1This is doc/cpp.info, produced by makeinfo version 4.8 from 2/scratch/mitchell/gcc-releases/gcc-4.2.1/gcc-4.2.1/gcc/doc/cpp.texi. 3 4 Copyright (C) 1987, 1989, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 51998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software 6Foundation, Inc. 7 8 Permission is granted to copy, distribute and/or modify this document 9under the terms of the GNU Free Documentation License, Version 1.1 or 10any later version published by the Free Software Foundation. A copy of 11the license is included in the section entitled "GNU Free Documentation 12License". 13 14 This manual contains no Invariant Sections. The Front-Cover Texts 15are (a) (see below), and the Back-Cover Texts are (b) (see below). 16 17 (a) The FSF's Front-Cover Text is: 18 19 A GNU Manual 20 21 (b) The FSF's Back-Cover Text is: 22 23 You have freedom to copy and modify this GNU Manual, like GNU 24software. Copies published by the Free Software Foundation raise 25funds for GNU development. 26 27INFO-DIR-SECTION Software development 28START-INFO-DIR-ENTRY 29* Cpp: (cpp). The GNU C preprocessor. 30END-INFO-DIR-ENTRY 31 32 33File: cpp.info, Node: Top, Next: Overview, Up: (dir) 34 35The C Preprocessor 36****************** 37 38The C preprocessor implements the macro language used to transform C, 39C++, and Objective-C programs before they are compiled. It can also be 40useful on its own. 41 42* Menu: 43 44* Overview:: 45* Header Files:: 46* Macros:: 47* Conditionals:: 48* Diagnostics:: 49* Line Control:: 50* Pragmas:: 51* Other Directives:: 52* Preprocessor Output:: 53* Traditional Mode:: 54* Implementation Details:: 55* Invocation:: 56* Environment Variables:: 57* GNU Free Documentation License:: 58* Index of Directives:: 59* Option Index:: 60* Concept Index:: 61 62 --- The Detailed Node Listing --- 63 64Overview 65 66* Character sets:: 67* Initial processing:: 68* Tokenization:: 69* The preprocessing language:: 70 71Header Files 72 73* Include Syntax:: 74* Include Operation:: 75* Search Path:: 76* Once-Only Headers:: 77* Computed Includes:: 78* Wrapper Headers:: 79* System Headers:: 80 81Macros 82 83* Object-like Macros:: 84* Function-like Macros:: 85* Macro Arguments:: 86* Stringification:: 87* Concatenation:: 88* Variadic Macros:: 89* Predefined Macros:: 90* Undefining and Redefining Macros:: 91* Directives Within Macro Arguments:: 92* Macro Pitfalls:: 93 94Predefined Macros 95 96* Standard Predefined Macros:: 97* Common Predefined Macros:: 98* System-specific Predefined Macros:: 99* C++ Named Operators:: 100 101Macro Pitfalls 102 103* Misnesting:: 104* Operator Precedence Problems:: 105* Swallowing the Semicolon:: 106* Duplication of Side Effects:: 107* Self-Referential Macros:: 108* Argument Prescan:: 109* Newlines in Arguments:: 110 111Conditionals 112 113* Conditional Uses:: 114* Conditional Syntax:: 115* Deleted Code:: 116 117Conditional Syntax 118 119* Ifdef:: 120* If:: 121* Defined:: 122* Else:: 123* Elif:: 124 125Implementation Details 126 127* Implementation-defined behavior:: 128* Implementation limits:: 129* Obsolete Features:: 130* Differences from previous versions:: 131 132Obsolete Features 133 134* Assertions:: 135* Obsolete once-only headers:: 136 137 Copyright (C) 1987, 1989, 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1381998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software 139Foundation, Inc. 140 141 Permission is granted to copy, distribute and/or modify this document 142under the terms of the GNU Free Documentation License, Version 1.1 or 143any later version published by the Free Software Foundation. A copy of 144the license is included in the section entitled "GNU Free Documentation 145License". 146 147 This manual contains no Invariant Sections. The Front-Cover Texts 148are (a) (see below), and the Back-Cover Texts are (b) (see below). 149 150 (a) The FSF's Front-Cover Text is: 151 152 A GNU Manual 153 154 (b) The FSF's Back-Cover Text is: 155 156 You have freedom to copy and modify this GNU Manual, like GNU 157software. Copies published by the Free Software Foundation raise 158funds for GNU development. 159 160 161File: cpp.info, Node: Overview, Next: Header Files, Prev: Top, Up: Top 162 1631 Overview 164********** 165 166The C preprocessor, often known as "cpp", is a "macro processor" that 167is used automatically by the C compiler to transform your program 168before compilation. It is called a macro processor because it allows 169you to define "macros", which are brief abbreviations for longer 170constructs. 171 172 The C preprocessor is intended to be used only with C, C++, and 173Objective-C source code. In the past, it has been abused as a general 174text processor. It will choke on input which does not obey C's lexical 175rules. For example, apostrophes will be interpreted as the beginning of 176character constants, and cause errors. Also, you cannot rely on it 177preserving characteristics of the input which are not significant to 178C-family languages. If a Makefile is preprocessed, all the hard tabs 179will be removed, and the Makefile will not work. 180 181 Having said that, you can often get away with using cpp on things 182which are not C. Other Algol-ish programming languages are often safe 183(Pascal, Ada, etc.) So is assembly, with caution. `-traditional-cpp' 184mode preserves more white space, and is otherwise more permissive. Many 185of the problems can be avoided by writing C or C++ style comments 186instead of native language comments, and keeping macros simple. 187 188 Wherever possible, you should use a preprocessor geared to the 189language you are writing in. Modern versions of the GNU assembler have 190macro facilities. Most high level programming languages have their own 191conditional compilation and inclusion mechanism. If all else fails, 192try a true general text processor, such as GNU M4. 193 194 C preprocessors vary in some details. This manual discusses the GNU 195C preprocessor, which provides a small superset of the features of ISO 196Standard C. In its default mode, the GNU C preprocessor does not do a 197few things required by the standard. These are features which are 198rarely, if ever, used, and may cause surprising changes to the meaning 199of a program which does not expect them. To get strict ISO Standard C, 200you should use the `-std=c89' or `-std=c99' options, depending on which 201version of the standard you want. To get all the mandatory 202diagnostics, you must also use `-pedantic'. *Note Invocation::. 203 204 This manual describes the behavior of the ISO preprocessor. To 205minimize gratuitous differences, where the ISO preprocessor's behavior 206does not conflict with traditional semantics, the traditional 207preprocessor should behave the same way. The various differences that 208do exist are detailed in the section *Note Traditional Mode::. 209 210 For clarity, unless noted otherwise, references to `CPP' in this 211manual refer to GNU CPP. 212 213* Menu: 214 215* Character sets:: 216* Initial processing:: 217* Tokenization:: 218* The preprocessing language:: 219 220 221File: cpp.info, Node: Character sets, Next: Initial processing, Up: Overview 222 2231.1 Character sets 224================== 225 226Source code character set processing in C and related languages is 227rather complicated. The C standard discusses two character sets, but 228there are really at least four. 229 230 The files input to CPP might be in any character set at all. CPP's 231very first action, before it even looks for line boundaries, is to 232convert the file into the character set it uses for internal 233processing. That set is what the C standard calls the "source" 234character set. It must be isomorphic with ISO 10646, also known as 235Unicode. CPP uses the UTF-8 encoding of Unicode. 236 237 The character sets of the input files are specified using the 238`-finput-charset=' option. 239 240 All preprocessing work (the subject of the rest of this manual) is 241carried out in the source character set. If you request textual output 242from the preprocessor with the `-E' option, it will be in UTF-8. 243 244 After preprocessing is complete, string and character constants are 245converted again, into the "execution" character set. This character 246set is under control of the user; the default is UTF-8, matching the 247source character set. Wide string and character constants have their 248own character set, which is not called out specifically in the 249standard. Again, it is under control of the user. The default is 250UTF-16 or UTF-32, whichever fits in the target's `wchar_t' type, in the 251target machine's byte order.(1) Octal and hexadecimal escape sequences 252do not undergo conversion; '\x12' has the value 0x12 regardless of the 253currently selected execution character set. All other escapes are 254replaced by the character in the source character set that they 255represent, then converted to the execution character set, just like 256unescaped characters. 257 258 Unless the experimental `-fextended-identifiers' option is used, GCC 259does not permit the use of characters outside the ASCII range, nor `\u' 260and `\U' escapes, in identifiers. Even with that option, characters 261outside the ASCII range can only be specified with the `\u' and `\U' 262escapes, not used directly in identifiers. 263 264 ---------- Footnotes ---------- 265 266 (1) UTF-16 does not meet the requirements of the C standard for a 267wide character set, but the choice of 16-bit `wchar_t' is enshrined in 268some system ABIs so we cannot fix this. 269 270 271File: cpp.info, Node: Initial processing, Next: Tokenization, Prev: Character sets, Up: Overview 272 2731.2 Initial processing 274====================== 275 276The preprocessor performs a series of textual transformations on its 277input. These happen before all other processing. Conceptually, they 278happen in a rigid order, and the entire file is run through each 279transformation before the next one begins. CPP actually does them all 280at once, for performance reasons. These transformations correspond 281roughly to the first three "phases of translation" described in the C 282standard. 283 284 1. The input file is read into memory and broken into lines. 285 286 Different systems use different conventions to indicate the end of 287 a line. GCC accepts the ASCII control sequences `LF', `CR LF' and 288 `CR' as end-of-line markers. These are the canonical sequences 289 used by Unix, DOS and VMS, and the classic Mac OS (before OSX) 290 respectively. You may therefore safely copy source code written 291 on any of those systems to a different one and use it without 292 conversion. (GCC may lose track of the current line number if a 293 file doesn't consistently use one convention, as sometimes happens 294 when it is edited on computers with different conventions that 295 share a network file system.) 296 297 If the last line of any input file lacks an end-of-line marker, 298 the end of the file is considered to implicitly supply one. The C 299 standard says that this condition provokes undefined behavior, so 300 GCC will emit a warning message. 301 302 2. If trigraphs are enabled, they are replaced by their corresponding 303 single characters. By default GCC ignores trigraphs, but if you 304 request a strictly conforming mode with the `-std' option, or you 305 specify the `-trigraphs' option, then it converts them. 306 307 These are nine three-character sequences, all starting with `??', 308 that are defined by ISO C to stand for single characters. They 309 permit obsolete systems that lack some of C's punctuation to use 310 C. For example, `??/' stands for `\', so '??/n' is a character 311 constant for a newline. 312 313 Trigraphs are not popular and many compilers implement them 314 incorrectly. Portable code should not rely on trigraphs being 315 either converted or ignored. With `-Wtrigraphs' GCC will warn you 316 when a trigraph may change the meaning of your program if it were 317 converted. *Note Wtrigraphs::. 318 319 In a string constant, you can prevent a sequence of question marks 320 from being confused with a trigraph by inserting a backslash 321 between the question marks, or by separating the string literal at 322 the trigraph and making use of string literal concatenation. 323 "(??\?)" is the string `(???)', not `(?]'. Traditional C 324 compilers do not recognize these idioms. 