) as a decimal number (00–53). [tm_year, tm_wday, tm_yday] %x Locale’s appropriate date representation. Chapter 44.
Specifier Description %X Locale’s appropriate time representation. %y Last 2 digits of the year as a decimal number (00–99). [tm_year] %Y Year as a decimal number (e.g., 1997). [tm_year] %z Offset from UTC in the ISO 8601 format "-0430" (meaning 4 hours 30 minutes behind UTC, west of Greenwich), or by no characters if no time zone is determinable. [tm_isdst] %Z Locale’s time zone name or abbreviation, or by no characters if no time zone is determinable. [tm_isdst] %% A plain ol’ %
Phew. That’s love. %G, %g, and %v are a little funky in that they use something called the ISO 8601 week-based year. I’d never heard of it. But, again stealing from the spec, these are the rules: %g, %G, and %V give values according to the ISO 8601 week-based year. In this system, weeks begin on a Monday and week 1 of the year is the week that includes January 4th, which is also the week that includes the first Thursday of the year, and is also the first week that contains at least four days in the year. If the first Monday of January is the 2nd, 3rd, or 4th, the preceding days are part of the last week of the preceding year; thus, for Saturday 2nd January 1999, %G is replaced by 1998 and %V is replaced by 53. If December 29th, 30th, or 31st is a Monday, it and any following days are part of week 1 of the following year. Thus, for Tuesday 30th December 1997, %G is replaced by 1998 and %V is replaced by 01. Learn something new every day! If you want to know more, Wikipedia has a page on it7. If you’re in the “C” locale, the specifiers produce the following (again, stolen from the spec):
Specifier Description %a The first three characters of %A. %A One of Sunday, Monday,…, Saturday. %b The first three characters of %B. %B One of January, February,…, December. %c Equivalent to %a %b %e %T %Y. %p One of AM or PM. %r Equivalent to %I:%M:%S %p. %x Equivalent to %m/%d/%y. %X Equivalent to %T. %Z Implementation-defined.
There are additional variants of the format specifiers that indicate you want to use a locale’s alternative format. These don’t exist for all locales. It’s one of the format specifies above, with either an E or O prefix: %Ec %EC %Ex %EX %Ey %EY %Od %Oe %OH %OI %Om %OM %OS %Ou %OU %OV %Ow %OW %Oy
The E and O prefixes are ignored in the “C” locale.
Return Value Returns the total number of bytes put into the result string, not including the NUL terminator. If the result doesn’t fit in the string, zero is returned and the value in s is indeterminate.
7https://en.wikipedia.org/wiki/ISO_week_date Chapter 44.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 char s[128];
7 time_t now = time(NULL);
8
9 // %c: print date as per current locale
10 strftime(s, sizeof s, "%c", localtime(&now));
11 puts(s); // Sun Feb 28 22:29:00 2021
12
13 // %A: full weekday name
14 // %B: full month name
15 // %d: day of the month
16 strftime(s, sizeof s, "%A, %B %d", localtime(&now));
17 puts(s); // Sunday, February 28
18
19 // %I: hour (12 hour clock)
20 // %M: minute
21 // %S: second
22 // %p: AM or PM
23 strftime(s, sizeof s, "It's %I:%M:%S %p", localtime(&now));
24 puts(s); // It's 10:29:00 PM
25
26 // %F: ISO 8601 yyyy-mm-dd
27 // %T: ISO 8601 hh:mm:ss
28 // %z: ISO 8601 timezone offset
29 strftime(s, sizeof s, "ISO 8601: %FT%T%z", localtime(&now));
30 puts(s); // ISO 8601: 2021-02-28T22:29:00-0800
31 }
See Also
ctime(), asctime() Chapter 45
Function Description isalnum() Tests if a character is alphabetic or is a digit isalpha() Returns true if a character is alphabetic isblank() Tests if a character is word-separating whitespace iscntrl() Test if a character is a control character isdigit() Tests if a character is a digit isgraph() Tests if the character is printable and not a space islower() Tests if a character is lowercase isprint() Tests if a character is printable ispunct() Test if a character is punctuation isspace() Test if a character is whitespace isupper() Tests if a character is uppercase isxdigit() Tests if a character is a hexadecimal digit tolower() Convert a letter to lowercase toupper() Convert a letter to uppercase
This collection of macros is good for testing characters to see if they’re of a certain class, such as alphabetic, numeric, control characters, etc. Surprisingly, they take int arguments instead of some kind of char. This is so you can feed EOF in for convenience if you have an integer representation of that. If not EOF, the value passed in has to be representable in an unsigned char. Otherwise it’s (dun dun DUUNNNN) undefined behavior. So you can forget about passing in your UTF-8 multibyte characters. Also, the behavior of these functions varies based on locale. In many of the pages in this section, I give some examples. These are from the “C” locale, and might vary if you’ve set a different locale. Note that wide characters have their own set of classification functions, so don’t try to use these on wchar_ts. Or else!
45.1 isalnum()
Tests if a character is alphabetic or is a digit
406 Chapter 45.
Synopsis
#include
int isalnum(int c);
Description
Tests if a character is alphabetic (A-Z or a-z) or a digit (0-9). Is equivalent to: isalpha(c) || isdigit(c)
Return Value
Returns true if a character is alphabetic (A-Z or a-z) or a digit (0-9).
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isalnum('a')? "yes": "no"); // yes
9 printf("%s\n", isalnum('B')? "yes": "no"); // yes
10 printf("%s\n", isalnum('5')? "yes": "no"); // yes
11 printf("%s\n", isalnum('?')? "yes": "no"); // no
12 }
See Also
isalpha(), isdigit()
45.2 isalpha()
Returns true if a character is alphabetic
Synopsis
#include
int isalpha(int c);
Description
Returns true for alphabetic characters (A-Z or a-z). Technically (and in the “C” locale) equivalent to: isupper(c) || islower(c)
Extra super technically, because I know you’re dying for this to be extra unnecessarily complex, it can also include some locale-specific characters for which this is true: Chapter 45.
!iscntrl(c) && !isdigit(c) && !ispunct(c) && !isspace(c)
and this is true: isupper(c) || islower(c)
Return Value
Returns true for alphabetic characters (A-Z or a-z). Or for any of the other crazy stuff in the description, above.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isalpha('a')? "yes": "no"); // yes
9 printf("%s\n", isalpha('B')? "yes": "no"); // yes
10 printf("%s\n", isalpha('5')? "yes": "no"); // no
11 printf("%s\n", isalpha('?')? "yes": "no"); // no
12 }
See Also
isalnum()
45.3 isblank()
Tests if a character is word-separating whitespace
Synopsis
#include
int isblank(int c);
Description True if the character is a whitespace character used to separate words in a single line. For example, space ('') or horizontal tab ('\t'). Other locales might define other blank characters.
Return Value Returns true if the character is a whitespace character used to separate words in a single line.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isblank('')? "yes": "no"); // yes
9 printf("%s\n", isblank('\t')? "yes": "no"); // yes
10 printf("%s\n", isblank('\n')? "yes": "no"); // no
11 printf("%s\n", isblank('a')? "yes": "no"); // no
12 printf("%s\n", isblank('?')? "yes": "no"); // no
13 }
See Also
isspace()
45.4 iscntrl()
Test if a character is a control character
Synopsis
#include
int iscntrl(int c);
Description A control character is a locale-specific non-printing character. For the “C” locale, this means control characters are in the range 0x00 to 0x1F (the character right before SPACE) and 0x7F (the DEL character). Basically if it’s not an ASCII (or Unicode less than 128) printable character, it’s a control character in the “C” locale.
Return Value
Returns true if c is a control character.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", iscntrl('\t')? "yes": "no"); // yes (tab)
9 printf("%s\n", iscntrl('\n')? "yes": "no"); // yes (newline)
10 printf("%s\n", iscntrl('\r')? "yes": "no"); // yes (return)
11 printf("%s\n", iscntrl('\a')? "yes": "no"); // yes (bell)
12 printf("%s\n", iscntrl('')? "yes": "no"); // no
13 printf("%s\n", iscntrl('a')? "yes": "no"); // no Chapter 45.
14 printf("%s\n", iscntrl('?')? "yes": "no"); // no
15 }
See Also
isgraph(), isprint()
45.5 isdigit()
Tests if a character is a digit
Synopsis
#include
int isdigit(int c);
Description
Tests if c is a digit in the range 0-9.
Return Value Returns true if the character is a digit, unsurprisingly.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isdigit('0')? "yes": "no"); // yes
9 printf("%s\n", isdigit('5')? "yes": "no"); // yes
10 printf("%s\n", isdigit('a')? "yes": "no"); // no
11 printf("%s\n", isdigit('B')? "yes": "no"); // no
12 printf("%s\n", isdigit('?')? "yes": "no"); // no
13 }
See Also
isalnum(), isxdigit()
45.6 isgraph()
Tests if the character is printable and not a space Chapter 45.
Synopsis
#include
int isgraph(int c);
Description
Tests if c is any printable character that isn’t a space ('').
Return Value
Returns true if c is any printable character that isn’t a space ('').
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isgraph('0')? "yes": "no"); // yes
9 printf("%s\n", isgraph('a')? "yes": "no"); // yes
10 printf("%s\n", isgraph('B')? "yes": "no"); // yes
11 printf("%s\n", isgraph('?')? "yes": "no"); // yes
12 printf("%s\n", isgraph('')? "yes": "no"); // no
13 printf("%s\n", isgraph('\n')? "yes": "no"); // no
14 }
See Also
iscntrl(), isprint()
45.7 islower()
Tests if a character is lowercase
Synopsis
#include
int islower(int c);
Description
Tests if a character is lowercase, in the range a-z. In other locales, there could be other lowercase characters. In all cases, to be lowercase, the following must be true: !iscntrl(c) && !isdigit(c) && !ispunct(c) && !isspace(c) Chapter 45.
Return Value Returns true if the character is lowercase.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", islower('c')? "yes": "no"); // yes
9 printf("%s\n", islower('0')? "yes": "no"); // no
10 printf("%s\n", islower('B')? "yes": "no"); // no
11 printf("%s\n", islower('?')? "yes": "no"); // no
12 printf("%s\n", islower('')? "yes": "no"); // no
13 }
See Also
isupper(), isalpha(), toupper(), tolower()
45.8 isprint()
Tests if a character is printable
Synopsis
#include
int isprint(int c);
Description
Tests if a character is printable, including space (''). So like isgraph(), except space isn’t left out in the cold.
Return Value
Returns true if the character is printable, including space ('').
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isprint('c')? "yes": "no"); // yes
9 printf("%s\n", isprint('0')? "yes": "no"); // yes
10 printf("%s\n", isprint('')? "yes": "no"); // yes Chapter 45.
11 printf("%s\n", isprint('\r')? "yes": "no"); // no
12 }
See Also
isgraph(), iscntrl()
45.9 ispunct()
Test if a character is punctuation
Synopsis
#include
int ispunct(int c);
Description Tests if a character is punctuation. In the “C” locale, this means: !isspace(c) && !isalnum(c)
In other locales, there could be other punctuation characters (but they also can’t be space or alphanumeric).
Return Value True if the character is punctuation.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", ispunct(',')? "yes": "no"); // yes
9 printf("%s\n", ispunct('!')? "yes": "no"); // yes
10 printf("%s\n", ispunct('c')? "yes": "no"); // no
11 printf("%s\n", ispunct('0')? "yes": "no"); // no
12 printf("%s\n", ispunct('')? "yes": "no"); // no
13 printf("%s\n", ispunct('\n')? "yes": "no"); // no
14 }
See Also
isspace(), isalnum() Chapter 45.
45.10 isspace()
Test if a character is whitespace
Synopsis
#include
int isspace(int c);
Description
Tests if c is a whitespace character. These are: • Space ('') • Formfeed ('\f') • Newline ('\n') • Carriage Return ('\r') • Horizontal Tab ('\t') • Vertical Tab ('\v') Other locales might specify other whitespace characters. isalnum() is false for all whitespace characters.
Return Value True if the character is whitespace.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isspace('')? "yes": "no"); // yes
9 printf("%s\n", isspace('\n')? "yes": "no"); // yes
10 printf("%s\n", isspace('\t')? "yes": "no"); // yes
11 printf("%s\n", isspace(',')? "yes": "no"); // no
12 printf("%s\n", isspace('!')? "yes": "no"); // no
13 printf("%s\n", isspace('c')? "yes": "no"); // no
14 }
See Also
isblank()
45.11 isupper()
Tests if a character is uppercase
Synopsis Chapter 45.
#include
int isupper(int c);
Description
Tests if a character is uppercase, in the range A-Z. In other locales, there could be other uppercase characters. In all cases, to be uppercase, the following must be true: !iscntrl(c) && !isdigit(c) && !ispunct(c) && !isspace(c)
Return Value Returns true if the character is uppercase.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isupper('B')? "yes": "no"); // yes
9 printf("%s\n", isupper('c')? "yes": "no"); // no
10 printf("%s\n", isupper('0')? "yes": "no"); // no
11 printf("%s\n", isupper('?')? "yes": "no"); // no
12 printf("%s\n", isupper('')? "yes": "no"); // no
13 }
See Also
islower(), isalpha(), toupper(), tolower()
45.12 isxdigit()
Tests if a character is a hexadecimal digit
Synopsis
#include
int isxdigit(int c);
Description
Returns true if the character is a hexadecimal digit. Namely if it’s 0-9, a-f, or A-F.
Return Value True if the character is a hexadecimal digit. Chapter 45.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 printf("%s\n", isxdigit('B')? "yes": "no"); // yes
9 printf("%s\n", isxdigit('c')? "yes": "no"); // yes
10 printf("%s\n", isxdigit('2')? "yes": "no"); // yes
11 printf("%s\n", isxdigit('G')? "yes": "no"); // no
12 printf("%s\n", isxdigit('?')? "yes": "no"); // no
13 }
See Also
isdigit()
45.13 tolower()
Convert a letter to lowercase
Synopsis
#include
int tolower(int c);
Description
If the character is uppercase (i.e. isupper(c) is true), this function returns the corresponding lowercase letter. Different locales might have different upper- and lowercase letters.
Return Value Returns the lowercase value for an uppercase letter. If the letter isn’t uppercase, returns it unchanged.
Example
1 // changing this char
2 // v
3 printf("%c\n", tolower('B')); // b (made lowercase!)
4 printf("%c\n", tolower('e')); // e (unchanged)
5 printf("%c\n", tolower('!')); // ! (unchanged)
See Also
toupper(), islower(), isupper() Chapter 45.
45.14 toupper()
Convert a letter to uppercase
Synopsis
#include
int toupper(int c);
Description
If the character is lower (i.e. islower(c) is true), this function returns the corresponding uppercase letter. Different locales might have different upper- and lowercase letters.
Return Value Returns the uppercase value for a lowercase letter. If the letter isn’t lowercase, returns it unchanged.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // changing this char
7 // v
8 printf("%c\n", toupper('B')); // B (unchanged)
9 printf("%c\n", toupper('e')); // E (made uppercase!)
10 printf("%c\n", toupper('!')); // ! (unchanged)
11 }
See Also
tolower(), islower(), isupper() Chapter 46
Function Description call_once() Call a function one time no matter how many threads try cnd_broadcast() Wake up all threads waiting on a condition variable cnd_destroy() Free up resources from a condition variable cnd_init() Initialize a condition variable to make it ready for use cnd_signal() Wake up a thread waiting on a condition variable cnd_timedwait() Wait on a condition variable with a timeout cnd_wait() Wait for a signal on a condition variable mtx_destroy() Cleanup a mutex when done with it mtx_init() Initialize a mutex for use mtx_lock() Acquire a lock on a mutex mtx_timedlock() Lock a mutex allowing for timeout mtx_trylock() Try to lock a mutex, returning if not possible mtx_unlock() Free a mutex when you’re done with the critical section thrd_create() Create a new thread of execution thrd_current() Get the ID of the calling thread thrd_detach() Automatically clean up threads when they exit thrd_equal() Compare two thread descriptors for equality thrd_exit() Stop and exit this thread thrd_join() Wait for a thread to exit thrd_yield() Stop running that other threads might run tss_create() Create new thread-specific storage tss_delete() Clean up a thread-specific storage variable tss_get() Get thread-specific data tss_set() Set thread-specific data
We have a bunch of good things at our disposal with this one: • Threads • Mutexes • Condition Variables • Thread-Specific Storage • And, last but not least, the always-fun call_once() function! Enjoy!
418 Chapter 46.
46.1 call_once()
Call a function one time no matter how many threads try
Synopsis
#include
void call_once(once_flag *flag, void (*func)(void));
Description If you have a bunch of threads running over the same piece of code that calls a function, but you only want that function to run one time, call_once() can help you out. The catch is the function that is called doesn’t return anything and takes no arguments. If you need more than that, you’ll have to set a threadsafe flag such as atomic_flag, or one that you protect with a mutex. To use this, you need to pass it a pointer to a function to execute, func, and also a pointer to a flag of type once_flag. once_flag is an opaque type, so all you need to know is that you initialize it to the value ONCE_FLAG_INIT.
Return Value Returns nothing.
Example
1 #include
2 #include
3
4 once_flag of = ONCE_FLAG_INIT; // Initialize it like this
5
6 void run_once_function(void)
7 {
8 printf("I'll only run once!\n");
9 }
10
11 int run(void *arg)
12 {
13 (void)arg;
14
15 printf("Thread running!\n");
16
17 call_once(&of, run_once_function);
18
19 return 0;
20 }
21
22 #define THREAD_COUNT 5
23
24 int main(void)
25 {
26 thrd_t t[THREAD_COUNT];
27 Chapter 46.
28 for (int i = 0; i < THREAD_COUNT; i++)
29 thrd_create(t + i, run, NULL);
30
31 for (int i = 0; i < THREAD_COUNT; i++)
32 thrd_join(t[i], NULL);
33 }
Output (might vary per run): Thread running! Thread running! I'll only run once! Thread running! Thread running! Thread running!
46.2 cnd_broadcast()
Wake up all threads waiting on a condition variable
Synopsis
#include
int cnd_broadcast(cnd_t *cond);
Description
This is just like cnd_signal() in that it wakes up threads that are waiting on a condition variable…. except instead of just rousing one thread, it wakes them all. Of course, only one will get the mutex, and the rest will have to wait their turn. But instead of being asleep waiting for a signal, they’ll be asleep waiting to reacquire the mutex. They’re rearin’ to go, in other words. This can make a difference in a specific set of circumstances where cnd_signal() might leave you hanging. If you’re relying on subsequent threads to issue the next cnd_signal(), but you have the cnd_wait() in a while loop1 that doesn’t allow any threads to escape, you’ll be stuck. No more threads will be woken up from the wait. But if you cnd_broadcast(), all the threads will be woken, and presumably at least one of them will be allowed to escape the while loop, freeing it up to broadcast the next wakeup when its work is done.
Return Value
Returns thrd_success or thrd_error depending on how well things went.
Example In the example below, we launch a bunch of threads, but they’re only allowed to run if their ID matches the current ID. If it doesn’t, they go back to waiting. If you cnd_signal() to wake the next thread, it might not be the one with the proper ID to run. If it’s not, it goes back to sleep and we hang (because no thread is awake to hit cnd_signal() again).
1Which you should because of spurious wakeups. Chapter 46.
But if you cnd_broadcast() to wake them all, then they’ll all try (one after another) to get out of the while loop. And one of them will make it. Try switching the cnd_broadcast() to cnd_signal() to see likely deadlocks. It doesn’t happen every time, but usually does.
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 int id = *(int*)arg;
10
11 static int current_id = 0;
12
13 mtx_lock(&mutex);
14
15 while (id != current_id) {
16 printf("THREAD %d: waiting\n", id);
17 cnd_wait(&condvar, &mutex);
18
19 if (id != current_id)
20 printf("THREAD %d: woke up, but it's not my turn!\n", id);
21 else
22 printf("THREAD %d: woke up, my turn! Let's go!\n", id);
23 }
24
25 current_id++;
26
27 printf("THREAD %d: signaling thread %d to run\n", id, current_id);
28
29 //cnd_signal(&condvar);
30 cnd_broadcast(&condvar);
31 mtx_unlock(&mutex);
32
33 return 0;
34 }
35
36 #define THREAD_COUNT 5
37
38 int main(void)
39 {
40 thrd_t t[THREAD_COUNT];
41 int id[] = {4, 3, 2, 1, 0};
42
43 mtx_init(&mutex, mtx_plain);
44 cnd_init(&condvar);
45
46 for (int i = 0; i < THREAD_COUNT; i++)
47 thrd_create(t + i, run, id + i);
48
49 for (int i = 0; i < THREAD_COUNT; i++)
50 thrd_join(t[i], NULL);
51
52 mtx_destroy(&mutex);
53 cnd_destroy(&condvar); Chapter 46.
54 }
Example run with cnd_broadcast(): THREAD 4: waiting THREAD 1: waiting THREAD 3: waiting THREAD 2: waiting THREAD 0: signaling thread 1 to run THREAD 2: woke up, but it's not my turn! THREAD 2: waiting THREAD 4: woke up, but it's not my turn! THREAD 4: waiting THREAD 3: woke up, but it's not my turn! THREAD 3: waiting THREAD 1: woke up, my turn! Let's go! THREAD 1: signaling thread 2 to run THREAD 4: woke up, but it's not my turn! THREAD 4: waiting THREAD 3: woke up, but it's not my turn! THREAD 3: waiting THREAD 2: woke up, my turn! Let's go! THREAD 2: signaling thread 3 to run THREAD 4: woke up, but it's not my turn! THREAD 4: waiting THREAD 3: woke up, my turn! Let's go! THREAD 3: signaling thread 4 to run THREAD 4: woke up, my turn! Let's go! THREAD 4: signaling thread 5 to run
Example run with cnd_signal(): THREAD 4: waiting THREAD 1: waiting THREAD 3: waiting THREAD 2: waiting THREAD 0: signaling thread 1 to run THREAD 4: woke up, but it's not my turn! THREAD 4: waiting
[deadlock at this point]
See how THREAD 0 signaled that it was THREAD 1’s turn? But—bad news—it was THREAD 4 that got woken up. So no one continued the process. cnd_broadcast() would have woken them all, so eventually THREAD 1 would have run, gotten out of the while, and broadcast for the next thread to run.
See Also
cnd_signal(), mtx_lock(), mtx_unlock()
46.3 cnd_destroy()
Free up resources from a condition variable Chapter 46.
Synopsis
#include
void cnd_destroy(cnd_t *cond);
Description
This is the opposite of cnd_init() and should be called when all threads are done using a condition variable.
Return Value Returns nothing!
Example
General-purpose condition variable example here, but you can see the cnd_destroy() down at the end.
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 mtx_lock(&mutex);
12
13 printf("Thread: waiting...\n");
14 cnd_wait(&condvar, &mutex);
15 printf("Thread: running again!\n");
16
17 mtx_unlock(&mutex);
18
19 return 0;
20 }
21
22 int main(void)
23 {
24 thrd_t t;
25
26 mtx_init(&mutex, mtx_plain);
27 cnd_init(&condvar);
28
29 printf("Main creating thread\n");
30 thrd_create(&t, run, NULL);
31
32 // Sleep 0.1s to allow the other thread to wait
33 thrd_sleep(&(struct timespec){.tv_nsec=100000000L}, NULL);
34
35 mtx_lock(&mutex);
36 printf("Main: signaling thread\n");
37 cnd_signal(&condvar);
38 mtx_unlock(&mutex);
39 Chapter 46.
40 thrd_join(t, NULL);
41
42 mtx_destroy(&mutex);
43 cnd_destroy(&condvar); // <-- DESTROY CONDITION VARIABLE
44 }
Output: Main creating thread Thread: waiting... Main: signaling thread Thread: running again!
See Also
cnd_init()
46.4 cnd_init()
Initialize a condition variable to make it ready for use
Synopsis
#include
int cnd_init(cnd_t *cond);
Description
This is the opposite of cnd_destroy(). This prepares a condition variable for use, doing behind-the- scenes work on it. Don’t use a condition variable without calling this first!
Return Value
If all goes well, returns thrd_success. It all doesn’t go well, it could return thrd_nomem if the system is out of memory, or thread_error in the case of any other error.
Example
General-purpose condition variable example here, but you can see the cnd_init() down at the start of main().
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 mtx_lock(&mutex);
12 Chapter 46.
13 printf("Thread: waiting...\n");
14 cnd_wait(&condvar, &mutex);
15 printf("Thread: running again!\n");
16
17 mtx_unlock(&mutex);
18
19 return 0;
20 }
21
22 int main(void)
23 {
24 thrd_t t;
25
26 mtx_init(&mutex, mtx_plain);
27 cnd_init(&condvar); // <-- INITIALIZE CONDITION VARIABLE
28
29 printf("Main creating thread\n");
30 thrd_create(&t, run, NULL);
31
32 // Sleep 0.1s to allow the other thread to wait
33 thrd_sleep(&(struct timespec){.tv_nsec=100000000L}, NULL);
34
35 mtx_lock(&mutex);
36 printf("Main: signaling thread\n");
37 cnd_signal(&condvar);
38 mtx_unlock(&mutex);
39
40 thrd_join(t, NULL);
41
42 mtx_destroy(&mutex);
43 cnd_destroy(&condvar);
44 }
Output: Main creating thread Thread: waiting... Main: signaling thread Thread: running again!
See Also
cnd_destroy()
46.5 cnd_signal()
Wake up a thread waiting on a condition variable
Synopsis
#include
int cnd_signal(cnd_t *cond); Chapter 46.
Description If you have a thread (or a bunch of threads) waiting on a condition variable, this function will wake one of them up to run. Compare to cnd_broadcast() that wakes up all the threads. See the cnd_broadcast() page for more information on when you’re want to use that versus this.
Return Value
Returns thrd_success or thrd_error depending on how happy your program is.
Example
General-purpose condition variable example here, but you can see the cnd_signal() in the middle of main().
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 mtx_lock(&mutex);
12
13 printf("Thread: waiting...\n");
14 cnd_wait(&condvar, &mutex);
15 printf("Thread: running again!\n");
16
17 mtx_unlock(&mutex);
18
19 return 0;
20 }
21
22 int main(void)
23 {
24 thrd_t t;
25
26 mtx_init(&mutex, mtx_plain);
27 cnd_init(&condvar);
28
29 printf("Main creating thread\n");
30 thrd_create(&t, run, NULL);
31
32 // Sleep 0.1s to allow the other thread to wait
33 thrd_sleep(&(struct timespec){.tv_nsec=100000000L}, NULL);
34
35 mtx_lock(&mutex);
36 printf("Main: signaling thread\n");
37 cnd_signal(&condvar); // <-- SIGNAL CHILD THREAD HERE!
38 mtx_unlock(&mutex);
39
40 thrd_join(t, NULL);
41
42 mtx_destroy(&mutex); Chapter 46.
43 cnd_destroy(&condvar);
44 }
Output: Main creating thread Thread: waiting... Main: signaling thread Thread: running again!
See Also
cnd_init(), cnd_destroy()
46.6 cnd_timedwait()
Wait on a condition variable with a timeout
Synopsis
#include
int cnd_timedwait(cnd_t *restrict cond, mtx_t *restrict mtx, const struct timespec *restrict ts);
Description
This is like cnd_wait() except we get to specify a timeout, as well. Note that the thread still must reacquire the mutex to get more work done even after the timeout. The the main difference is that regular cnd_wait() will only try to get the mutex after a cnd_signal() or cnd_broadcast(), whereas cnd_timedwait() will do that, too, and try to get the mutex after the timeout. The timeout is specified as an absolute UTC time since Epoch. You can get this with the timespec_get() function and then add values on to the result to timeout later than now, as shown in the example. Beware that you can’t have more than 999999999 nanoseconds in the tv_nsec field of the struct time- spec. Mod those so they stay in range.
Return Value
If the thread wakes up for a non-timeout reason (e.g. signal or broadcast), returns thrd_success. If woken up due to timeout, returns thrd_timedout. Otherwise returns thrd_error.
Example This example has a thread wait on a condition variable for a maximum of 1.75 seconds. And it always times out because no one ever sends a signal. Tragic.
1 #include
2 #include
3 #include
4
5 cnd_t condvar;
6 mtx_t mutex;
7 Chapter 46.
8 int run(void *arg)
9 {
10 (void)arg;
11
12 mtx_lock(&mutex);
13
14 struct timespec ts;
15
16 // Get the time now
17 timespec_get(&ts, TIME_UTC);
18
19 // Add on 1.75 seconds from now
20 ts.tv_sec += 1;
21 ts.tv_nsec += 750000000L;
22
23 // Handle nsec overflow
24 ts.tv_sec += ts.tv_nsec / 1000000000L;
25 ts.tv_nsec = ts.tv_nsec % 1000000000L;
26
27 printf("Thread: waiting...\n");
28 int r = cnd_timedwait(&condvar, &mutex, &ts);
29
30 switch (r) {
31 case thrd_success:
32 printf("Thread: signaled!\n");
33 break;
34
35 case thrd_timedout:
36 printf("Thread: timed out!\n");
37 return 1;
38
39 case thrd_error:
40 printf("Thread: Some kind of error\n");
41 return 2;
42 }
43
44 mtx_unlock(&mutex);
45
46 return 0;
47 }
48
49 int main(void)
50 {
51 thrd_t t;
52
53 mtx_init(&mutex, mtx_plain);
54 cnd_init(&condvar);
55
56 printf("Main creating thread\n");
57 thrd_create(&t, run, NULL);
58
59 // Sleep 3s to allow the other thread to timeout
60 thrd_sleep(&(struct timespec){.tv_sec=3}, NULL);
61
62 thrd_join(t, NULL);
63
64 mtx_destroy(&mutex);
65 cnd_destroy(&condvar); Chapter 46.
66 }
Output: Main creating thread Thread: waiting... Thread: timed out!
See Also
cnd_wait(), timespec_get()
46.7 cnd_wait()
Wait for a signal on a condition variable
Synopsis
#include
int cnd_wait(cnd_t *cond, mtx_t *mtx);
Description
This puts the calling thread to sleep until it is awakened by a call to cnd_signal() or cnd_broadcast().
Return Value
If everything’s fantastic, returns thrd_success. Otherwise it returns thrd_error to report that some- thing has gone fantastically, horribly awry.
Example
General-purpose condition variable example here, but you can see the cnd_wait() in the run() function.
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 mtx_lock(&mutex);
12
13 printf("Thread: waiting...\n");
14 cnd_wait(&condvar, &mutex); // <-- WAIT HERE!
15 printf("Thread: running again!\n");
16
17 mtx_unlock(&mutex);
18
19 return 0;
20 }
21 Chapter 46.
22 int main(void)
23 {
24 thrd_t t;
25
26 mtx_init(&mutex, mtx_plain);
27 cnd_init(&condvar);
28
29 printf("Main creating thread\n");
30 thrd_create(&t, run, NULL);
31
32 // Sleep 0.1s to allow the other thread to wait
33 thrd_sleep(&(struct timespec){.tv_nsec=100000000L}, NULL);
34
35 mtx_lock(&mutex);
36 printf("Main: signaling thread\n");
37 cnd_signal(&condvar); // <-- SIGNAL CHILD THREAD HERE!