325 326 The nine trigraphs and their replacements are 327 328 Trigraph: ??( ??) ??< ??> ??= ??/ ??' ??! ??- 329 Replacement: [ ] { } # \ ^ | ~ 330 331 3. Continued lines are merged into one long line. 332 333 A continued line is a line which ends with a backslash, `\'. The 334 backslash is removed and the following line is joined with the 335 current one. No space is inserted, so you may split a line 336 anywhere, even in the middle of a word. (It is generally more 337 readable to split lines only at white space.) 338 339 The trailing backslash on a continued line is commonly referred to 340 as a "backslash-newline". 341 342 If there is white space between a backslash and the end of a line, 343 that is still a continued line. However, as this is usually the 344 result of an editing mistake, and many compilers will not accept 345 it as a continued line, GCC will warn you about it. 346 347 4. All comments are replaced with single spaces. 348 349 There are two kinds of comments. "Block comments" begin with `/*' 350 and continue until the next `*/'. Block comments do not nest: 351 352 /* this is /* one comment */ text outside comment 353 354 "Line comments" begin with `//' and continue to the end of the 355 current line. Line comments do not nest either, but it does not 356 matter, because they would end in the same place anyway. 357 358 // this is // one comment 359 text outside comment 360 361 It is safe to put line comments inside block comments, or vice versa. 362 363 /* block comment 364 // contains line comment 365 yet more comment 366 */ outside comment 367 368 // line comment /* contains block comment */ 369 370 But beware of commenting out one end of a block comment with a line 371comment. 372 373 // l.c. /* block comment begins 374 oops! this isn't a comment anymore */ 375 376 Comments are not recognized within string literals. "/* blah */" is 377the string constant `/* blah */', not an empty string. 378 379 Line comments are not in the 1989 edition of the C standard, but they 380are recognized by GCC as an extension. In C++ and in the 1999 edition 381of the C standard, they are an official part of the language. 382 383 Since these transformations happen before all other processing, you 384can split a line mechanically with backslash-newline anywhere. You can 385comment out the end of a line. You can continue a line comment onto the 386next line with backslash-newline. You can even split `/*', `*/', and 387`//' onto multiple lines with backslash-newline. For example: 388 389 /\ 390 * 391 */ # /* 392 */ defi\ 393 ne FO\ 394 O 10\ 395 20 396 397is equivalent to `#define FOO 1020'. All these tricks are extremely 398confusing and should not be used in code intended to be readable. 399 400 There is no way to prevent a backslash at the end of a line from 401being interpreted as a backslash-newline. This cannot affect any 402correct program, however. 403 404 405File: cpp.info, Node: Tokenization, Next: The preprocessing language, Prev: Initial processing, Up: Overview 406 4071.3 Tokenization 408================ 409 410After the textual transformations are finished, the input file is 411converted into a sequence of "preprocessing tokens". These mostly 412correspond to the syntactic tokens used by the C compiler, but there are 413a few differences. White space separates tokens; it is not itself a 414token of any kind. Tokens do not have to be separated by white space, 415but it is often necessary to avoid ambiguities. 416 417 When faced with a sequence of characters that has more than one 418possible tokenization, the preprocessor is greedy. It always makes 419each token, starting from the left, as big as possible before moving on 420to the next token. For instance, `a+++++b' is interpreted as 421`a ++ ++ + b', not as `a ++ + ++ b', even though the latter 422tokenization could be part of a valid C program and the former could 423not. 424 425 Once the input file is broken into tokens, the token boundaries never 426change, except when the `##' preprocessing operator is used to paste 427tokens together. *Note Concatenation::. For example, 428 429 #define foo() bar 430 foo()baz 431 ==> bar baz 432 _not_ 433 ==> barbaz 434 435 The compiler does not re-tokenize the preprocessor's output. Each 436preprocessing token becomes one compiler token. 437 438 Preprocessing tokens fall into five broad classes: identifiers, 439preprocessing numbers, string literals, punctuators, and other. An 440"identifier" is the same as an identifier in C: any sequence of 441letters, digits, or underscores, which begins with a letter or 442underscore. Keywords of C have no significance to the preprocessor; 443they are ordinary identifiers. You can define a macro whose name is a 444keyword, for instance. The only identifier which can be considered a 445preprocessing keyword is `defined'. *Note Defined::. 446 447 This is mostly true of other languages which use the C preprocessor. 448However, a few of the keywords of C++ are significant even in the 449preprocessor. *Note C++ Named Operators::. 450 451 In the 1999 C standard, identifiers may contain letters which are not 452part of the "basic source character set", at the implementation's 453discretion (such as accented Latin letters, Greek letters, or Chinese 454ideograms). This may be done with an extended character set, or the 455`\u' and `\U' escape sequences. The implementation of this feature in 456GCC is experimental; such characters are only accepted in the `\u' and 457`\U' forms and only if `-fextended-identifiers' is used. 458 459 As an extension, GCC treats `$' as a letter. This is for 460compatibility with some systems, such as VMS, where `$' is commonly 461used in system-defined function and object names. `$' is not a letter 462in strictly conforming mode, or if you specify the `-$' option. *Note 463Invocation::. 464 465 A "preprocessing number" has a rather bizarre definition. The 466category includes all the normal integer and floating point constants 467one expects of C, but also a number of other things one might not 468initially recognize as a number. Formally, preprocessing numbers begin 469with an optional period, a required decimal digit, and then continue 470with any sequence of letters, digits, underscores, periods, and 471exponents. Exponents are the two-character sequences `e+', `e-', `E+', 472`E-', `p+', `p-', `P+', and `P-'. (The exponents that begin with `p' 473or `P' are new to C99. They are used for hexadecimal floating-point 474constants.) 475 476 The purpose of this unusual definition is to isolate the preprocessor 477from the full complexity of numeric constants. It does not have to 478distinguish between lexically valid and invalid floating-point numbers, 479which is complicated. The definition also permits you to split an 480identifier at any position and get exactly two tokens, which can then be 481pasted back together with the `##' operator. 482 483 It's possible for preprocessing numbers to cause programs to be 484misinterpreted. For example, `0xE+12' is a preprocessing number which 485does not translate to any valid numeric constant, therefore a syntax 486error. It does not mean `0xE + 12', which is what you might have 487intended. 488 489 "String literals" are string constants, character constants, and 490header file names (the argument of `#include').(1) String constants 491and character constants are straightforward: "..." or '...'. In either 492case embedded quotes should be escaped with a backslash: '\'' is the 493character constant for `''. There is no limit on the length of a 494character constant, but the value of a character constant that contains 495more than one character is implementation-defined. *Note 496Implementation Details::. 497 498 Header file names either look like string constants, "...", or are 499written with angle brackets instead, <...>. In either case, backslash 500is an ordinary character. There is no way to escape the closing quote 501or angle bracket. The preprocessor looks for the header file in 502different places depending on which form you use. *Note Include 503Operation::. 504 505 No string literal may extend past the end of a line. Older versions 506of GCC accepted multi-line string constants. You may use continued 507lines instead, or string constant concatenation. *Note Differences 508from previous versions::. 509 510 "Punctuators" are all the usual bits of punctuation which are 511meaningful to C and C++. All but three of the punctuation characters in 512ASCII are C punctuators. The exceptions are `@', `$', and ``'. In 513addition, all the two- and three-character operators are punctuators. 514There are also six "digraphs", which the C++ standard calls 515"alternative tokens", which are merely alternate ways to spell other 516punctuators. This is a second attempt to work around missing 517punctuation in obsolete systems. It has no negative side effects, 518unlike trigraphs, but does not cover as much ground. The digraphs and 519their corresponding normal punctuators are: 520 521 Digraph: <% %> <: :> %: %:%: 522 Punctuator: { } [ ] # ## 523 524 Any other single character is considered "other". It is passed on to 525the preprocessor's output unmolested. The C compiler will almost 526certainly reject source code containing "other" tokens. In ASCII, the 527only other characters are `@', `$', ``', and control characters other 528than NUL (all bits zero). (Note that `$' is normally considered a 529letter.) All characters with the high bit set (numeric range 5300x7F-0xFF) are also "other" in the present implementation. This will 531change when proper support for international character sets is added to 532GCC. 533 534 NUL is a special case because of the high probability that its 535appearance is accidental, and because it may be invisible to the user 536(many terminals do not display NUL at all). Within comments, NULs are 537silently ignored, just as any other character would be. In running 538text, NUL is considered white space. For example, these two directives 539have the same meaning. 540 541 #define X^@1 542 #define X 1 543 544(where `^@' is ASCII NUL). Within string or character constants, NULs 545are preserved. In the latter two cases the preprocessor emits a 546warning message. 547 548 ---------- Footnotes ---------- 549 550 (1) The C standard uses the term "string literal" to refer only to 551what we are calling "string constants". 552 553 554File: cpp.info, Node: The preprocessing language, Prev: Tokenization, Up: Overview 555 5561.4 The preprocessing language 557============================== 558 559After tokenization, the stream of tokens may simply be passed straight 560to the compiler's parser. However, if it contains any operations in the 561"preprocessing language", it will be transformed first. This stage 562corresponds roughly to the standard's "translation phase 4" and is what 563most people think of as the preprocessor's job. 564 565 The preprocessing language consists of "directives" to be executed 566and "macros" to be expanded. Its primary capabilities are: 567 568 * Inclusion of header files. These are files of declarations that 569 can be substituted into your program. 570 571 * Macro expansion. You can define "macros", which are abbreviations 572 for arbitrary fragments of C code. The preprocessor will replace 573 the macros with their definitions throughout the program. Some 574 macros are automatically defined for you. 575 576 * Conditional compilation. You can include or exclude parts of the 577 program according to various conditions. 578 579 * Line control. If you use a program to combine or rearrange source 580 files into an intermediate file which is then compiled, you can 581 use line control to inform the compiler where each source line 582 originally came from. 583 584 * Diagnostics. You can detect problems at compile time and issue 585 errors or warnings. 586 587 There are a few more, less useful, features. 588 589 Except for expansion of predefined macros, all these operations are 590triggered with "preprocessing directives". Preprocessing directives 591are lines in your program that start with `#'. Whitespace is allowed 592before and after the `#'. The `#' is followed by an identifier, the 593"directive name". It specifies the operation to perform. Directives 594are commonly referred to as `#NAME' where NAME is the directive name. 595For example, `#define' is the directive that defines a macro. 596 597 The `#' which begins a directive cannot come from a macro expansion. 598Also, the directive name is not macro expanded. Thus, if `foo' is 599defined as a macro expanding to `define', that does not make `#foo' a 600valid preprocessing directive. 601 602 The set of valid directive names is fixed. Programs cannot define 603new preprocessing directives. 604 605 Some directives require arguments; these make up the rest of the 606directive line and must be separated from the directive name by 607whitespace. For example, `#define' must be followed by a macro name 608and the intended expansion of the macro. 609 610 A preprocessing directive cannot cover more than one line. The line 611may, however, be continued with backslash-newline, or by a block comment 612which extends past the end of the line. In either case, when the 613directive is processed, the continuations have already been merged with 614the first line to make one long line. 615 616 617File: cpp.info, Node: Header Files, Next: Macros, Prev: Overview, Up: Top 618 6192 Header Files 620************** 621 622A header file is a file containing C declarations and macro definitions 623(*note Macros::) to be shared between several source files. You request 624the use of a header file in your program by "including" it, with the C 625preprocessing directive `#include'. 626 627 Header files serve two purposes. 628 629 * System header files declare the interfaces to parts of the 630 operating system. You include them in your program to supply the 631 definitions and declarations you need to invoke system calls and 632 libraries. 633 634 * Your own header files contain declarations for interfaces between 635 the source files of your program. Each time you have a group of 636 related declarations and macro definitions all or most of which 637 are needed in several different source files, it is a good idea to 638 create a header file for them. 639 640 Including a header file produces the same results as copying the 641header file into each source file that needs it. Such copying would be 642time-consuming and error-prone. With a header file, the related 643declarations appear in only one place. If they need to be changed, they 644can be changed in one place, and programs that include the header file 645will automatically use the new version when next recompiled. The header 646file eliminates the labor of finding and changing all the copies as well 647as the risk that a failure to find one copy will result in 648inconsistencies within a program. 649 650 In C, the usual convention is to give header files names that end 651with `.h'. It is most portable to use only letters, digits, dashes, and 652underscores in header file names, and at most one dot. 653 654* Menu: 655 656* Include Syntax:: 657* Include Operation:: 658* Search Path:: 659* Once-Only Headers:: 660* Computed Includes:: 661* Wrapper Headers:: 662* System Headers:: 663 664 665File: cpp.info, Node: Include Syntax, Next: Include Operation, Up: Header Files 666 6672.1 Include Syntax 668================== 669 670Both user and system header files are included using the preprocessing 671directive `#include'. It has two variants: 672 673`#include <FILE>' 674 This variant is used for system header files. It searches for a 675 file named FILE in a standard list of system directories. You can 676 prepend directories to this list with the `-I' option (*note 677 Invocation::). 678 679`#include "FILE"' 680 This variant is used for header files of your own program. It 681 searches for a file named FILE first in the directory containing 682 the current file, then in the quote directories and then the same 683 directories used for `<FILE>'. You can prepend directories to the 684 list of quote directories with the `-iquote' option. 685 686 The argument of `#include', whether delimited with quote marks or 687angle brackets, behaves like a string constant in that comments are not 688recognized, and macro names are not expanded. Thus, `#include <x/*y>' 689specifies inclusion of a system header file named `x/*y'. 690 691 However, if backslashes occur within FILE, they are considered 692ordinary text characters, not escape characters. None of the character 693escape sequences appropriate to string constants in C are processed. 694Thus, `#include "x\n\\y"' specifies a filename containing three 695backslashes. (Some systems interpret `\' as a pathname separator. All 696of these also interpret `/' the same way. It is most portable to use 697only `/'.) 698 699 It is an error if there is anything (other than comments) on the line 700after the file name. 701 702 703File: cpp.info, Node: Include Operation, Next: Search Path, Prev: Include Syntax, Up: Header Files 704 7052.2 Include Operation 706===================== 707 708The `#include' directive works by directing the C preprocessor to scan 709the specified file as input before continuing with the rest of the 710current file. The output from the preprocessor contains the output 711already generated, followed by the output resulting from the included 712file, followed by the output that comes from the text after the 713`#include' directive. For example, if you have a header file 714`header.h' as follows, 715 716 char *test (void); 717 718and a main program called `program.c' that uses the header file, like 719this, 720 721 int x; 722 #include "header.h" 723 724 int 725 main (void) 726 { 727 puts (test ()); 728 } 729 730the compiler will see the same token stream as it would if `program.c' 731read 732 733 int x; 734 char *test (void); 735 736 int 737 main (void) 738 { 739 puts (test ()); 740 } 741 742 Included files are not limited to declarations and macro definitions; 743those are merely the typical uses. Any fragment of a C program can be 744included from another file. The include file could even contain the 745beginning of a statement that is concluded in the containing file, or 746the end of a statement that was started in the including file. However, 747an included file must consist of complete tokens. Comments and string 748literals which have not been closed by the end of an included file are 749invalid. For error recovery, they are considered to end at the end of 750the file. 751 752 To avoid confusion, it is best if header files contain only complete 753syntactic units--function declarations or definitions, type 754declarations, etc. 755 756 The line following the `#include' directive is always treated as a 757separate line by the C preprocessor, even if the included file lacks a 758final newline. 759 760 761File: cpp.info, Node: Search Path, Next: Once-Only Headers, Prev: Include Operation, Up: Header Files 762 7632.3 Search Path 764=============== 765 766GCC looks in several different places for headers. On a normal Unix 767system, if you do not instruct it otherwise, it will look for headers 768requested with `#include <FILE>' in: 769 770 /usr/local/include 771 LIBDIR/gcc/TARGET/VERSION/include 772 /usr/TARGET/include 773 /usr/include 774 775 For C++ programs, it will also look in `/usr/include/g++-v3', first. 776In the above, TARGET is the canonical name of the system GCC was 777configured to compile code for; often but not always the same as the 778canonical name of the system it runs on. VERSION is the version of GCC 779in use. 780 781 You can add to this list with the `-IDIR' command line option. All 782the directories named by `-I' are searched, in left-to-right order, 783_before_ the default directories. The only exception is when `dir' is 784already searched by default. In this case, the option is ignored and 785the search order for system directories remains unchanged. 786 787 Duplicate directories are removed from the quote and bracket search 788chains before the two chains are merged to make the final search chain. 789Thus, it is possible for a directory to occur twice in the final search 790chain if it was specified in both the quote and bracket chains. 791 792 You can prevent GCC from searching any of the default directories 793with the `-nostdinc' option. This is useful when you are compiling an 794operating system kernel or some other program that does not use the 795standard C library facilities, or the standard C library itself. `-I' 796options are not ignored as described above when `-nostdinc' is in 797effect. 798 799 GCC looks for headers requested with `#include "FILE"' first in the 800directory containing the current file, then in the directories as 801specified by `-iquote' options, then in the same places it would have 802looked for a header requested with angle brackets. For example, if 803`/usr/include/sys/stat.h' contains `#include "types.h"', GCC looks for 804`types.h' first in `/usr/include/sys', then in its usual search path. 805 806 `#line' (*note Line Control::) does not change GCC's idea of the 807directory containing the current file. 808 809 You may put `-I-' at any point in your list of `-I' options. This 810has two effects. First, directories appearing before the `-I-' in the 811list are searched only for headers requested with quote marks. 812Directories after `-I-' are searched for all headers. Second, the 813directory containing the current file is not searched for anything, 814unless it happens to be one of the directories named by an `-I' switch. 815`-I-' is deprecated, `-iquote' should be used instead. 816 817 `-I. -I-' is not the same as no `-I' options at all, and does not 818cause the same behavior for `<>' includes that `""' includes get with 819no special options. `-I.' searches the compiler's current working 820directory for header files. That may or may not be the same as the 821directory containing the current file. 822 823 If you need to look for headers in a directory named `-', write 824`-I./-'. 825 826 There are several more ways to adjust the header search path. They 827are generally less useful. *Note Invocation::. 828 829 830File: cpp.info, Node: Once-Only Headers, Next: Computed Includes, Prev: Search Path, Up: Header Files 831 8322.4 Once-Only Headers 833===================== 834 835If a header file happens to be included twice, the compiler will process 836its contents twice. This is very likely to cause an error, e.g. when 837the compiler sees the same structure definition twice. Even if it does 838not, it will certainly waste time. 839 840 The standard way to prevent this is to enclose the entire real 841contents of the file in a conditional, like this: 842 843 /* File foo. */ 844 #ifndef FILE_FOO_SEEN 845 #define FILE_FOO_SEEN 846 847 THE ENTIRE FILE 848 849 #endif /* !FILE_FOO_SEEN */ 850 851 This construct is commonly known as a "wrapper #ifndef". When the 852header is included again, the conditional will be false, because 853`FILE_FOO_SEEN' is defined. The preprocessor will skip over the entire 854contents of the file, and the compiler will not see it twice. 855 856 CPP optimizes even further. It remembers when a header file has a 857wrapper `#ifndef'. If a subsequent `#include' specifies that header, 858and the macro in the `#ifndef' is still defined, it does not bother to 859rescan the file at all. 860 861 You can put comments outside the wrapper. They will not interfere 862with this optimization. 863 864 The macro `FILE_FOO_SEEN' is called the "controlling macro" or 865"guard macro". In a user header file, the macro name should not begin 866with `_'. In a system header file, it should begin with `__' to avoid 867conflicts with user programs. In any kind of header file, the macro 868name should contain the name of the file and some additional text, to 869avoid conflicts with other header files. 870 871 872File: cpp.info, Node: Computed Includes, Next: Wrapper Headers, Prev: Once-Only Headers, Up: Header Files 873 8742.5 Computed Includes 875===================== 876 877Sometimes it is necessary to select one of several different header 878files to be included into your program. They might specify 879configuration parameters to be used on different sorts of operating 880systems, for instance. You could do this with a series of conditionals, 881 882 #if SYSTEM_1 883 # include "system_1.h" 884 #elif SYSTEM_2 885 # include "system_2.h" 886 #elif SYSTEM_3 887 ... 