38 mtx_unlock(&mutex);
39
40 thrd_join(t, NULL);
41
42 mtx_destroy(&mutex);
43 cnd_destroy(&condvar);
44 }
Output: Main creating thread Thread: waiting... Main: signaling thread Thread: running again!
See Also
cnd_timedwait()
46.8 mtx_destroy()
Cleanup a mutex when done with it
Synopsis
#include
void mtx_destroy(mtx_t *mtx);
Description
The opposite of mtx_init(), this function frees up any resources associated with the given mutex. You should call this when all threads are done using the mutex.
Return Value Returns nothing, the selfish ingrate! Chapter 46.
Example
General-purpose mutex example here, but you can see the mtx_destroy() down at the end.
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 static int count = 0;
12
13 mtx_lock(&mutex);
14
15 printf("Thread: I got %d!\n", count);
16 count++;
17
18 mtx_unlock(&mutex);
19
20 return 0;
21 }
22
23 #define THREAD_COUNT 5
24
25 int main(void)
26 {
27 thrd_t t[THREAD_COUNT];
28
29 mtx_init(&mutex, mtx_plain);
30
31 for (int i = 0; i < THREAD_COUNT; i++)
32 thrd_create(t + i, run, NULL);
33
34 for (int i = 0; i < THREAD_COUNT; i++)
35 thrd_join(t[i], NULL);
36
37 mtx_destroy(&mutex); // <-- DESTROY THE MUTEX HERE
38 }
Output: Thread: I got 0! Thread: I got 1! Thread: I got 2! Thread: I got 3! Thread: I got 4!
See Also
mtx_init()
46.9 mtx_init()
Initialize a mutex for use Chapter 46.
Synopsis
#include
int mtx_init(mtx_t *mtx, int type);
Description Before you can use a mutex variable, you have to initialize it with this call to get it all prepped and ready to go. But wait! It’s not quite that simple. You have to tell it what type of mutex you want to create.
Type Description mtx_plain Regular ol’ mutex mtx_timed Mutex that supports timeouts mtx_plain|mtx_recursive Recursive mutex mtx_timed|mtx_recursive Recursive mutex that supports timeouts
As you can see, you can make a plain or timed mutex recursive by bitwise-ORing the value with mtx_recursive. “Recursive” means that the holder of a lock can call mtx_lock() multiple times on the same lock. (They have to unlock it an equal number of times before anyone else can take the mutex.) This might ease coding from time to time, especially if you call a function that needs to lock the mutex when you already hold the mutex. And the timeout gives a thread a chance to try to get the lock for a while, but then bail out if it can’t get it in that timeframe. You use the mtx_timedlock() function with mtx_timed mutexes.
Return Value
Returns thrd_success in a perfect world, and potentially thrd_error in an imperfect one.
Example
General-purpose mutex example here, but you can see the mtx_init() down at the top of main():
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 static int count = 0;
12
13 mtx_lock(&mutex);
14
15 printf("Thread: I got %d!\n", count);
16 count++;
17
18 mtx_unlock(&mutex);
19
20 return 0; Chapter 46.
21 }
22
23 #define THREAD_COUNT 5
24
25 int main(void)
26 {
27 thrd_t t[THREAD_COUNT];
28
29 mtx_init(&mutex, mtx_plain); // <-- CREATE THE MUTEX HERE
30
31 for (int i = 0; i < THREAD_COUNT; i++)
32 thrd_create(t + i, run, NULL);
33
34 for (int i = 0; i < THREAD_COUNT; i++)
35 thrd_join(t[i], NULL);
36
37 mtx_destroy(&mutex); // <-- DESTROY THE MUTEX HERE
38 }
Output: Thread: I got 0! Thread: I got 1! Thread: I got 2! Thread: I got 3! Thread: I got 4!
See Also
mtx_destroy()
46.10 mtx_lock()
Acquire a lock on a mutex
Synopsis
#include
int mtx_lock(mtx_t *mtx);
Description If you’re a thread and want to enter a critical section, do I have the function for you! A thread that calls this function will wait until it can acquire the mutex, then it will grab it, wake up, and run! If the mutex is recursive and is already locked by this thread, it will be locked again and the lock count will increase. If the mutex is not recursive and the thread already holds it, this call will error out.
Return Value
Returns thrd_success on goodness and thrd_error on badness. Chapter 46.
Example
General-purpose mutex example here, but you can see the mtx_lock() in the run() function:
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 static int count = 0;
12
13 mtx_lock(&mutex); // <-- LOCK HERE
14
15 printf("Thread: I got %d!\n", count);
16 count++;
17
18 mtx_unlock(&mutex);
19
20 return 0;
21 }
22
23 #define THREAD_COUNT 5
24
25 int main(void)
26 {
27 thrd_t t[THREAD_COUNT];
28
29 mtx_init(&mutex, mtx_plain); // <-- CREATE THE MUTEX HERE
30
31 for (int i = 0; i < THREAD_COUNT; i++)
32 thrd_create(t + i, run, NULL);
33
34 for (int i = 0; i < THREAD_COUNT; i++)
35 thrd_join(t[i], NULL);
36
37 mtx_destroy(&mutex); // <-- DESTROY THE MUTEX HERE
38 }
Output: Thread: I got 0! Thread: I got 1! Thread: I got 2! Thread: I got 3! Thread: I got 4!
See Also
mtx_unlock(), mtx_trylock(), mtx_timedlock()
46.11 mtx_timedlock()
Lock a mutex allowing for timeout Chapter 46.
Synopsis
#include
int mtx_timedlock(mtx_t *restrict mtx, const struct timespec *restrict ts);
Description
This is just like mtx_lock() except you can add a timeout if you don’t want to wait forever. The timeout is specified as an absolute UTC time since Epoch. You can get this with the timespec_get() function and then add values on to the result to timeout later than now, as shown in the example. Beware that you can’t have more than 999999999 nanoseconds in the tv_nsec field of the struct time- spec. Mod those so they stay in range.
Return Value
If everything works and the mutex is obtained, returns thrd_success. If a timeout happens first, returns thrd_timedout. Otherwise, returns thrd_error. Because if nothing is right, everything is wrong.
Example This example has a thread wait on a mutex for a maximum of 1.75 seconds. And it always times out because no one ever sends a signal.
1 #include
2 #include
3 #include
4
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 struct timespec ts;
12
13 // Get the time now
14 timespec_get(&ts, TIME_UTC);
15
16 // Add on 1.75 seconds from now
17 ts.tv_sec += 1;
18 ts.tv_nsec += 750000000L;
19
20 // Handle nsec overflow
21 ts.tv_sec += ts.tv_nsec / 1000000000L;
22 ts.tv_nsec = ts.tv_nsec % 1000000000L;
23
24 printf("Thread: waiting for lock...\n");
25 int r = mtx_timedlock(&mutex, &ts);
26
27 switch (r) {
28 case thrd_success:
29 printf("Thread: grabbed lock!\n");
30 break;
31
32 case thrd_timedout: Chapter 46.
33 printf("Thread: timed out!\n");
34 break;
35
36 case thrd_error:
37 printf("Thread: Some kind of error\n");
38 break;
39 }
40
41 mtx_unlock(&mutex);
42
43 return 0;
44 }
45
46 int main(void)
47 {
48 thrd_t t;
49
50 mtx_init(&mutex, mtx_plain);
51
52 mtx_lock(&mutex);
53
54 printf("Main creating thread\n");
55 thrd_create(&t, run, NULL);
56
57 // Sleep 3s to allow the other thread to timeout
58 thrd_sleep(&(struct timespec){.tv_sec=3}, NULL);
59
60 mtx_unlock(&mutex);
61
62 thrd_join(t, NULL);
63
64 mtx_destroy(&mutex);
65 }
Output: Main creating thread Thread: waiting for lock... Thread: timed out!
See Also
mtx_lock(), mtx_trylock(), timespec_get()
46.12 mtx_trylock()
Try to lock a mutex, returning if not possible
Synopsis
#include
int mtx_trylock(mtx_t *mtx); Chapter 46.
Description
This works just like mtx_lock except that it returns instantly if a lock can’t be obtained. The spec notes that there’s a chance that mtx_trylock() might spuriously fail with thrd_busy even if there are no other threads holding the lock. I’m not sure why this is, but you should defensively code against it.
Return Value
Returns thrd_success if all’s well. Or thrd_busy if some other thread holds the lock. Or thrd_error, which means something went right. I mean “wrong”.
Example
1 #include
2 #include
3 #include
4
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 int id = *(int*)arg;
10
11 int r = mtx_trylock(&mutex); // <-- TRY TO GRAB THE LOCK
12
13 switch (r) {
14 case thrd_success:
15 printf("Thread %d: grabbed lock!\n", id);
16 break;
17
18 case thrd_busy:
19 printf("Thread %d: lock already taken :(\n", id);
20 return 1;
21
22 case thrd_error:
23 printf("Thread %d: Some kind of error\n", id);
24 return 2;
25 }
26
27 mtx_unlock(&mutex);
28
29 return 0;
30 }
31
32 #define THREAD_COUNT 5
33
34 int main(void)
35 {
36 thrd_t t[THREAD_COUNT];
37 int id[THREAD_COUNT];
38
39 mtx_init(&mutex, mtx_plain);
40
41 for (int i = 0; i < THREAD_COUNT; i++) {
42 id[i] = i;
43 thrd_create(t + i, run, id + i);
44 } Chapter 46.
45
46 for (int i = 0; i < THREAD_COUNT; i++)
47 thrd_join(t[i], NULL);
48
49 mtx_destroy(&mutex);
50 }
Output (varies by run): Thread 0: grabbed lock! Thread 1: lock already taken :( Thread 4: lock already taken :( Thread 3: grabbed lock! Thread 2: lock already taken :(
See Also
mtx_lock(), mtx_timedlock(), mtx_unlock()
46.13 mtx_unlock()
Free a mutex when you’re done with the critical section
Synopsis
#include
int mtx_unlock(mtx_t *mtx);
Description After you’ve done all the dangerous stuff you have to do, wherein the involved threads should not be stepping on each other’s toes… you can free up your stranglehold on the mutex by calling mtx_unlock().
Return Value
Returns thrd_success on success. Or thrd_error on error. It’s not very original in this regard.
Example
General-purpose mutex example here, but you can see the mtx_unlock() in the run() function:
1 #include
2 #include
3
4 cnd_t condvar;
5 mtx_t mutex;
6
7 int run(void *arg)
8 {
9 (void)arg;
10
11 static int count = 0;
12
13 mtx_lock(&mutex);
14 Chapter 46.
15 printf("Thread: I got %d!\n", count);
16 count++;
17
18 mtx_unlock(&mutex); // <-- UNLOCK HERE
19
20 return 0;
21 }
22
23 #define THREAD_COUNT 5
24
25 int main(void)
26 {
27 thrd_t t[THREAD_COUNT];
28
29 mtx_init(&mutex, mtx_plain);
30
31 for (int i = 0; i < THREAD_COUNT; i++)
32 thrd_create(t + i, run, NULL);
33
34 for (int i = 0; i < THREAD_COUNT; i++)
35 thrd_join(t[i], NULL);
36
37 mtx_destroy(&mutex);
38 }
Output: Thread: I got 0! Thread: I got 1! Thread: I got 2! Thread: I got 3! Thread: I got 4!
See Also
mtx_lock(), mtx_timedlock(), mtx_trylock()
46.14 thrd_create()
Create a new thread of execution
Synopsis
#include
int thrd_create(thrd_t *thr, thrd_start_t func, void *arg);
Description Now you have the POWER! Right? This is how you launch new threads to make your program do multiple things at once2!
2Well, as at least as many things as you have free cores. Your OS will schedule them as it can. Chapter 46.
In order to make this happen, you need to pass a pointer to a thrd_t that will be used to represent the thread you’re spawning. That thread will start running the function you pass a pointer to in func. This is a value of type thrd_start_t, which is a pointer to a function that returns an int and takes a single void* as a parameter, i.e.: int thread_run_func(void *arg)
And, as you might have guessed, the pointer you pass to thrd_create() for the arg parameter is passed on to the func function. This is how you can give additional information to the thread when it starts up. Of course, for arg, you have to be sure to pass a pointer to an object that is thread-safe or per-thread. If the thread returns from the function, it exits just as if it had called thrd_exit(). Finally, the value that the func function returns can be picked up by the parent thread with thrd_join().
Return Value
In the case of goodness, returns thrd_success. If you’re out of memory, will return thrd_nomem. Oth- erwise, thrd_error.
Example
1 #include
2 #include
3
4 int run(void *arg)
5 {
6 int id = *(int*)arg;
7
8 printf("Thread %d: I'm alive!!\n", id);
9
10 return id;
11 }
12
13 #define THREAD_COUNT 5
14
15 int main(void)
16 {
17 thrd_t t[THREAD_COUNT];
18 int id[THREAD_COUNT]; // One of these per thread
19
20 for (int i = 0; i < THREAD_COUNT; i++) {
21 id[i] = i; // Let's pass in the thread number as the ID
22 thrd_create(t + i, run, id + i);
23 }
24
25 for (int i = 0; i < THREAD_COUNT; i++) {
26 int res;
27
28 thrd_join(t[i], &res);
29
30 printf("Main: thread %d exited with code %d\n", i, res);
31 }
32 }
Output (might vary from run to run): Chapter 46.
Thread 1: I'm alive!! Thread 0: I'm alive!! Thread 3: I'm alive!! Thread 2: I'm alive!! Main: thread 0 exited with code 0 Main: thread 1 exited with code 1 Main: thread 2 exited with code 2 Main: thread 3 exited with code 3 Thread 4: I'm alive!! Main: thread 4 exited with code 4
See Also thrd_exit(), thrd_join()
46.15 thrd_current()
Get the ID of the calling thread
Synopsis
#include
Description
Each thread has an opaque ID of type thrd_t. This is the value we see get initialized when we call thrd_create(). But what if you want to get the ID of the currently running thread? No problem! Just call this function and it will be returned to you. Why? Who knows! Well, to be honest, I could see it being used a couple places. 1. You could use it to have a thread detach itself with thrd_detach(). I’m not sure why you’d want to do this, however. 2. You could use it to compare this thread’s ID with another you have stored in a variable somewhere by using the thrd_equal() function. Seems like the most legit use. 3. … 4. Profit! If anyone has another use, please let me know.
Return Value Returns the calling thread’s ID.
Example Here’s a general example that shows getting the current thread ID and comparing it to a previously- recorded thread ID and taking exciting action based on the result! Starring Arnold Schwarzenegger! Chapter 46.
1 #include
2 #include
3
4 thrd_t first_thread_id;
5
6 int run(void *arg)
7 {
8 (void)arg;
9
10 thrd_t my_id = thrd_current(); // <-- GET MY THREAD ID
11
12 if (thrd_equal(my_id, first_thread_id))
13 printf("I'm the first thread!\n");
14 else
15 printf("I'm not the first!\n");
16
17 return 0;
18 }
19
20 int main(void)
21 {
22 thrd_t t;
23
24 thrd_create(&first_thread_id, run, NULL);
25 thrd_create(&t, run, NULL);
26
27 thrd_join(first_thread_id, NULL);
28 thrd_join(t, NULL);
29 }
Output: Come on, you got what you want, Cohaagen! Give deez people ay-ah!
No, wait, that’s an Arnold Schwarzenegger quote from Total Recall, one of the best science fiction films of all time. Watch it now and then come back to finish this reference page. Man–what an ending! And Johnny Cab? So excellent. Anyway! Output: I'm the first thread! I'm not the first!
See Also
thrd_equal(), thrd_detach()
46.16 thrd_detach()
Automatically clean up threads when they exit
Synopsis
#include
int thrd_detach(thrd_t thr); Chapter 46.
Description
Normally you have to thrd_join() to get resources associated with a deceased thread cleaned up. (Most notably, its exit status is still floating around waiting to get picked up.) But if you call thrd_detach() on the thread first, manual cleanup isn’t necessary. They just exit and are cleaned up by the OS. (Note that when the main thread dies, all the threads die in any case.)
Return Value
thrd_success if the thread successfully detaches, thrd_error otherwise.
Example
1 #include
2 #include
3
4 thrd_t first_thread_id;
5
6 int run(void *arg)
7 {
8 (void)arg;
9
10 printf("Thread running!\n");
11
12 return 0;
13 }
14
15 #define THREAD_COUNT 5
16
17 int main(void)
18 {
19 thrd_t t;
20
21 for (int i = 0; i < THREAD_COUNT; i++) {
22 thrd_create(&t, run, NULL);
23 thrd_detach(t);
24 }
25
26 // No need to thrd_join()!
27
28 // Sleep a quarter second to let them all finish
29 thrd_sleep(&(struct timespec){.tv_nsec=250000000}, NULL);
30 }
See Also
thrd_join(), thrd_exit()
46.17 thrd_equal()
Compare two thread descriptors for equality Chapter 46.
Synopsis
#include
int thrd_equal(thrd_t thr0, thrd_t thr1);
Description
If you have two thread descriptors in thrd_t variables, you can test them for equality with this function. For example, maybe one of the threads has special powers the others don’t, and the run function needs to be able to tell them apart, as in the example.
Return Value
Returns non-zero if the threads are equal. Returns 0 if they’re not.
Example Here’s a general example that shows getting the current thread ID and comparing it to a previously- recorded thread ID and taking boring action based on the result.
1 #include
2 #include
3
4 thrd_t first_thread_id;
5
6 int run(void *arg)
7 {
8 (void)arg;
9
10 thrd_t my_id = thrd_current();
11
12 if (thrd_equal(my_id, first_thread_id)) // <-- COMPARE!
13 printf("I'm the first thread!\n");
14 else
15 printf("I'm not the first!\n");
16
17 return 0;
18 }
19
20 int main(void)
21 {
22 thrd_t t;
23
24 thrd_create(&first_thread_id, run, NULL);
25 thrd_create(&t, run, NULL);
26
27 thrd_join(first_thread_id, NULL);
28 thrd_join(t, NULL);
29 }
Output: I'm the first thread! I'm not the first!
See Also
thrd_current() Chapter 46.
46.18 thrd_exit()
Stop and exit this thread
Synopsis
#include
_Noreturn void thrd_exit(int res);
Description A thread commonly exits by returning from its run function. But if it wants to exit early (perhaps from deeper in the call stack), this function will get that done. The res code can be picked up by a thread calling thrd_join(), and is equivalent to returning a value from the run function. Like with returning from the run function, this will also properly clean up all the thread-specific storage associated with this thread—all the destructors for the threads TSS variables will be called. If there are any remaining TSS variables with destructors after the first round of destruction3, the remaining destructors will be called. This happens repeatedly until there are no more, or the number of rounds of carnage reaches TSS_DTOR_ITERATIONS. If the main thread calls this, it’s as if you called exit(EXIT_SUCCESS).
Return Value This function never returns because the thread calling it is killed in the process. Trippy!
Example
Threads in this example exit early with result 22 if they get a NULL value for arg.
1 #include
2 #include
3
4 thrd_t first_thread_id;
5
6 int run(void *arg)
7 {
8 (void)arg;
9
10 if (arg == NULL)
11 thrd_exit(22);
12
13 return 0;
14 }
15
16 #define THREAD_COUNT 5
17
18 int main(void)
19 {
20 thrd_t t[THREAD_COUNT];
21
3For example, if a destructor caused more variables to be set. Chapter 46.
22 for (int i = 0; i < THREAD_COUNT; i++)
23 thrd_create(t + i, run, i == 2? NULL: "spatula");
24
25
26 for (int i = 0; i < THREAD_COUNT; i++) {
27 int res;
28 thrd_join(t[i], &res);
29
30 printf("Thread %d exited with code %d\n", i, res);
31 }
32 }
Output: Thread 0 exited with code 0 Thread 1 exited with code 0 Thread 2 exited with code 22 Thread 3 exited with code 0 Thread 4 exited with code 0
See Also
thrd_join()
46.19 thrd_join()
Wait for a thread to exit
Synopsis
#include
int thrd_join(thrd_t thr, int *res);
Description When a parent thread fires off some child threads, it can wait for them to complete with this call
Return Value Example
Threads in this example exit early with result 22 if they get a NULL value for arg. The parent thread picks up this result code with thrd_join().
1 #include
2 #include
3
4 thrd_t first_thread_id;
5
6 int run(void *arg)
7 {
8 (void)arg;
9
10 if (arg == NULL)
11 thrd_exit(22); Chapter 46.
12
13 return 0;
14 }
15
16 #define THREAD_COUNT 5
17
18 int main(void)
19 {
20 thrd_t t[THREAD_COUNT];
21
22 for (int i = 0; i < THREAD_COUNT; i++)
23 thrd_create(t + i, run, i == 2? NULL: "spatula");
24
25
26 for (int i = 0; i < THREAD_COUNT; i++) {
27 int res;
28 thrd_join(t[i], &res);
29
30 printf("Thread %d exited with code %d\n", i, res);
31 }
32 }
Output: Thread 0 exited with code 0 Thread 1 exited with code 0 Thread 2 exited with code 22 Thread 3 exited with code 0 Thread 4 exited with code 0
See Also
thrd_exit()
46.20 thrd_sleep()
Sleep for a specific number of seconds and nanoseconds
Synopsis
#include
int thrd_sleep(const struct timespec *duration, struct timespec *remaining);
Description This function puts the current thread to sleep for a while4 allowing other threads to run. The calling thread will wake up after the time has elapsed, or if it gets interrupted by a signal or something. If it doesn’t get interrupted, it’ll sleep at least as long as you asked. Maybe a tad longer. You know how hard it can be to get out of bed. The structure looks like this: struct timespec { time_t tv_sec; // Seconds
4Unix-like systems have a sleep() syscall that sleeps for an integer number of seconds. But thrd_sleep() is likely more portable and gives subsecond resolution, besides! Chapter 46.
long tv_nsec; // Nanoseconds (billionths of a second) };
Don’t set tv_nsec greater than 999,999,999. I can’t see what officially happens if you do, but on my system thrd_sleep() returns -2 and fails.
Return Value
Returns 0 on timeout, or -1 if interrupted by a signal. Or any negative value on some other error. Weirdly, the spec allows this “other error negative value” to also be -1, so good luck with that.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // Sleep for 3.25 seconds
7 thrd_sleep(&(struct timespec){.tv_sec=3, .tv_nsec=250000000}, NULL);
8
9 return 0;
10 }
See Also
thrd_yield()
46.21 thrd_yield()
Stop running that other threads might run
Synopsis
#include
void thrd_yield(void);
Description If you have a thread that’s hogging the CPU and you want to give your other threads time to run, you can call thrd_yield(). If the system sees fit, it will put the calling thread to sleep and one of the other threads will run instead. It’s a good way to be “polite” to the other threads in your program if you want the encourage them to run instead.
Return Value Returns nothing!
Example This example’s kinda poor because the OS is probably going to reschedule threads on the output anyway, but it gets the point across. Chapter 46.
The main thread is giving other threads a chance to run after every block of dumb work it does.
1 #include
2 #include
3
4 int run(void *arg)
5 {
6 int main_thread = arg != NULL;
7
8 if (main_thread) {
9 long int total = 0;
10
11 for (int i = 0; i < 10; i++) {
12 for (long int j = 0; j < 1000L; j++)
13 total++;
14
15 printf("Main thread yielding\n");
16 thrd_yield(); // <-- YIELD HERE
17 }
18 } else
19 printf("Other thread running!\n");
20
21 return 0;
22 }
23
24 #define THREAD_COUNT 10
25
26 int main(void)
27 {
28 thrd_t t[THREAD_COUNT];
29
30 for (int i = 0; i < THREAD_COUNT; i++)
31 thrd_create(t + i, run, i == 0? "main": NULL);
32
33 for (int i = 0; i < THREAD_COUNT; i++)
34 thrd_join(t[i], NULL);
35
36 return 0;
37 }
The output will vary from run to run. Notice that even after thrd_yield() other threads might not yet be ready to run and the main thread will continue. Main thread yielding Main thread yielding Main thread yielding Other thread running! Other thread running! Other thread running! Other thread running! Main thread yielding Other thread running! Other thread running! Main thread yielding Main thread yielding Main thread yielding Other thread running! Main thread yielding Main thread yielding Chapter 46.
Main thread yielding Other thread running! Other thread running!
See Also
thrd_sleep()
46.22 tss_create()
Create new thread-specific storage
Synopsis
#include
int tss_create(tss_t *key, tss_dtor_t dtor);
Description This helps when you need per-thread storage of different values. A common place this comes up is if you have a file scope variable that is shared between a bunch of functions and often returned. That’s not threadsafe. One way to refactor is to replace it with thread- specific storage so that each thread gets their own code and doesn’t step on other thread’s toes. To make this work, you pass in a pointer to a tss_t key—this is the variable you will use in subsequent tss_set() and tss_get() calls to set and get the value associated with the key. The interesting part of this is the dtor destructor pointer of type tss_dtor_t. This is actually a pointer to a function that takes a void* argument and returns void, i.e. void dtor(void *p) { ... }
This function will be called per thread when the thread exits with thrd_exit() (or returns from the run function). It’s unspecified behavior to call this function while other threads’ destructors are running.
Return Value Returns nothing!
Example
This is a general-purpose TSS example. Note the TSS variable is created near the top of main().
1 #include
2 #include
3 #include
4
5 tss_t str;
6
7 void some_function(void)
8 {
9 // Retrieve the per-thread value of this string
10 char *tss_string = tss_get(str);
11 Chapter 46.
12 // And print it
13 printf("TSS string: %s\n", tss_string);
14 }
15
16 int run(void *arg)
17 {
18 int serial = *(int*)arg; // Get this thread's serial number
19 free(arg);
20
21 // malloc() space to hold the data for this thread
22 char *s = malloc(64);
23 sprintf(s, "thread %d! :)", serial); // Happy little string
24
25 // Set this TSS variable to point at the string
26 tss_set(str, s);
27
28 // Call a function that will get the variable
29 some_function();
30
31 return 0; // Equivalent to thrd_exit(0); fires destructors
32 }
33
34 #define THREAD_COUNT 15
35
36 int main(void)
37 {
38 thrd_t t[THREAD_COUNT];
39
40 // Make a new TSS variable, the free() function is the destructor
41 tss_create(&str, free); // <-- CREATE TSS VAR!
42
43 for (int i = 0; i < THREAD_COUNT; i++) {
44 int *n = malloc(sizeof *n); // Holds a thread serial number
45 *n = i;
46 thrd_create(t + i, run, n);
47 }
48
49 for (int i = 0; i < THREAD_COUNT; i++) {
50 thrd_join(t[i], NULL);
51 }
52
53 // And all threads are done, so let's free this
54 tss_delete(str);
55 }
Output: TSS string: thread 0! :) TSS string: thread 2! :) TSS string: thread 1! :) TSS string: thread 5! :) TSS string: thread 3! :) TSS string: thread 6! :) TSS string: thread 4! :) TSS string: thread 7! :) TSS string: thread 8! :) TSS string: thread 9! :) TSS string: thread 10! :) Chapter 46.
TSS string: thread 13! :) TSS string: thread 12! :) TSS string: thread 11! :) TSS string: thread 14! :)
See Also
tss_delete(), tss_set(), tss_get(), thrd_exit()
46.23 tss_delete()
Clean up a thread-specific storage variable
Synopsis
#include
void tss_delete(tss_t key);
Description
This is the opposite of tss_create(). You create (initialize) the TSS variable before using it, then, when all the threads are done that need it, you delete (deinitialize/free) it with this. This doesn’t call any destructors! Those are all called by thrd_exit()!
Return Value Returns nothing!
Example
This is a general-purpose TSS example. Note the TSS variable is deleted near the bottom of main().
1 #include
2 #include
3 #include
4
5 tss_t str;
6
7 void some_function(void)
8 {
9 // Retrieve the per-thread value of this string
10 char *tss_string = tss_get(str);
11
12 // And print it
13 printf("TSS string: %s\n", tss_string);
14 }
15
16 int run(void *arg)
17 {
18 int serial = *(int*)arg; // Get this thread's serial number
19 free(arg);
20
21 // malloc() space to hold the data for this thread
22 char *s = malloc(64); Chapter 46.
23 sprintf(s, "thread %d! :)", serial); // Happy little string
24
25 // Set this TSS variable to point at the string
26 tss_set(str, s);
27
28 // Call a function that will get the variable
29 some_function();
30
31 return 0; // Equivalent to thrd_exit(0); fires destructors
32 }
33
34 #define THREAD_COUNT 15
35
36 int main(void)
37 {
38 thrd_t t[THREAD_COUNT];
39
40 // Make a new TSS variable, the free() function is the destructor
41 tss_create(&str, free);
42
43 for (int i = 0; i < THREAD_COUNT; i++) {
44 int *n = malloc(sizeof *n); // Holds a thread serial number
45 *n = i;
46 thrd_create(t + i, run, n);
47 }
48
49 for (int i = 0; i < THREAD_COUNT; i++) {
50 thrd_join(t[i], NULL);
51 }
52
53 // And all threads are done, so let's free this
54 tss_delete(str); // <-- DELETE TSS VARIABLE!
55 }
Output: TSS string: thread 0! :) TSS string: thread 2! :) TSS string: thread 1! :) TSS string: thread 5! :) TSS string: thread 3! :) TSS string: thread 6! :) TSS string: thread 4! :) TSS string: thread 7! :) TSS string: thread 8! :) TSS string: thread 9! :) TSS string: thread 10! :) TSS string: thread 13! :) TSS string: thread 12! :) TSS string: thread 11! :) TSS string: thread 14! :)
See Also
tss_create(), tss_set(), tss_get(), thrd_exit() Chapter 46.
46.24 tss_get()
Get thread-specific data
Synopsis
#include
void *tss_get(tss_t key);
Description
Once you’ve set a variable with tss_set(), you can retrieve the value with tss_get()—just pass in the key and you’ll get a pointer to the value back. Don’t call this from a destructor.
Return Value
Returns the value stored for the given key, or NULL if there’s trouble.
Example
This is a general-purpose TSS example. Note the TSS variable is retrieved in some_function(), below.
1 #include
2 #include
3 #include
4
5 tss_t str;
6
7 void some_function(void)
8 {
9 // Retrieve the per-thread value of this string
10 char *tss_string = tss_get(str); // <-- GET THE VALUE
11
12 // And print it
13 printf("TSS string: %s\n", tss_string);
14 }
15
16 int run(void *arg)
17 {
18 int serial = *(int*)arg; // Get this thread's serial number
19 free(arg);
20
21 // malloc() space to hold the data for this thread
22 char *s = malloc(64);
23 sprintf(s, "thread %d! :)", serial); // Happy little string
24
25 // Set this TSS variable to point at the string
26 tss_set(str, s);
27
28 // Call a function that will get the variable
29 some_function();
30
31 return 0; // Equivalent to thrd_exit(0); fires destructors
32 }
33
34 #define THREAD_COUNT 15 Chapter 46.