888 #endif 889 890 That rapidly becomes tedious. Instead, the preprocessor offers the 891ability to use a macro for the header name. This is called a "computed 892include". Instead of writing a header name as the direct argument of 893`#include', you simply put a macro name there instead: 894 895 #define SYSTEM_H "system_1.h" 896 ... 897 #include SYSTEM_H 898 899`SYSTEM_H' will be expanded, and the preprocessor will look for 900`system_1.h' as if the `#include' had been written that way originally. 901`SYSTEM_H' could be defined by your Makefile with a `-D' option. 902 903 You must be careful when you define the macro. `#define' saves 904tokens, not text. The preprocessor has no way of knowing that the macro 905will be used as the argument of `#include', so it generates ordinary 906tokens, not a header name. This is unlikely to cause problems if you 907use double-quote includes, which are close enough to string constants. 908If you use angle brackets, however, you may have trouble. 909 910 The syntax of a computed include is actually a bit more general than 911the above. If the first non-whitespace character after `#include' is 912not `"' or `<', then the entire line is macro-expanded like running 913text would be. 914 915 If the line expands to a single string constant, the contents of that 916string constant are the file to be included. CPP does not re-examine 917the string for embedded quotes, but neither does it process backslash 918escapes in the string. Therefore 919 920 #define HEADER "a\"b" 921 #include HEADER 922 923looks for a file named `a\"b'. CPP searches for the file according to 924the rules for double-quoted includes. 925 926 If the line expands to a token stream beginning with a `<' token and 927including a `>' token, then the tokens between the `<' and the first 928`>' are combined to form the filename to be included. Any whitespace 929between tokens is reduced to a single space; then any space after the 930initial `<' is retained, but a trailing space before the closing `>' is 931ignored. CPP searches for the file according to the rules for 932angle-bracket includes. 933 934 In either case, if there are any tokens on the line after the file 935name, an error occurs and the directive is not processed. It is also 936an error if the result of expansion does not match either of the two 937expected forms. 938 939 These rules are implementation-defined behavior according to the C 940standard. To minimize the risk of different compilers interpreting your 941computed includes differently, we recommend you use only a single 942object-like macro which expands to a string constant. This will also 943minimize confusion for people reading your program. 944 945 946File: cpp.info, Node: Wrapper Headers, Next: System Headers, Prev: Computed Includes, Up: Header Files 947 9482.6 Wrapper Headers 949=================== 950 951Sometimes it is necessary to adjust the contents of a system-provided 952header file without editing it directly. GCC's `fixincludes' operation 953does this, for example. One way to do that would be to create a new 954header file with the same name and insert it in the search path before 955the original header. That works fine as long as you're willing to 956replace the old header entirely. But what if you want to refer to the 957old header from the new one? 958 959 You cannot simply include the old header with `#include'. That will 960start from the beginning, and find your new header again. If your 961header is not protected from multiple inclusion (*note Once-Only 962Headers::), it will recurse infinitely and cause a fatal error. 963 964 You could include the old header with an absolute pathname: 965 #include "/usr/include/old-header.h" 966 This works, but is not clean; should the system headers ever move, 967you would have to edit the new headers to match. 968 969 There is no way to solve this problem within the C standard, but you 970can use the GNU extension `#include_next'. It means, "Include the 971_next_ file with this name". This directive works like `#include' 972except in searching for the specified file: it starts searching the 973list of header file directories _after_ the directory in which the 974current file was found. 975 976 Suppose you specify `-I /usr/local/include', and the list of 977directories to search also includes `/usr/include'; and suppose both 978directories contain `signal.h'. Ordinary `#include <signal.h>' finds 979the file under `/usr/local/include'. If that file contains 980`#include_next <signal.h>', it starts searching after that directory, 981and finds the file in `/usr/include'. 982 983 `#include_next' does not distinguish between `<FILE>' and `"FILE"' 984inclusion, nor does it check that the file you specify has the same 985name as the current file. It simply looks for the file named, starting 986with the directory in the search path after the one where the current 987file was found. 988 989 The use of `#include_next' can lead to great confusion. We 990recommend it be used only when there is no other alternative. In 991particular, it should not be used in the headers belonging to a specific 992program; it should be used only to make global corrections along the 993lines of `fixincludes'. 994 995 996File: cpp.info, Node: System Headers, Prev: Wrapper Headers, Up: Header Files 997 9982.7 System Headers 999================== 1000 1001The header files declaring interfaces to the operating system and 1002runtime libraries often cannot be written in strictly conforming C. 1003Therefore, GCC gives code found in "system headers" special treatment. 1004All warnings, other than those generated by `#warning' (*note 1005Diagnostics::), are suppressed while GCC is processing a system header. 1006Macros defined in a system header are immune to a few warnings 1007wherever they are expanded. This immunity is granted on an ad-hoc 1008basis, when we find that a warning generates lots of false positives 1009because of code in macros defined in system headers. 1010 1011 Normally, only the headers found in specific directories are 1012considered system headers. These directories are determined when GCC 1013is compiled. There are, however, two ways to make normal headers into 1014system headers. 1015 1016 The `-isystem' command line option adds its argument to the list of 1017directories to search for headers, just like `-I'. Any headers found 1018in that directory will be considered system headers. 1019 1020 All directories named by `-isystem' are searched _after_ all 1021directories named by `-I', no matter what their order was on the 1022command line. If the same directory is named by both `-I' and 1023`-isystem', the `-I' option is ignored. GCC provides an informative 1024message when this occurs if `-v' is used. 1025 1026 There is also a directive, `#pragma GCC system_header', which tells 1027GCC to consider the rest of the current include file a system header, 1028no matter where it was found. Code that comes before the `#pragma' in 1029the file will not be affected. `#pragma GCC system_header' has no 1030effect in the primary source file. 1031 1032 On very old systems, some of the pre-defined system header 1033directories get even more special treatment. GNU C++ considers code in 1034headers found in those directories to be surrounded by an `extern "C"' 1035block. There is no way to request this behavior with a `#pragma', or 1036from the command line. 1037 1038 1039File: cpp.info, Node: Macros, Next: Conditionals, Prev: Header Files, Up: Top 1040 10413 Macros 1042******** 1043 1044A "macro" is a fragment of code which has been given a name. Whenever 1045the name is used, it is replaced by the contents of the macro. There 1046are two kinds of macros. They differ mostly in what they look like 1047when they are used. "Object-like" macros resemble data objects when 1048used, "function-like" macros resemble function calls. 1049 1050 You may define any valid identifier as a macro, even if it is a C 1051keyword. The preprocessor does not know anything about keywords. This 1052can be useful if you wish to hide a keyword such as `const' from an 1053older compiler that does not understand it. However, the preprocessor 1054operator `defined' (*note Defined::) can never be defined as a macro, 1055and C++'s named operators (*note C++ Named Operators::) cannot be 1056macros when you are compiling C++. 1057 1058* Menu: 1059 1060* Object-like Macros:: 1061* Function-like Macros:: 1062* Macro Arguments:: 1063* Stringification:: 1064* Concatenation:: 1065* Variadic Macros:: 1066* Predefined Macros:: 1067* Undefining and Redefining Macros:: 1068* Directives Within Macro Arguments:: 1069* Macro Pitfalls:: 1070 1071 1072File: cpp.info, Node: Object-like Macros, Next: Function-like Macros, Up: Macros 1073 10743.1 Object-like Macros 1075====================== 1076 1077An "object-like macro" is a simple identifier which will be replaced by 1078a code fragment. It is called object-like because it looks like a data 1079object in code that uses it. They are most commonly used to give 1080symbolic names to numeric constants. 1081 1082 You create macros with the `#define' directive. `#define' is 1083followed by the name of the macro and then the token sequence it should 1084be an abbreviation for, which is variously referred to as the macro's 1085"body", "expansion" or "replacement list". For example, 1086 1087 #define BUFFER_SIZE 1024 1088 1089defines a macro named `BUFFER_SIZE' as an abbreviation for the token 1090`1024'. If somewhere after this `#define' directive there comes a C 1091statement of the form 1092 1093 foo = (char *) malloc (BUFFER_SIZE); 1094 1095then the C preprocessor will recognize and "expand" the macro 1096`BUFFER_SIZE'. The C compiler will see the same tokens as it would if 1097you had written 1098 1099 foo = (char *) malloc (1024); 1100 1101 By convention, macro names are written in uppercase. Programs are 1102easier to read when it is possible to tell at a glance which names are 1103macros. 1104 1105 The macro's body ends at the end of the `#define' line. You may 1106continue the definition onto multiple lines, if necessary, using 1107backslash-newline. When the macro is expanded, however, it will all 1108come out on one line. For example, 1109 1110 #define NUMBERS 1, \ 1111 2, \ 1112 3 1113 int x[] = { NUMBERS }; 1114 ==> int x[] = { 1, 2, 3 }; 1115 1116The most common visible consequence of this is surprising line numbers 1117in error messages. 1118 1119 There is no restriction on what can go in a macro body provided it 1120decomposes into valid preprocessing tokens. Parentheses need not 1121balance, and the body need not resemble valid C code. (If it does not, 1122you may get error messages from the C compiler when you use the macro.) 1123 1124 The C preprocessor scans your program sequentially. Macro 1125definitions take effect at the place you write them. Therefore, the 1126following input to the C preprocessor 1127 1128 foo = X; 1129 #define X 4 1130 bar = X; 1131 1132produces 1133 1134 foo = X; 1135 bar = 4; 1136 1137 When the preprocessor expands a macro name, the macro's expansion 1138replaces the macro invocation, then the expansion is examined for more 1139macros to expand. For example, 1140 1141 #define TABLESIZE BUFSIZE 1142 #define BUFSIZE 1024 1143 TABLESIZE 1144 ==> BUFSIZE 1145 ==> 1024 1146 1147`TABLESIZE' is expanded first to produce `BUFSIZE', then that macro is 1148expanded to produce the final result, `1024'. 1149 1150 Notice that `BUFSIZE' was not defined when `TABLESIZE' was defined. 1151The `#define' for `TABLESIZE' uses exactly the expansion you 1152specify--in this case, `BUFSIZE'--and does not check to see whether it 1153too contains macro names. Only when you _use_ `TABLESIZE' is the 1154result of its expansion scanned for more macro names. 1155 1156 This makes a difference if you change the definition of `BUFSIZE' at 1157some point in the source file. `TABLESIZE', defined as shown, will 1158always expand using the definition of `BUFSIZE' that is currently in 1159effect: 1160 1161 #define BUFSIZE 1020 1162 #define TABLESIZE BUFSIZE 1163 #undef BUFSIZE 1164 #define BUFSIZE 37 1165 1166Now `TABLESIZE' expands (in two stages) to `37'. 1167 1168 If the expansion of a macro contains its own name, either directly or 1169via intermediate macros, it is not expanded again when the expansion is 1170examined for more macros. This prevents infinite recursion. *Note 1171Self-Referential Macros::, for the precise details. 