35
36 int main(void)
37 {
38 thrd_t t[THREAD_COUNT];
39
40 // Make a new TSS variable, the free() function is the destructor
41 tss_create(&str, free);
42
43 for (int i = 0; i < THREAD_COUNT; i++) {
44 int *n = malloc(sizeof *n); // Holds a thread serial number
45 *n = i;
46 thrd_create(t + i, run, n);
47 }
48
49 for (int i = 0; i < THREAD_COUNT; i++) {
50 thrd_join(t[i], NULL);
51 }
52
53 // And all threads are done, so let's free this
54 tss_delete(str);
55 }
Output: TSS string: thread 0! :) TSS string: thread 2! :) TSS string: thread 1! :) TSS string: thread 5! :) TSS string: thread 3! :) TSS string: thread 6! :) TSS string: thread 4! :) TSS string: thread 7! :) TSS string: thread 8! :) TSS string: thread 9! :) TSS string: thread 10! :) TSS string: thread 13! :) TSS string: thread 12! :) TSS string: thread 11! :) TSS string: thread 14! :)
See Also
tss_set()
46.25 tss_set()
Set thread-specific data
Synopsis
#include
int tss_set(tss_t key, void *val); Chapter 46.
Description
Once you’ve set up your TSS variable with tss_create(), you can set it on a per thread basis with tss_set(). key is the identifier for this data, and val is a pointer to it. The destructor specified in tss_create() will be called for the value set when the thread exits. Also, if there’s a destructor and there is already at value for this key in place, the destructor will not be called for the already-existing value. In fact, this function will never cause a destructor to be called. So you’re on your own, there—best clean up the old value before overwriting it with the new one.
Return Value
Returns thrd_success when happy, and thrd_error when not.
Example
This is a general-purpose TSS example. Note the TSS variable is set in run(), below.
1 #include
2 #include
3 #include
4
5 tss_t str;
6
7 void some_function(void)
8 {
9 // Retrieve the per-thread value of this string
10 char *tss_string = tss_get(str);
11
12 // And print it
13 printf("TSS string: %s\n", tss_string);
14 }
15
16 int run(void *arg)
17 {
18 int serial = *(int*)arg; // Get this thread's serial number
19 free(arg);
20
21 // malloc() space to hold the data for this thread
22 char *s = malloc(64);
23 sprintf(s, "thread %d! :)", serial); // Happy little string
24
25 // Set this TSS variable to point at the string
26 tss_set(str, s); // <-- SET THE TSS VARIABLE
27
28 // Call a function that will get the variable
29 some_function();
30
31 return 0; // Equivalent to thrd_exit(0); fires destructors
32 }
33
34 #define THREAD_COUNT 15
35
36 int main(void)
37 {
38 thrd_t t[THREAD_COUNT];
39
40 // Make a new TSS variable, the free() function is the destructor Chapter 46.
41 tss_create(&str, free);
42
43 for (int i = 0; i < THREAD_COUNT; i++) {
44 int *n = malloc(sizeof *n); // Holds a thread serial number
45 *n = i;
46 thrd_create(t + i, run, n);
47 }
48
49 for (int i = 0; i < THREAD_COUNT; i++) {
50 thrd_join(t[i], NULL);
51 }
52
53 // And all threads are done, so let's free this
54 tss_delete(str);
55 }
Output: TSS string: thread 0! :) TSS string: thread 2! :) TSS string: thread 1! :) TSS string: thread 5! :) TSS string: thread 3! :) TSS string: thread 6! :) TSS string: thread 4! :) TSS string: thread 7! :) TSS string: thread 8! :) TSS string: thread 9! :) TSS string: thread 10! :) TSS string: thread 13! :) TSS string: thread 12! :) TSS string: thread 11! :) TSS string: thread 14! :)
See Also
tss_get() Chapter 47
Variable Description errno Holds the error status of the last call
This header defines a single variable1, errno, that can be checked to see if an error has occurred. errno is set to 0 on startup, but no library function sets it to 0. If you’re going to use solely it to check for errors, set it to 0 before the call and then check it after. Not only that, but if there’s no error, all library functions will leave the value of errno unchanged. Often, though, you’ll get some error indication from the function you’re calling then check errno to see what went wrong. This is commonly used in conjunction with perror() to get a human-readable error message that corre- sponds to the specific error. Important Safety Tip: You should never make your own variable called errno—that’s undefined behavior. Note that the C Spec defines less than a handful of values errno can take on. Unix defines a bunch more2, as does Windows3.
47.1 errno
Holds the error status of the last call
Synopsis errno // Type is undefined, but it's assignable
Description Indicates the error status of the last call (note that not all calls will set this value).
Value Description 0 No error EDOM Domain error (from math) EILSEQ Encoding error (from character conversion)
1Really it’s just required to be a modifiable lvalue, so not necessarily a variable. But you can treat it as such. 2https://man.archlinux.org/man/errno.3.en 3https://docs.microsoft.com/en-us/cpp/c-runtime-library/errno-constants?view=msvc-160
458 Chapter 47.
Value Description ERANGE Range error (from math)
If you’re doing a number of math functions, you might come across EDOM or ERANGE. With multibyte/wide character conversion functions, you might see EILSEQ. And your system might define any other number of values that errno could be set to, all of which will begin with the letter E. Fun Fact: you can use EDOM, EILSEQ, and ERANGE with preprocessor directives such as #ifdef. But, frankly, I’m not sure why you’d do that other than to test their existence.
Example
The following prints an error message, since passing 2.0 to acos() is outside the function’s domain.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 double x;
8
9 errno = 0; // Make sure this is clear before the call
10
11 x = acos(2.0); // Invalid argument to acos()
12
13 if (errno == EDOM)
14 perror("acos");
15 else
16 printf("Answer is %f\n", x);
17
18 return 0;
19 }
Output: acos: Numerical argument out of domain
The following prints an error message (on my system), since passing 1e+30 to exp() produces a result that’s outside the range of a double.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 double x;
8
9 errno = 0; // Make sure this is clear before the call
10
11 x = exp(1e+30); // Pass in some too-huge number
12
13 if (errno == ERANGE)
14 perror("exp");
15 else
16 printf("Answer is %f\n", x); Chapter 47.
17
18 return 0;
19 }
Output: exp: Numerical result out of range
This example tries to convert an invalid character into a wide character, failing. This sets errno to EILSEQ. We then use perror() to print an error message.
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 char *bad_str = "\xff"; // Probably invalid char in C locale
9 wchar_t wc;
10 size_t result;
11 mbstate_t ps;
12
13 memset(&ps, 0, sizeof ps);
14
15 result = mbrtowc(&wc, bad_str, 1, &ps);
16
17 if (result == (size_t)(-1))
18 perror("mbrtowc");
19 else
20 printf("Converted to L'%lc'\n", wc);
21
22 return 0;
23 }
Output: mbrtowc: Invalid or incomplete multibyte or wide character
See Also
perror(), mbrtoc16(), c16rtomb(), mbrtoc32(), c32rtomb(), fgetwc(), fputwc(), mbrtowc(), wcrtomb(), mbsrtowcs(), wcsrtombs(),
Function Description btowc() Convert a single byte character to a wide character fgetwc() Get a wide character from a wide stream fgetws() Read a wide string from a wide stream fputwc() Write a wide character to a wide stream fputws() Write a wide string to a wide stream fwide() Get or set the orientation of the stream fwprintf() Formatted wide output to a wide stream fwscanf() Formatted wide input from a wide stream getwchar() Get a wide character from stdin getwc() Get a wide character from stdin mbrlen() Compute the number of bytes in a multibyte character restartably mbrtowc() Convert multibyte to wide characters restartably mbsinit() Test if an mbstate_t is in the initial conversion state mbsrtowcs() Convert a multibyte string to a wide character string restartably putwchar() Write a wide character to stdout putwc() Write a wide character to stdout swprintf() Formatted wide output to a wide string swscanf() Formatted wide input from a wide string ungetwc() Pushes a wide character back into the input stream vfwprintf() Variadic formatted wide output to a wide stream vfwscanf() Variadic formatted wide input from a wide stream vswprintf() Variadic formatted wide output to a wide string vswscanf() Variadic formatted wide input from a wide string vwprintf() Variadic formatted wide output vwscanf() Variadic formatted wide input wcscat() Concatenate wide strings dangerously wcschr() Find a wide character in a wide string wcscmp() Compare wide strings wcscoll() Compare two wide strings accounting for locale wcscpy() Copy a wide string dangerously wcscspn() Count characters not from a start at the front of a wide string wcsftime() Formatted date and time output wcslen() Returns the length of a wide string wcsncat() Concatenate wide strings more safely wcsncmp() Compare wide strings, length limited wcsncpy() Copy a wide string more safely wcspbrk() Search a wide string for one of a set of wide characters wcsrchr() Find a wide character in a wide string from the end wcsrtombs() Convert a wide character string to a multibyte string restartably wcsspn() Count characters from a set at the front of a wide string
461 Chapter 48.
Function Description wcsstr() Find a wide string in another wide string wcstod() Convert a wide string to a double wcstof() Convert a wide string to a float wcstok() Tokenize a wide string wcstold() Convert a wide string to a long double wcstoll() Convert a wide string to a long long wcstol() Convert a wide string to a long wcstoull() Convert a wide string to an unsigned long long wcstoul() Convert a wide string to an unsigned long wcsxfrm() Transform a wide string for comparing based on locale wctob() Convert a wide character to a single byte character wctombr() Convert wide to multibyte characters restartably wmemcmp() Compare wide characters in memory wmemcpy() Copy wide character memory wmemmove() Copy wide character memory, potentially overlapping wprintf() Formatted wide output wscanf() Formatted wide input
These are the wide character variants of the functions found in
This header also introduces a type wint_t that is used by the character I/O functions. It’s a type that can hold any single wide character, but also the macro WEOF to indicate wide end-of-file.
48.1 Restartable Functions
Finally, a note on the “restartable” functions that are included here. When conversion is happening, some encodings require C to keep track of some state about the progress of the conversion so far. For a lot of the functions, C uses an internal variable for the state that is shared between function calls. The problem is if you’re writing multithreaded code, this state might get trampled by other threads. To avoid this, each thread needs to maintain its own state in a variable of the opaque type mbstate_t. And the “restartable” functions allow you to pass in this state so that each thread can use their own.
48.2 wprintf(), fwprintf(), swprintf()
Formatted output with a wide string
Synopsis
#include
int fwprintf(FILE * restrict stream, const wchar_t * restrict format, ...);
int swprintf(wchar_t * restrict s, size_t n, const wchar_t * restrict format, ...);
Description
These are the wide versions of printf(), fprintf()](#man-printf), and [sprintf()‘. See those pages for exact substantial usage. These are the same except the format string is a wide character string instead of a multibyte string. And that swprintf() is analogous to snprintf() in that they both take the size of the destination array as an argument. And one more thing: the precision specified for a %s specifier corresponds to the number of wide characters printed, not the number of bytes. If you know of other difference, let me know.
Return Value
Returns the number of wide characters outputted, or -1 if there’s an error.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 char *mbs = "multibyte";
7 wchar_t *ws = L"wide";
8
9 wprintf(L"We're all wide for %s and %ls!\n", mbs, ws);
10
11 double pi = 3.14159265358979;
12 wprintf(L"pi = %f\n", pi);
13 }
Output: We're all wide for multibyte and wide! pi = 3.141593
See Also
printf(), vwprintf()
48.3 wscanf() fwscanf() swscanf()
Scan a wide stream or wide string for formatted input
Synopsis
#include
int wscanf(const wchar_t * restrict format, ...);
int fwscanf(FILE * restrict stream, const wchar_t * restrict format, ...);
int swscanf(const wchar_t * restrict s, const wchar_t * restrict format, ...);
Description
These are the wide variants of scanf(), fscanf(), and sscanf(). See the scanf() page for all the details.
Return Value
Returns the number of items successfully scanned, or EOF on some kind of input failure.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 int quantity;
7 wchar_t item[100];
8
9 wprintf(L"Enter \"quantity: item\"\n");
10
11 if (wscanf(L"%d:%99ls", &quantity, item) != 2)
12 wprintf(L"Malformed input!\n");
13 else
14 wprintf(L"You entered: %d %ls\n", quantity, item);
15 }
Output (input of 12: apples): Enter "quantity: item" 12: apples You entered: 12 apples
See Also
scanf(), vwscanf()
48.4 vwprintf() vfwprintf() vswprintf()
wprintf() variants using variable argument lists (va_list)
Synopsis
#include
int vwprintf(const wchar_t * restrict format, va_list arg); Chapter 48.
int vswprintf(wchar_t * restrict s, size_t n, const wchar_t * restrict format, va_list arg);
int vfwprintf(FILE * restrict stream, const wchar_t * restrict format, va_list arg);
Description
These functions are the wide character variants of the vprintf(), functions. You can refer to that refer- ence page for more details.
Return Value Returns the number of wide characters stored, or a negative value on error.
Example
In this example, we make our own version of wprintf() called wlogger() that timestamps output. No- tice how the calls to wlogger() have all the bells and whistles of wprintf().
1 #include
2 #include
3 #include
4
5 int wlogger(wchar_t *format, ...)
6 {
7 va_list va;
8 time_t now_secs = time(NULL);
9 struct tm *now = gmtime(&now_secs);
10
11 // Output timestamp in format "YYYY-MM-DD hh:mm:ss : "
12 wprintf(L"%04d-%02d-%02d %02d:%02d:%02d : ",
13 now->tm_year + 1900, now->tm_mon + 1, now->tm_mday,
14 now->tm_hour, now->tm_min, now->tm_sec);
15
16 va_start(va, format);
17 int result = vwprintf(format, va);
18 va_end(va);
19
20 wprintf(L"\n");
21
22 return result;
23 }
24
25 int main(void)
26 {
27 int x = 12;
28 float y = 3.2;
29
30 wlogger(L"Hello!");
31 wlogger(L"x = %d and y = %.2f", x, y);
32 }
Output: 2021-03-30 04:25:49 : Hello! 2021-03-30 04:25:49 : x = 12 and y = 3.20 Chapter 48.
See Also
printf(), vprintf()
48.5 vwscanf(), vfwscanf(), vswscanf()
wscanf() variants using variable argument lists (va_list)
Synopsis
#include
int vwscanf(const wchar_t * restrict format, va_list arg);
int vfwscanf(FILE * restrict stream, const wchar_t * restrict format, va_list arg);
int vswscanf(const wchar_t * restrict s, const wchar_t * restrict format, va_list arg);
Description
These are the wide counterparts to the vscanf() collection of functions. See their reference page for details.
Return Value
Returns the number of items successfully scanned, or EOF on some kind of input failure.
Example I have to admit I was wracking my brain to think of when you’d ever want to use this. The best example I could find was one on Stack Overflow1 that error-checks the return value from scanf() against the expected. A variant of that is shown below.
1 #include
2 #include
3 #include
4
5 int error_check_wscanf(int expected_count, wchar_t *format, ...)
6 {
7 va_list va;
8
9 va_start(va, format);
10 int count = vwscanf(format, va);
11 va_end(va);
12
13 // This line will crash the program if the condition is false:
14 assert(count == expected_count);
15
16 return count;
17 }
18
1https://stackoverflow.com/questions/17017331/c99-vscanf-for-dummies/17018046#17018046 Chapter 48.
19 int main(void)
20 {
21 int a, b;
22 float c;
23
24 error_check_wscanf(3, L"%d, %d/%f", &a, &b, &c);
25 error_check_wscanf(2, L"%d", &a);
26 }
See Also
wscanf()
48.6 getwc() fgetwc() getwchar()
Get a wide character from an input stream
Synopsis
#include
wint_t getwchar(void);
wint_t getwc(FILE *stream);
wint_t fgetwc(FILE *stream);
Description
These are the wide variants of fgetc(). fgetwc() and getwc() are identical except that getwc() might be implemented as a macro and is al- lowed to evaluate stream multiple times. getwchar() is identical to getwc() with stream set to stdin. I don’t know why you’d ever use getwc() instead of fgetwc(), but if anyone knows, drop me a line.
Return Value
Returns the next wide character in the input stream. Return WEOF on end-of-file or error. If an I/O error occurs, the error flag is also set on the stream. If an invalid byte sequence is encountered, errno is set to ILSEQ.
Example
1 // read all characters from a file, outputting only the letter 'b's
2 // it finds in the file
3
4 #include
5 #include
6
7 int main(void)
8 { Chapter 48.
9 FILE *fp;
10 wint_t c;
11
12 fp = fopen("datafile.txt", "r"); // error check this!
13
14 // this while-statement assigns into c, and then checks against EOF:
15
16 while((c = fgetc(fp)) != WEOF)
17 if (c == L'b')
18 fputwc(c, stdout);
19
20 fclose(fp);
21 }
See Also
fputwc, fgetws, errno
48.7 fgetws()
Read a wide string from a file
Synopsis
#include
wchar_t *fgetws(wchar_t * restrict s, int n, FILE * restrict stream);
Description
This is the wide version of fgets(). See its reference page for details. A wide NUL character is used to terminate the string.
Return Value
Returns s on success, or a NULL pointer on end-of-file or error.
Example The following example reads lines from a file and prepends them with numbers:
1 #include
2 #include
3
4 #define BUF_SIZE 1024
5
6 int main(void)
7 {
8 FILE *fp;
9 wchar_t buf[BUF_SIZE];
10
11 fp = fopen("textfile.txt", "r"); // error check this!
12
13 int line_count = 0; Chapter 48.
14
15 while ((fgetws(buf, BUF_SIZE, fp)) != NULL)
16 wprintf(L"%04d: %ls", ++line_count, buf);
17
18 fclose(fp);
19 }
Example output for a file with these lines in them (without the prepended numbers): 0001: line 1 0002: line 2 0003: something 0004: line 4
See Also
fgetwc(), fgets()
48.8 putwchar() putwc() fputwc()
Write a single wide character to the console or to a file
Synopsis
#include
wint_t putwchar(wchar_t c);
wint_t putwc(wchar_t c, FILE *stream);
wint_t fputwc(wchar_t c, FILE *stream);
Description These are the wide character equivalents to the ‘fputc()’ group of functions. You can find more information ‘in that reference section’. fputwc() and putwc() are identical except that putwc() might be implemented as a macro and is al- lowed to evaluate stream multiple times. putwchar() is identical to putwc() with stream set to stdin. I don’t know why you’d ever use putwc() instead of fputwc(), but if anyone knows, drop me a line.
Return Value
Returns the wide character written, or WEOF on error. If it’s an I/O error, the error flag will be set for the stream. If it’s an encoding error, errno will be set to EILSEQ.
Example
1 // read all characters from a file, outputting only the letter 'b's
2 // it finds in the file
3 Chapter 48.
4 #include
5 #include
6
7 int main(void)
8 {
9 FILE *fp;
10 wint_t c;
11
12 fp = fopen("datafile.txt", "r"); // error check this!
13
14 // this while-statement assigns into c, and then checks against EOF:
15
16 while((c = fgetc(fp)) != WEOF)
17 if (c == L'b')
18 fputwc(c, stdout);
19
20 fclose(fp);
21 }
See Also
fgetwc(), fputc(), errno
48.9 fputws()
Write a wide string to a file
Synopsis
#include
int fputws(const wchar_t * restrict s, FILE * restrict stream);
Description
This is the wide version of fputs(). Pass in a wide string and an output stream, and it will so be written.
Return Value
Returns a non-negative value on success, or EOF on error.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 fputws(L"Hello, world!\n", stdout);
7 } Chapter 48.
See Also
fputwc() fputs()
48.10 fwide()
Get or set the orientation of the stream
Synopsis
#include
int fwide(FILE *stream, int mode);
Description Streams can be either wide-oriented (meaning the wide functions are in use) or byte-oriented (that the regular multibyte functions are in use). Or, before an orientation is chosen, unoriented. There are two ways to set the orientation of an unoriented stream: • Implicitly: just use a function like printf() (byte oriented) or wprintf() (wide oriented), and the orientation will be set. • Explicitly: use this function to set it. You can set the orientation for the stream by passing different numbers to mode:
mode Description 0 Do not alter the orientation -1 Set stream to byte-oriented 1 Set stream to wide-oriented
(I said -1 and 1 there, but really it could be any positive or negative number.) Most people choose the wide or byte functions (printf() or wprintf()) and just start using them and never use fwide() to set the orientation. And once the orientation is set, you can’t change it. So you can’t use fwide() for that, either. So what can you use it for? You can test to see what orientation a stream is in by passing 0 as the mode and checking the return value.
Return Value Returns greater than zero if the stream is wide-oriented. Returns less than zero if the stream is byte-oriented. Returns zero if the stream is unoriented.
Example Example setting to byte-oriented:
1 #include
2 #include
3 Chapter 48.
4 int main(void)
5 {
6 printf("Hello world!\n"); // Implicitly set to byte
7
8 int mode = fwide(stdout, 0);
9
10 printf("Stream is %s-oriented\n", mode < 0? "byte": "wide");
11 }
Output: Hello world! Stream is byte-oriented
Example setting to wide-oriented:
1 #include
2 #include
3
4 int main(void)
5 {
6 wprintf(L"Hello world!\n"); // Implicitly set to wide
7
8 int mode = fwide(stdout, 0);
9
10 wprintf(L"Stream is %ls-oriented\n", mode < 0? L"byte": L"wide");
11 }
Output: Hello world! Stream is wide-oriented
48.11 ungetwc()
Pushes a wide character back into the input stream
Synopsis
#include
wint_t ungetwc(wint_t c, FILE *stream);
Description
This is the wide character variant of ungetc(). It performs the reverse operation of fgetwc(), pushing a character back on the input stream. The spec guarantees you can do this one time in a row. You can probably do it more times, but it’s up to the implementation. If you do too many calls without an intervening read, an error could be returned. Setting the file position discards any characters pushed by ungetwc() without being subsequently read. The end-of-file flag is cleared after a successful call. Chapter 48.
Return Value
Returns the value of the pushed character on success, or WEOF on failure.
Example This example reads a piece of punctuation, then everything after it up to the next piece of punctuation. It returns the leading punctuation, and stores the rest in a string.
1 #include
2 #include
3 #include
4
5 wint_t read_punctstring(FILE *fp, wchar_t *s)
6 {
7 wint_t origpunct, c;
8
9 origpunct = fgetwc(fp);
10
11 if (origpunct == WEOF) // return EOF on end-of-file
12 return WEOF;
13
14 while (c = fgetwc(fp), !iswpunct(c) && c != WEOF)
15 *s++ = c; // save it in the string
16
17 *s = L'\0'; // nul-terminate the string
18
19 // if we read punctuation last, ungetc it so we can fgetc it next
20 // time:
21 if (iswpunct(c))
22 ungetwc(c, fp);
23
24 return origpunct;
25 }
26
27 int main(void)
28 {
29 wchar_t s[128];
30 wint_t c;
31
32 while ((c = read_punctstring(stdin, s)) != WEOF) {
33 wprintf(L"%lc: %ls\n", c, s);
34 }
35 }
Sample Input: !foo#bar*baz
Sample output: !: foo #: bar *: baz
See Also
fgetwc(), ungetc() Chapter 48.
48.12 wcstod() wcstof() wcstold()
Convert a wide string to a floating point number
Synopsis
#include
double wcstod(const wchar_t * restrict nptr, wchar_t ** restrict endptr);
float wcstof(const wchar_t * restrict nptr, wchar_t ** restrict endptr);
long double wcstold(const wchar_t * restrict nptr, wchar_t ** restrict endptr);
Description
These are the wide counterparts to the strtod() family of functions. See their reference pages for details.
Return Value Returns the string converted to a floating point value. Returns 0 if there’s no valid number in the string. On overflow, returns an apporpriately-signed HUGE_VAL, HUGE_VALF. or HUGE_VALL depending on the return type, and errno is set to ERANGE. On underflow, returns a number no greater than the smallest normalized positive number, appropriately signed. The implemention might set errno to ERANGE.
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t *inp = L" 123.4567beej";
6 wchar_t *badchar;
7
8 double val = wcstod(inp, &badchar);
9
10 wprintf(L"Converted string to %f\n", val);
11 wprintf(L"Encountered bad characters: %ls\n", badchar);
12
13 val = wcstod(L"987.654321beej", NULL);
14 wprintf(L"Ignoring bad chars: %f\n", val);
15
16 val = wcstod(L"11.2233", &badchar);
17
18 if (*badchar == L'\0')
19 wprintf(L"No bad chars: %f\n", val);
20 else
21 wprintf(L"Found bad chars: %f, %ls\n", val, badchar);
22 }
Output: Converted string to 123.456700 Encountered bad characters: beej Chapter 48.
Ignoring bad chars: 987.654321 No bad chars: 11.223300
See Also
wcstol(), strtod(), errno
48.13 wcstol() wcstoll() wcstoul() wcstoull()
Convert a wide string to an integer value
Synopsis
#include
long int wcstol(const wchar_t * restrict nptr, wchar_t ** restrict endptr, int base);
long long int wcstoll(const wchar_t * restrict nptr, wchar_t ** restrict endptr, int base);
unsigned long int wcstoul(const wchar_t * restrict nptr, wchar_t ** restrict endptr, int base);
unsigned long long int wcstoull(const wchar_t * restrict nptr, wchar_t ** restrict endptr, int base);
Description
These are the wide counterparts to the strtol() family of functions, so see their reference pages for the details.
Return Value Returns the integer value of the string. If nothing can be found, 0 is returned. If the result is out of range, the value returned is one of LONG_MIN, LONG_MAX, LLONG_MIN, LLONG_MAX, ULONG_MAX or ULLONG_MAX, as appropriate. And errno is set to ERANGE.
Example
1 #include
2
3 int main(void)
4 {
5 // All output in decimal (base 10)
6
7 wprintf(L"%ld\n", wcstol(L"123", NULL, 0)); // 123
8 wprintf(L"%ld\n", wcstol(L"123", NULL, 10)); // 123
9 wprintf(L"%ld\n", wcstol(L"101010", NULL, 2)); // binary, 42
10 wprintf(L"%ld\n", wcstol(L"123", NULL, 8)); // octal, 83
11 wprintf(L"%ld\n", wcstol(L"123", NULL, 16)); // hex, 291
12 Chapter 48.
13 wprintf(L"%ld\n", wcstol(L"0123", NULL, 0)); // octal, 83
14 wprintf(L"%ld\n", wcstol(L"0x123", NULL, 0)); // hex, 291
15
16 wchar_t *badchar;
17 long int x = wcstol(L" 1234beej", &badchar, 0);
18
19 wprintf(L"Value is %ld\n", x); // Value is 1234
20 wprintf(L"Bad chars at \"%ls\"\n", badchar); // Bad chars at "beej"
21 }
Output: 123 123 42 83 291 83 291 Value is 1234 Bad chars at "beej"
See Also
wcstod(), strtol(), errno, wcstoimax(), wcstoumax()
48.14 wcscpy() wcsncpy()
Copy a wide string
Synopsis
#include
wchar_t *wcscpy(wchar_t * restrict s1, const wchar_t * restrict s2);
wchar_t *wcsncpy(wchar_t * restrict s1, const wchar_t * restrict s2, size_t n);
Description
These are the wide versions of strcpy() and strncpy(). They’ll copy a string up to a wide NUL. Or, in the case of the safer wcsncpy(), until then or until n wide characters are copied. If the string in s1 is shorter than n, wcsncpy() will pad s2 with wide NUL characters until the nth wide character is reached. Even though wcsncpy() is safer because it will never overrun the end of s2 (assuming you set n correctly), it’s still unsafe a NUL is not found in s1 in the first n characters. In that case, s2 will not be NUL- terminated. Always make sure n is greater than the string length of s1!
Return Value
Returns s1. Chapter 48.
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t *s1 = L"Hello!";
6 wchar_t s2[10];
7
8 wcsncpy(s2, s1, 10);
9
10 wprintf(L"\"%ls\"\n", s2); // "Hello!"
11 }
See Also
wmemcpy(), wmemmove() strcpy(), strncpy()
48.15 wmemcpy() wmemmove()
Copy wide characters
Synopsis
#include
wchar_t *wmemcpy(wchar_t * restrict s1, const wchar_t * restrict s2, size_t n);
wchar_t *wmemmove(wchar_t *s1, const wchar_t *s2, size_t n);
Description
These are the wide versions of memcpy() and memmove(). They copy n wide characters from s2 to s1. They’re the same except that wmemmove() is guaranteed to work with overlapping memory regions, and wmemcpy() is not.
Return Value
Both functions return the pointer s1.
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t s[100] = L"Goats";
6 wchar_t t[100];
7
8 wmemcpy(t, s, 6); // Copy non-overlapping memory
9
10 wmemmove(s + 2, s, 6); // Copy overlapping memory Chapter 48.
11
12 wprintf(L"s is \"%ls\"\n", s);
13 wprintf(L"t is \"%ls\"\n", t);
14 }
Output: s is "GoGoats" t is "Goats"
See Also
wcscpy(), wcsncpy(), memcpy(), memmove()
48.16 wcscat() wcsncat()
Concatenate wide strings
Synopsis
#include
wchar_t *wcscat(wchar_t * restrict s1, const wchar_t * restrict s2);
wchar_t *wcsncat(wchar_t * restrict s1, const wchar_t * restrict s2, size_t n);
Description
These are the wide variants of strcat() and strncat(). They concatenate s2 onto the end of s1. They’re the same except wcsncat() gives you the option to limit the number of wide characters appended. Note that wcsncat() always adds a NUL terminator to the end, even if n characters were appended. So be sure to leave room for that.
Return Value
Both functions return the pointer s1.
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t dest[30] = L"Hello";
6 wchar_t *src = L", World!";
7 wchar_t numbers[] = L"12345678";
8
9 wprintf(L"dest before strcat: \"%ls\"\n", dest); // "Hello"
10
11 wcscat(dest, src);
12 wprintf(L"dest after strcat: \"%ls\"\n", dest); // "Hello, world!"
13 Chapter 48.
14 wcsncat(dest, numbers, 3); // strcat first 3 chars of numbers
15 wprintf(L"dest after strncat: \"%ls\"\n", dest); // "Hello, world!123"
16 }
See Also
strcat(), strncat()
48.17 wcscmp(), wcsncmp(), wmemcmp()
Compare wide strings or memory
Synopsis
#include
int wcscmp(const wchar_t *s1, const wchar_t *s2);
int wcsncmp(const wchar_t *s1, const wchar_t *s2, size_t n);
int wmemcmp(const wchar_t *s1, const wchar_t *s2, size_t n);
Description
These are the wide variants of memcmp(), strcmp(), and strncmp(). wcscmp() and wcsncmp() both compare strings until a NUL character. wcsncmp() also has the additional restriction that it will only compare the first n characters. wmemcmp() is like wcsncmp() except it won’t stop at a NUL. The comparison is done against the character value (which might (or might not) be its Unicode code point).
Return Value Returns zero if both regions are equal. Returns a negative number if the region pointed to by s1 is less than s2. Returns a positive number if the region pointed to by s1 is greater than s2.
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t *s1 = L"Muffin";
6 wchar_t *s2 = L"Muffin Sandwich";
7 wchar_t *s3 = L"Muffin";
8
9 wprintf(L"%d\n", wcscmp(L"Biscuits", L"Kittens")); // <0 since 'B' < 'K'
10 wprintf(L"%d\n", wcscmp(L"Kittens", L"Biscuits")); // >0 since 'K' > 'B'
11
12 if (wcscmp(s1, s2) == 0)
13 wprintf(L"This won't get printed because the strings differ\n"); Chapter 48.