1172 1173 1174File: cpp.info, Node: Function-like Macros, Next: Macro Arguments, Prev: Object-like Macros, Up: Macros 1175 11763.2 Function-like Macros 1177======================== 1178 1179You can also define macros whose use looks like a function call. These 1180are called "function-like macros". To define a function-like macro, 1181you use the same `#define' directive, but you put a pair of parentheses 1182immediately after the macro name. For example, 1183 1184 #define lang_init() c_init() 1185 lang_init() 1186 ==> c_init() 1187 1188 A function-like macro is only expanded if its name appears with a 1189pair of parentheses after it. If you write just the name, it is left 1190alone. This can be useful when you have a function and a macro of the 1191same name, and you wish to use the function sometimes. 1192 1193 extern void foo(void); 1194 #define foo() /* optimized inline version */ 1195 ... 1196 foo(); 1197 funcptr = foo; 1198 1199 Here the call to `foo()' will use the macro, but the function 1200pointer will get the address of the real function. If the macro were to 1201be expanded, it would cause a syntax error. 1202 1203 If you put spaces between the macro name and the parentheses in the 1204macro definition, that does not define a function-like macro, it defines 1205an object-like macro whose expansion happens to begin with a pair of 1206parentheses. 1207 1208 #define lang_init () c_init() 1209 lang_init() 1210 ==> () c_init()() 1211 1212 The first two pairs of parentheses in this expansion come from the 1213macro. The third is the pair that was originally after the macro 1214invocation. Since `lang_init' is an object-like macro, it does not 1215consume those parentheses. 1216 1217 1218File: cpp.info, Node: Macro Arguments, Next: Stringification, Prev: Function-like Macros, Up: Macros 1219 12203.3 Macro Arguments 1221=================== 1222 1223Function-like macros can take "arguments", just like true functions. 1224To define a macro that uses arguments, you insert "parameters" between 1225the pair of parentheses in the macro definition that make the macro 1226function-like. The parameters must be valid C identifiers, separated 1227by commas and optionally whitespace. 1228 1229 To invoke a macro that takes arguments, you write the name of the 1230macro followed by a list of "actual arguments" in parentheses, separated 1231by commas. The invocation of the macro need not be restricted to a 1232single logical line--it can cross as many lines in the source file as 1233you wish. The number of arguments you give must match the number of 1234parameters in the macro definition. When the macro is expanded, each 1235use of a parameter in its body is replaced by the tokens of the 1236corresponding argument. (You need not use all of the parameters in the 1237macro body.) 1238 1239 As an example, here is a macro that computes the minimum of two 1240numeric values, as it is defined in many C programs, and some uses. 1241 1242 #define min(X, Y) ((X) < (Y) ? (X) : (Y)) 1243 x = min(a, b); ==> x = ((a) < (b) ? (a) : (b)); 1244 y = min(1, 2); ==> y = ((1) < (2) ? (1) : (2)); 1245 z = min(a + 28, *p); ==> z = ((a + 28) < (*p) ? (a + 28) : (*p)); 1246 1247(In this small example you can already see several of the dangers of 1248macro arguments. *Note Macro Pitfalls::, for detailed explanations.) 1249 1250 Leading and trailing whitespace in each argument is dropped, and all 1251whitespace between the tokens of an argument is reduced to a single 1252space. Parentheses within each argument must balance; a comma within 1253such parentheses does not end the argument. However, there is no 1254requirement for square brackets or braces to balance, and they do not 1255prevent a comma from separating arguments. Thus, 1256 1257 macro (array[x = y, x + 1]) 1258 1259passes two arguments to `macro': `array[x = y' and `x + 1]'. If you 1260want to supply `array[x = y, x + 1]' as an argument, you can write it 1261as `array[(x = y, x + 1)]', which is equivalent C code. 1262 1263 All arguments to a macro are completely macro-expanded before they 1264are substituted into the macro body. After substitution, the complete 1265text is scanned again for macros to expand, including the arguments. 1266This rule may seem strange, but it is carefully designed so you need 1267not worry about whether any function call is actually a macro 1268invocation. You can run into trouble if you try to be too clever, 1269though. *Note Argument Prescan::, for detailed discussion. 1270 1271 For example, `min (min (a, b), c)' is first expanded to 1272 1273 min (((a) < (b) ? (a) : (b)), (c)) 1274 1275and then to 1276 1277 ((((a) < (b) ? (a) : (b))) < (c) 1278 ? (((a) < (b) ? (a) : (b))) 1279 : (c)) 1280 1281(Line breaks shown here for clarity would not actually be generated.) 1282 1283 You can leave macro arguments empty; this is not an error to the 1284preprocessor (but many macros will then expand to invalid code). You 1285cannot leave out arguments entirely; if a macro takes two arguments, 1286there must be exactly one comma at the top level of its argument list. 1287Here are some silly examples using `min': 1288 1289 min(, b) ==> (( ) < (b) ? ( ) : (b)) 1290 min(a, ) ==> ((a ) < ( ) ? (a ) : ( )) 1291 min(,) ==> (( ) < ( ) ? ( ) : ( )) 1292 min((,),) ==> (((,)) < ( ) ? ((,)) : ( )) 1293 1294 min() error--> macro "min" requires 2 arguments, but only 1 given 1295 min(,,) error--> macro "min" passed 3 arguments, but takes just 2 1296 1297 Whitespace is not a preprocessing token, so if a macro `foo' takes 1298one argument, `foo ()' and `foo ( )' both supply it an empty argument. 1299Previous GNU preprocessor implementations and documentation were 1300incorrect on this point, insisting that a function-like macro that 1301takes a single argument be passed a space if an empty argument was 1302required. 1303 1304 Macro parameters appearing inside string literals are not replaced by 1305their corresponding actual arguments. 1306 1307 #define foo(x) x, "x" 1308 foo(bar) ==> bar, "x" 1309 1310 1311File: cpp.info, Node: Stringification, Next: Concatenation, Prev: Macro Arguments, Up: Macros 1312 13133.4 Stringification 1314=================== 1315 1316Sometimes you may want to convert a macro argument into a string 1317constant. Parameters are not replaced inside string constants, but you 1318can use the `#' preprocessing operator instead. When a macro parameter 1319is used with a leading `#', the preprocessor replaces it with the 1320literal text of the actual argument, converted to a string constant. 1321Unlike normal parameter replacement, the argument is not macro-expanded 1322first. This is called "stringification". 1323 1324 There is no way to combine an argument with surrounding text and 1325stringify it all together. Instead, you can write a series of adjacent 1326string constants and stringified arguments. The preprocessor will 1327replace the stringified arguments with string constants. The C 1328compiler will then combine all the adjacent string constants into one 1329long string. 1330 1331 Here is an example of a macro definition that uses stringification: 1332 1333 #define WARN_IF(EXP) \ 1334 do { if (EXP) \ 1335 fprintf (stderr, "Warning: " #EXP "\n"); } \ 1336 while (0) 1337 WARN_IF (x == 0); 1338 ==> do { if (x == 0) 1339 fprintf (stderr, "Warning: " "x == 0" "\n"); } while (0); 1340 1341The argument for `EXP' is substituted once, as-is, into the `if' 1342statement, and once, stringified, into the argument to `fprintf'. If 1343`x' were a macro, it would be expanded in the `if' statement, but not 1344in the string. 1345 1346 The `do' and `while (0)' are a kludge to make it possible to write 1347`WARN_IF (ARG);', which the resemblance of `WARN_IF' to a function 1348would make C programmers want to do; see *Note Swallowing the 1349Semicolon::. 1350 1351 Stringification in C involves more than putting double-quote 1352characters around the fragment. The preprocessor backslash-escapes the 1353quotes surrounding embedded string constants, and all backslashes 1354within string and character constants, in order to get a valid C string 1355constant with the proper contents. Thus, stringifying `p = "foo\n";' 1356results in "p = \"foo\\n\";". However, backslashes that are not inside 1357string or character constants are not duplicated: `\n' by itself 1358stringifies to "\n". 1359 1360 All leading and trailing whitespace in text being stringified is 1361ignored. Any sequence of whitespace in the middle of the text is 1362converted to a single space in the stringified result. Comments are 1363replaced by whitespace long before stringification happens, so they 1364never appear in stringified text. 1365 1366 There is no way to convert a macro argument into a character 1367constant. 1368 1369 If you want to stringify the result of expansion of a macro argument, 1370you have to use two levels of macros. 1371 1372 #define xstr(s) str(s) 1373 #define str(s) #s 1374 #define foo 4 1375 str (foo) 1376 ==> "foo" 1377 xstr (foo) 1378 ==> xstr (4) 1379 ==> str (4) 1380 ==> "4" 1381 1382 `s' is stringified when it is used in `str', so it is not 1383macro-expanded first. But `s' is an ordinary argument to `xstr', so it 1384is completely macro-expanded before `xstr' itself is expanded (*note 1385Argument Prescan::). Therefore, by the time `str' gets to its 1386argument, it has already been macro-expanded. 1387 1388 1389File: cpp.info, Node: Concatenation, Next: Variadic Macros, Prev: Stringification, Up: Macros 1390 13913.5 Concatenation 1392================= 1393 1394It is often useful to merge two tokens into one while expanding macros. 1395This is called "token pasting" or "token concatenation". The `##' 1396preprocessing operator performs token pasting. When a macro is 1397expanded, the two tokens on either side of each `##' operator are 1398combined into a single token, which then replaces the `##' and the two 1399original tokens in the macro expansion. Usually both will be 1400identifiers, or one will be an identifier and the other a preprocessing 1401number. When pasted, they make a longer identifier. This isn't the 1402only valid case. It is also possible to concatenate two numbers (or a 1403number and a name, such as `1.5' and `e3') into a number. Also, 1404multi-character operators such as `+=' can be formed by token pasting. 1405 1406 However, two tokens that don't together form a valid token cannot be 1407pasted together. For example, you cannot concatenate `x' with `+' in 1408either order. If you try, the preprocessor issues a warning and emits 1409the two tokens. Whether it puts white space between the tokens is 1410undefined. It is common to find unnecessary uses of `##' in complex 1411macros. If you get this warning, it is likely that you can simply 1412remove the `##'. 1413 1414 Both the tokens combined by `##' could come from the macro body, but 1415you could just as well write them as one token in the first place. 1416Token pasting is most useful when one or both of the tokens comes from a 1417macro argument. If either of the tokens next to an `##' is a parameter 1418name, it is replaced by its actual argument before `##' executes. As 1419with stringification, the actual argument is not macro-expanded first. 1420If the argument is empty, that `##' has no effect. 1421 1422 Keep in mind that the C preprocessor converts comments to whitespace 1423before macros are even considered. Therefore, you cannot create a 1424comment by concatenating `/' and `*'. You can put as much whitespace 1425between `##' and its operands as you like, including comments, and you 1426can put comments in arguments that will be concatenated. However, it 1427is an error if `##' appears at either end of a macro body. 1428 1429 Consider a C program that interprets named commands. There probably 1430needs to be a table of commands, perhaps an array of structures declared 1431as follows: 1432 1433 struct command 1434 { 1435 char *name; 1436 void (*function) (void); 1437 }; 1438 1439 struct command commands[] = 1440 { 1441 { "quit", quit_command }, 1442 { "help", help_command }, 1443 ... 1444 }; 1445 1446 It would be cleaner not to have to give each command name twice, 1447once in the string constant and once in the function name. A macro 1448which takes the name of a command as an argument can make this 1449unnecessary. The string constant can be created with stringification, 1450and the function name by concatenating the argument with `_command'. 1451Here is how it is done: 1452 1453 #define COMMAND(NAME) { #NAME, NAME ## _command } 1454 1455 struct command commands[] = 1456 { 1457 COMMAND (quit), 1458 COMMAND (help), 1459 ... 1460 }; 1461 1462 1463File: cpp.info, Node: Variadic Macros, Next: Predefined Macros, Prev: Concatenation, Up: Macros 1464 14653.6 Variadic Macros 1466=================== 1467 1468A macro can be declared to accept a variable number of arguments much as 1469a function can. The syntax for defining the macro is similar to that of 1470a function. Here is an example: 1471 1472 #define eprintf(...) fprintf (stderr, __VA_ARGS__) 1473 1474 This kind of macro is called "variadic". When the macro is invoked, 1475all the tokens in its argument list after the last named argument (this 1476macro has none), including any commas, become the "variable argument". 1477This sequence of tokens replaces the identifier `__VA_ARGS__' in the 1478macro body wherever it appears. Thus, we have this expansion: 1479 1480 eprintf ("%s:%d: ", input_file, lineno) 1481 ==> fprintf (stderr, "%s:%d: ", input_file, lineno) 1482 1483 The variable argument is completely macro-expanded before it is 1484inserted into the macro expansion, just like an ordinary argument. You 1485may use the `#' and `##' operators to stringify the variable argument 1486or to paste its leading or trailing token with another token. (But see 1487below for an important special case for `##'.) 1488 1489 If your macro is complicated, you may want a more descriptive name 1490for the variable argument than `__VA_ARGS__'. CPP permits this, as an 1491extension. You may write an argument name immediately before the 1492`...'; that name is used for the variable argument. The `eprintf' 1493macro above could be written 1494 1495 #define eprintf(args...) fprintf (stderr, args) 1496 1497using this extension. You cannot use `__VA_ARGS__' and this extension 1498in the same macro. 1499 1500 You can have named arguments as well as variable arguments in a 1501variadic macro. We could define `eprintf' like this, instead: 1502 1503 #define eprintf(format, ...) fprintf (stderr, format, __VA_ARGS__) 1504 1505This formulation looks more descriptive, but unfortunately it is less 1506flexible: you must now supply at least one argument after the format 1507string. In standard C, you cannot omit the comma separating the named 1508argument from the variable arguments. Furthermore, if you leave the 1509variable argument empty, you will get a syntax error, because there 1510will be an extra comma after the format string. 1511 1512 eprintf("success!\n", ); 1513 ==> fprintf(stderr, "success!\n", ); 1514 1515 GNU CPP has a pair of extensions which deal with this problem. 1516First, you are allowed to leave the variable argument out entirely: 1517 1518 eprintf ("success!\n") 1519 ==> fprintf(stderr, "success!\n", ); 1520 1521Second, the `##' token paste operator has a special meaning when placed 1522between a comma and a variable argument. If you write 1523 1524 #define eprintf(format, ...) fprintf (stderr, format, ##__VA_ARGS__) 1525 1526and the variable argument is left out when the `eprintf' macro is used, 1527then the comma before the `##' will be deleted. This does _not_ happen 1528if you pass an empty argument, nor does it happen if the token 1529preceding `##' is anything other than a comma. 1530 1531 eprintf ("success!\n") 1532 ==> fprintf(stderr, "success!\n"); 1533 1534The above explanation is ambiguous about the case where the only macro 1535parameter is a variable arguments parameter, as it is meaningless to 1536try to distinguish whether no argument at all is an empty argument or a 1537missing argument. In this case the C99 standard is clear that the 1538comma must remain, however the existing GCC extension used to swallow 1539the comma. So CPP retains the comma when conforming to a specific C 1540standard, and drops it otherwise. 1541 1542 C99 mandates that the only place the identifier `__VA_ARGS__' can 1543appear is in the replacement list of a variadic macro. It may not be 1544used as a macro name, macro argument name, or within a different type 1545of macro. It may also be forbidden in open text; the standard is 1546ambiguous. We recommend you avoid using it except for its defined 1547purpose. 1548 1549 Variadic macros are a new feature in C99. GNU CPP has supported them 1550for a long time, but only with a named variable argument (`args...', 1551not `...' and `__VA_ARGS__'). If you are concerned with portability to 1552previous versions of GCC, you should use only named variable arguments. 1553On the other hand, if you are concerned with portability to other 1554conforming implementations of C99, you should use only `__VA_ARGS__'. 1555 1556 Previous versions of CPP implemented the comma-deletion extension 1557much more generally. We have restricted it in this release to minimize 1558the differences from C99. To get the same effect with both this and 1559previous versions of GCC, the token preceding the special `##' must be 1560a comma, and there must be white space between that comma and whatever 1561comes immediately before it: 1562 1563 #define eprintf(format, args...) fprintf (stderr, format , ##args) 1564 1565*Note Differences from previous versions::, for the gory details. 1566 1567 1568File: cpp.info, Node: Predefined Macros, Next: Undefining and Redefining Macros, Prev: Variadic Macros, Up: Macros 1569 15703.7 Predefined Macros 1571===================== 1572 1573Several object-like macros are predefined; you use them without 1574supplying their definitions. They fall into three classes: standard, 1575common, and system-specific. 1576 1577 In C++, there is a fourth category, the named operators. They act 1578like predefined macros, but you cannot undefine them. 1579 1580* Menu: 1581 1582* Standard Predefined Macros:: 1583* Common Predefined Macros:: 1584* System-specific Predefined Macros:: 1585* C++ Named Operators:: 1586 1587 1588File: cpp.info, Node: Standard Predefined Macros, Next: Common Predefined Macros, Up: Predefined Macros 1589 15903.7.1 Standard Predefined Macros 1591-------------------------------- 1592 1593The standard predefined macros are specified by the relevant language 1594standards, so they are available with all compilers that implement 1595those standards. Older compilers may not provide all of them. Their 1596names all start with double underscores. 1597 1598`__FILE__' 1599 This macro expands to the name of the current input file, in the 1600 form of a C string constant. This is the path by which the 1601 preprocessor opened the file, not the short name specified in 1602 `#include' or as the input file name argument. For example, 1603 `"/usr/local/include/myheader.h"' is a possible expansion of this 1604 macro. 1605 1606`__LINE__' 1607 This macro expands to the current input line number, in the form 1608 of a decimal integer constant. While we call it a predefined 1609 macro, it's a pretty strange macro, since its "definition" changes 1610 with each new line of source code. 1611 1612 `__FILE__' and `__LINE__' are useful in generating an error message 1613to report an inconsistency detected by the program; the message can 1614state the source line at which the inconsistency was detected. For 1615example, 1616 1617 fprintf (stderr, "Internal error: " 1618 "negative string length " 1619 "%d at %s, line %d.", 1620 length, __FILE__, __LINE__); 1621 1622 An `#include' directive changes the expansions of `__FILE__' and 1623`__LINE__' to correspond to the included file. At the end of that 1624file, when processing resumes on the input file that contained the 1625`#include' directive, the expansions of `__FILE__' and `__LINE__' 1626revert to the values they had before the `#include' (but `__LINE__' is 1627then incremented by one as processing moves to the line after the 1628`#include'). 1629 1630 A `#line' directive changes `__LINE__', and may change `__FILE__' as 1631well. *Note Line Control::. 1632 1633 C99 introduces `__func__', and GCC has provided `__FUNCTION__' for a 1634long time. Both of these are strings containing the name of the 1635current function (there are slight semantic differences; see the GCC 1636manual). Neither of them is a macro; the preprocessor does not know the 1637name of the current function. They tend to be useful in conjunction 1638with `__FILE__' and `__LINE__', though. 1639 1640`__DATE__' 1641 This macro expands to a string constant that describes the date on 1642 which the preprocessor is being run. The string constant contains 1643 eleven characters and looks like `"Feb 12 1996"'. If the day of 1644 the month is less than 10, it is padded with a space on the left. 1645 1646 If GCC cannot determine the current date, it will emit a warning 1647 message (once per compilation) and `__DATE__' will expand to 1648 `"??? ?? ????"'. 1649 1650`__TIME__' 1651 This macro expands to a string constant that describes the time at 1652 which the preprocessor is being run. The string constant contains 1653 eight characters and looks like `"23:59:01"'. 1654 1655 If GCC cannot determine the current time, it will emit a warning 1656 message (once per compilation) and `__TIME__' will expand to 1657 `"??:??:??"'. 1658 1659`__STDC__' 1660 In normal operation, this macro expands to the constant 1, to 1661 signify that this compiler conforms to ISO Standard C. If GNU CPP 1662 is used with a compiler other than GCC, this is not necessarily 1663 true; however, the preprocessor always conforms to the standard 1664 unless the `-traditional-cpp' option is used. 1665 1666 This macro is not defined if the `-traditional-cpp' option is used. 1667 1668 On some hosts, the system compiler uses a different convention, 1669 where `__STDC__' is normally 0, but is 1 if the user specifies 1670 strict conformance to the C Standard. CPP follows the host 1671 convention when processing system header files, but when 1672 processing user files `__STDC__' is always 1. This has been 1673 reported to cause problems; for instance, some versions of Solaris 1674 provide X Windows headers that expect `__STDC__' to be either 1675 undefined or 1. *Note Invocation::. 1676 1677`__STDC_VERSION__' 1678 This macro expands to the C Standard's version number, a long 1679 integer constant of the form `YYYYMML' where YYYY and MM are the 1680 year and month of the Standard version. This signifies which 1681 version of the C Standard the compiler conforms to. Like 1682 `__STDC__', this is not necessarily accurate for the entire 1683 implementation, unless GNU CPP is being used with GCC. 1684 1685 The value `199409L' signifies the 1989 C standard as amended in 1686 1994, which is the current default; the value `199901L' signifies 1687 the 1999 revision of the C standard. Support for the 1999 1688 revision is not yet complete. 1689 1690 This macro is not defined if the `-traditional-cpp' option is 1691 used, nor when compiling C++ or Objective-C. 1692 1693`__STDC_HOSTED__' 1694 This macro is defined, with value 1, if the compiler's target is a 1695 "hosted environment". A hosted environment has the complete 1696 facilities of the standard C library available. 1697 1698`__cplusplus' 1699 This macro is defined when the C++ compiler is in use. You can use 1700 `__cplusplus' to test whether a header is compiled by a C compiler 1701 or a C++ compiler. This macro is similar to `__STDC_VERSION__', in 1702 that it expands to a version number. A fully conforming 1703 implementation of the 1998 C++ standard will define this macro to 1704 `199711L'. The GNU C++ compiler is not yet fully conforming, so 1705 it uses `1' instead. It is hoped to complete the implementation 1706 of standard C++ in the near future. 1707 1708`__OBJC__' 1709 This macro is defined, with value 1, when the Objective-C compiler 1710 is in use. You can use `__OBJC__' to test whether a header is 1711 compiled by a C compiler or a Objective-C compiler. 1712 1713`__ASSEMBLER__' 1714 This macro is defined with value 1 when preprocessing assembly 1715 language. 1716 1717 1718 1719File: cpp.info, Node: Common Predefined Macros, Next: System-specific Predefined Macros, Prev: Standard Predefined Macros, Up: Predefined Macros 1720 17213.7.2 Common Predefined Macros 1722------------------------------ 1723 1724The common predefined macros are GNU C extensions. They are available 1725with the same meanings regardless of the machine or operating system on 1726which you are using GNU C. Their names all start with double 1727underscores. 