14
15 if (wcscmp(s1, s3) == 0)
16 wprintf(L"This will print because s1 and s3 are the same\n");
17
18 // this is a little weird...but if the strings are the same, it'll
19 // return zero, which can also be thought of as "false". Not-false
20 // is "true", so (!wcscmp()) will be true if the strings are the
21 // same. yes, it's odd, but you see this all the time in the wild
22 // so you might as well get used to it:
23
24 if (!wcscmp(s1, s3))
25 wprintf(L"The strings are the same!\n");
26
27 if (!wcsncmp(s1, s2, 6))
28 wprintf(L"The first 6 characters of s1 and s2 are the same\n");
29 }
Output: -1 1 This will print because s1 and s3 are the same The strings are the same! The first 6 characters of s1 and s2 are the same
See Also
wcscoll(), memcmp(), strcmp(), strncmp()
48.18 wcscoll()
Compare two wide strings accounting for locale
Synopsis
#include
int wcscoll(const wchar_t *s1, const wchar_t *s2);
Description
This is the wide version of strcoll(). See that reference page for details. This is slower than wcscmp(), so only use it if you need the locale-specific compare.
Return Value Returns zero if both regions are equal in this locale. Returns a negative number if the region pointed to by s1 is less than s2 in this locale. Returns a positive number if the region pointed to by s1 is greater than s2 in this locale.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 setlocale(LC_ALL, "");
7
8 // If your source character set doesn't support "é" in a string
9 // you can replace it with `\u00e9`, the Unicode code point
10 // for "é".
11
12 wprintf(L"%d\n", wcscmp(L"é", L"f")); // Reports é > f, yuck.
13 wprintf(L"%d\n", wcscoll(L"é", L"f")); // Reports é < f, yay!
14 }
See Also
wcscmp(), wcsxfrm(), strcoll()
48.19 wcsxfrm()
Transform a wide string for comparing based on locale
Synopsis
#include
size_t wcsxfrm(wchar_t * restrict s1, const wchar_t * restrict s2, size_t n);
Description
This is the wide variant of strxfrm(). See that reference page for details.
Return Value Returns the length of the transformed wide string in wide characters. If the return value is greater than n, all bets are off for the result in s1.
Example
1 #include
2 #include
3 #include
4
5 // Transform a string for comparison, returning a malloc'd
6 // result
7 wchar_t *get_xfrm_str(wchar_t *s)
8 {
9 int len = wcsxfrm(NULL, s, 0) + 1;
10 wchar_t *d = malloc(len * sizeof(wchar_t));
11
12 wcsxfrm(d, s, len);
13
14 return d;
15 } Chapter 48.
16
17 // Does half the work of a regular wcscoll() because the second
18 // string arrives already transformed.
19 int half_wcscoll(wchar_t *s1, wchar_t *s2_transformed)
20 {
21 wchar_t *s1_transformed = get_xfrm_str(s1);
22
23 int result = wcscmp(s1_transformed, s2_transformed);
24
25 free(s1_transformed);
26
27 return result;
28 }
29
30 int main(void)
31 {
32 setlocale(LC_ALL, "");
33
34 // Pre-transform the string to compare against
35 wchar_t *s = get_xfrm_str(L"éfg");
36
37 // Repeatedly compare against "éfg"
38 wprintf(L"%d\n", half_wcscoll(L"fgh", s)); // "fgh" > "éfg"
39 wprintf(L"%d\n", half_wcscoll(L"àbc", s)); // "àbc" < "éfg"
40 wprintf(L"%d\n", half_wcscoll(L"ĥij", s)); // "ĥij" > "éfg"
41
42 free(s);
43 }
Output: 1 -1 1
See Also
wcscmp(), wcscoll(), strxfrm()
48.20 wcschr() wcsrchr()
Find a wide character in a wide string
Synopsis
#include
wchar_t *wcschr(const wchar_t *s, wchar_t c);
wchar_t *wcsrchr(const wchar_t *s, wchar_t c);
wchar_t *wmemchr(const wchar_t *s, wchar_t c, size_t n);
Description
These are the wide equivalents to strchr(), strrchr(), and memchr(). Chapter 48.
They search for wide characters in a wide string from the front (wcschr()), the end (wcsrchr()) or for an arbitrary number of wide characters (wmemchr()).
Return Value
All three functions return a pointer to the wide character found, or NULL if the character, sadly, isn’t found.
Example
1 #include
2
3 int main(void)
4 {
5 // "Hello, world!"
6 // ^ ^ ^
7 // ABC
8
9 wchar_t *str = L"Hello, world!";
10 wchar_t *p;
11
12 p = wcschr(str, ','); // p now points at position A
13 p = wcsrchr(str, 'o'); // p now points at position B
14
15 p = wmemchr(str, '!', 13); // p now points at position C
16
17 // repeatedly find all occurrences of the letter 'B'
18 str = L"A BIG BROWN BAT BIT BEEJ";
19
20 for(p = wcschr(str, 'B'); p != NULL; p = wcschr(p + 1, 'B')) {
21 wprintf(L"Found a 'B' here: %ls\n", p);
22 }
23
24 // output is:
25 //
26 // Found a 'B' here: BIG BROWN BAT BIT BEEJ
27 // Found a 'B' here: BROWN BAT BIT BEEJ
28 // Found a 'B' here: BAT BIT BEEJ
29 // Found a 'B' here: BIT BEEJ
30 // Found a 'B' here: BEEJ
31 }
See Also
strchr(), strrchr(), memchr()
48.21 wcsspn() wcscspn()
Return the length of a wide string consisting entirely of a set of wide characters, or of not a set of wide characters
Synopsis
#include
size_t wcsspn(const wchar_t *s1, const wchar_t *s2); Chapter 48.
size_t wcscspn(const wchar_t *s1, const wchar_t *s2);
Description
The are the wide character counterparts to [strspn()] (#man-strspn)and strcspn(). They compute the length of the string pointed to by s1 consisting entirely of the characters found in s2. Or, in the case of wcscspn(), the characters not found in s2.
Return Value
The length of the string pointed to by s1 consisting solely of the characters in s2 (in the case of wcsspn()) or of the characters not in s2 (in th ecase of wcscspn()).
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t str1[] = L"a banana";
6 wchar_t str2[] = L"the bolivian navy on maneuvers in the south pacific";
7 int n;
8
9 // how many letters in str1 until we reach something that's not a vowel?
10 n = wcsspn(str1, L"aeiou");
11 wprintf(L"%d\n", n); // n == 1, just "a"
12
13 // how many letters in str1 until we reach something that's not a, b,
14 // or space?
15 n = wcsspn(str1, L"ab ");
16 wprintf(L"%d\n", n); // n == 4, "a ba"
17
18 // how many letters in str2 before we get a "y"?
19 n = wcscspn(str2, L"y");
20 wprintf(L"%d\n", n); // n = 16, "the bolivian nav"
21 }
See Also
wcschr(), wcsrchr(), strspn()
48.22 wcspbrk()
Search a wide string for one of a set of wide characters
Synopsis
#include
wchar_t *wcspbrk(const wchar_t *s1, const wchar_t *s2); Chapter 48.
Description
This is the wide character variant of strpbrk(). It finds the first occurrance of any of a set of wide characters in a wide string.
Return Value
Returns a pointer to the first character in the string s1 that exists in the string s2. Or NULL if none of the characters in s2 can be found in s1.
Example
1 #include
2
3 int main(void)
4 {
5 // p points here after wcspbrk
6 // v
7 wchar_t *s1 = L"Hello, world!";
8 wchar_t *s2 = L"dow!"; // Match any of these chars
9
10 wchar_t *p = wcspbrk(s1, s2); // p points to the o
11
12 wprintf(L"%ls\n", p); // "o, world!"
13 }
See Also
wcschr(), wmemchr(), strpbrk()
48.23 wcsstr()
Find a wide string in another wide string
Synopsis
#include
wchar_t *wcsstr(const wchar_t *s1, const wchar_t *s2);
Description
This is the wide variant of strstr(). It locates a substring in a string.
Return Value
Returns a pointer to the location in s1 that contains s2. Or NULL if s2 cannot be found in s1. Chapter 48.
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t *str = L"The quick brown fox jumped over the lazy dogs.";
6 wchar_t *p;
7
8 p = wcsstr(str, L"lazy");
9 wprintf(L"%ls\n", p == NULL? L"null": p); // "lazy dogs."
10
11 // p is NULL after this, since the string "wombat" isn't in str:
12 p = wcsstr(str, L"wombat");
13 wprintf(L"%ls\n", p == NULL? L"null": p); // "null"
14 }
See Also
wcschr(), wcsrchr(), wcsspn(), wcscspn(), strstr()
48.24 wcstok()
Tokenize a wide string
Synopsis
#include
Description
This is the wide version of strtok(). And, like that one, it modifies the string s1. So make a copy of it first if you want to preserve the original. One key difference is that wcstok() can be threadsafe because you pass in the pointer ptr to the current state of the transformation. This gets initializers for you when s1 is initially passed in as non-NULL. (Subsequent calls with a NULL s1 cause the state to update.)
Return Value Example
1 #include
2
3 int main(void)
4 {
5 // break up the string into a series of space or
6 // punctuation-separated words
7 wchar_t str[] = L"Where is my bacon, dude?";
8 wchar_t *token;
9 wchar_t *state;
10
11 // Note that the following if-do-while construct is very very Chapter 48.
12 // very very very common to see when using strtok().
13
14 // grab the first token (making sure there is a first token!)
15 if ((token = wcstok(str, L".,?! ", &state)) != NULL) {
16 do {
17 wprintf(L"Word: \"%ls\"\n", token);
18
19 // now, the while continuation condition grabs the
20 // next token (by passing NULL as the first param)
21 // and continues if the token's not NULL:
22 } while ((token = wcstok(NULL, L".,?! ", &state)) != NULL);
23 }
24
25 // output is:
26 //
27 // Word: "Where"
28 // Word: "is"
29 // Word: "my"
30 // Word: "bacon"
31 // Word: "dude"
32 //
33 }
See Also
strtok()
48.25 wcslen()
Returns the length of a wide string
Synopsis
#include
size_t wcslen(const wchar_t *s);
Description
This is the wide counterpart to strlen().
Return Value Returns the number of wide characters before the wide NUL terminator.
Example
1 #include
2
3 int main(void)
4 {
5 wchar_t *s = L"Hello, world!"; // 13 characters
6
7 // prints "The string is 13 characters long.": Chapter 48.
8
9 wprintf(L"The string is %zu characters long.\n", wcslen(s));
10 }
See Also
strlen()
48.26 wcsftime()
Formatted date and time output
Synopsis
#include
size_t wcsftime(wchar_t * restrict s, size_t maxsize, const wchar_t * restrict format, const struct tm * restrict timeptr);
Description
This is the wide equivalent to strftime(). See that reference page for details. maxsize here refers to the maximum number of wide characters that can be in the result string.
Return Value If successful, returns the number of wide characters written. If not successful because the result couldn’t fit in the space alloted, 0 is returned and the contents of the string could be anything.
Example
1 #include
2 #include
3
4 #define BUFSIZE 128
5
6 int main(void)
7 {
8 wchar_t s[BUFSIZE];
9 time_t now = time(NULL);
10
11 // %c: print date as per current locale
12 wcsftime(s, BUFSIZE, L"%c", localtime(&now));
13 wprintf(L"%ls\n", s); // Sun Feb 28 22:29:00 2021
14
15 // %A: full weekday name
16 // %B: full month name
17 // %d: day of the month
18 wcsftime(s, BUFSIZE, L"%A, %B %d", localtime(&now));
19 wprintf(L"%ls\n", s); // Sunday, February 28 Chapter 48.
20
21 // %I: hour (12 hour clock)
22 // %M: minute
23 // %S: second
24 // %p: AM or PM
25 wcsftime(s, BUFSIZE, L"It's %I:%M:%S %p", localtime(&now));
26 wprintf(L"%ls\n", s); // It's 10:29:00 PM
27
28 // %F: ISO 8601 yyyy-mm-dd
29 // %T: ISO 8601 hh:mm:ss
30 // %z: ISO 8601 timezone offset
31 wcsftime(s, BUFSIZE, L"ISO 8601: %FT%T%z", localtime(&now));
32 wprintf(L"%ls\n", s); // ISO 8601: 2021-02-28T22:29:00-0800
33 }
See Also
strftime()
48.27 btowc() wctob()
Convert a single byte character to a wide character
Synopsis
#include
wint_t btowc(int c);
int wctob(wint_t c);
Description These functions convert between single byte characters and wide characters, and vice-versa. Even though ints are involved, don’t let this mislead you; they’re effectively converted to unsigned chars internally. The characters in the basic character set are guaranteed to be a single byte.
Return Value
btowc() returns the single-byte character as a wide character. Returns WEOF if EOF is passed in, or if the byte doesn’t correspond to a valid wide character. wctob() returns the wide character as a single-byte character. Returns EOF if WEOF is passed in, or if the wide character doesn’t correspond to a value single-byte character. See mbtowc() and wctomb() for multibyte to wide character conversion.
Example
1 #include
2
3 int main(void)
4 { Chapter 48.
5 wint_t wc = btowc('B'); // Convert single byte to wide char
6
7 wprintf(L"Wide character: %lc\n", wc);
8
9 unsigned char c = wctob(wc); // Convert back to single byte
10
11 wprintf(L"Single-byte character: %c\n", c);
12 }
Output: Wide character: B Single-byte character: B
See Also
mbtowc(), wctomb()
48.28 mbsinit()
Test if an mbstate_t is in the initial conversion state
Synopsis
#include
int mbsinit(const mbstate_t *ps);
Description
For a given conversion state in a mbstate_t variable, this function determines if it’s in the initial conver- sion state.
Return Value
Returns non-zero if the value pointed to by ps is in the initial conversion state, or if ps is NULL. Returns 0 if the value pointed to by ps is not in the initial conversion state.
Example
For me, this example doesn’t do anything exciting, saying that the mbstate_t variable is always in the initial state. Yay. But if have a stateful encoding like 2022-JP, try messing around with this to see if you can get into an intermediate state. This program has a bit of code at the top that reports if your locale’s encoding requires any state.
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 mbstate_t state; Chapter 48.
9 wchar_t wc[128];
10
11 setlocale(LC_ALL, "");
12
13 int is_state_dependent = mbtowc(NULL, NULL, 0);
14
15 wprintf(L"Is encoding state dependent? %d\n", is_state_dependent);
16
17 memset(&state, 0, sizeof state); // Set to initial state
18
19 wprintf(L"In initial conversion state? %d\n", mbsinit(&state));
20
21 mbrtowc(wc, "B", 5, &state);
22
23 wprintf(L"In initial conversion state? %d\n", mbsinit(&state));
24 }
See Also
mbtowc(), wctomb(), mbrtowc(), wcrtomb()
48.29 mbrlen()
Compute the number of bytes in a multibyte character, restartably
Synopsis
#include
size_t mbrlen(const char * restrict s, size_t n, mbstate_t * restrict ps);
Description
This is the restartable version of mblen(). It inspects at most n bytes of the string s to see how many bytes in this character. The conversion state is stored in ps. This function doesn’t have the functionality of mblen() that allowed you to query if this character encod- ing was stateful and to reset the internal state.
Return Value Returns the number of bytes required for this multibyte character. Returns (size_t)(-1) if the data in s is not a valid multibyte character. Returns (size_t)(-2) if the data is s is a valid but not complete multibyte character.
Example
If your character set doesn’t support the Euro symbol “€”, substitute the Unicode escape sequence \u20ac, below.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 mbstate_t state;
8 int len;
9
10 setlocale(LC_ALL, "");
11
12 memset(&state, 0, sizeof state); // Set to initial state
13
14 len = mbrlen("B", 5, &state);
15
16 wprintf(L"Length of 'B' is %d byte(s)\n", len);
17
18 len = mbrlen("€", 5, &state);
19
20 wprintf(L"Length of '€' is %d byte(s)\n", len);
21 }
Output: Length of 'B' is 1 byte(s) Length of '€' is 3 byte(s)
See Also
mblen()
48.30 mbrtowc()
Convert multibyte to wide characters restartably
Synopsis
#include
size_t mbrtowc(wchar_t * restrict pwc, const char * restrict s, size_t n, mbstate_t * restrict ps);
Description
This is the restartable counterpart to mbtowc(). It converts individual characters from multibyte to wide, tracking the conversion state in the variable pointed to by ps. At most n bytes are inspected for conversion to a wide character. These two variants are identical and cause the state pointed to by ps to be set to the initial conversion state: mbrtowc(NULL, NULL, 0, &state); mbrtowc(NULL, "", 1, &state);
Also, if you’re just interested in the length in bytes of the multibyte character, you can pass NULL for pwc and nothing will be stored for the wide character: Chapter 48.
int len = mbrtowc(NULL, "€", 5, &state);
This function doesn’t have the functionality of mbtowc() that allowed you to query if this character en- coding was stateful and to reset the internal state.
Return Value On success, returns a positive number corresponding to the number of bytes in the multibyte character. Returns 0 if the character encoded is a wide NUL character. Returns (size_t)(-1) if the data in s is not a valid multibyte character. Returns (size_t)(-2) if the data is s is a valid but not complete multibyte character.
Example
If your character set doesn’t support the Euro symbol “€”, substitute the Unicode escape sequence \u20ac, below.
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 mbstate_t state;
9
10 memset(&state, 0, sizeof state);
11
12 setlocale(LC_ALL, "");
13
14 wprintf(L"State dependency: %d\n", mbtowc(NULL, NULL, 0));
15
16 wchar_t wc;
17 int bytes;
18
19 bytes = mbrtowc(&wc, "€", 5, &state);
20
21 wprintf(L"L'%lc' takes %d bytes as multibyte char '€'\n", wc, bytes);
22 }
Output on my system: State dependency: 0 L'€' takes 3 bytes as multibyte char '€'
See Also
mbtowc(), wcrtomb()
48.31 wctombr()
Convert wide to multibyte characters restartably Chapter 48.
Synopsis
#include
size_t wcrtomb(char * restrict s, wchar_t wc, mbstate_t * restrict ps);
Description
This is the restartable counterpart to wctomb(). It converts individual characters from wide to multibyte, tracking the conversion state in the variable pointed to by ps. The destination array s should be at least MB_CUR_MAX2 bytes in size—you won’t get anything bigger back from this function. Note that the values in this result array won’t be NUL-terminated. If you pass a wide NUL character in, the result will contain any bytes needed to restore the conversion state to its initial state followed by a NUL character, and the state pointed to by ps will be reset to its initial state: // Reset state wcrtomb(mb, L'\0', &state)
If you don’t care about the results (i.e. you’re just interested in resetting the state or getting the return value), you can do this by passing NULL for s: wcrtomb(NULL, L'\0', &state); // Reset state
int byte_count = wctomb(NULL, "X", &state); // Count bytes in 'X'
This function doesn’t have the functionality of wctomb() that allowed you to query if this character en- coding was stateful and to reset the internal state.
Return Value On success, returns the number of bytes needed to encode this wide character in the current locale. If the input is an invalid wide character, errno will be set to EILSEQ and the function returns (size_t)(- 1). If this happens, all bets are off for the conversion state, so you might as well reset it.
Example
If your character set doesn’t support the Euro symbol “€”, substitute the Unicode escape sequence \u20ac, below.
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 mbstate_t state;
9
10 memset(&state, 0, sizeof state);
11
12 setlocale(LC_ALL, "");
13
14 wprintf(L"State dependency: %d\n", mbtowc(NULL, NULL, 0));
2This is a variable, not a macro, so if you use it to define an array, it’ll be a variable-length array. Chapter 48.
15
16 char mb[10] = {0};
17 int bytes = wcrtomb(mb, L'€', &state);
18
19 wprintf(L"L'€' takes %d bytes as multibyte char '%s'\n", bytes, mb);
20 }
See Also
mbrtowc(), wctomb(), errno
48.32 mbsrtowcs()
Convert a multibyte string to a wide character string restartably
Synopsis
#include
size_t mbsrtowcs(wchar_t * restrict dst, const char ** restrict src, size_t len, mbstate_t * restrict ps);
Description
This is the restartable version of mbstowcs(). It converts a multibyte string to a wide character string. The result is put in the buffer pointed to by dst, and the pointer src is updated to indicate how much of the string was consumed (unless dst is NULL). At most len wide characters will be stored. This also takes a pointer to its own mbstate_t variable in ps for holding the conversion state. You can set dst to NULL if you only care about the return value. This could be useful for getting the number of characters in a multibyte string. In the normal case, the src string will be consumed up to the NUL character, and the results will be stored in the dst buffer, including the wide NUL character. In this case, the pointer pointed to by src will be set to NULL. And the conversion state will be set to the initial conversion state. If things go wrong because the source string isn’t a valid sequence of characters, conversion will stop and the pointer pointed to by src will be set to the address just after the last successfully-translated multibyte character.
Return Value If successful, returns the number of characters converted, not including any NUL terminator. If the multibyte sequence is invalid, the function returns (size_t)(-1) and errno is set to EILSEQ.
Example Here we’ll convert the string “€5 ± π” into a wide character string:
1 ginclude
2 #include
3 #include
4
5 #define WIDE_STR_SIZE 10
6
7 int main(void)
8 {
9 const char *mbs = "€5 ± π"; // That's the exact price range
10
11 wchar_t wcs[WIDE_STR_SIZE];
12
13 setlocale(LC_ALL, "");
14
15 mbstate_t state;
16 memset(&state, 0, sizeof state);
17
18 size_t count = mbsrtowcs(wcs, &mbs, WIDE_STR_SIZE, &state);
19
20 wprintf(L"Wide string L\"%ls\" is %d characters\n", wcs, count);
21 }
Output: Wide string L"€5 ± π" is 6 characters
Here’s another example of using mbsrtowcs() to get the length in characters of a multibyte string even if the string is full of multibyte characters. This is in contrast to strlen(), which returns the total number of bytes in the string.
1 #include
2 #include
3
4 #include
5 #include
6 #include
7
8 size_t mbstrlen(const char *mbs)
9 {
10 mbstate_t state;
11
12 memset(&state, 0, sizeof state);
13
14 return mbsrtowcs(NULL, &mbs, SIZE_MAX, &state);
15 }
16
17 int main(void)
18 {
19 setlocale(LC_ALL, "");
20
21 char *mbs = "€5 ± π"; // That's the exact price range
22
23 printf("\"%s\" is %zu characters...\n", mbs, mbstrlen(mbs));
24 printf("but it's %zu bytes!\n", strlen(mbs));
25 }
Output on my system: "€5 ± π" is 6 characters... but it's 10 bytes! Chapter 48.
See Also
mbrtowc(), mbstowcs(), wcsrtombs(), strlen(), errno
48.33 wcsrtombs()
Convert a wide character string to a multibyte string restartably
Synopsis
#include
size_t wcsrtombs(char * restrict dst, const wchar_t ** restrict src, size_t len, mbstate_t * restrict ps);
Description If you have a wide character string, you can convert it to a multibyte character string in the current locale using this function. At most len bytes of data will be stored in the buffer pointed to by dst. Conversion will stop just after the NUL terminator is copied, or len bytes get copied, or some other error occurs. If dst is a NULL pointer, no result is stored. You might do this if you’re just interested in the return value (nominally the number of bytes this would use in a multibyte string, not including the NUL terminator). If dst is not a NULL pointer, the pointer pointed to by src will get modified to indicate how much of the data was copied. If it contains NULL at the end, it means everything went well. In this case, the state ps will be set to the initial conversion state. If len was reached or an error occurred, it’ll point one address past dst+len.
Return Value If everything goes well, returns the number of bytes needed for the multibyte string, not counting the NUL terminator. If any character in the string doesn’t correspond to a valid multibyte character in the currently locale, it returns (size_t)(-1) and EILSEQ is stored in errno.
Example Here we’ll convert the wide string “€5 ± π” into a multibyte character string:
1 #include
2 #include
3 #include
4
5 #define MB_STR_SIZE 20
6
7 int main(void)
8 {
9 const wchar_t *wcs = L"€5 ± π"; // That's the exact price range
10
11 char mbs[MB_STR_SIZE];
12
13 setlocale(LC_ALL, "");
14
15 mbstate_t state; Chapter 48.
16 memset(&state, 0, sizeof state);
17
18 size_t count = wcsrtombs(mbs, &wcs, MB_STR_SIZE, &state);
19
20 wprintf(L"Multibyte string \"%s\" is %d bytes\n", mbs, count);
21 }
Here’s another example helper function that malloc()s just enough memory to hold the converted string, then returns the result. (Which must later be freed, of course, to prevent leaking memory.)
1 #include
2 #include
3 #include
4 #include
5 #include
6
7 char *get_mb_string(const wchar_t *wcs)
8 {
9 setlocale(LC_ALL, "");
10
11 mbstate_t state;
12 memset(&state, 0, sizeof state);
13
14 // Need a copy of this because wcsrtombs changes it
15 const wchar_t *p = wcs;
16
17 // Compute the number of bytes needed to hold the result
18 size_t bytes_needed = wcsrtombs(NULL, &p, SIZE_MAX, &state);
19
20 // If we didn't get a good full conversion, forget it
21 if (bytes_needed == (size_t)(-1))
22 return NULL;
23
24 // Allocate space for result
25 char *mbs = malloc(bytes_needed + 1); // +1 for NUL terminator
26
27 // Set conversion state to initial state
28 memset(&state, 0, sizeof state);
29
30 // Convert and store result
31 wcsrtombs(mbs, &wcs, bytes_needed + 1, &state);
32
33 // Make sure things went well
34 if (wcs != NULL) {
35 free(mbs);
36 return NULL;
37 }
38
39 // Success!
40 return mbs;
41 }
42
43 int main(void)
44 {
45 char *mbs = get_mb_string(L"€5 ± π");
46
47 wprintf(L"Multibyte result: \"%s\"\n", mbs);
48 Chapter 48.
49 free(mbs);
50 }
See Also
wcrtomb(), wcstombs(), mbsrtowcs(), errno Chapter 49
Function Description iswalnum() Test if a wide character is alphanumeric. iswalpha() Tests if a wide character is alphabetic iswblank() Tests if this is a wide blank character iswcntrl() Tests if this is a wide control character. iswctype() Determine wide character classification iswdigit() Test if this wide character is a digit iswgraph() Test to see if a wide character is a printable non-space iswlower() Tests if a wide character is lowercase iswprint() Tests if a wide character is printable iswpunct() Test if a wide character is punctuation iswspace() Test if a wide character is whitespace iswupper() Tests if a wide character is uppercase iswxdigit() Tests if a wide character is a hexadecimal digit towctrans() Convert wide characters to upper or lowercase towlower() Convert an uppercase wide character to lowercase towupper() Convert a lowercase wide character to uppercase wctrans() Helper function for towctrans() wctype() Helper function for iswctype()
This is like
49.1 iswalnum()
Test if a wide character is alphanumeric.
Synopsis
#include
500 Chapter 49.
Description
Basically tests if a character is alphabetic (A-Z or a-z) or a digit (0-9). But some other characters might also qualify based on the locale. This is equivalent to testing if iswalpha() or iswdigit() is true.
Return Value Returns true if the character is alphanumeric.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswalnum(L'a')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswalnum(L'B')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswalnum(L'5')? L"yes": L"no"); // yes
11 wprintf(L"%ls\n", iswalnum(L'?')? L"yes": L"no"); // no
12 }
See Also
iswalpha(), iswdigit(), isalnum()
49.2 iswalpha()
Tests if a wide character is alphabetic
Synopsis
#include
int iswalpha(wint_t wc);
Description
Basically tests if a character is alphabetic (A-Z or a-z). But some other characters might also qualify based on the locale. (If other characters qualify, they won’t be control characters, digits, punctuation, or spaces.) This is the same as testing for iswupper() or iswlower().
Return Value Returns true if the character is alphabetic.
Example
1 #include
2 #include
3 Chapter 49.
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswalpha(L'a')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswalpha(L'B')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswalpha(L'5')? L"yes": L"no"); // no
11 wprintf(L"%ls\n", iswalpha(L'?')? L"yes": L"no"); // no
12 }
See Also
iswalnum(), isalpha()
49.3 iswblank()
Tests if this is a wide blank character
Synopsis
#include
int iswblank(wint_t wc);
Description Blank characters are whitespace that are also used as word separators on the same line. In the “C” locale, the only blank characters are space and tab. Other locales might define other blank characters.
Return Value Returns true if this is a blank character.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswblank(L'')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswblank(L'\t')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswblank(L'\n')? L"yes": L"no"); // no
11 wprintf(L"%ls\n", iswblank(L'a')? L"yes": L"no"); // no
12 wprintf(L"%ls\n", iswblank(L'?')? L"yes": L"no"); // no
13 }
See Also
iswspace(), isblank() Chapter 49.
49.4 iswcntrl()
Tests if this is a wide control character.
Synopsis
#include
int iswcntrl(wint_t wc);
Description The spec is pretty barren, here. But I’m just going to assume that it works like the non-wide version. So let’s look at that. A control character is a locale-specific non-printing character. For the “C” locale, this means control characters are in the range 0x00 to 0x1F (the character right before SPACE) and 0x7F (the DEL character). Basically if it’s not an ASCII (or Unicode less than 128) printable character, it’s a control character in the “C” locale. Probably.
Return Value Returns true if this is a control character.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswcntrl(L'\t')? L"yes": L"no"); // yes (tab)
9 wprintf(L"%ls\n", iswcntrl(L'\n')? L"yes": L"no"); // yes (newline)
10 wprintf(L"%ls\n", iswcntrl(L'\r')? L"yes": L"no"); // yes (return)
11 wprintf(L"%ls\n", iswcntrl(L'\a')? L"yes": L"no"); // yes (bell)
12 wprintf(L"%ls\n", iswcntrl(L'')? L"yes": L"no"); // no
13 wprintf(L"%ls\n", iswcntrl(L'a')? L"yes": L"no"); // no
14 wprintf(L"%ls\n", iswcntrl(L'?')? L"yes": L"no"); // no
15 }
See Also
iscntrl()
49.5 iswdigit()
Test if this wide character is a digit Chapter 49.
Synopsis
#include
int iswdigit(wint_t wc);
Description
Tests if the wide character is a digit (0-9).
Return Value Returns true if the character is a digit.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswdigit(L'0')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswdigit(L'5')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswdigit(L'a')? L"yes": L"no"); // no
11 wprintf(L"%ls\n", iswdigit(L'B')? L"yes": L"no"); // no
12 wprintf(L"%ls\n", iswdigit(L'?')? L"yes": L"no"); // no
13 }
See Also
iswalnum(), isdigit()
49.6 iswgraph()
Test to see if a wide character is a printable non-space
Synopsis
#include
int iswgraph(wint_t wc);
Description Returns true if this is a printable (non-control) character and also not a whitespace character. Basically if iswprint() is true and iswspace() is false.