1728 1729`__GNUC__' 1730`__GNUC_MINOR__' 1731`__GNUC_PATCHLEVEL__' 1732 These macros are defined by all GNU compilers that use the C 1733 preprocessor: C, C++, and Objective-C. Their values are the major 1734 version, minor version, and patch level of the compiler, as integer 1735 constants. For example, GCC 3.2.1 will define `__GNUC__' to 3, 1736 `__GNUC_MINOR__' to 2, and `__GNUC_PATCHLEVEL__' to 1. These 1737 macros are also defined if you invoke the preprocessor directly. 1738 1739 `__GNUC_PATCHLEVEL__' is new to GCC 3.0; it is also present in the 1740 widely-used development snapshots leading up to 3.0 (which identify 1741 themselves as GCC 2.96 or 2.97, depending on which snapshot you 1742 have). 1743 1744 If all you need to know is whether or not your program is being 1745 compiled by GCC, or a non-GCC compiler that claims to accept the 1746 GNU C dialects, you can simply test `__GNUC__'. If you need to 1747 write code which depends on a specific version, you must be more 1748 careful. Each time the minor version is increased, the patch 1749 level is reset to zero; each time the major version is increased 1750 (which happens rarely), the minor version and patch level are 1751 reset. If you wish to use the predefined macros directly in the 1752 conditional, you will need to write it like this: 1753 1754 /* Test for GCC > 3.2.0 */ 1755 #if __GNUC__ > 3 || \ 1756 (__GNUC__ == 3 && (__GNUC_MINOR__ > 2 || \ 1757 (__GNUC_MINOR__ == 2 && \ 1758 __GNUC_PATCHLEVEL__ > 0)) 1759 1760 Another approach is to use the predefined macros to calculate a 1761 single number, then compare that against a threshold: 1762 1763 #define GCC_VERSION (__GNUC__ * 10000 \ 1764 + __GNUC_MINOR__ * 100 \ 1765 + __GNUC_PATCHLEVEL__) 1766 ... 1767 /* Test for GCC > 3.2.0 */ 1768 #if GCC_VERSION > 30200 1769 1770 Many people find this form easier to understand. 1771 1772`__GNUG__' 1773 The GNU C++ compiler defines this. Testing it is equivalent to 1774 testing `(__GNUC__ && __cplusplus)'. 1775 1776`__STRICT_ANSI__' 1777 GCC defines this macro if and only if the `-ansi' switch, or a 1778 `-std' switch specifying strict conformance to some version of ISO 1779 C, was specified when GCC was invoked. It is defined to `1'. 1780 This macro exists primarily to direct GNU libc's header files to 1781 restrict their definitions to the minimal set found in the 1989 C 1782 standard. 1783 1784`__BASE_FILE__' 1785 This macro expands to the name of the main input file, in the form 1786 of a C string constant. This is the source file that was specified 1787 on the command line of the preprocessor or C compiler. 1788 1789`__INCLUDE_LEVEL__' 1790 This macro expands to a decimal integer constant that represents 1791 the depth of nesting in include files. The value of this macro is 1792 incremented on every `#include' directive and decremented at the 1793 end of every included file. It starts out at 0, it's value within 1794 the base file specified on the command line. 1795 1796`__ELF__' 1797 This macro is defined if the target uses the ELF object format. 1798 1799`__VERSION__' 1800 This macro expands to a string constant which describes the 1801 version of the compiler in use. You should not rely on its 1802 contents having any particular form, but it can be counted on to 1803 contain at least the release number. 1804 1805`__OPTIMIZE__' 1806`__OPTIMIZE_SIZE__' 1807`__NO_INLINE__' 1808 These macros describe the compilation mode. `__OPTIMIZE__' is 1809 defined in all optimizing compilations. `__OPTIMIZE_SIZE__' is 1810 defined if the compiler is optimizing for size, not speed. 1811 `__NO_INLINE__' is defined if no functions will be inlined into 1812 their callers (when not optimizing, or when inlining has been 1813 specifically disabled by `-fno-inline'). 1814 1815 These macros cause certain GNU header files to provide optimized 1816 definitions, using macros or inline functions, of system library 1817 functions. You should not use these macros in any way unless you 1818 make sure that programs will execute with the same effect whether 1819 or not they are defined. If they are defined, their value is 1. 1820 1821`__GNUC_GNU_INLINE__' 1822 GCC defines this macro if functions declared `inline' will be 1823 handled in GCC's traditional gnu89 mode. In this mode an `extern 1824 inline' function will never be compiled as a standalone function, 1825 and an `inline' function which is neither `extern' nor `static' 1826 will always be compiled as a standalone function. 1827 1828`__GNUC_STDC_INLINE__' 1829 GCC defines this macro if functions declared `inline' will be 1830 handled according to the ISO C99 standard. In this mode an 1831 `extern inline' function will always be compiled as a standalone 1832 externally visible function, and an `inline' function which is 1833 neither `extern' nor `static' will never be compiled as a 1834 standalone function. 1835 1836 If this macro is defined, GCC supports the `gnu_inline' function 1837 attribute as a way to always get the gnu89 behaviour. Support for 1838 this and `__GNUC_GNU_INLINE__' was added in GCC 4.1.3. If neither 1839 macro is defined, an older version of GCC is being used: `inline' 1840 functions will be compiled in gnu89 mode, and the `gnu_inline' 1841 function attribute will not be recognized. 1842 1843`__CHAR_UNSIGNED__' 1844 GCC defines this macro if and only if the data type `char' is 1845 unsigned on the target machine. It exists to cause the standard 1846 header file `limits.h' to work correctly. You should not use this 1847 macro yourself; instead, refer to the standard macros defined in 1848 `limits.h'. 1849 1850`__WCHAR_UNSIGNED__' 1851 Like `__CHAR_UNSIGNED__', this macro is defined if and only if the 1852 data type `wchar_t' is unsigned and the front-end is in C++ mode. 1853 1854`__REGISTER_PREFIX__' 1855 This macro expands to a single token (not a string constant) which 1856 is the prefix applied to CPU register names in assembly language 1857 for this target. You can use it to write assembly that is usable 1858 in multiple environments. For example, in the `m68k-aout' 1859 environment it expands to nothing, but in the `m68k-coff' 1860 environment it expands to a single `%'. 1861 1862`__USER_LABEL_PREFIX__' 1863 This macro expands to a single token which is the prefix applied to 1864 user labels (symbols visible to C code) in assembly. For example, 1865 in the `m68k-aout' environment it expands to an `_', but in the 1866 `m68k-coff' environment it expands to nothing. 1867 1868 This macro will have the correct definition even if 1869 `-f(no-)underscores' is in use, but it will not be correct if 1870 target-specific options that adjust this prefix are used (e.g. the 1871 OSF/rose `-mno-underscores' option). 1872 1873`__SIZE_TYPE__' 1874`__PTRDIFF_TYPE__' 1875`__WCHAR_TYPE__' 1876`__WINT_TYPE__' 1877`__INTMAX_TYPE__' 1878`__UINTMAX_TYPE__' 1879 These macros are defined to the correct underlying types for the 1880 `size_t', `ptrdiff_t', `wchar_t', `wint_t', `intmax_t', and 1881 `uintmax_t' typedefs, respectively. They exist to make the 1882 standard header files `stddef.h' and `wchar.h' work correctly. 1883 You should not use these macros directly; instead, include the 1884 appropriate headers and use the typedefs. 1885 1886`__CHAR_BIT__' 1887 Defined to the number of bits used in the representation of the 1888 `char' data type. It exists to make the standard header given 1889 numerical limits work correctly. You should not use this macro 1890 directly; instead, include the appropriate headers. 1891 1892`__SCHAR_MAX__' 1893`__WCHAR_MAX__' 1894`__SHRT_MAX__' 1895`__INT_MAX__' 1896`__LONG_MAX__' 1897`__LONG_LONG_MAX__' 1898`__INTMAX_MAX__' 1899 Defined to the maximum value of the `signed char', `wchar_t', 1900 `signed short', `signed int', `signed long', `signed long long', 1901 and `intmax_t' types respectively. They exist to make the 1902 standard header given numerical limits work correctly. You should 1903 not use these macros directly; instead, include the appropriate 1904 headers. 1905 1906`__DEPRECATED' 1907 This macro is defined, with value 1, when compiling a C++ source 1908 file with warnings about deprecated constructs enabled. These 1909 warnings are enabled by default, but can be disabled with 1910 `-Wno-deprecated'. 1911 1912`__EXCEPTIONS' 1913 This macro is defined, with value 1, when compiling a C++ source 1914 file with exceptions enabled. If `-fno-exceptions' was used when 1915 compiling the file, then this macro will not be defined. 1916 1917`__USING_SJLJ_EXCEPTIONS__' 1918 This macro is defined, with value 1, if the compiler uses the old 1919 mechanism based on `setjmp' and `longjmp' for exception handling. 1920 1921`__GXX_WEAK__' 1922 This macro is defined when compiling a C++ source file. It has the 1923 value 1 if the compiler will use weak symbols, COMDAT sections, or 1924 other similar techniques to collapse symbols with "vague linkage" 1925 that are defined in multiple translation units. If the compiler 1926 will not collapse such symbols, this macro is defined with value 1927 0. In general, user code should not need to make use of this 1928 macro; the purpose of this macro is to ease implementation of the 1929 C++ runtime library provided with G++. 1930 1931`__NEXT_RUNTIME__' 1932 This macro is defined, with value 1, if (and only if) the NeXT 1933 runtime (as in `-fnext-runtime') is in use for Objective-C. If 1934 the GNU runtime is used, this macro is not defined, so that you 1935 can use this macro to determine which runtime (NeXT or GNU) is 1936 being used. 1937 1938`__LP64__' 1939`_LP64' 1940 These macros are defined, with value 1, if (and only if) the 1941 compilation is for a target where `long int' and pointer both use 1942 64-bits and `int' uses 32-bit. 1943 1944`__SSP__' 1945 This macro is defined, with value 1, when `-fstack-protector' is in 1946 use. 1947 1948`__SSP_ALL__' 1949 This macro is defined, with value 2, when `-fstack-protector-all' 1950 is in use. 1951 1952`__TIMESTAMP__' 1953 This macro expands to a string constant that describes the date 1954 and time of the last modification of the current source file. The 1955 string constant contains abbreviated day of the week, month, day 1956 of the month, time in hh:mm:ss form, year and looks like 1957 `"Sun Sep 16 01:03:52 1973"'. If the day of the month is less 1958 than 10, it is padded with a space on the left. 1959 1960 If GCC cannot determine the current date, it will emit a warning 1961 message (once per compilation) and `__TIMESTAMP__' will expand to 1962 `"??? ??? ?? ??:??:?? ????"'. 1963 1964 1965 1966File: cpp.info, Node: System-specific Predefined Macros, Next: C++ Named Operators, Prev: Common Predefined Macros, Up: Predefined Macros 1967 19683.7.3 System-specific Predefined Macros 1969--------------------------------------- 1970 1971The C preprocessor normally predefines several macros that indicate what 1972type of system and machine is in use. They are obviously different on 1973each target supported by GCC. This manual, being for all systems and 1974machines, cannot tell you what their names are, but you can use `cpp 1975-dM' to see them all. *Note Invocation::. All system-specific 1976predefined macros expand to the constant 1, so you can test them with 1977either `#ifdef' or `#if'. 1978 1979 The C standard requires that all system-specific macros be part of 1980the "reserved namespace". All names which begin with two underscores, 1981or an underscore and a capital letter, are reserved for the compiler and 1982library to use as they wish. However, historically system-specific 1983macros have had names with no special prefix; for instance, it is common 1984to find `unix' defined on Unix systems. For all such macros, GCC 1985provides a parallel macro with two underscores added at the beginning 1986and the end. If `unix' is defined, `__unix__' will be defined too. 1987There will never be more than two underscores; the parallel of `_mips' 1988is `__mips__'. 1989 1990 When the `-ansi' option, or any `-std' option that requests strict 1991conformance, is given to the compiler, all the system-specific 1992predefined macros outside the reserved namespace are suppressed. The 1993parallel macros, inside the reserved namespace, remain defined. 