Return Value Returns true if this is a printable non-space character. Chapter 49.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswgraph(L'0')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswgraph(L'a')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswgraph(L'B')? L"yes": L"no"); // yes
11 wprintf(L"%ls\n", iswgraph(L'?')? L"yes": L"no"); // yes
12 wprintf(L"%ls\n", iswgraph(L'')? L"yes": L"no"); // no
13 wprintf(L"%ls\n", iswgraph(L'\n')? L"yes": L"no"); // no
14 }
See Also
iswprint(), iswspace(), isgraph()
49.7 iswlower()
Tests if a wide character is lowercase
Synopsis
#include
int iswlower(wint_t wc);
Description
Tests if a character is lowercase, in the range a-z. In other locales, there could be other lowercase characters. In all cases, to be lowercase, the following must be true: !iswcntrl(c) && !iswdigit(c) && !iswpunct(c) && !iswspace(c)
Return Value Returns true if the wide character is lowercase.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswlower(L'c')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswlower(L'0')? L"yes": L"no"); // no Chapter 49.
10 wprintf(L"%ls\n", iswlower(L'B')? L"yes": L"no"); // no
11 wprintf(L"%ls\n", iswlower(L'?')? L"yes": L"no"); // no
12 wprintf(L"%ls\n", iswlower(L'')? L"yes": L"no"); // no
13 }
See Also
islower(), iswupper(), iswalpha(), towupper(), towlower()
49.8 iswprint()
Tests if a wide character is printable
Synopsis
#include
int iswprint(wint_t wc);
Description
Tests if a wide character is printable, including space (''). So like isgraph(), except space isn’t left out in the cold.
Return Value
Returns true if the wide character is printable, including space ('').
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswprint(L'c')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswprint(L'0')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswprint(L'')? L"yes": L"no"); // yes
11 wprintf(L"%ls\n", iswprint(L'\r')? L"yes": L"no"); // no
12 }
See Also
isprint(), iswgraph(), iswcntrl()
49.9 iswpunct()
Test if a wide character is punctuation Chapter 49.
Synopsis
#include
int iswpunct(wint_t wc);
Description Tests if a wide character is punctuation. This means for any given locale: !isspace(c) && !isalnum(c)
Return Value True if the wide character is punctuation.
Example Results may vary based on locale.
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswpunct(L',')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswpunct(L'!')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswpunct(L'c')? L"yes": L"no"); // no
11 wprintf(L"%ls\n", iswpunct(L'0')? L"yes": L"no"); // no
12 wprintf(L"%ls\n", iswpunct(L'')? L"yes": L"no"); // no
13 wprintf(L"%ls\n", iswpunct(L'\n')? L"yes": L"no"); // no
14 }
See Also
ispunct(), iswspace(), iswalnum()
49.10 iswspace()
Test if a wide character is whitespace
Synopsis
#include
int iswspace(wint_t wc);
Description
Tests if c is a whitespace character. These are probably: • Space ('') • Formfeed ('\f') Chapter 49.
• Newline ('\n') • Carriage Return ('\r') • Horizontal Tab ('\t') • Vertical Tab ('\v') Other locales might specify other whitespace characters. iswalnum(), iswgraph(), and iswpunct() are all false for all whitespace characters.
Return Value True if the character is whitespace.
Example Results may vary based on locale.
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswspace(L'')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswspace(L'\n')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswspace(L'\t')? L"yes": L"no"); // yes
11 wprintf(L"%ls\n", iswspace(L',')? L"yes": L"no"); // no
12 wprintf(L"%ls\n", iswspace(L'!')? L"yes": L"no"); // no
13 wprintf(L"%ls\n", iswspace(L'c')? L"yes": L"no"); // no
14 }
See Also
isspace(), iswblank()
49.11 iswupper()
Tests if a wide character is uppercase
Synopsis
#include
int iswupper(wint_t wc);
Description Tests if a character is uppercase in the current locale. To be uppercase, the following must be true: !iscntrl(c) && !isdigit(c) && !ispunct(c) && !isspace(c)
Return Value Returns true if the wide character is uppercase. Chapter 49.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswupper(L'B')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswupper(L'c')? L"yes": L"no"); // no
10 wprintf(L"%ls\n", iswupper(L'0')? L"yes": L"no"); // no
11 wprintf(L"%ls\n", iswupper(L'?')? L"yes": L"no"); // no
12 wprintf(L"%ls\n", iswupper(L'')? L"yes": L"no"); // no
13 }
See Also
isupper(), iswlower(), iswalpha(), towupper(), towlower()
49.12 iswxdigit()
Tests if a wide character is a hexadecimal digit
Synopsis
#include
int iswxdigit(wint_t wc);
Description
Returns true if the wide character is a hexadecimal digit. Namely if it’s 0-9, a-f, or A-F.
Return Value True if the character is a hexadecimal digit.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // testing this char
7 // v
8 wprintf(L"%ls\n", iswxdigit(L'B')? L"yes": L"no"); // yes
9 wprintf(L"%ls\n", iswxdigit(L'c')? L"yes": L"no"); // yes
10 wprintf(L"%ls\n", iswxdigit(L'2')? L"yes": L"no"); // yes
11 wprintf(L"%ls\n", iswxdigit(L'G')? L"yes": L"no"); // no
12 wprintf(L"%ls\n", iswxdigit(L'?')? L"yes": L"no"); // no
13 } Chapter 49.
See Also isxdigit(), iswdigit()
49.13 iswctype()
Determine wide character classification
Synopsis
#include
Description This is the Swiss Army knife of classification functions; it’s all the other ones rolled into one. You call it with something like this: if (iswctype(c, wctype("digit"))) // or "alpha" or "space" or... and it behaves just like you’d called: if (iswdigit(c))
The difference is that you can specify the type of matching you want to do as a string at runtime, which might be convenient. iswctype() relies on the return value from the wctype() call to get its work done. Stolen from the spec, here are the iswctype() calls and their equivalents:
iswctype() call Hard-coded equivalent iswctype(c, wctype("alnum")) iswalnum(c) iswctype(c, wctype("alpha")) iswalpha(c) iswctype(c, wctype("blank")) iswblank(c) iswctype(c, wctype("cntrl")) iswcntrl(c) iswctype(c, wctype("digit")) iswdigit(c) iswctype(c, wctype("graph")) iswgraph(c) iswctype(c, wctype("lower")) iswlower(c) iswctype(c, wctype("print")) iswprint(c) iswctype(c, wctype("punct")) iswpunct(c) iswctype(c, wctype("space")) iswspace(c) iswctype(c, wctype("upper")) iswupper(c) iswctype(c, wctype("xdigit")) iswxdigit(c)
See the wctype() documentation for how that helper function works.
Return Value
Returns true if the wide character wc matches the character class in desc.
Example Test for a given character classification at when the classification isn’t known at compile time: Chapter 49.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 wchar_t c; // Holds a single wide character (to test)
8 char desc[128]; // Holds the character class
9
10 // Get the character and classification from the user
11 wprintf(L"Enter a character and character class: ");
12 fflush(stdout);
13 wscanf(L"%lc %s", &c, desc);
14
15 // Compute the type from the given class
16 wctype_t t = wctype(desc);
17
18 if (t == 0)
19 // If the type is 0, it's an unknown class
20 wprintf(L"Unknown character class: \"%s\"\n", desc);
21 else {
22 // Otherwise, let's test the character and see if its that
23 // classification
24 if (iswctype(c, t))
25 wprintf(L"Yes! '%lc' is %s!\n", c, desc);
26 else
27 wprintf(L"Nope! '%lc' is not %s.\n", c, desc);
28 }
29 }
Output: Enter a character and character class: 5 digit Yes! '5' is digit!
Enter a character and character class: b digit Nope! 'b' is not digit.
Enter a character and character class: x alnum Yes! 'x' is alnum!
See Also
wctype()
49.14 wctype()
Helper function for iswctype()
Synopsis
#include
wctype_t wctype(const char *property); Chapter 49.
Description
This function returns an opaque value for the given property that is meant to be passed as the second argument to iswctype(). The returned value is of type wctype_t. Valid properties in all locales are: "alnum" "alpha" "blank" "cntrl" "digit" "graph" "lower" "print" "punct" "space" "upper" "xdigit"
Other properties might be defined as determined by the LC_CTYPE category of the current locale. See the iswctype() reference page for more usage details.
Return Value
Returns the wctype_t value associated with the given property. If an invalid value is passed for property, returns 0.
Example Test for a given character classification at when the classification isn’t known at compile time:
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 wchar_t c; // Holds a single wide character (to test)
8 char desc[128]; // Holds the character class
9
10 // Get the character and classification from the user
11 wprintf(L"Enter a character and character class: ");
12 fflush(stdout);
13 wscanf(L"%lc %s", &c, desc);
14
15 // Compute the type from the given class
16 wctype_t t = wctype(desc);
17
18 if (t == 0)
19 // If the type is 0, it's an unknown class
20 wprintf(L"Unknown character class: \"%s\"\n", desc);
21 else {
22 // Otherwise, let's test the character and see if its that
23 // classification
24 if (iswctype(c, t))
25 wprintf(L"Yes! '%lc' is %s!\n", c, desc);
26 else
27 wprintf(L"Nope! '%lc' is not %s.\n", c, desc);
28 }
29 }
Output: Enter a character and character class: 5 digit Yes! '5' is digit!
Enter a character and character class: b digit Chapter 49.
Nope! 'b' is not digit.
Enter a character and character class: x alnum Yes! 'x' is alnum!
See Also
iswctype()
49.15 towlower()
Convert an uppercase wide character to lowercase
Synopsis
#include
wint_t towlower(wint_t wc);
Description
If the character is upper (i.e. iswupper(c) is true), this function returns the corresponding lowercase letter. Different locales might have different upper and lowercase letters.
Return Value
If the letter wc is uppercase, a lowercase version of that letter will be returned according to the current locale. If the letter is not uppercase, wc is returned unchanged.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // changing this char
7 // v
8 wprintf(L"%lc\n", towlower(L'B')); // b (made lowercase!)
9 wprintf(L"%lc\n", towlower(L'e')); // e (unchanged)
10 wprintf(L"%lc\n", towlower(L'!')); // ! (unchanged)
11 }
See Also
tolower(), towupper(), iswlower(), iswupper() Chapter 49.
49.16 towupper()
Convert a lowercase wide character to uppercase
Synopsis
#include
wint_t towupper(wint_t wc);
Description
If the character is upper (i.e. iswupper(c) is true), this function returns the corresponding lowercase letter. Different locales might have different upper and lowercase letters.
Return Value
If the letter wc is lowercase, an uppercase version of that letter will be returned according to the current locale. If the letter is not lowercase, wc is returned unchanged.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 // changing this char
7 // v
8 wprintf(L"%lc\n", towupper(L'B')); // B (unchanged)
9 wprintf(L"%lc\n", towupper(L'e')); // E (made uppercase!)
10 wprintf(L"%lc\n", towupper(L'!')); // ! (unchanged)
11 }
See Also
toupper(), towlower(), iswlower(), iswupper()
49.17 towctrans()
Convert wide characters to upper or lowercase
Synopsis
#include
wint_t towctrans(wint_t wc, wctrans_t desc); Chapter 49.
Description This is the Swiss Army knife of character conversion functions; it’s all the other ones rolled into one. And by “all the other ones” I mean towupper() and towlower(), since those are the only ones there are. You call it with something like this: if (towctrans(c, wctrans("toupper"))) // or "tolower"
and it behaves just like you’d called: towupper(c);
The difference is that you can specify the type of conversion you want to do as a string at runtime, which might be convenient. towctrans() relies on the return value from the wctrans() call to get its work done.
towctrans() call Hard-coded equivalent towctrans(c, wctrans("toupper")) towupper(c) towctrans(c, wctrans("tolower")) towlower(c)
See the wctrans() documentation for how that helper function works.
Return Value
Returns the character wc as if run through towupper() or towlower(), depending on the value of desc. If the character already matches the classification, it is returned as-is.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 wchar_t c; // Holds a single wide character (to test)
8 char desc[128]; // Holds the conversion type
9
10 // Get the character and conversion type from the user
11 wprintf(L"Enter a character and conversion type: ");
12 fflush(stdout);
13 wscanf(L"%lc %s", &c, desc);
14
15 // Compute the type from the given conversion type
16 wctrans_t t = wctrans(desc);
17
18 if (t == 0)
19 // If the type is 0, it's an unknown conversion type
20 wprintf(L"Unknown conversion: \"%s\"\n", desc);
21 else {
22 // Otherwise, let's do the conversion
23 wint_t result = towctrans(c, t);
24 wprintf(L"'%lc' -> %s -> '%lc'\n", c, desc, result);
25 }
26 }
Output on my system: Chapter 49.
Enter a character and conversion type: b toupper 'b' -> toupper -> 'B'
Enter a character and conversion type: B toupper 'B' -> toupper -> 'B'
Enter a character and conversion type: B tolower 'B' -> tolower -> 'b'
Enter a character and conversion type: ! toupper '!' -> toupper -> '!'
See Also
wctrans(), towupper(), towlower()
49.18 wctrans()
Helper function for towctrans()
Synopsis
#include
wctrans_t wctrans(const char *property);
Description
This is a helper function for generating the second argument to towctrans(). You can pass in one of two things for the property: • toupper to make towctrans() behave like towupper() • tolower to make towctrans() behave like towlower()
Return Value
On success, returns a value that can be used as the desc argument to towctrans(). Otherwise, if the property isn’t recognized, returns 0.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 wchar_t c; // Holds a single wide character (to test)
8 char desc[128]; // Holds the conversion type
9
10 // Get the character and conversion type from the user
11 wprintf(L"Enter a character and conversion type: ");
12 fflush(stdout);
13 wscanf(L"%lc %s", &c, desc); Chapter 49.
14
15 // Compute the type from the given conversion type
16 wctrans_t t = wctrans(desc);
17
18 if (t == 0)
19 // If the type is 0, it's an unknown conversion type
20 wprintf(L"Unknown conversion: \"%s\"\n", desc);
21 else {
22 // Otherwise, let's do the conversion
23 wint_t result = towctrans(c, t);
24 wprintf(L"'%lc' -> %s -> '%lc'\n", c, desc, result);
25 }
26 }
Output on my system: Enter a character and conversion type: b toupper 'b' -> toupper -> 'B'
Enter a character and conversion type: B toupper 'B' -> toupper -> 'B'
Enter a character and conversion type: B tolower 'B' -> tolower -> 'b'
Enter a character and conversion type: ! toupper '!' -> toupper -> '!'
See Also
towctrans() Chapter 50
Function Description signal() Set a signal handler for a given signal raise() Cause a signal to be raised
Handle signals in a portable way, kind of! These signals get raised for a variety of reasons such as CTRL-C being hit, requests to terminate for external programs, memory access violations, and so on. Your OS likely defines a plethora of other signals, as well. This system is pretty limited, as seen below. If you’re on Unix, it’s almost certain your OS has far superior signal handling capabilities than the C standard library. Check out sigaction1.
50.1 signal()
Set a signal handler for a given signal
Synopsis
#include
Description How’s that for a function declaration? Let’s ignore it for a moment and just talk about what this function does. When a signal is raised, something is going to happen. This function lets you decide to do one of these things when the signal is raised: • Ignore the signal • Perform the default action • Have a specific function called The signal() function takes two arguments. The first, sig, is the name of the signal to handle.
1https://man.archlinux.org/man/sigaction.2.en
518 Chapter 50.
Signal Description SIGABRT Raised when abort() is called SIGFPE Floating-point arithmetic exception SIGILL CPU tried to execute an illegal instruction SIGINT Interrupt signal, as if CTRL-C were pressed SIGSEGV Segmention Violation: attempted to access restricted memory SIGTERM Termination request2
So that’s the first bit when you call signal()—tell it the signal in question: signal(SIGINT, ...
But what’s that func parameter? For spoilers, it’s a pointer to a function that takes an int argument and returns void. We can use this to call an arbitrary function when the signal occurs. Before we do that, though, let’s look at the easy ones: telling the system to ignore the signal or perform the default action (which it does by default if you never call signal()). You can set func to one of two special values to make this happen:
func Description SIG_DFL Perform the default action on this signal SIG_IGN Ignore this signal
For example: signal(SIGTERM, SIG_DFL); // Default action on SIGTERM signal(SIGINT, SIG_IGN); // Ignore SIGINT
But what if you want to have your own handler do something instead of the default or ignoring it? You can pass in your own function to be called. That’s what the crazy function signature is partially about. It’s saying that the argument can be a pointer to a function that takes an int argument and returns void. So if you wanted to call your handler, you could have code like this: int handler(int sig) { // Handle the signal } int main(void) { signal(SIGINT, handler);
What can you do in the signal handler? Not much. If the signal is due to abort() or raise(), the handler can’t call raise(). If the signal is not due to abort() or raise(), you’re only allowed to call these functions from the standard library (though the spec doesn’t prohibit calling other non-library functions): • abort() • _Exit() • quick_exit() • Functions in
2As if might be sent from Unix’s kill command.] Chapter 50.
In addition, if the signal was not due to abort() or raise(), the handler can’t access any object with static or thread-storage duration unless it’s lock-free. An exception is that you can assign to (but not read from!) a variable of type volatile sig_atomic_t. It’s up to the implementation, but the signal handler might be reset to SIG_DFL just before the handler is called. It’s undefined behavior to call signal() in a multithreaded program. It’s undefined behavior to return from the handler for SIGFPE, SIGILL, SIGSEGV, or any implementation- defined value. You must exit. The implementation might or might not prevent other signals from arising while in the signal handler.
Return Value
On success, signal() returns a pointer to the previous signal handler set by a call to signal() for that particular signal number. If you haven’t called it set, returns SIG_DFL. On failure, SIG_ERR is returned and errno is set to a positive value.
Example
Here’s a program that causes SIGINT to be ignored. Commonly you trigger this signal by hitting CTRL-C.
1 #include
2 #include
3
4 int main(void)
5 {
6 signal(SIGINT, SIG_IGN);
7
8 printf("You can't hit CTRL-C to exit this program. Try it!\n\n");
9 printf("Press return to exit, instead.");
10 fflush(stdout);
11 getchar();
12 }
Output: You can't hit CTRL-C to exit this program. Try it!
Press return to exit, instead.^C^C^C^C^C^C^C^C^C^C^C
This program sets the signal handler, then raises the signal. The signal handler fires.
1 #include
2 #include
3
4 void handler(int sig)
5 {
6 // Undefined behavior to call printf() if this handler was not
7 // as the result of a raise(), i.e. if you hit CTRL-C.
8
9 printf("Got signal %d!\n", sig);
10
11 // Common to reset the handler just in case the implementation set
12 // it to SIG_DFL when the signal occurred.
13
14 signal(sig, handler);
15 }
16 Chapter 50.
17 int main(void)
18 {
19 signal(SIGINT, handler);
20
21 raise(SIGINT);
22 raise(SIGINT);
23 raise(SIGINT);
24 }
Output: Got signal 2! Got signal 2! Got signal 2!
This example catches SIGINT but then sets a flag to 1. Then the main loop sees the flag and exits.
1 #include
2 #include
3
4 volatile sig_atomic_t x;
5
6 void handler(int sig)
7 {
8 x = 1;
9 }
10
11 int main(void)
12 {
13 signal(SIGINT, handler);
14
15 printf("Hit CTRL-C to exit\n");
16 while (x != 1);
17 }
See Also
raise(), abort()
50.2 raise()
Cause a signal to be raised
Synopsis
#include
int raise(int sig);
Description
Causes the signal handler for the signal sig to be called. If the handler is SIG_DFL or SIG_IGN, then the default action or no action happens. raise() returns after the signal handler has finished running. Chapter 50.
Interestingly, if you cause a signal to happen with raise(), you can call library functions from within the signal handler without causing undefined behavior. I’m not sure how this fact is practically useful, though.
Return Value
Returns 0 on success. Nonzero otherwise.
Example This program sets the signal handler, then raises the signal. The signal handler fires.
1 #include
2 #include
3
4 void handler(int sig)
5 {
6 // Undefined behavior to call printf() if this handler was not
7 // as the result of a raise(), i.e. if you hit CTRL-C.
8
9 printf("Got signal %d!\n", sig);
10
11 // Common to reset the handler just in case the implementation set
12 // it to SIG_DFL when the signal occurred.
13
14 signal(sig, handler);
15 }
16
17 int main(void)
18 {
19 signal(SIGINT, handler);
20
21 raise(SIGINT);
22 raise(SIGINT);
23 raise(SIGINT);
24 }
Output: Got signal 2! Got signal 2! Got signal 2!
See Also
signal() Chapter 51
Function Description setlocale() Set the locale localeconv() Get information about the current locale
The “locale” is the details of how the program should run given its physical location on the planet. For example, in one locale, a unit of money might be printed as $123, and in another €123. Or one locale might use ASCII encoding and another UTF-8 encoding. By default, the program runs in the “C” locale. It has a basic set of characters with a single-byte encoding. If you try to print UTF-8 characters in the C locale, nothing will print. You have to switch to a proper locale.
51.1 setlocale()
Set the locale
Synopsis
#include
Description
Sets the locale for the given category. Category is one of the following:
Category Description LC_ALL All of the following categories LC_COLLATE Affects the strcoll() and strxfrm() functions LC_CTYPE Affects the functions in
Category Description LC_TIME Affects the strftime() and wcsftime() functions
And there are three portable things you can pass in for locale; any other string passed in is implementation-defined and non-portable.
Locale Description "C" Set the program to the C locale "" (Empty string) Set the program to the native locale of this system NULL Change nothing; just return the current locale Other Set the program to an implementation-defined locale
The most common call, I’d wager, is this: // Set all locale settings to the local, native locale
setlocale(LC_ALL, "");
Handily, setlocale() returns the locale that was just set, so you could see what the actual locale is on your system.
Return Value On success, returns a pointer to the string representing the current locale. You may not modify this string, and it might be changed by subsequent calls to setlocale(). On failure, returns NULL.
Example Here we get the current locale. Then we set it to the native locale, and print out what that is.
1 #include
2 #include
3
4 int main(void)
5 {
6 char *loc;
7
8 // Get the current locale
9 loc = setlocale(LC_ALL, NULL);
10
11 printf("Starting locale: %s\n", loc);
12
13 // Set (and get) the locale to native locale
14 loc = setlocale(LC_ALL, "");
15
16 printf("Native locale: %s\n", loc);
17 }
Output on my system: Starting locale: C Native locale: en_US.UTF-8
Note that my native locale (on a Linux box) might be different from what you see. Nevertheless, I can explicitly set it on my system without a problem, or to any other locale I have installed: Chapter 51.
13 loc = setlocale(LC_ALL, "en_US.UTF-8"); // Non-portable
But again, your system might have different locales defined.
See Also
localeconv(), strcoll(), strxfrm(), strftime(), wcsftime(), printf(), scanf(),
51.2 localeconv()
Get information about the current locale
Synopsis
#include
struct lconv *localeconv(void);
Description
This function just returns a pointer to a struct lconv, but is still a bit of a powerhouse. The returned structure contains tons of information about the locale. Here are the fields of struct lconv and their meanings. First, some conventions. An _p_ means “positive”, and _n_ means “negative”, and int_ means “inter- national”. Though a lot of these are type char or char*, most (or the strings they point to) are actually treated as integers1. Before we go further, know that CHAR_MAX (from
Field Description char *mon_decimal_point Decimal pointer character for money, e.g. ".". char *mon_thousands_sep Thousands separator character for money, e.g. ",". char *mon_grouping Grouping description for money (see below). char *positive_sign Positive sign for money, e.g. "+" or "". char *negative_sign Negative sign for money, e.g. "-". char *currency_symbol Currency symbol, e.g. "$". char frac_digits When printing monetary amounts, how many digits to print past the decimal point, e.g. 2. char p_cs_precedes 1 if the currency_symbol comes before the value for a non-negative monetary amount, 0 if after. char n_cs_precedes 1 if the currency_symbol comes before the value for a negative monetary amount, 0 if after. char p_sep_by_space Determines the separation of the currency symbol from the value for non-negative amounts (see below). char n_sep_by_space Determines the separation of the currency symbol from the value for negative amounts (see below). char p_sign_posn Determines the positive_sign position for non-negative values. char p_sign_posn Determines the positive_sign position for negative values. char *int_curr_symbol International currency symbol, e.g. "USD ". char int_frac_digits International value for frac_digits. char int_p_cs_precedes International value for p_cs_precedes.
1Remember that char is just a byte-sized integer. Chapter 51.
Field Description char int_n_cs_precedes International value for n_cs_precedes. char int_p_sep_by_space International value for p_sep_by_space. char int_n_sep_by_space International value for n_sep_by_space. char int_p_sign_posn International value for p_sign_posn. char int_n_sign_posn International value for n_sign_posn.
Even though many of these have char type, the value stored within is meant to be accessed as an integer. All the sep_by_space variants deal with spacing around the currency sign. Valid values are:
Value Description 0 No space between currency symbol and value. 1 Separate the currency symbol (and sign, if any) from the value with a space. 2 Separate the sign symbol from the currency symbol (if adjacent) with a space, otherwise separate the sign symbol from the value with a space.
The sign_posn variants are determined by the following values:
Value Description 0 Put parens around the value and the currency symbol. 1 Put the sign string in front of the currency symbol and value. 2 Put the sign string after the currency symbol and value. 3 Put the sign string directly in front of the currency symbol. 4 Put the sign string directly behind the currency symbol.
For more information on the mon_grouping field, see “Monetary Digit Grouping” in the “Locale and Internationalization” chapter.
Return Value Returns a pointer to the structure containing the locale information. The program may not modify this structure. Subsequent calls to localeconv() may overwrite this structure, as might calls to setlocale() with LC_ALL, LC_MONETARY, or LC_NUMERIC.
Example Here’s a program to print the locale information for the native locale.
1 #include
2 #include
3 #include
4
5 void print_grouping(char *mg)
6 {
7 int done = 0;
8
9 while (!done) {
10 if (*mg == CHAR_MAX)
11 printf("CHAR_MAX ");
12 else
13 printf("%c ", *mg + '0'); Chapter 51.
14 done = *mg == CHAR_MAX || *mg == 0;
15 mg++;
16 }
17 }
18
19 int main(void)
20 {
21 setlocale(LC_ALL, "");
22
23 struct lconv *lc = localeconv();
24
25 printf("mon_decimal_point : %s\n", lc->mon_decimal_point);
26 printf("mon_thousands_sep : %s\n", lc->mon_thousands_sep);
27 printf("mon_grouping :");
28 print_grouping(lc->mon_grouping);
29 printf("\n");
30 printf("positive_sign : %s\n", lc->positive_sign);
31 printf("negative_sign : %s\n", lc->negative_sign);
32 printf("currency_symbol : %s\n", lc->currency_symbol);
33 printf("frac_digits : %c\n", lc->frac_digits);
34 printf("p_cs_precedes : %c\n", lc->p_cs_precedes);
35 printf("n_cs_precedes : %c\n", lc->n_cs_precedes);
36 printf("p_sep_by_space : %c\n", lc->p_sep_by_space);
37 printf("n_sep_by_space : %c\n", lc->n_sep_by_space);
38 printf("p_sign_posn : %c\n", lc->p_sign_posn);
39 printf("p_sign_posn : %c\n", lc->p_sign_posn);
40 printf("int_curr_symbol : %s\n", lc->int_curr_symbol);
41 printf("int_frac_digits : %c\n", lc->int_frac_digits);
42 printf("int_p_cs_precedes : %c\n", lc->int_p_cs_precedes);
43 printf("int_n_cs_precedes : %c\n", lc->int_n_cs_precedes);
44 printf("int_p_sep_by_space: %c\n", lc->int_p_sep_by_space);
45 printf("int_n_sep_by_space: %c\n", lc->int_n_sep_by_space);
46 printf("int_p_sign_posn : %c\n", lc->int_p_sign_posn);
47 printf("int_n_sign_posn : %c\n", lc->int_n_sign_posn);
48 }
Output on my system: mon_decimal_point : . mon_thousands_sep : , mon_grouping : 3 3 0 positive_sign : negative_sign :- currency_symbol : $ frac_digits : 2 p_cs_precedes : 1 n_cs_precedes : 1 p_sep_by_space : 0 n_sep_by_space : 0 p_sign_posn : 1 p_sign_posn : 1 int_curr_symbol : USD int_frac_digits : 2 int_p_cs_precedes : 1 int_n_cs_precedes : 1 int_p_sep_by_space: 1 int_n_sep_by_space: 1 int_p_sign_posn : 1 Chapter 51.
See Also setlocale() Chapter 52
The complex functions in this reference section come in three flavors each: double complex, float complex, and long double complex. The float variants end with f and the long double variants end with l, e.g. for complex cosine: ccos() double complex ccosf() float complex ccosl() long double complex
The table below only lists the double complex version for brevity.
Function Description cabs() Compute the complex absolute value cacos() Compute the complex arc-cosine cacosh() Compute the complex arc hyperbolic cosine carg() Compute the complex argument casin() Compute the complex arc-sine casinh() Compute the complex arc hyperbolic sine catan() Compute the complex arc-tangent catanh() Compute the complex arc hyperbolic tangent ccos() Compute the complex cosine ccosh() Compute the complex hyperbolic cosine cexp() Compute the complex base-푒 exponential cimag() Returns the imaginary part of a complex number clog() Compute the complex logarithm CMPLX() Build a complex value from real and imaginary types conj() Compute the conjugate of a complex number cproj() Compute the projection of a complex number creal() Returns the real part of a complex number csin() Compute the complex sine csinh() Compute the complex hyperbolic sine csqrt() Compute the complex square root ctan() Compute the complex tangent ctanh() Compute the complex hyperbolic tangent
You can test for complex number support by looking at the __STDC_NO_COMPLEX__ macro. If it’s defined, complex numbers aren’t available. There are possibly two types of numbers defined: complex and imaginary. No system I’m currently aware of implements imaginary types.
529 Chapter 52.
The complex types, which are a real value plus a multiple of 푖, are: float complex double complex long double complex
The imaginary types, which hold a multiple of 푖, are: float imaginary double imaginary long double imaginary √ The mathematical value 푖 = −1 is represented by the symbol _Complex_I or _Imaginary_I, if it exists. The The macro I will be preferentially set to _Imaginary_I (if it exists), or to _Complex_I otherwise. You can write imaginary literals (if supported) using this notation: double imaginary x = 3.4 * I;
You can write complex literals using regular complex notation: double complex x = 1.2 + 3.4 * I; or build them with the CMPLX() macro: double complex x = CMPLX(1.2, 3.4); // Like 1.2 + 3.4 * I
The latter has the advantage of handing special cases of complex numbers correctly (like those involving infinity or signed zeroes) as if _Imaginary_I were present, even if it’s not. All angular values are in radians. Some functions have discontinuities called branch cuts. Now, I’m no mathematician so I can’t really talk sensibly about this, but if you’re here, I like to think you know what you’re doing when it comes to this side of things. If you system has signed zeroes, you can tell which side of the cut you’re on by the sign. And you can’t if you don’t. The spec elaborates: Implementations that do not support a signed zero […] cannot distinguish the sides of branch cuts. These implementations shall map a cut so the function is continuous as the cut is approached coming around the finite endpoint of the cut in a counter clockwise direction. (Branch cuts for the functions specified here have just one finite endpoint.) For example, for the square root function, coming counter clockwise around the finite endpoint of the cut along the negative real axis approaches the cut from above, so the cut maps to the positive imaginary axis. Finally, there’s a pragma called CX_LIMITED_RANGE that can be turned on and off (default is off). You can turn it on with: #pragma STDC CX_LIMITED_RANGE ON
It allows for certain intermediate operations to underflow, overflow, or deal badly with infinity, presumably for a tradeoff in speed. If you’re sure these types of errors won’t occur with the numbers you’re using AND you’re trying to get as much speed out as you can, you could turn this macro on. The spec also elaborates here: The purpose of the pragma is to allow the implementation to use the formulas: (푥 + 푖푦) × (푢 + 푖푣) = (푥푢 − 푦푣) + 푖(푦푢 + 푥푣) (푥 + 푖푦)/(푢 + 푖푣) = [(푥푢 + 푦푣) + 푖(푦푢 − 푥푣)]/(푢2 + 푣2) |푥 + 푖푦| = √푥2 + 푦2 where the programmer can determine they are safe. Chapter 52.
52.1 cacos() cacosf() cacosl()
Compute the complex arc-cosine
Synopsis
#include
double complex cacos(double complex z);
float complex cacosf(float complex z);
long double complex cacosl(long double complex z);
Description Computes the complex arc-cosine of a complex number. The complex number z will have an imaginary component in the range [0, 휋], and the real component is unbounded. There are branch cuts outside the interval [−1, +1] on the real axis.
Return Value
Returns the complex arc-cosine of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = cacos(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 0.195321 + -2.788006i
See Also
ccos(), casin(), catan()
52.2 casin() casinf() casinl()
Compute the complex arc-sine Chapter 52.
Synopsis
#include
double complex casin(double complex z);
float complex casinf(float complex z);
long double complex casinl(long double complex z);
Description Computes the complex arc-sine of a complex number. The complex number z will have an imaginary component in the range [−휋/2, +휋/2], and the real component is unbounded. There are branch cuts outside the interval [−1, +1] on the real axis.
Return Value
Returns the complex arc-sine of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = casin(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 1.375476 + 2.788006i
See Also
csin(), cacos(), catan()
52.3 catan() catanf() catanl()
Compute the complex arc-tangent
Synopsis
#include
double complex catan(double complex z);
float complex catanf(float complex z); Chapter 52.
long double complex catanl(long double complex z);
Description Computes the complex arc-tangent of a complex number. The complex number z will have an real component in the range [−휋/2, +휋/2], and the imaginary component is unbounded. There are branch cuts outside the interval [−푖, +푖] on the imaginary axis.
Return Value
Returns the complex arc-tangent of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double wheat = 8;
7 double sheep = 1.5708;
8
9 double complex x = wheat + sheep * I;
10
11 double complex y = catan(x);
12
13 printf("Result: %f + %fi\n", creal(y), cimag(y));
14 }
Output: Result: 1.450947 + 0.023299i
See Also
ctan(), cacos(), casin()
52.4 ccos() ccosf() ccosl()
Compute the complex cosine
Synopsis
#include
double complex ccos(double complex z);
float complex ccosf(float complex z);
long double complex ccosl(long double complex z); Chapter 52.
Description Computes the complex cosine of a complex number.
Return Value
Returns the complex cosine of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = ccos(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: -0.365087 + -2.276818i
See Also
csin(), ctan(), cacos()
52.5 csin() csinf() csinl()
Compute the complex sine
Synopsis
#include
double complex csin(double complex z);
float complex csinf(float complex z);
long double complex csinl(long double complex z);
Description Computes the complex sine of a complex number.
Return Value
Returns the complex sine of z.
Example Chapter 52.
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = csin(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 2.482485 + -0.334840i
See Also
ccos(), ctan(), casin()
52.6 ctan() ctanf() ctanl()
Compute the complex tangent
Synopsis
#include
double complex ctan(double complex z);
float complex ctanf(float complex z);
long double complex ctanl(long double complex z);
Description Computes the complex tangent of a complex number.
Return Value
Returns the complex tangent of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = ctan(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 } Chapter 52.
Output: Result: -0.027073 + 1.085990i
See Also
ccos(), csin(), catan()
52.7 cacosh() cacoshf() cacoshl()
Compute the complex arc hyperbolic cosine
Synopsis
#include
double complex cacosh(double complex z);
float complex cacoshf(float complex z);
long double complex cacoshl(long double complex z);
Description Computes the complex arc hyperbolic cosine of a complex number. There is a branch cut at values less than 1 on the real axis. The return value will be non-negative on the real number axis, and in the range [−푖휋, +푖휋] on the imagi- nary axis.
Return Value
Returns the complex arc hyperbolic cosine of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = cacosh(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 2.788006 + 0.195321i
See Also
casinh(), catanh(), acosh() Chapter 52.
52.8 casinh() casinhf() casinhl()
Compute the complex arc hyperbolic sine
Synopsis
#include
double complex casinh(double complex z);
float complex casinhf(float complex z);
long double complex casinhl(long double complex z);
Description Computes the complex arc hyperbolic sine of a complex number. There are branch cuts outside [−푖, +푖] on the imaginary axis. The return value will be unbounded on the real number axis, and in the range [−푖휋/2, +푖휋/2] on the imaginary axis.
Return Value
Returns the complex arc hyperbolic sine of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = casinh(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 2.794970 + 0.192476i
See Also
cacosh(), catanh(), asinh()
52.9 catanh() catanhf() catanhl()
Compute the complex arc hyperbolic tangent
Synopsis Chapter 52.
#include
double complex catanh(double complex z);
float complex catanhf(float complex z);
long double complex catanhl(long double complex z);
Description Computes the complex arc hyperbolic tangent of a complex number. There are branch cuts outside [−1, +1] on the real axis. The return value will be unbounded on the real number axis, and in the range [−푖휋/2, +푖휋/2] on the imaginary axis.
Return Value
Returns the complex arc hyperbolic tangent of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = catanh(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 0.120877 + 1.546821i
See Also
cacosh(), casinh(), atanh()
52.10 ccosh() ccoshf() ccoshl()
Compute the complex hyperbolic cosine
Synopsis
#include
double complex ccosh(double complex z);
float complex ccoshf(float complex z);
long double complex ccoshl(long double complex z); Chapter 52.
Description Computes the complex hyperbolic cosine of a complex number.
Return Value
Returns the complex hyperbolic cosine of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = ccosh(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: -0.005475 + 1490.478826i
See Also
csinh(), ctanh(), ccos()
52.11 csinh() csinhf() csinhl()
Compute the complex hyperbolic sine
Synopsis
#include
double complex csinh(double complex z);
float complex csinhf(float complex z);
long double complex csinhl(long double complex z);
Description Computes the complex hyperbolic sine of a complex number.
Return Value
Returns the complex hyperbolic sine of z.
Example Chapter 52.
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = csinh(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: -0.005475 + 1490.479161i
See Also
ccosh(), ctanh(), csin()
52.12 ctanh() ctanhf() ctanhl()
Compute the complex hyperbolic tangent
Synopsis
#include
double complex ctanh(double complex z);
float complex ctanhf(float complex z);
long double complex ctanhl(long double complex z);
Description Computes the complex hyperbolic tangent of a complex number.
Return Value
Returns the complex hyperbolic tangent of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 8 + 1.5708 * I;
7
8 double complex y = ctanh(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 } Chapter 52.
Output: Result: 1.000000 + -0.000000i
See Also
ccosh(), csinh(), ctan()
52.13 cexp() cexpf() cexpl()
Compute the complex base-푒 exponential
Synopsis
#include
double complex cexp(double complex z);
float complex cexpf(float complex z);
long double complex cexpl(long double complex z);
Description
Computes the complex base-푒 exponential of z.
Return Value
Returns the complex base-푒 exponential of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double complex y = cexp(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: -1.131204 + 2.471727i
See Also
cpow(), clog(), exp() Chapter 52.
52.14 clog() clogf() clogl()
Compute the complex logarithm
Synopsis
#include
double complex clog(double complex z);
float complex clogf(float complex z);
long double complex clogl(long double complex z);
Description
Compute the base-푒 complex logarithm of z. There is a branch cut on the negative real axis. The returns value is unbounded on the real axis and in the range [−푖휋, +푖휋] on the imaginary axis.
Return Value
Returns the base-푒 complex logarithm of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double complex y = clog(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 0.804719 + 1.107149i
See Also
cexp(), log()
52.15 cabs() cabsf() cabsl()
Compute the complex absolute value
Synopsis
#include
double cabs(double complex z); Chapter 52.
float cabsf(float complex z);
long double cabsl(long double complex z);
Description
Computes the complex absolute value of z.
Return Value
Returns the complex absolute value of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double complex y = cabs(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 2.236068 + 0.000000i
See Also
fabs(), abs()
52.16 cpow() cpowf() cpowl()
Compute complex power
Synopsis
#include
double complex cpow(double complex x, double complex y);
float complex cpowf(float complex x, float complex y);
long double complex cpowl(long double complex x, long double complex y);
Description Computes the complex 푥푦. There is a branch cut for x along the negative real axis. Chapter 52.
Return Value Returns the complex 푥푦.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7 double complex y = 3 + 4 * I;
8
9 double r = cpow(x, y);
10
11 printf("Result: %f + %fi\n", creal(r), cimag(r));
12 }
Result: Result: 0.129010 + 0.000000i
See Also
csqrt(), cexp()
52.17 csqrt() csqrtf() csqrtl()
Compute the complex square root
Synopsis
#include
double complex csqrt(double complex z);
float complex csqrtf(float complex z);
long double complex csqrtl(long double complex z);
Description
Computes the complex square root of z. There is a branch cut along the negative real axis. The return value is in the right half of the complex plane and includes the imaginary axis.
Return Value
Returns the complex square root of z.
Example Chapter 52.
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double complex y = csqrt(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 1.272020 + 0.786151i
See Also
cpow(), sqrt()
52.18 carg() cargf() cargl()
Compute the complex argument
Synopsis
#include
double carg(double complex z);
float cargf(float complex z);
long double cargl(long double complex z);
Description
Computes the complex argument (AKA phase angle) of z. There is a branch cut along the negative real axis. Returns a value in the range [−휋, +휋].
Return Value
Returns the complex argument of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double y = carg(x); Chapter 52.
9
10 printf("Result: %f\n", y);
11 }
Output: Result: 1.107149
52.19 cimag() cimagf() cimagl()
Returns the imaginary part of a complex number
Synopsis
#include
double cimag(double complex z);
float cimagf(float complex z);
long double cimagl(long double complex z);
Description
Returns the imaginary part of z. As a footnote, the spec points out that any complex number x is part of the following equivalency: x == creal(x) + cimag(x) * I;
Return Value
Returns the imaginary part of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double y = cimag(x);
9
10 printf("Result: %f\n", y);
11 }
Output—just the imaginary part: Result: 2.000000
See Also
creal() Chapter 52.
52.20 CMPLX() CMPLXF() CMPLXL()
Build a complex value from real and imaginary types
Synopsis
#include
double complex CMPLX(double x, double y);
float complex CMPLXF(float x, float y);
long double complex CMPLXL(long double x, long double y);
Description These macros build a complex value from real and imaginary types. Now I know what you’re thinking. “But I can already build a complex value from real and imaginary types using the I macro, like in the example you’re about to give us.” double complex x = 1 + 2 * I;
And that’s true. But the reality of the matter is weird and complex. Maybe I got undefined, or maybe you redefined it. Or maybe I was defined as _Complex_I which doesn’t necessarily preserve the sign of a zero value. As the spec points out, these macros build complex numbers as if _Imaginary_I were defined (thus preserving your zero sign) even if it’s not. That is, they are defined equivalently to: #define CMPLX(x, y) ((double complex)((double)(x) + \ _Imaginary_I * (double)(y)))
#define CMPLXF(x, y) ((float complex)((float)(x) + \ _Imaginary_I * (float)(y)))
#define CMPLXL(x, y) ((long double complex)((long double)(x) + \ _Imaginary_I * (long double)(y)))
Return Value
Returns the complex number for the given real x and imaginary y components.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = CMPLX(1, 2); // Like 1 + 2 * I
7
8 printf("Result: %f + %fi\n", creal(x), cimag(x));
9 }
Output: Chapter 52.
Result: 1.000000 + 2.000000i
See Also
creal(), cimag()
52.21 conj() conjf() conjl()
Compute the conjugate of a complex number
Synopsis
#include
double complex conj(double complex z);
float complex conjf(float complex z);
long double complex conjl(long double complex z);
Description
This function computes the complex conjugate1 of z. Apparently it does this by reversing the sign of the imaginary part, but dammit, I’m a programmer not a mathematician, Jim!
Return Value
Returns the complex conjugate of z
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double complex y = conj(x);
9
10 printf("Result: %f + %fi\n", creal(y), cimag(y));
11 }
Output: Result: 1.000000 + -2.000000i
52.22 cproj() cproj() cproj()
Compute the projection of a complex number
1https://en.wikipedia.org/wiki/Complex_conjugate Chapter 52.
Synopsis
#include
double complex cproj(double complex z);
float complex cprojf(float complex z);
long double complex cprojl(long double complex z);
Description
Computes the projection of z onto a Riemann sphere2. Now we’re really outside my expertise. The spec has this to say, which I’m quoting verbatim because I’m not knowledgable enough to rewrite it sensibly. Hopefully it makes sense to anyone who would need to use this function. z projects to z except that all complex infinities (even those with one infinite part and one NaN part) project to positive infinity on the real axis. If z has an infinite part, then cproj(z) is equivalent to
INFINITY + I * copysign(0.0, cimag(z)) So there you have it.
Return Value
Returns the projection of z onto a Riemann sphere.
Example Fingers crossed this is a remotely sane example…
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 double complex x = 1 + 2 * I;
8
9 double complex y = cproj(x);
10
11 printf("Result: %f + %fi\n", creal(y), cimag(y));
12
13 x = INFINITY + 2 * I;
14 y = cproj(x);
15
16 printf("Result: %f + %fi\n", creal(y), cimag(y));
17 }
Output: Result: 1.000000 + 2.000000i Result: inf + 0.000000i
2https://en.wikipedia.org/wiki/Riemann_sphere Chapter 52.
52.23 creal() crealf() creall()
Returns the real part of a complex number
Synopsis
#include
double creal(double complex z);
float crealf(float complex z);
long double creall(long double complex z);
Description
Returns the real part of z. As a footnote, the spec points out that any complex number x is part of the following equivalency: x == creal(x) + cimag(x) * I;
Return Value
Returns the real part of z.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 double complex x = 1 + 2 * I;
7
8 double y = creal(x);
9
10 printf("Result: %f\n", y);
11 }
Output—just the real part: Result: 1.000000
See Also
cimag() Chapter 53
Function Description c16rtomb() Convert a char16_t to a multibyte character c32rtomb() Convert a char32_t to a multibyte character mbrtoc16() Convert a multibyte character to a char16_t mbrtoc32() Convert a multibyte character to a char32_t
These functions are restartable, meaning multiple threads can safely call them at once. They handle this by having their own conversion state variable (of type mbstate_t) per call.
53.1 Types
This header file defines four types.
Type Description char16_t Type to hold 16-bit characters char32_t Type to hold 32-bit characters mbstate_t Holds the conversion state for restartable funcitons (also defined in
String literals for the character types are u for char16_t and U for char32_t. char16_t *str1 = u"Hello, world!"; char32_t *str2 = U"Hello, world!"; char16_t *chr1 = u'A'; char32_t *chr2 = U'B';
Note that char16_t and char32_t might contain Unicode. Or not. If __STDC_UTF_16__ or __STDC_UTF_32__ is defined as 1, then char16_t and char32_t use Unicode, respectively. Otherwise they don’t and the actual value stored depend on the locale. And if you’re not using Unicode, you have my commiserations.
53.2 OS X issue
This header file doesn’t exist on OS X—bummer. If you just want the types, you can: #include
551 Chapter 53.
But if you also want the functions, that’s all on you.
53.3 mbrtoc16() mbrtoc32()
Convert a multibyte character to a char16_t or char32_t restartably
Synopsis
#include
Description
Given a source string s and a destination buffer pc16 (or pc32 for mbrtoc32()), convert the first character of the source to char16_ts (or char32_ts for mbrtoc32()). Basically you have a regular character and you want it as char16_t or char32_t. Use these functions to do it. Note that only one character is converted no matter how many characters in s. As the functions scan s, you don’t want them to overrun the end. So you pass in n as the maximum number of bytes to inspect. The functions will quit after that many bytes or when they have a complete multibyte character, whichever comes first. Since they’re restartable, pass in a conversion state variable for the functions to do their work. And the result will be placed in pc16 (or pc32 for mbrtoc32()).
Return Value
When successful this function returns a number between 1 and n inclusive representing the number of bytes that made up the multibyte character. Or, also in the success category, they can return 0 if the source character is the NUL character (value 0). When not entirely successful, they can return a variety of codes. These are all of type size_t, but negative values cast to that type.
Return Value Description (size_t)(-1) Encoding error—this isn’t a valid sequence of bytes. errno is set to EILSEQ. (size_t)(-2) n bytes were examined and were a partial valid character, but not a complete one. (size_t)(-3) A subsequent value of a character that can’t be represented as a single value. See below.
Case (size_t)(-3) is an odd one. Basically there are some characters that can’t be represented with 16 bits and so can’t be stored in a char16_t. These characters are store in something called (in the Unicode world) surrogate pairs. That is, there are two 16-bit values back to back that represent a larger Unicode value. Chapter 53.
For example, if you want to read the Unicode character \U0001fbc5 (which is a stick figure1—I’m just not putting it in the text because my font doesn’t render it) that’s more than 16 bits. But each call to mbrtoc16() only returns a single char16_t! So subsequent calls to mbrtoc16() resolves the next value in the surrogate pair and returns (size_t)(- 3) to let you know this has happened. You can also pass NULL for pc16 or pc32. This will cause no result to be stored, but you can use it if you’re only interested in the return value from the functions. Finally, if you pass NULL for s, the call is equivalent to: mbrtoc16(NULL, "", 1, ps)
Since the character is a NUL in that case, this has the effect of setting the state in ps to the initial conversion state.
Example
Normal use case example where we get the first two character values from the multibyte string "€Zil- lion": #include
int main(void) { char *s = "\u20acZillion"; // 20ac is "€" char16_t pc16; size_t r; mbstate_t mbs;
setlocale(LC_ALL, ""); memset(&mbs, 0, sizeof mbs);
// Examine the next 8 bytes to see if there's a character in there r = mbrtoc16(&pc16, s, 8, &mbs);
printf("%zu\n", r); // Prints a value >= 1 (3 in UTF-8 locale) printf("%#x\n", pc16); // Prints 0x20ac for "€"
s += r; // Move to next character
// Examine the next 8 bytes to see if there's a character in there r = mbrtoc16(&pc16, s, 8, &mbs);
printf("%zu\n", r); // Prints 1 printf("%#x\n", pc16); // Prints 0x5a for "Z" }
Example with a surrogate pair. In this case we read plenty to get the entire character, but the result must be stored in two char16_ts, requiring two calls to get them both.
1 #include
2 #include
3 #include
4 #include
5
1https://en.wikipedia.org/wiki/Symbols_for_Legacy_Computing Chapter 53.
6 int main(void)
7 {
8 char *s = "\U0001fbc5*"; // Stick figure glyph, more than 16 bits
9 char16_t pc16;
10 mbstate_t mbs;
11 size_t r;
12
13 setlocale(LC_ALL, "");
14 memset(&mbs, 0, sizeof mbs);
15
16 r = mbrtoc16(&pc16, s, 8, &mbs);
17
18 printf("%zd\n", r); // r is 4 bytes in UTF-8 locale
19 printf("%#x\n", pc16); // First value of surrogate pair
20
21 s += r; // Move to next character
22
23 r = mbrtoc16(&pc16, s, 8, &mbs);
24
25 printf("%zd\n", r); // r is (size_t)(-3) here to indicate...
26 printf("%#x\n", pc16); // ...Second value of surrogate pair
27
28 // Since r is -3, it means we're still processing the same
29 // character, so DON'T move to the next character this time
30 //s += r; // Commented out
31
32 r = mbrtoc16(&pc16, s, 8, &mbs);
33
34 printf("%zd\n", r); // 1 byte for "*"
35 printf("%#x\n", pc16); // 0x2a for "*"
36 }
Output on my system, indicating the first character is represented by the pair (0xd83e, 0xdfc5) and the second character is represented by 0x2a: 4 0xd83e -3 0xdfc5 1 0x2a
See Also
c16rtomb(), c32rtomb()
53.4 c16rtomb() c32rtomb()
Convert a char16_t or char32_t to a multibyte character restartably
Synopsis
#include
size_t c16rtomb(char * restrict s, char16_t c16, mbstate_t * restrict ps); Chapter 53.
size_t c32rtomb(char * restrict s, char32_t c32, mbstate_t * restrict ps);
Description
If you have a character in a char16_t or char32_t, use these functions to convert them into a multibyte character. These functions figure out how many bytes are needed for the multibyte character in the current locale and stores them in the buffer pointed to by s. But how big to make that buffer? Luckily there is a macro to help: it needs be no larger than MB_CUR_MAX. As a special case, if s is NULL, it’s the same as calling c16rtomb(buf, L'\0', ps); // or... c32rtomb(buf, L'\0', ps);
where buf is a buffer maintained by the system that you don’t have access to. This has the effect of setting the ps state to the initial state. Finally for surrogate pairs (where the character has been split into two char16_ts), you call this once with the first of the pair—at this point, the function will return 0. Then you call it again with the second of the pair, and the function will return the number of bytes and store the result in the array s.
Return Value
Returns the number of bytes stored in the array pointed to by s. Returns 0 if processing is not yet complete for the current character, as in the case of surrogate pairs. If there is an encoding error, the functions return (size_t)(-1) and errno is set to EILSEQ.
Example
1 #include
2 #include
3 #include
4 #include
5 #include
6
7 int main(void)
8 {
9 char16_t c16 = 0x20ac; // Unicode for Euro symbol
10 char dest[MB_CUR_MAX];
11 size_t r;
12 mbstate_t mbs;
13
14 setlocale(LC_ALL, "");
15 memset(&mbs, 0, sizeof mbs); // Reset conversion state
16
17 // Convert
18 r = c16rtomb(dest, c16, &mbs);
19
20 printf("r == %zd\n", r); // r == 3 on my system
21
22 // And this should print a Euro symbol
23 printf("dest == \"%s\"\n", dest);
24 }
Output on my system: Chapter 53.
r == 3 dest == "€"
This is a more complex example that converts a large-valued character in a multibyte string into a surrogate pair (as in the mbrtoc16() example, above) and then converts it back again into a multibyte string to print.
1 #include
2 #include
3 #include
4 #include
5 #include
6
7 int main(void)
8 {
9 char *src = "\U0001fbc5*"; // Stick figure glyph, more than 16 bits
10 char dest[MB_CUR_MAX];
11 char16_t surrogate0, surrogate1;
12 mbstate_t mbs;
13 size_t r;
14
15 setlocale(LC_ALL, "");
16 memset(&mbs, 0, sizeof mbs); // Reset conversion state
17
18 // Get first surrogate character
19 r = mbrtoc16(&surrogate0, src, 8, &mbs);
20
21 // Get next surrogate character
22 src += r; // Move to next character
23 r = mbrtoc16(&surrogate1, src, 8, &mbs);
24
25 printf("Surrogate pair: %#x, %#x\n", surrogate0, surrogate1);
26
27 // Now reverse it
28 memset(&mbs, 0, sizeof mbs); // Reset conversion state
29
30 // Process first surrogate character
31 r = c16rtomb(dest, surrogate0, &mbs);
32
33 // r should be 0 at this point, because the character hasn't been
34 // processed yet. And dest won't have anything useful... yet!
35 printf("r == %zd\n", r); // r == 0
36
37 // Process second surrogate character
38 r = c16rtomb(dest, surrogate1, &mbs);
39
40 // Now we should be in business. r should have the number of
41 // bytes, and dest should hold the character.
42 printf("r == %zd\n", r); // r == 4 on my system
43
44 // And this should print a stick figure, if your font supports it
45 printf("dest == \"%s\"\n", dest);
46 }
See Also
mbrtoc16(), mbrtoc32() Chapter 54
Macro Description assert() Runtime assertion static_assert() Compile-time assertion
This functionality has to do with things that Should Never Happen™. If you have something that should never be true and you want your program to bomb out because it happened, this is the header file for you. There are two types of assertions: compile-time assertions (called “static assertions”) and runtime asser- tions. If the assertion fails (i.e. the thing that you need to be true is not true) then the program will bomb out either at compile-time or runtime.
54.1 Macros
If you define the macro NDEBUG before you include
54.2 assert()
Bomb out at runtime if a condition fails
Synopsis
#include
Description You pass in an expression to this macro. If it evaluates to false, the program will crash with an assertion failure (by calling the abort() function).
557 Chapter 54.
Basically, you’re saying, “Hey, I’m assuming this condition is true, and if it’s not, I don’t want to continue running.” This is used while debugging to make sure no unexpected conditions arise. And if you find during de- velopment that the condition does arise, maybe you should modify the code to handle it before going to production. If you’ve defined the macro NDEBUG to any value before
Return Value
This macro doesn’t return (since it calls abort() which never returns). If NDEBUG is set, the macro evaluates to ((void)0), which does nothing.
Example
Here’s a function that divides the size of our goat herd. But we’re assuming we’ll never get a 0 passed to us. So we assert that amount != 0… and if it is, the program aborts/
1 // #define NDEBUG 1 // uncomment this to disable the assert
2
3 #include
4
5 void divide_goat_herd_by(int amount)
6 {
7 assert(amount != 0);
8
9 goat_count /= amount;
10 }
When I run this and pass 0 to the function, I get the following on my system (the exact output may vary): Assertion failed: (amount != 0), function divide_goat_herd_by, file foo.c, line 7.
See Also
static_assert(), abort()
54.3 static_assert()
Bomb out at compile-time if a condition fails
Synopsis
#include
static_assert(constant-expression, string-literal);
Description This macro prevents your program from even compiling if a condition isn’t true. And it prints the string literal you give it. Chapter 54.
Basically if constant-expression is false, then compilation will cease and the string-literal will be printed. The constant expression must be truly constant–just values, no variables. And the same is true for the string literal: no variables, just a literal string in double quotes. (It has to be this way since the program’s not running at this point.)
Return Value Not applicable, as this is a compile-time feature.
Example Here’s a partial example with an algorithm that presumably has poor performance or memory issues if the size of the local array is too large. We prevent that eventuality at compile-time by catching it with the static_assert().
1 #include
2
3 #define ARRAY_SIZE 5150
4
5 void some_algorithm(void)
6 {
7 static_assert(ARRAY_SIZE <= 32, "ARRAY_SIZE too large");
8
9 int array[ARRAY_SIZE];
10
11 // ...
On my system, when I try to compile it, this prints (your output may vary): foo.c:7:5: error: static_assert failed due to requirement '5150 <= 32' "ARRAY_SIZE too large" static_assert(ARRAY_SIZE <= 32, "ARRAY_SIZE too large"); ^ ~~~~~~~~~~~~~~~~
See Also
assert() Chapter 55
Function Description feclearexcept() Clear floating point exceptions fegetexceptflag() Save the floating point exception flags fesetexceptflag() Restore the floating point exception flags feraiseexcept() Raise a floating point exception through software fetestexcept() Test to see if an exception has occurred fegetround() Get the rounding direction fesetround() Set the rounding direction fegetenv() Save the entire floating point environment fesetenv() Restore the entire floating point environment feholdexcept() Save floating point state and install non-stop mode feupdateenv() Restore floating point environment and apply recent exceptions
55.1 Types and Macros
There are two types defined in this header:
Type Description fenv_t The entire floating point environment fexcept_t A set of floating point exceptions
The “environment” can be thought of as the status at this moment of the floating point processing system: this includes the exceptions, rounding, etc. It’s an opaque type, so you won’t be able to access it directly, and it must be done through the proper functions. If the functions in question exist on your system (they might not be!), then you’ll also have these macros defined to represent different exceptions:
Macro Description FE_DIVBYZERO Division by zero FE_INEXACT Result was not exact, was rounded FE_INVALID Domain error FE_OVERFLOW Numeric overflow FE_UNDERFLOW Numeric underflow FE_ALL_EXCEPT All of the above combined
560 Chapter 55.
The idea is that you can bitwise-OR these together to represent multiple exceptions, e.g. FE_INVALID|FE_OVERFLOW. The functions, below, that have an excepts parameter will take these values. See
55.2 Pragmas
Normally C is free to optimize all kinds of stuff that might cause the flags to not look like you might expect. So if you’re going to use this stuff, be sure to set this pragma: #pragma STDC FENV_ACCESS ON
If you do this at global scope, it remains in effect until you turn it off: #pragma STDC FENV_ACCESS OFF
If you do it in block scope, it has to come before any statements or declarations. In this case, it has effect until the block ends (or until it is explicitly turned off.) A caveat: this program isn’t supported on either of the compilers I have (gcc and clang) as of this writing, so though I have built the code, below, it’s not particularly well-tested.
55.3 feclearexcept()
Clear floating point exceptions
Synopsis
#include
int feclearexcept(int excepts);
Description If a floating point exception has occurred, this function can clear it. Set excepts to a bitwise-OR list of exceptions to clear. Passing 0 has no effect.
Return Value
Returns 0 on success and non-zero on failure.
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 #pragma STDC FENV_ACCESS ON
8
9 double f = sqrt(-1);
10
11 int r = feclearexcept(FE_INVALID); Chapter 55.
12
13 printf("%d %f\n", r, f);
14 }
See Also
feraiseexcept(), fetestexcept()
55.4 fegetexceptflag() fesetexceptflag()
Save or restore the floating point exception flags
Synopsis
#include
int fegetexceptflag(fexcept_t *flagp, int excepts);
int fesetexceptflag(fexcept_t *flagp, int excepts);
Description Use these functions to save or restore the current floating point environment in a variable. Set excepts to the set of exceptions you want to save or restore the state of. Setting it to FE_ALL_EXCEPT will save or restore the entire state. Note that fexcept_t is an opaque type—you don’t know what’s in it. excepts can be set to zero for no effect.
Return Value
Returns 0 on success or if excepts is zero. Returns non-zero on failure.
Example
This program√ saves the state (before any error has happened), then deliberately causes a domain error by trying to take −1. After that, it restores the floating point state to before the error had occurred, thereby clearing it.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 #pragma STDC FENV_ACCESS ON
8
9 fexcept_t flag;
10
11 fegetexceptflag(&flag, FE_ALL_EXCEPT); // Save state
12
13 double f = sqrt(-1); // I imagine this won't work
14 printf("%f\n", f); // "nan" Chapter 55.
15
16 if (fetestexcept(FE_INVALID))
17 printf("1: Domain error\n"); // This prints!
18 else
19 printf("1: No domain error\n");
20
21 fesetexceptflag(&flag, FE_ALL_EXCEPT); // Restore to before error
22
23 if (fetestexcept(FE_INVALID))
24 printf("2: Domain error\n");
25 else
26 printf("2: No domain error\n"); // This prints!
27 }
55.5 feraiseexcept()
Raise a floating point exception through software
Synopsis
#include
int feraiseexcept(int excepts);
Description This attempts to raise a floating point exception as if it had happened. You can specify multiple exceptions to raise. If either FE_UNDERFLOW or FE_OVERFLOW is raised, C might also raise FE_INEXACT. If either FE_UNDERFLOW or FE_OVERFLOW is raised at the same time as FE_INEXACT, then FE_UNDERFLOW or FE_OVERFLOW will be raised before FE_INEXACT behind the scenes. The order the other exceptions are raised is undefined.
Return Value
Returns 0 if all the exceptions were raised or if excepts is 0. Returns non-zero otherwise.
Example This code deliberately raises a division-by-zero exception and then detects it.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 #pragma STDC FENV_ACCESS ON
8
9 feraiseexcept(FE_DIVBYZERO);
10
11 if (fetestexcept(FE_DIVBYZERO) == FE_DIVBYZERO) Chapter 55.
12 printf("Detected division by zero\n"); // This prints!!
13 else
14 printf("This is fine.\n");
15 }
See Also
feclearexcept(), fetestexcept()
55.6 fetestexcept()
Test to see if an exception has occurred
Synopsis
#include
int fetestexcept(int excepts);
Description
Put the exceptions you want to test in excepts, bitwise-ORing them together.
Return Value Returns the bitwise-OR of the exceptions that have been raised.
Example This code deliberately raises a division-by-zero exception and then detects it.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 #pragma STDC FENV_ACCESS ON
8
9 feraiseexcept(FE_DIVBYZERO);
10
11 if (fetestexcept(FE_DIVBYZERO) == FE_DIVBYZERO)
12 printf("Detected division by zero\n"); // This prints!!
13 else
14 printf("This is fine.\n");
15 }
See Also
feclearexcept(), feraiseexcept() Chapter 55.
55.7 fegetround() fesetround()
Get or set the rounding direction
Synopsis
#include
int fegetround(void);
int fesetround(int round);
Description Use these to get or set the rounding direction used by a variety of math functions. Basically when a function “rounds” a number, it wants to know how to do it. By default, it does it how we tend to expect: if the fractional part is less than 0.5, it rounds down closer to zero, otherwise up farther from zero.
Macro Description FE_TONEAREST Round to the nearest whole number, the default FE_TOWARDZERO Round toward zero always FE_DOWNWARD Round toward the next lesser whole number FE_UPWARD Round toward the next greater whole number
Some implementations don’t support rounding. If it does, the above macros will be defined. Note that the round() function is always “to-nearest” and doesn’t pay attention to the rounding mode.
Return Value
fegetround() returns the current rounding direction, or a negative value on error. fesetround() returns zero on success, or non-zero on failure.
Example This rounds some numbers
1 #include
2 #include
3 #include
4
5 // Helper function to print the rounding mode
6 const char *rounding_mode_str(int mode)
7 {
8 switch (mode) {
9 case FE_TONEAREST: return "FE_TONEAREST";
10 case FE_TOWARDZERO: return "FE_TOWARDZERO";
11 case FE_DOWNWARD: return "FE_DOWNWARD";
12 case FE_UPWARD: return "FE_UPWARD";
13 }
14
15 return "Unknown";
16 }
17
18 int main(void) Chapter 55.
19 {
20 #pragma STDC FENV_ACCESS ON
21
22 int rm;
23
24 rm = fegetround();
25
26 printf("%s\n", rounding_mode_str(rm)); // Print current mode
27 printf("%f %f\n", rint(2.1), rint(2.7)); // Try rounding
28
29 fesetround(FE_TOWARDZERO); // Set the mode
30
31 rm = fegetround();
32
33 printf("%s\n", rounding_mode_str(rm)); // Print it
34 printf("%f %f\n", rint(2.1), rint(2.7)); // Try it now!
35 }
Output: FE_TONEAREST 2.000000 3.000000 FE_TOWARDZERO 2.000000 2.000000
See Also
nearbyint(), nearbyintf(), nearbyintl(), rint(), rintf(), rintl(), lrint(), lrintf(), lrintl(), llrint(), llrintf(), llrintl()
55.8 fegetenv() fesetenv()
Save or restore the entire floating point environment
Synopsis
#include
int fegetenv(fenv_t *envp); int fesetenv(const fenv_t *envp);
Description
You can save the environment (exceptions, rounding direction, etc.) by calling fegetenv() and restore it with fesetenv(). Use this if you want to restore the state after a function call, i.e. hide from the caller that some floating point exceptions or changes occurred.
Return Value
fegetenv() and fesetenv() return 0 on success, and non-zero otherwise.
Example This example saves the environment, messes with the rounding and exceptions, then restores it. After the environment is restored, we see that the rounding is back to default and the exception is cleared. Chapter 55.
1 #include
2 #include
3 #include
4
5 void show_status(void)
6 {
7 printf("Rounding is FE_TOWARDZERO: %d\n",
8 fegetround() == FE_TOWARDZERO);
9
10 printf("FE_DIVBYZERO is set: %d\n",
11 fetestexcept(FE_DIVBYZERO) != 0);
12 }
13
14 int main(void)
15 {
16 #pragma STDC FENV_ACCESS ON
17
18 fenv_t env;
19
20 fegetenv(&env); // Save the environment
21
22 fesetround(FE_TOWARDZERO); // Change rounding
23 feraiseexcept(FE_DIVBYZERO); // Raise an exception
24
25 show_status();
26
27 fesetenv(&env); // Restore the environment
28
29 show_status();
30 }
Output: Rounding is FE_TOWARDZERO: 1 FE_DIVBYZERO is set: 1 Rounding is FE_TOWARDZERO: 0 FE_DIVBYZERO is set: 0
See Also
feholdexcept(), feupdateenv()
55.9 feholdexcept()
Save floating point state and install non-stop mode
Synopsis
#include
int feholdexcept(fenv_t *envp);
Description
This is just like fegetenv() except that it updates the current environment to be in non-stop mode, namely it won’t halt on any exceptions. Chapter 55.
It remains in this state until you restore the state with fesetenv() or feupdateenv().
Return Value Example This example saves the environment and goes into non-stop mode, messes with the rounding and excep- tions, then restores it. After the environment is restored, we see that the rounding is back to default and the exception is cleared. We’ll also be out of non-stop mode.
1 #include
2 #include
3 #include
4
5 void show_status(void)
6 {
7 printf("Rounding is FE_TOWARDZERO: %d\n",
8 fegetround() == FE_TOWARDZERO);
9
10 printf("FE_DIVBYZERO is set: %d\n",
11 fetestexcept(FE_DIVBYZERO) != 0);
12 }
13
14 int main(void)
15 {
16 #pragma STDC FENV_ACCESS ON
17
18 fenv_t env;
19
20 // Save the environment and don't stop on exceptions
21 feholdexcept(&env);
22
23 fesetround(FE_TOWARDZERO); // Change rounding
24 feraiseexcept(FE_DIVBYZERO); // Raise an exception
25
26 show_status();
27
28 fesetenv(&env); // Restore the environment
29
30 show_status();
31 }
See Also
fegetenv(), fesetenv(), feupdateenv()
55.10 feupdateenv()
Restore floating point environment and apply recent exceptions
Synopsis
#include
int feupdateenv(const fenv_t *envp); Chapter 55.
Description
This is like fesetenv() except that it modifies the passed-in environment so that it is updated with exceptions that have happened in the meantime. So let’s say you had a function that might raise exceptions, but you wanted to hide those in the caller. One option might be to: 1. Save the environment with fegetenv() or feholdexcept(). 2. Do whatever you do that might raise exceptions. 3. Restore the environment with fesetenv(), thereby hiding the exceptions that happened in step 2. But that hides all exceptions. What if you just wanted to hide some of them? You could use feupda- teenv() like this: 1. Save the environment with fegetenv() or feholdexcept(). 2. Do whatever you do that might raise exceptions. 3. Call feclearexcept() to clear the exceptions you want to hide from the caller. 4. Call feupdateenv() to restore the previous environment and update it with the other exceptions that have occurred. So it’s like a more capable way of restoring the environment than simply fegetenv()/fesetenv().
Return Value
Returns 0 on success, non-zero otherwise.
Example This program saves state, raises some exceptions, then clears one of the exceptions, then restores and updates the state.
1 #include
2 #include
3 #include
4
5 void show_status(void)
6 {
7 printf("FE_DIVBYZERO: %d\n", fetestexcept(FE_DIVBYZERO) != 0);
8 printf("FE_INVALID : %d\n", fetestexcept(FE_INVALID) != 0);
9 printf("FE_OVERFLOW : %d\n\n", fetestexcept(FE_OVERFLOW) != 0);
10 }
11
12 int main(void)
13 {
14 #pragma STDC FENV_ACCESS ON
15
16 fenv_t env;
17
18 feholdexcept(&env); // Save the environment
19
20 // Pretend some bad math happened here:
21 feraiseexcept(FE_DIVBYZERO); // Raise an exception
22 feraiseexcept(FE_INVALID); // Raise an exception
23 feraiseexcept(FE_OVERFLOW); // Raise an exception
24
25 show_status();
26
27 feclearexcept(FE_INVALID);
28
29 feupdateenv(&env); // Restore the environment
30 Chapter 55.
31 show_status();
32 }
In the output, at first we have no exceptions. Then we have the three we raised. Then after we re- store/update the environment, we see the one we cleared (FE_INVALID) hasn’t been applied: FE_DIVBYZERO: 0 FE_INVALID : 0 FE_OVERFLOW : 0
FE_DIVBYZERO: 1 FE_INVALID : 1 FE_OVERFLOW : 1
FE_DIVBYZERO: 1 FE_INVALID : 0 FE_OVERFLOW : 1
See Also
fegetenv(), fesetenv(), feholdexcept(), feclearexcept() Chapter 56
Important note: the “minimum magnitude” in the table below is the minimum allowed by the spec. It’s very likely that the values on your bad-ass system exceed those, below.
Macro Minimum Magnitude Description CHAR_BIT 8 Number of bits in a byte SCHAR_MIN -127 Minimum value of a signed char SCHAR_MAX 127 Maximum value of a signed char UCHAR_MAX 255 Maximum value of an unsigned char[^The minimum value of an unsigned char is 0. Same fo an unsigned short and unsigned long. Or any unsigned type, for that matter.] CHAR_MIN 0 or SCHAR_MIN More detail below CHAR_MAX SCHAR_MAX or More detail below UCHAR_MAX MB_LEN_MAX 1 Maximum number of bytes in a multibyte character on any locale SHRT_MIN -32767 Minimum value of a short SHRT_MAX 32767 Maximum value of a short USHRT_MAX 65535 Maximum value of an unsigned short INT_MIN -32768 Minimum vale of an int INT_MAX 32768 Maximum value of an int UINT_MAX 65535 Maximum value of an unsigned int LONG_MIN -2147483647 Minimum value of a long LONG_MAX 2147483647 Maximum value of a long ULONG_MAX 4294967295 Maximum value of an unsigned long LLONG_MIN - Minimum value of a long long 9223372036854775807 LLONG_MAX 9223372036854775807 Maximum value of a long long ULLONG_MAX 18446744073709551615 Maximum value of an unsigned long long
56.1 CHAR_MIN and CHAR_MAX
When it comes to the CHAR_MIN and CHAR_MAX macros, it all depends on if your char type is signed or unsigned by default. Remember that C leaves that up to the implementation? No? Well, it does. So if it’s signed, the values of CHAR_MIN and CHAR_MAX are the same as SCHAR_MIN and SCHAR_MAX. And if it’s unsigned, the values of CHAR_MIN and CHAR_MAX are the same as 0 and UCHAR_MAX. Side benefit: you can tell at runtime if the system has signed or unsigned chars by checking to see if CHAR_MIN is 0.
571 Chapter 56.
1 #include
2 #include
3
4 int main(void)
5 {
6 printf("chars are %ssigned\n", CHAR_MIN == 0? "un": "");
7 }
On my system, chars are signed.
56.2 Choosing the Correct Type
If you want to be super portable, choose a type you know will be at least as big as you need by the table, above. That said, a lot of code, for better or (likely) worse, assumes ints are 32-bits, when in actuality it’s only guaranteed to be 16. If you need a guaranteed bit size, check out the int_leastN_t types in
56.3 Whither Two’s Complement?
If you were looking closely and have a priori knowledge of the matter, you might have thought I erred in the minimum values of the macros, above. “short goes from 32767 to -32767? Shouldn’t it go to -32768?” No, I have it right. The spec list the minimum magnitudes for those macros, and some old-timey systems might have used a different encoding for their signed values that could only go that far. Virtually every modern system uses Two’s Complement1 for signed numbers, and those would go from 32767 to -32768 for a short. Your system probably does, too.
56.4 Demo Program
Here’s a program to print out the values of the macros:
1 #include
2 #include
3
4 int main(void)
5 {
6 printf("CHAR_BIT = %d\n", CHAR_BIT);
7 printf("SCHAR_MIN = %d\n", SCHAR_MIN);
8 printf("SCHAR_MAX = %d\n", SCHAR_MAX);
9 printf("UCHAR_MAX = %d\n", UCHAR_MAX);
10 printf("CHAR_MIN = %d\n", CHAR_MIN);
11 printf("CHAR_MAX = %d\n", CHAR_MAX);
12 printf("MB_LEN_MAX = %d\n", MB_LEN_MAX);
13 printf("SHRT_MIN = %d\n", SHRT_MIN);
14 printf("SHRT_MAX = %d\n", SHRT_MAX);
15 printf("USHRT_MAX = %u\n", USHRT_MAX);
16 printf("INT_MIN = %d\n", INT_MIN);
17 printf("INT_MAX = %d\n", INT_MAX);
18 printf("UINT_MAX = %u\n", UINT_MAX);
19 printf("LONG_MIN = %ld\n", LONG_MIN);
20 printf("LONG_MAX = %ld\n", LONG_MAX);
1https://en.wikipedia.org/wiki/Two%27s_complement Chapter 56.
21 printf("ULONG_MAX = %lu\n", ULONG_MAX);
22 printf("LLONG_MIN = %lld\n", LLONG_MIN);
23 printf("LLONG_MAX = %lld\n", LLONG_MAX);
24 printf("ULLONG_MAX = %llu\n", ULLONG_MAX);
25 }
On my 64-bit Intel system with clang, this outputs: CHAR_BIT = 8 SCHAR_MIN = -128 SCHAR_MAX = 127 UCHAR_MAX = 255 CHAR_MIN = -128 CHAR_MAX = 127 MB_LEN_MAX = 6 SHRT_MIN = -32768 SHRT_MAX = 32767 USHRT_MAX = 65535 INT_MIN = -2147483648 INT_MAX = 2147483647 UINT_MAX = 4294967295 LONG_MIN = -9223372036854775808 LONG_MAX = 9223372036854775807 ULONG_MAX = 18446744073709551615 LLONG_MIN = -9223372036854775808 LLONG_MAX = 9223372036854775807 ULLONG_MAX = 18446744073709551615
Looks like my system probably uses two’s-complement encoding for signed numbers, my chars are signed, and my ints are 32-bit. Chapter 57
These functions enable you to rewind the call stack to an earlier point, with a bunch of gotchas. See the chapter on setjmp()/longjmp() for more info.
Function Description longjmp() Return to the previously-placed bookmark setjmp() Bookmark this place to return to later
There’s also a new opaque type, jmp_buf, that holds all the information needed to pull off this magic trick. If you want your automatic local variables to be correct after a call to longjmp(). declare them as volatile where you called setjmp().
57.1 setjmp()
Save this location as one to return to later
Synopsis
#include
Description
This is how you save your position so you can longjmp() back it, later. Think of it as setting up a warp destination for later use. Basically, you call this, giving it an env it can fill in with all the information it needs to come back here later. This env is one you’ll pass to longjmp() later when you want to teleport back here. And the really funky part is this can return two different ways: 1. It can return 0 from the call where you set up the jump destination. 2. If can return non-zero when you actually warp back here as the result of a call to longjmp(). What you can do is check the return value to see which case has occurred. You’re only allowed to call setjmp() in a limited number of circumstances.
574 Chapter 57.
1. As a standalone expression: setjmp(env);
You can also cast it to (void) if you really wanted to do such a thing. 2. As the complete controlling expression in an if or switch. if (setjmp(env)) { ... }
switch (setjmp(env)) { ... }
But not this as it’s not the complete controlling expression in this case: if (x == 2 && setjmp()) { ... } // Undefined behavior
3. The same as (2), above, except with a comparison to an integer constant: if (setjmp(env) == 0) { ... }
if (setjmp(env) > 2) { ... }
4. As the operand to the not (!) operator: if (!setjmp(env)) { ... }
Anything else is (you guessed it) undefined behavior! This can be a macro or a function, but you’ll treat it the same way in any case.
Return Value This one is funky. It returns one of two things: Returns 0 if this was the call to setjmp() to set it up. Returns non-zero if being here was the result of a call to longjmp(). (Namely, it returns the value passed into the longjmp() function.)
Example
Here’s a function that calls setjmp() to set things up (where it returns 0), then calls a couple levels deep into functions, and finally short-circuits the return path by longjmp()ing back to the place where setjmp() was called, earlier. This time, it passes 3490 as a value, which setjmp() returns.
1 #include
2 #include
3
4 jmp_buf env;
5
6 void depth2(void)
7 {
8 printf("Entering depth 2\n");
9 longjmp(env, 3490); // Jump back to setjmp()!!
10 printf("Leaving depth 2\n"); // This won't happen
11 }
12
13 void depth1(void)
14 {
15 printf("Entering depth 1\n");
16 depth2();
17 printf("Leaving depth 1\n"); // This won't happen
18 }
19
20 int main(void) Chapter 57.
21 {
22 switch (setjmp(env)) {
23 case 0:
24 printf("Calling into functions, setjmp() returned 0\n");
25 depth1();
26 printf("Returned from functions\n"); // This won't happen
27 break;
28
29 case 3490:
30 printf("Bailed back to main, setjmp() returned 3490\n");
31 break;
32 }
33 }
When run, this outputs: Calling into functions, setjmp() returned 0 Entering depth 1 Entering depth 2 Bailed back to main, setjmp() returned 3490
Notice that the second printf() in case 0 didn’t run; it got jumped over by longjmp()!
See Also
longjmp()
57.2 longjmp()
Return to the previous setjmp() location
Synopsis
#include
_Noreturn void longjmp(jmp_buf env, int val);
Description
This returns to a previous call to setjmp() back in the call history. setjmp() will return the val passed into longjmp(). The env passed to setjmp() should be the same one you pass into longjmp(). There are a bunch of potential issues with doing this, so you’ll want to be careful that you avoid undefined behavior by not doing the following: 1. Don’t call longjmp() if the corresponding setjmp() was in a different thread. 2. Don’t call longjmp() if you didn’t call setjmp() first. 3. Don’t call longjmp() if the function that called setjmp() has completed. 4. Don’t call longjmp() if the call to setjmp() had a variable length array (VLA) in scope and the scope has ended. 5. Don’t call longjmp() if there are any VLAs in any active scopes between the setjmp() and the longjmp(). A good rule of thumb here is to not mix VLAs and longjmp(). Chapter 57.
Though longjmp() attempts to restore the machine to the state at the setjmp(), including local variables, there are some things that aren’t brought back to life: • Non-volatile local variables that might have changed • Floating point status flags • Open files • Any other component of the abstract machine
Return Value This one is also funky in that it is one of the few functions in C that never returns!
Example
Here’s a function that calls setjmp() to set things up (where it returns 0), then calls a couple levels deep into functions, and finally short-circuits the return path by longjmp()ing back to the place where setjmp() was called, earlier. This time, it passes 3490 as a value, which setjmp() returns.
1 #include
2 #include
3
4 jmp_buf env;
5
6 void depth2(void)
7 {
8 printf("Entering depth 2\n");
9 longjmp(env, 3490); // Jump back to setjmp()!!
10 printf("Leaving depth 2\n"); // This won't happen
11 }
12
13 void depth1(void)
14 {
15 printf("Entering depth 1\n");
16 depth2();
17 printf("Leaving depth 1\n"); // This won't happen
18 }
19
20 int main(void)
21 {
22 switch (setjmp(env)) {
23 case 0:
24 printf("Calling into functions, setjmp() returned 0\n");
25 depth1();
26 printf("Returned from functions\n"); // This won't happen
27 break;
28
29 case 3490:
30 printf("Bailed back to main, setjmp() returned 3490\n");
31 break;
32 }
33 }
When run, this outputs: Calling into functions, setjmp() returned 0 Entering depth 1 Entering depth 2 Bailed back to main, setjmp() returned 3490
Notice that the second printf() in case 0 didn’t run; it got jumped over by longjmp()! Chapter 57.
See Also setjmp() Chapter 58
Macro Minimum Magnitude Description FLT_ROUNDS Current rounding mode FLT_EVAL_METHOD Types used for evaluation FLT_HAS_SUBNORM Subnormal support for float DBL_HAS_SUBNORM Subnormal support for double LDBL_HAS_SUBNORM Subnormal support for long double FLT_RADIX 2 Floating point radix (base) FLT_MANT_DIG Number of base FLT_RADIX digits in a float DBL_MANT_DIG Number of base FLT_RADIX digits in a double LDBL_MANT_DIG Number of base FLT_RADIX digits in a long double FLT_DECIMAL_DIG 6 Number of decimal digits required to encode a float DBL_DECIMAL_DIG 10 Number of decimal digits required to encode a double LDBL_DECIMAL_DIG 10 Number of decimal digits required to encode a long double DECIMAL_DIG 10 Number of decimal digits required to encode the the widest floating point number supported FLT_DIG 6 Number of decimal digits that can be safely stored in a float DBL_DIG 10 Number of decimal digits that can be safely stored in a double LDBL_DIG 10 Number of decimal digits that can be safely stored in a long double FLT_MIN_EXP FLT_RADIX to the FLT_MIN_EXP-1 power is the smallest normalized float DBL_MIN_EXP FLT_RADIX to the DBL_MIN_EXP-1 power is the smallest normalized double LDBL_MIN_EXP FLT_RADIX to the LDBL_MIN_EXP-1 power is the smallest normalized long double FLT_MIN_10_EXP -37 Minimum exponent such that 10 to this number is a normalized float DBL_MIN_10_EXP -37 Minimum exponent such that 10 to this number is a normalized double LDBL_MIN_10_EXP -37 Minimum exponent such that 10 to this number is a normalized long_double FLT_MAX_EXP FLT_RADIX to the FLT_MAX_EXP-1 power is the largest finite float DBL_MAX_EXP FLT_RADIX to the DBL_MAX_EXP-1 power is the largest finite double LDBL_MAX_EXP FLT_RADIX to the LDBL_MAX_EXP-1 power is the largest finite long double
579 Chapter 58.
Macro Minimum Magnitude Description FLT_MAX_10_EXP -37 Minimum exponent such that 10 to this number is a finite float DBL_MAX_10_EXP -37 Minimum exponent such that 10 to this number is a finite double LDBL_MAX_10_EXP -37 Minimum exponent such that 10 to this number is a finite long_double FLT_MAX 1E+37 Largest finite float DBL_MAX 1E+37 Largest finite double LDBL_MAX 1E+37 Largest finite long double
Macro Maximum Value Description FLT_EPSILON 1E-5 Difference between 1 and the next biggest representable float DBL_EPSILON 1E-9 Difference between 1 and the next biggest representable double LDBL_EPSILON 1E-9 Difference between 1 and the next biggest representable long double FLT_MIN 1E-37 Minimum positive normalized float DBL_MIN 1E-37 Minimum positive normalized double LDBL_MIN 1E-37 Minimum positive normalized long double FLT_TRUE_MIN 1E-37 Minimum positive float DBL_TRUE_MIN 1E-37 Minimum positive double LDBL_TRUE_MIN 1E-37 Minimum positive long double
The minimum and maximum values here are from the spec—they should what you can at least expect across all platforms. Your super dooper machine might do better, still!
58.1 Background
The spec allows a lot of leeway when it comes to how C represents floating point numbers. This header file spells out the limits on those numbers. It gives a model that can describe any floating point number that I know you’re going to absolutely love. It looks like this: 푝 푒 −푘 푥 = 푠푏 ∑ 푓푘푏 , 푒푚푖푛 ≤ 푒 ≤ 푒푚푎푥 푘=1 where:
Variable Meaning 푠 Sign, −1 or 1 푏 Base (radix), probably 2 on your system 푒 Exponent 푝 Precision: how many base-푏 digits in the number 푓푘 The individual digits of the number, the significand
But let’s blissfully ignore all that for a second. Let’s assume your computer uses base 2 for it’s floating point (it probably does). And that in the example below the 1s-and-0s numbers are in binary, and the rest are in decimal. The short of it is you could have floating point numbers like shown in this example: Chapter 58.
−0.10100101 × 25 = −10100.101 = −20.625 That’s your fractional part multiplied by the base to the exponent’s power. The exponent controls where the decimal point is. It “floats” around!
58.2 FLT_ROUNDS Details
This tells you the rounding mode. It can be changed with a call to fesetround().
Mode Description -1 Indeterminable 0 Toward zero 1 To nearest 2 Toward positive infinity 3 Toward negative infinity… and beyond!
Unlike every other macro in this here header, FLT_ROUNDS might not be a constant expression.
58.3 FLT_EVAL_METHOD Details
This basically tells you how floating point values are promoted to different types in expressions.
Method Description -1 Indeterminable 0 Evaluate all operations and constants to the precision of their respective types 1 Evaluate float and double operations as double and long double ops as long double 2 Evaluate all operations and constants as long double
58.4 Subnormal Numbers
The macros FLT_HAS_SUBNORM, DBL_HAS_SUBNORM, and LDBL_HAS_SUBNORM all let you know if those types support subnormal numbers1.
Value Description -1 Indeterminable 0 Subnormals not supported for this type 1 Subnormals supported for this type
58.5 How Many Decimal Places Can I Use?
It depends on what you want to do. The safe thing is if you never use more than FLT_DIG base-10 digits in your float, you’re good. (Same for DBL_DIG and LDBL_DIG for their types.) And by “use” I mean print out, have in code, read from the keyboard, etc. You can print out that many decimal places with printf() and the %g format specifier:
1https://en.wikipedia.org/wiki/Subnormal_number Chapter 58.
1 #include
2 #include
3
4 int main(void)
5 {
6 float pi = 3.1415926535897932384626433832795028841971;
7
8 // With %g or %G, the precision refers to the number of significant
9 // digits:
10
11 printf("%.*g\n", FLT_DIG, pi); // For me: 3.14159
12
13 // But %f prints too many, since the precision is the number of
14 // digits to the right of the decimal--it doesn't count the digits
15 // to the left of it:
16
17 printf("%.*f\n", FLT_DIG, pi); // For me: 3.14159... 3 ???
18 }
That’s the end, but stay tuned for the exciting conclusion of “How Many Decimal Places Can I Use?” Because base 10 and base 2 (your typical FLT_RADIX) don’t mix very well, you can actually have more than FLT_DIG in your float; the bits of storage go out a little farther. But these might round in a way you don’t expect. But if you want to convert a floating point number to base 10 and then be able to convert it back again to the exact same floating point number, you’ll need FLT_DECIMAL_DIG digits from your float to make sure you get those extra bits of storage represented. Here’s some example output that shows how the value stored might have some extra decimal places at the end.
1 #include
2 #include
3 #include
4 #include
5
6 int main(void)
7 {
8 printf("FLT_DIG = %d\n", FLT_DIG);
9 printf("FLT_DECIMAL_DIG = %d\n\n", FLT_DECIMAL_DIG);
10
11 assert(FLT_DIG == 6); // Code below assumes this
12
13 for (float x = 0.123456; x < 0.12346; x += 0.000001) {
14 printf("As written: %.*g\n", FLT_DIG, x);
15 printf("As stored: %.*g\n\n", FLT_DECIMAL_DIG, x);
16 }
17 }
And the output on my machine, starting at 0.123456 and incrementing by 0.000001 each time: FLT_DIG = 6 FLT_DECIMAL_DIG = 9
As written: 0.123456 As stored: 0.123456001
As written: 0.123457 As stored: 0.123457 Chapter 58.
As written: 0.123458 As stored: 0.123457998
As written: 0.123459 As stored: 0.123458996
As written: 0.12346 As stored: 0.123459995
You can see that the value stored isn’t always the value we’re expecting since base-2 can’t represent all base-10 fractions exactly. The best it can do is store more places and then round. Also notice that even though we tried to stop the for loop before 0.123460, it actually ran including that value since the stored version of that number was 0.123459995, which is still less than 0.123460. Aren’t floating point numbers fun?
58.6 Comprehensive Example
Here’s a program that prints out the details for a particular machine:
1 #include
2 #include
3
4 int main(void)
5 {
6 printf("FLT_RADIX: %d\n", FLT_RADIX);
7 printf("FLT_ROUNDS: %d\n", FLT_ROUNDS);
8 printf("FLT_EVAL_METHOD: %d\n", FLT_EVAL_METHOD);
9 printf("DECIMAL_DIG: %d\n\n", DECIMAL_DIG);
10
11 printf("FLT_HAS_SUBNORM: %d\n", FLT_HAS_SUBNORM);
12 printf("FLT_MANT_DIG: %d\n", FLT_MANT_DIG);
13 printf("FLT_DECIMAL_DIG: %d\n", FLT_DECIMAL_DIG);
14 printf("FLT_DIG: %d\n", FLT_DIG);
15 printf("FLT_MIN_EXP: %d\n", FLT_MIN_EXP);
16 printf("FLT_MIN_10_EXP: %d\n", FLT_MIN_10_EXP);
17 printf("FLT_MAX_EXP: %d\n", FLT_MAX_EXP);
18 printf("FLT_MAX_10_EXP: %d\n", FLT_MAX_10_EXP);
19 printf("FLT_MIN: %.*e\n", FLT_DECIMAL_DIG, FLT_MIN);
20 printf("FLT_MAX: %.*e\n", FLT_DECIMAL_DIG, FLT_MAX);
21 printf("FLT_EPSILON: %.*e\n", FLT_DECIMAL_DIG, FLT_EPSILON);
22 printf("FLT_TRUE_MIN: %.*e\n\n", FLT_DECIMAL_DIG, FLT_TRUE_MIN);
23
24 printf("DBL_HAS_SUBNORM: %d\n", DBL_HAS_SUBNORM);
25 printf("DBL_MANT_DIG: %d\n", DBL_MANT_DIG);
26 printf("DBL_DECIMAL_DIG: %d\n", DBL_DECIMAL_DIG);
27 printf("DBL_DIG: %d\n", DBL_DIG);
28 printf("DBL_MIN_EXP: %d\n", DBL_MIN_EXP);
29 printf("DBL_MIN_10_EXP: %d\n", DBL_MIN_10_EXP);
30 printf("DBL_MAX_EXP: %d\n", DBL_MAX_EXP);
31 printf("DBL_MAX_10_EXP: %d\n", DBL_MAX_10_EXP);
32 printf("DBL_MIN: %.*e\n", DBL_DECIMAL_DIG, DBL_MIN);
33 printf("DBL_MAX: %.*e\n", DBL_DECIMAL_DIG, DBL_MAX);
34 printf("DBL_EPSILON: %.*e\n", DBL_DECIMAL_DIG, DBL_EPSILON);
35 printf("DBL_TRUE_MIN: %.*e\n\n", DBL_DECIMAL_DIG, DBL_TRUE_MIN);
36
37 printf("LDBL_HAS_SUBNORM: %d\n", LDBL_HAS_SUBNORM);
38 printf("LDBL_MANT_DIG: %d\n", LDBL_MANT_DIG); Chapter 58.
39 printf("LDBL_DECIMAL_DIG: %d\n", LDBL_DECIMAL_DIG);
40 printf("LDBL_DIG: %d\n", LDBL_DIG);
41 printf("LDBL_MIN_EXP: %d\n", LDBL_MIN_EXP);
42 printf("LDBL_MIN_10_EXP: %d\n", LDBL_MIN_10_EXP);
43 printf("LDBL_MAX_EXP: %d\n", LDBL_MAX_EXP);
44 printf("LDBL_MAX_10_EXP: %d\n", LDBL_MAX_10_EXP);
45 printf("LDBL_MIN: %.*Le\n", LDBL_DECIMAL_DIG, LDBL_MIN);
46 printf("LDBL_MAX: %.*Le\n", LDBL_DECIMAL_DIG, LDBL_MAX);
47 printf("LDBL_EPSILON: %.*Le\n", LDBL_DECIMAL_DIG, LDBL_EPSILON);
48 printf("LDBL_TRUE_MIN: %.*Le\n\n", LDBL_DECIMAL_DIG, LDBL_TRUE_MIN);
49
50 printf("sizeof(float): %zu\n", sizeof(float));
51 printf("sizeof(double): %zu\n", sizeof(double));
52 printf("sizeof(long double): %zu\n", sizeof(long double));
53 }
And here’s the output on my machine: FLT_RADIX: 2 FLT_ROUNDS: 1 FLT_EVAL_METHOD: 0 DECIMAL_DIG: 21
FLT_HAS_SUBNORM: 1 FLT_MANT_DIG: 24 FLT_DECIMAL_DIG: 9 FLT_DIG: 6 FLT_MIN_EXP: -125 FLT_MIN_10_EXP: -37 FLT_MAX_EXP: 128 FLT_MAX_10_EXP: 38 FLT_MIN: 1.175494351e-38 FLT_MAX: 3.402823466e+38 FLT_EPSILON: 1.192092896e-07 FLT_TRUE_MIN: 1.401298464e-45
DBL_HAS_SUBNORM: 1 DBL_MANT_DIG: 53 DBL_DECIMAL_DIG: 17 DBL_DIG: 15 DBL_MIN_EXP: -1021 DBL_MIN_10_EXP: -307 DBL_MAX_EXP: 1024 DBL_MAX_10_EXP: 308 DBL_MIN: 2.22507385850720138e-308 DBL_MAX: 1.79769313486231571e+308 DBL_EPSILON: 2.22044604925031308e-16 DBL_TRUE_MIN: 4.94065645841246544e-324
LDBL_HAS_SUBNORM: 1 LDBL_MANT_DIG: 64 LDBL_DECIMAL_DIG: 21 LDBL_DIG: 18 LDBL_MIN_EXP: -16381 LDBL_MIN_10_EXP: -4931 LDBL_MAX_EXP: 16384 LDBL_MAX_10_EXP: 4932 LDBL_MIN: 3.362103143112093506263e-4932 Chapter 58.
LDBL_MAX: 1.189731495357231765021e+4932 LDBL_EPSILON: 1.084202172485504434007e-19 LDBL_TRUE_MIN: 3.645199531882474602528e-4951 sizeof(float): 4 sizeof(double): 8 sizeof(long double): 16 Chapter 59
Macro Description va_arg() Get the next variable argument va_copy() Copy a va_list and the work done so far va_end() Signify we’re done processing variable arguments va_start() Initialize a va_list to start variable argument processing
This header file is what allows you to write functions that take a variable number of arguments. In addition to the macros, you get a new type that helps C keep track of where it is in the variable-number- of-arguments-processing: va_list. This type is opaque, and you’ll be passing it around to the various macros to help get at the arguments. Note that every variadic function requires at least one non-variable parameter. You need this to kick off processing with va_start().
59.1 va_arg()
Get the next variable argument
Synopsis
#include
Description
If you have a variable argument list you’ve initialized with va_start(), pass it to this one along with the type of argument you’re trying to get, e.g. int x = va_arg(args, int); float y = va_arg(args, float);
Return Value Evaluates to the value and type of the next variable argument.
586 Chapter 59.
Example Here’s a demo that adds together an arbitrary number of integers. The first argument is the number of integers to add together. We’ll make use of that to figure out how many times we have to call va_arg().
1 #include
2 #include
3
4 int add(int count, ...)
5 {
6 int total = 0;
7 va_list va;
8
9 va_start(va, count); // Start with arguments after "count"
10
11 for (int i = 0; i < count; i++) {
12 int n = va_arg(va, int); // Get the next int
13
14 total += n;
15 }
16
17 va_end(va); // All done
18
19 return total;
20 }
21
22 int main(void)
23 {
24 printf("%d\n", add(4, 6, 2, -4, 17)); // 6 + 2 - 4 + 17 = 21
25 printf("%d\n", add(2, 22, 44)); // 22 + 44 = 66
26 }
See Also
va_start(), va_end()
59.2 va_copy()
Copy a va_list and the work done so far
Synopsis
#include
void va_copy(va_list dest, va_list src);
Description The main intended use of this is to save your state partway through processing variable arguments so you can scan ahead and then rewind back to the save point. You pass in a src va_list and it copies it to dest. If you’ve already called this once for a particular dest, you can’t call it (or va_start()) again with the same dest unless you call va_end() on that dest first. Chapter 59.
va_copy(dest, src); va_copy(dest, src2); // BAD!
va_copy(dest, src); va_start(dest, var); // BAD!
va_copy(dest, src); va_end(dest); va_copy(dest, src2); // OK!
va_copy(dest, src); va_end(dest); va_start(dest, var); // OK!
Return Value Returns nothing.
Example Here’s an example where we’re adding together all the variable arguments, but then we want to go back and add on all the numbers past the first two, for example if the arguments are: 10 20 30 40
First we add them all for 100, and then we add on everything from the third number on, so add on 30+40 for a total of 170. We’ll do this by saving our place in the variable argument processing with va_copy and then using that later to reprocess the trailing arguments. (And yes, I know there’s a mathematical way to do this without all the rewinding, but I’m having an heck of a time coming up with a good example!)
1 #include
2 #include
3
4 // Add all the numbers together, but then add on all the numbers
5 // past the second one again.
6 int contrived_adder(int count, ...)
7 {
8 if (count < 3) return 0; // OK, I'm being lazy. You got me.
9
10 int total = 0;
11
12 va_list args, mid_args;
13
14 va_start(args, count);
15
16 for (int i = 0; i < count; i++) {
17
18 // If we're at the second number, save our place in
19 // mid_args:
20
21 if (i == 2)
22 va_copy(mid_args, args);
23
24 total += va_arg(args, int);
25 }
26 Chapter 59.
27 va_end(args); // Done with this
28
29 // But now let's start with mid_args and add all those on:
30 for (int i = 0; i < count - 2; i++)
31 total += va_arg(mid_args, int);
32
33 va_end(mid_args); // Done with this, too
34
35 return total;
36 }
37
38 int main(void)
39 {
40 // 10+20+30 + 30 == 90
41 printf("%d\n", contrived_adder(3, 10, 20, 30));
42
43 // 10+20+30+40+50 + 30+40+50 == 270
44 printf("%d\n", contrived_adder(5, 10, 20, 30, 40, 50));
45 }
See Also
va_start(), va_arg(), va_end()
59.3 va_end()
Signify we’re done processing variable arguments
Synopsis
#include
void va_end(va_list ap);
Description
After you’ve va_start()ed or va_copy’d a new va_list, you must call va_end() with it before it goes out of scope. You also have to do this if you’re going to call va_start() or va_copy() again on a variable you’ve already done that to. Them’s the rules if you want to avoid undefined behavior. But just think of it as cleanup. You called va_start(), so you’ll call va_end() when you’re done.
Return Value Returns nothing.
Example Here’s a demo that adds together an arbitrary number of integers. The first argument is the number of integers to add together. We’ll make use of that to figure out how many times we have to call va_arg(). Chapter 59.
1 #include
2 #include
3
4 int add(int count, ...)
5 {
6 int total = 0;
7 va_list va;
8
9 va_start(va, count); // Start with arguments after "count"
10
11 for (int i = 0; i < count; i++) {
12 int n = va_arg(va, int); // Get the next int
13
14 total += n;
15 }
16
17 va_end(va); // All done
18
19 return total;
20 }
21
22 int main(void)
23 {
24 printf("%d\n", add(4, 6, 2, -4, 17)); // 6 + 2 - 4 + 17 = 21
25 printf("%d\n", add(2, 22, 44)); // 22 + 44 = 66
26 }
See Also
va_start(), va_copy()
59.4 va_start()
Initialize a va_list to start variable argument processing
Synopsis
#include
void va_start(va_list ap, parmN);
Description
You’ve declared a variable of type va_list to keep track of the variable argument processing… now how to initialize it so you can start calling va_arg() to get those arguments? va_start() to the rescue! What you do is pass in your va_list, here shown as parameter ap. Just pass the list, not a pointer to it. Then for the second argument to va_start(), you give the name of the parameter that you want to start processing arguments after. This must be the parameter right before the ... in the argument list. If you’ve already called va_start() on a particular va_list and you want to call va_start() on it again, you must call va_end() first! Chapter 59.
Return Value Returns nothing!
Example Here’s a demo that adds together an arbitrary number of integers. The first argument is the number of integers to add together. We’ll make use of that to figure out how many times we have to call va_arg().
1 #include
2 #include
3
4 int add(int count, ...)
5 {
6 int total = 0;
7 va_list va;
8
9 va_start(va, count); // Start with arguments after "count"
10
11 for (int i = 0; i < count; i++) {
12 int n = va_arg(va, int); // Get the next int
13
14 total += n;
15 }
16
17 va_end(va); // All done
18
19 return total;
20 }
21
22 int main(void)
23 {
24 printf("%d\n", add(4, 6, 2, -4, 17)); // 6 + 2 - 4 + 17 = 21
25 printf("%d\n", add(2, 22, 44)); // 22 + 44 = 66
26 }
See Also
va_arg(), va_end() Chapter 60
This header gives us access to (potentially) types of a fixed number of bits, or, at the very least, types that are at least that many bits. It also gives us handy macros to use.
60.1 Specific-Width Integers
There are three main classes of types defined here, signed and unsigned: • Integers of exactly a certain size (intN_t, uintN_t) • Integers that are at least a certain size (int_leastN_t, uint_leastN_t) • Integers that are at least a certain size and are as fast as possible (int_fastN_t, uint_fastN_t) Where the N occurs, you substitute the number of bits, commonly multiples of 8, e.g. uint16_t. The following types are guaranteed to be defined: int_least8_t uint_least8_t int_least16_t uint_least16_t int_least32_t uint_least32_t int_least64_t uint_least64_t int_fast8_t uint_fast8_t int_fast16_t uint_fast16_t int_fast32_t uint_fast32_t int_fast64_t uint_fast64_t
Everything else is optional, but you’ll probably also have the following, which are required when a system has integers of these sizes with no padding and two’s-complement representation… which is the case for Macs and PCs and a lot of other systems. In short, you very likely have these: int8_t uint8_t int16_t uint16_t int32_t uint32_t int64_t uint64_t
Other numbers of bits can also be supported by an implementation if it wants to go all crazy with it. Examples: #include
592 Chapter 60.
// ...
60.2 Other Integer Types
There are a couple optional types that are integers capable of holding pointer types. intptr_t uintptr_t
You can convert a void* to one of these types, and back again. And the void*s will compare equal. The use case is any place you need an integer that represents a pointer for some reason. Also, there are a couple types that are just there to be the biggest possible integers your system supports: intmax_t uintmax_t
Fun fact: you can print these types with the "%jd" and "%ju" printf() format specifiers. There are also a bunch of macros in
60.3 Macros
The following macros define the minimum and maximum values for these types: INT8_MAX INT8_MIN UINT8_MAX INT16_MAX INT16_MIN UINT16_MAX INT32_MAX INT32_MIN UINT32_MAX INT64_MAX INT64_MIN UINT64_MAX
INT_LEAST8_MAX INT_LEAST8_MIN UINT_LEAST8_MAX INT_LEAST16_MAX INT_LEAST16_MIN UINT_LEAST16_MAX INT_LEAST32_MAX INT_LEAST32_MIN UINT_LEAST32_MAX INT_LEAST64_MAX INT_LEAST64_MIN UINT_LEAST64_MAX
INT_FAST8_MAX INT_FAST8_MIN UINT_FAST8_MAX INT_FAST16_MAX INT_FAST16_MIN UINT_FAST16_MAX INT_FAST32_MAX INT_FAST32_MIN UINT_FAST32_MAX INT_FAST64_MAX INT_FAST64_MIN UINT_FAST64_MAX
INTMAX_MAX INTMAX_MIN UINTMAX_MAX
INTPTR_MAX INTPTR_MIN UINTPTR_MAX
For the exact-bit-size signed types, the minimum is exactly −(2푁−1) and the maximum is exactly 2푁−1− 1. And for the exact-bit-size unsigned types, the max is exactly 2푁 − 1. For the signed “least” and “fast” variants, the magnitude and sign of the minimum is at least −(2푁−1 −1) and the maximum is at least 2푁−1 − 1. And for unsigned it’s at least 2푁 − 1. INTMAX_MAX is at least 263 − 1, INTMAX_MIN is at least −(263 − 1) in sign and magnitude. And UINT- MAX_MAX is at least 264 − 1. Finally, INTPTR_MAX is at least 215 − 1, INTPTR_MIN is at least −(215 − 1) in sign and magnitude. And UINTPTR_MAX is at least 216 − 1. Chapter 60.
60.4 Other Limits
There are a bunch of types in
Macro Description PTRDIFF_MIN Minimum ptrdiff_t value PTRDIFF_MAX Maximum ptrdiff_t value SIG_ATOMIC_MIN Minimum sig_atomic_t value SIG_ATOMIC_MAX Maximum sig_atomic_t value SIZE_MAX Maximum size_t value WCHAR_MIN Minimum wchar_t value WCHAR_MAX Maximum wchar_t value WINT_MIN Minimum wint_t value WINT_MAX Maximum wint_t value
The spec says that PTRDIFF_MIN will be at least -65535 in magnitude. And PTRDIFF_MAX and SIZE_MAX will be at least 65535. SIG_ATOMIC_MIN and MAX will be either -127 and 127 (if it’s signed) or 0 and 255 (if it’s unsigned). Same for WCHAR_MIN and MAX. WINT_MIN and MAX will be either -32767 and 32767 (if it’s signed) or 0 and 65535 (if it’s unsigned).
60.5 Macros for Declaring Constants
If you recall, you can specify a type for integer constants: int x = 12; long int y = 12L; unsigned long long int z = 12ULL;
You can use the macros INTN_C() and UINTN() where N is 8, 16, 32 or 64. uint_least16_t x = INT16_C(3490); uint_least64_t y = INT64_C(1122334455);
A variant on these is INTMAX_C() and UINTMAX_C(). They will make a constant suitable for storing in an intmax_t or uintmax_t. intmax_t x = INTMAX_C(3490); uintmax_t x = UINTMAX_C(1122334455); Chapter 61
Function Description imaxabs() Compute the absolute value of an intmax_t imaxdiv() Compute the quotient and remainder of intmax_ts strtoimax() Convert strings to type intmax_t strtoumax() Convert strings to type uintmax_t wcstoimax() Convert wide strings to type intmax_t wcstoumax() Convert wide strings to type uintmax_t
This header does conversions to maximum sized integers, division with maximum sized integers, and also provides format specifiers for printf() and scanf() for a variety of types defined in
61.1 Macros
These are to help with printf() and scanf() when you use a type such as int_least16_t… what format specifiers do you use? Let’s start with printf()—all these macros start with PRI and then are followed by the format specifier you’d typically use for that type. Lastly, the number of bits is added on. For example, the format specifier for a 64-bit integer is PRId64—the d is because you usually print integers with "%d". An unsigned 16-bit integer could be printed with PRIu16. These macros expand to string literals. We can take advantage of the fact that C automatically concatenates neighboring string literals and use these specifiers like this:
1 #include
2 #include
3
4 int main(void)
5 {
6 int16_t x = 32;
7
8 printf("The value is %" PRId16 "!\n", x);
9 }
There’s nothing magical happening on line 8, above. Indeed, if I print out the value of the macro:
595 Chapter 61.
printf("%s\n", PRId16);
we get this on my system: hd
which is a printf() format specifier meaning “short signed integer” . So back to line 8, after string literal concatenation, it’s just as if I’d typed:
8 printf("The value is %hd!\n", x);
Here’s a table of all the macros you can use for printf() format specifiers… substitute the number of bits for N, usually 8, 16, 32, or 64.
PRIdN PRIdLEASTN PRIdFASTN PRIdMAX PRIdPTR PRIiN PRIiLEASTN PRIiFASTN PRIiMAX PRIiPTR PRIoN PRIoLEASTN PRIoFASTN PRIoMAX PRIoPTR PRIuN PRIuLEASTN PRIuFASTN PRIuMAX PRIuPTR PRIxN PRIxLEASTN PRIxFASTN PRIxMAX PRIxPTR PRIXN PRIXLEASTN PRIXFASTN PRIXMAX PRIXPTR
Note again how the lowercase center letter represents the usual format specifiers you’d pass to printf(): d, i, o, u, x, and X. And we have a similar set of macros for scanf() for reading in these various types:
SCNdN SCNdLEASTN SCNdFASTN SCNdMAX SCNdPTR SCNiN SCNiLEASTN SCNiFASTN SCNiMAX SCNiPTR SCNoN SCNoLEASTN SCNoFASTN SCNoMAX SCNoPTR SCNuN SCNuLEASTN SCNuFASTN SCNuMAX SCNuPTR SCNxN SCNxLEASTN SCNxFASTN SCNxMAX SCNxPTR
The rule is that for each type defined in
61.2 imaxabs()
Compute the absolute value of an intmax_t
Synopsis
#include
intmax_t imaxabs(intmax_t j);
Description When you need the absolute value of the biggest integer type on the system, this is the function for you. The spec notes that if the absolute value of the number cannot be represented, the behavior is undefined. This would happen if you tried to take the absolute value of the smallest possible negative number in a two’s-complement system.
Return Value Returns the absolute value of the input, |푗|. Chapter 61.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 intmax_t j = -3490;
7
8 printf("%jd\n", imaxabs(j)); // 3490
9 }
See Also
fabs()
61.3 imaxdiv()
Compute the quotient and remainder of intmax_ts
Synopsis
#include
imaxdiv_t imaxdiv(intmax_t numer, intmax_t denom);
Description When you want to do integer division and remainder in a single operation, this function will do it for you. It computes numer/denom and numer%denom and returns the result in a structure of type imaxdiv_t. This structure has two imaxdiv_t fields, quot and rem, that you use to retrieve the sought-after values.
Return Value
Returns an imaxdiv_t containing the quotient and remainder of the operation.
Example
1 #include
2 #include
3
4 int main(void)
5 {
6 intmax_t numer = INTMAX_C(3490);
7 intmax_t denom = INTMAX_C(17);
8
9 imaxdiv_t r = imaxdiv(numer, denom);
10
11 printf("Quotient: %jd, remainder: %jd\n", r.quot, r.rem);
12 }
Output: Chapter 61.
Quotient: 205, remainder: 5
See Also
remquo()
61.4 strtoimax() strtoumax()
Convert strings to types intmax_t and uintmax_t
Synopsis
#include
intmax_t strtoimax(const char * restrict nptr, char ** restrict endptr, int base);
uintmax_t strtoumax(const char * restrict nptr, char ** restrict endptr, int base);
Description
These work just like the strtol() family of functions, except they return an intmax_t or uintmax_t. See the strtol() reference page for details.
Return Value
Returns the converted string as an intmax_t or uintmax_t. If the result is out of range, the returned value will be INTMAX_MAX, INTMAX_MIN, or UINTMAX_MAX, as appropriate. And the errno variable will be set to ERANGE.
Example
The following example converts a base-10 number to an intmax_t. Then it attempts to convert an invalid base-2 number, catching the error.
1 #include
2 #include
3
4 int main(void)
5 {
6 intmax_t r;
7 char *endptr;
8
9 // Valid base-10 number
10 r = strtoimax("123456789012345", &endptr, 10);
11
12 if (*endptr != '\0')
13 printf("Invalid digit: %c\n", *endptr);
14 else
15 printf("Value is %jd\n", r);
16
17 // The following binary number contains an invalid digit
18 r = strtoimax("0100102010101101", &endptr, 2); Chapter 61.
19
20 if (*endptr != '\0')
21 printf("Invalid digit: %c\n", *endptr);
22 else
23 printf("Value is %jd\n", r);
24 }
Output: Value is 123456789012345 Invalid digit: 2
See Also
strtol(), errno
61.5 wcstoimax() wcstoumax()
Convert wide strings to types intmax_t and uintmax_t
Synopsis
#include
intmax_t wcstoimax(const wchar_t * restrict nptr, wchar_t ** restrict endptr, int base);
uintmax_t wcstoumax(const wchar_t * restrict nptr, wchar_t ** restrict endptr, int base);
Description
These work just like the wcstol() family of functions, except they return an intmax_t or uintmax_t. See the wcstol() reference page for details.
Return Value
Returns the converted wide string as an intmax_t or uintmax_t. If the result is out of range, the returned value will be INTMAX_MAX, INTMAX_MIN, or UINTMAX_MAX, as appropriate. And the errno variable will be set to ERANGE.
Example
The following example converts a base-10 number to an intmax_t. Then it attempts to convert an invalid base-2 number, catching the error.
1 #include
2 #include
3
4 int main(void)
5 {
6 intmax_t r;
7 wchar_t *endptr;
8 Chapter 61.
9 // Valid base-10 number
10 r = wcstoimax(L"123456789012345", &endptr, 10);
11
12 if (*endptr != '\0')
13 wprintf(L"Invalid digit: %lc\n", *endptr);
14 else
15 wprintf(L"Value is %jd\n", r);
16
17 // The following binary number contains an invalid digit
18 r = wcstoimax(L"0100102010101101", &endptr, 2);
19
20 if (*endptr != '\0')
21 wprintf(L"Invalid digit: %lc\n", *endptr);
22 else
23 wprintf(L"Value is %jd\n", r);
24 }
Value is 123456789012345 Invalid digit: 2
See Also
wcstol(), errno Chapter 62
Despite its name, I’ve haven’t seen this frequently included. It includes several types and macros.
Name Description ptrdiff_t Signed integer difference between two pointers size_t Unsigned integer type returned by sizeof max_align_t Declare a type with the biggest possible alignment wchar_t Wide character type NULL NULL pointer, as defined a number of places offsetof Get the byte offsets of struct or union fields
62.1 ptrdiff_t
This holds the different between two pointers. You could store this in another type, but the result of a pointer subtraction is an implementation-defined type; you can be maximally portable by using ptrdiff_t. int cats[100]; int *f = cats + 20; int *g = cats + 60; ptrdiff_t d = g - f; // difference is 40
And you can print it by prefixing the integer format specifier with t: printf("%td\n", d); // Print decimal: 40 printf("%tX\n", d); // Print hex: 28
62.2 size_t
This is the type returned by sizeof and used in a few other places. It’s an unsigned integer. You can print it using the z prefix in printf(): size_t x = sizeof(int); printf("%zu\n", x);
601 Chapter 62.
Some functions return negative numbers cast to size_t as error values (such as mbrtoc16()). If you want to print these as negative values, you can do it with %zd: size_t x = mbrtoc16(&a, "b", 8, &ps);
printf("%zd\n", x);
62.3 max_align_t
As far as I can tell, this exists to allow the runtime computation of the maximum fundamental alignment1 on the current platform. Someone please mail me if there’s another use. Maybe you need this if you’re writing your own memory allocator or somesuch.
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 int max = alignof(max_align_t);
8
9 printf("Maximum fundamental alignment: %d\n", max);
10 }
On my system, this prints: Maximum fundamental alignment: 16
See also alignas, alignof.
62.4 wchar_t
This is analogous to char, except it’s for wide characters. It’s an integer type that has enough range to hold unique values for all characters in all supported locales. The value 0 is the wide NUL character. Finally, the values of character constants from the basic character set will be the same as their correspond- ing wchar_t values… unless __STDC_MB_MIGHT_NEQ_WC__ is defined.
62.5 offsetof
If you have a struct or union, you can use this to get the byte offset of fields within that type. Usage is: offsetof(type, fieldname);
The resulting value has type size_t. Here’s an example that prints the field offsets of a struct:
1 #include
2 #include
3
4 struct foo {
5 int a;
6 char b;
7 char c;
1https://en.wikipedia.org/wiki/Data_structure_alignment Chapter 62.
8 float d;
9 };
10
11 int main(void)
12 {
13 printf("a: %zu\n", offsetof(struct foo, a));
14 printf("b: %zu\n", offsetof(struct foo, b));
15 printf("c: %zu\n", offsetof(struct foo, c));
16 printf("d: %zu\n", offsetof(struct foo, d));
17 }
On my system, this outputs: a: 0 b: 4 c: 5 d: 8
And you can’t use offsetof on a bitfield, so don’t get your hopes up. Chapter 63
This is a small header file that defines a number of convenient Boolean macros. If you really need that kind of thing.
Macro Description bool Type for Boolean, expands to _Bool true True value, expands to 1 false False value, expands to 0
There’s on more macro that I’m not putting in the table because it’s such a long name it’ll blow up the table alignment: __bool_true_false_are_defined
which expands to 1.
63.1 Example
Here’s a lame example that shows off these macros.
1 #include
2 #include
3
4 int main(void)
5 {
6 bool x;
7
8 x = (3 > 2);
9
10 if (x == true)
11 printf("The universe still makes sense.\n");
12
13 x = false;
14
15 printf("x is now %d\n", x); // 0
16 }
Output: The universe still makes sense. x is now 0
604 Chapter 63.
63.2 _Bool?
What’s the deal with _Bool? Why didn’t they just make it bool? Well, there was a lot of C code out there where people had defined their own bool type and adding an official bool would have broken those typedefs. But C has already reserved all identifiers that start with an underscore followed by a capital letter, so it was clear to make up a new _Bool type and go with that. And, if you know your code can handle it, you can include this header to get all this juicy syntax. One more note on conversions: unlike converting to int, the only thing that converts to false in a _Bool is a scalar zero value. Anything at all that’s not zero, like -3490, 0.12, or NaN, converts to true. Chapter 64
If you’re coding up something low-level like a memory allocator that interfaces with your OS, you might need this header file. But most C devs go their careers without using it. Alignment is all about multiples of addresses on which objects can be stored. Can you store this at any address? Or must it be a starting address that’s divisible by 2? Or 8? Or 16?
Name Description alignas() Specify alignment, expands to _Alignas alignof() Get alignment, expands to _Alignof
These two additional macros are defined to be 1: __alignas_is_defined __alignof_is_defined
Quick note: alignments greater than that of max_align_t are known as overalignments and are implementation-defined.
64.1 alignas() _Alignas()
Force a variable to have a certain alignment
Synopsis
#include
_Alignas(type-name) _Alignas(constant-expression)
Description This isn’t a function. Rather, it’s an alignment specifier that you can use with a variable declaration. If you need your char to be aligned like an int, you can force it like this when you declare it: char alignas(int) c;
You can also pass a constant value or expression in for the alignment. This has to be something supported by the system, but the spec stops short of dictating what values you can put in there. Small powers of 2
606 Chapter 64.
(1, 2, 4, 8, and 16) are generally safe bets. char alignas(8) c; // align on 8-byte boundaries
For convenience, you can also specify 0 if you want the default alignment (as if you hadn’t said alignas() at all): char alignas(0) c; // use default alignment for this type
Example
1 #include
2 #include
3 #include
4
5 int main(void)
6 {
7 int i, j;
8 char alignas(max_align_t) a, b;
9 char alignas(int) c, d;
10 char e, f;
11
12 printf("i: %p\n", (void *)&i);
13 printf("j: %p\n\n", (void *)&j);
14 printf("a: %p\n", (void *)&a);
15 printf("b: %p\n\n", (void *)&b);
16 printf("c: %p\n", (void *)&c);
17 printf("d: %p\n\n", (void *)&d);
18 printf("e: %p\n", (void *)&e);
19 printf("f: %p\n", (void *)&f);
20 }
Output on my system follows. Notice the difference between the pairs of values. • i and j, both ints, are aligned on 4-byte boundaries. • a and b have been forced to the boundary of the type max_align_t, which is every 16 bytes on my system. • c and d have been forced to the same alignment as int, which is 4 bytes, just like with i and j. • e and f do not have an alignment specified, so they were stored with their default alignment of 1 byte. i: 0x7ffee7dfb4cc <-- difference of 4 bytes j: 0x7ffee7dfb4c8
a: 0x7ffee7dfb4c0 <-- difference of 16 bytes b: 0x7ffee7dfb4b0
c: 0x7ffee7dfb4ac <-- difference of 4 bytes d: 0x7ffee7dfb4a8
e: 0x7ffee7dfb4a7 <-- difference of 1 byte f: 0x7ffee7dfb4a6
See Also
alignof, max_align_t Chapter 64.
64.2 alignof() _Alignof()
Get the alignment of a type
Synopsis
#include
alignof(type-name)
_Alignof(type-name)
Description
This evaluates to a value of type size_t that gives the alignment of a particular type on your system.
Return Value Returns the alignment value, i.e. the address of the beginning of the given type of object must begin on an address boundary divisible by this number.
Example Print out the alignments of a variety of different types.
1 #include
2 #include
3 #include
4
5 struct t {
6 int a;
7 char b;
8 float c;
9 };
10
11 int main(void)
12 {
13 printf("char : %zu\n", alignof(char));
14 printf("short : %zu\n", alignof(short));
15 printf("int : %zu\n", alignof(int));
16 printf("long : %zu\n", alignof(long));
17 printf("long long : %zu\n", alignof(long long));
18 printf("double : %zu\n", alignof(double));
19 printf("long double: %zu\n", alignof(long double));
20 printf("struct t : %zu\n", alignof(long double));
21 printf("max_align_t: %zu\n", alignof(max_align_t));
22 }
Output on my system: char : 1 short : 2 int : 4 long : 8 long long : 8 double : 8 long double: 16 struct t : 16 max_align_t: 16 Chapter 64.
See Also alignas, max_align_t