1994 1995 We are slowly phasing out all predefined macros which are outside the 1996reserved namespace. You should never use them in new programs, and we 1997encourage you to correct older code to use the parallel macros whenever 1998you find it. We don't recommend you use the system-specific macros that 1999are in the reserved namespace, either. It is better in the long run to 2000check specifically for features you need, using a tool such as 2001`autoconf'. 2002 2003 2004File: cpp.info, Node: C++ Named Operators, Prev: System-specific Predefined Macros, Up: Predefined Macros 2005 20063.7.4 C++ Named Operators 2007------------------------- 2008 2009In C++, there are eleven keywords which are simply alternate spellings 2010of operators normally written with punctuation. These keywords are 2011treated as such even in the preprocessor. They function as operators in 2012`#if', and they cannot be defined as macros or poisoned. In C, you can 2013request that those keywords take their C++ meaning by including 2014`iso646.h'. That header defines each one as a normal object-like macro 2015expanding to the appropriate punctuator. 2016 2017 These are the named operators and their corresponding punctuators: 2018 2019Named Operator Punctuator 2020`and' `&&' 2021`and_eq' `&=' 2022`bitand' `&' 2023`bitor' `|' 2024`compl' `~' 2025`not' `!' 2026`not_eq' `!=' 2027`or' `||' 2028`or_eq' `|=' 2029`xor' `^' 2030`xor_eq' `^=' 2031 2032 2033File: cpp.info, Node: Undefining and Redefining Macros, Next: Directives Within Macro Arguments, Prev: Predefined Macros, Up: Macros 2034 20353.8 Undefining and Redefining Macros 2036==================================== 2037 2038If a macro ceases to be useful, it may be "undefined" with the `#undef' 2039directive. `#undef' takes a single argument, the name of the macro to 2040undefine. You use the bare macro name, even if the macro is 2041function-like. It is an error if anything appears on the line after 2042the macro name. `#undef' has no effect if the name is not a macro. 2043 2044 #define FOO 4 2045 x = FOO; ==> x = 4; 2046 #undef FOO 2047 x = FOO; ==> x = FOO; 2048 2049 Once a macro has been undefined, that identifier may be "redefined" 2050as a macro by a subsequent `#define' directive. The new definition 2051need not have any resemblance to the old definition. 2052 2053 However, if an identifier which is currently a macro is redefined, 2054then the new definition must be "effectively the same" as the old one. 2055Two macro definitions are effectively the same if: 2056 * Both are the same type of macro (object- or function-like). 2057 2058 * All the tokens of the replacement list are the same. 2059 2060 * If there are any parameters, they are the same. 2061 2062 * Whitespace appears in the same places in both. It need not be 2063 exactly the same amount of whitespace, though. Remember that 2064 comments count as whitespace. 2065 2066These definitions are effectively the same: 2067 #define FOUR (2 + 2) 2068 #define FOUR (2 + 2) 2069 #define FOUR (2 /* two */ + 2) 2070 but these are not: 2071 #define FOUR (2 + 2) 2072 #define FOUR ( 2+2 ) 2073 #define FOUR (2 * 2) 2074 #define FOUR(score,and,seven,years,ago) (2 + 2) 2075 2076 If a macro is redefined with a definition that is not effectively the 2077same as the old one, the preprocessor issues a warning and changes the 2078macro to use the new definition. If the new definition is effectively 2079the same, the redefinition is silently ignored. This allows, for 2080instance, two different headers to define a common macro. The 2081preprocessor will only complain if the definitions do not match. 2082 2083 2084File: cpp.info, Node: Directives Within Macro Arguments, Next: Macro Pitfalls, Prev: Undefining and Redefining Macros, Up: Macros 2085 20863.9 Directives Within Macro Arguments 2087===================================== 2088 2089Occasionally it is convenient to use preprocessor directives within the 2090arguments of a macro. The C and C++ standards declare that behavior in 2091these cases is undefined. 2092 2093 Versions of CPP prior to 3.2 would reject such constructs with an 2094error message. This was the only syntactic difference between normal 2095functions and function-like macros, so it seemed attractive to remove 2096this limitation, and people would often be surprised that they could 2097not use macros in this way. Moreover, sometimes people would use 2098conditional compilation in the argument list to a normal library 2099function like `printf', only to find that after a library upgrade 2100`printf' had changed to be a function-like macro, and their code would 2101no longer compile. So from version 3.2 we changed CPP to successfully 2102process arbitrary directives within macro arguments in exactly the same 2103way as it would have processed the directive were the function-like 2104macro invocation not present. 2105 2106 If, within a macro invocation, that macro is redefined, then the new 2107definition takes effect in time for argument pre-expansion, but the 2108original definition is still used for argument replacement. Here is a 2109pathological example: 2110 2111 #define f(x) x x 2112 f (1 2113 #undef f 2114 #define f 2 2115 f) 2116 2117which expands to 2118 2119 1 2 1 2 2120 2121with the semantics described above. 2122 2123 2124File: cpp.info, Node: Macro Pitfalls, Prev: Directives Within Macro Arguments, Up: Macros 2125 21263.10 Macro Pitfalls 2127=================== 2128 2129In this section we describe some special rules that apply to macros and 2130macro expansion, and point out certain cases in which the rules have 2131counter-intuitive consequences that you must watch out for. 2132 2133* Menu: 2134 2135* Misnesting:: 2136* Operator Precedence Problems:: 2137* Swallowing the Semicolon:: 2138* Duplication of Side Effects:: 2139* Self-Referential Macros:: 2140* Argument Prescan:: 2141* Newlines in Arguments:: 2142 2143 2144File: cpp.info, Node: Misnesting, Next: Operator Precedence Problems, Up: Macro Pitfalls 2145 21463.10.1 Misnesting 2147----------------- 2148 2149When a macro is called with arguments, the arguments are substituted 2150into the macro body and the result is checked, together with the rest of 2151the input file, for more macro calls. It is possible to piece together 2152a macro call coming partially from the macro body and partially from the 2153arguments. For example, 2154 2155 #define twice(x) (2*(x)) 2156 #define call_with_1(x) x(1) 2157 call_with_1 (twice) 2158 ==> twice(1) 2159 ==> (2*(1)) 2160 2161 Macro definitions do not have to have balanced parentheses. By 2162writing an unbalanced open parenthesis in a macro body, it is possible 2163to create a macro call that begins inside the macro body but ends 2164outside of it. For example, 2165 2166 #define strange(file) fprintf (file, "%s %d", 2167 ... 2168 strange(stderr) p, 35) 2169 ==> fprintf (stderr, "%s %d", p, 35) 2170 2171 The ability to piece together a macro call can be useful, but the 2172use of unbalanced open parentheses in a macro body is just confusing, 2173and should be avoided. 2174 2175 2176File: cpp.info, Node: Operator Precedence Problems, Next: Swallowing the Semicolon, Prev: Misnesting, Up: Macro Pitfalls 2177 21783.10.2 Operator Precedence Problems 2179----------------------------------- 2180 2181You may have noticed that in most of the macro definition examples shown 2182above, each occurrence of a macro argument name had parentheses around 2183it. In addition, another pair of parentheses usually surround the 2184entire macro definition. Here is why it is best to write macros that 2185way. 2186 2187 Suppose you define a macro as follows, 2188 2189 #define ceil_div(x, y) (x + y - 1) / y 2190 2191whose purpose is to divide, rounding up. (One use for this operation is 2192to compute how many `int' objects are needed to hold a certain number 2193of `char' objects.) Then suppose it is used as follows: 2194 2195 a = ceil_div (b & c, sizeof (int)); 2196 ==> a = (b & c + sizeof (int) - 1) / sizeof (int); 2197 2198This does not do what is intended. The operator-precedence rules of C 2199make it equivalent to this: 2200 2201 a = (b & (c + sizeof (int) - 1)) / sizeof (int); 2202 2203What we want is this: 2204 2205 a = ((b & c) + sizeof (int) - 1)) / sizeof (int); 2206 2207Defining the macro as 2208 2209 #define ceil_div(x, y) ((x) + (y) - 1) / (y) 2210 2211provides the desired result. 2212 2213 Unintended grouping can result in another way. Consider `sizeof 2214ceil_div(1, 2)'. That has the appearance of a C expression that would 2215compute the size of the type of `ceil_div (1, 2)', but in fact it means 2216something very different. Here is what it expands to: 2217 2218 sizeof ((1) + (2) - 1) / (2) 2219 2220This would take the size of an integer and divide it by two. The 2221precedence rules have put the division outside the `sizeof' when it was 2222intended to be inside. 2223 2224 Parentheses around the entire macro definition prevent such problems. 2225Here, then, is the recommended way to define `ceil_div': 2226 2227 #define ceil_div(x, y) (((x) + (y) - 1) / (y)) 2228 2229 2230File: cpp.info, Node: Swallowing the Semicolon, Next: Duplication of Side Effects, Prev: Operator Precedence Problems, Up: Macro Pitfalls 2231 22323.10.3 Swallowing the Semicolon 2233------------------------------- 2234 2235Often it is desirable to define a macro that expands into a compound 2236statement. Consider, for example, the following macro, that advances a 2237pointer (the argument `p' says where to find it) across whitespace 2238characters: 2239 2240 #define SKIP_SPACES(p, limit) \ 2241 { char *lim = (limit); \ 2242 while (p < lim) { \ 2243 if (*p++ != ' ') { \ 2244 p--; break; }}} 2245 2246Here backslash-newline is used to split the macro definition, which must 2247be a single logical line, so that it resembles the way such code would 2248be laid out if not part of a macro definition. 2249 2250 A call to this macro might be `SKIP_SPACES (p, lim)'. Strictly 2251speaking, the call expands to a compound statement, which is a complete 2252statement with no need for a semicolon to end it. However, since it 2253looks like a function call, it minimizes confusion if you can use it 2254like a function call, writing a semicolon afterward, as in `SKIP_SPACES 2255(p, lim);' 2256 2257 This can cause trouble before `else' statements, because the 2258semicolon is actually a null statement. Suppose you write 2259 2260 if (*p != 0) 2261 SKIP_SPACES (p, lim); 2262 else ... 2263 2264The presence of two statements--the compound statement and a null 2265statement--in between the `if' condition and the `else' makes invalid C 2266code. 2267 2268 The definition of the macro `SKIP_SPACES' can be altered to solve 2269this problem, using a `do ... while' statement. Here is how: 2270 2271 #define SKIP_SPACES(p, limit) \ 2272 do { char *lim = (limit); \ 2273 while (p < lim) { \ 2274 if (*p++ != ' ') { \ 2275 p--; break; }}} \ 2276 while (0) 2277 2278 Now `SKIP_SPACES (p, lim);' expands into 2279 2280 do {...} while (0); 2281 2282which is one statement. The loop executes exactly once; most compilers 2283generate no extra code for it. 2284 2285 2286File: cpp.info, Node: Duplication of Side Effects, Next: Self-Referential Macros, Prev: Swallowing the Semicolon, Up: Macro Pitfalls 2287 22883.10.4 Duplication of Side Effects 2289---------------------------------- 2290 2291Many C programs define a macro `min', for "minimum", like this: 2292 2293 #define min(X, Y) ((X) < (Y) ? (X) : (Y)) 2294 2295 When you use this macro with an argument containing a side effect, 2296as shown here, 2297 2298 next = min (x + y, foo (z)); 2299 2300it expands as follows: 2301 2302 next = ((x + y) < (foo (z)) ? (x + y) : (foo (z))); 2303 2304where `x + y' has been substituted for `X' and `foo (z)' for `Y'. 2305 2306 The function `foo' is used only once in the statement as it appears 2307in the program, but the expression `foo (z)' has been substituted twice 2308into the macro expansion. As a result, `foo' might be called two times 2309when the statement is executed. If it has side effects or if it takes 2310a long time to compute, the results might not be what you intended. We 2311say that `min' is an "unsafe" macro. 2312 2313 The best solution to this problem is to define `min' in a way that 2314computes the value of `foo (z)' only once. The C language offers no 2315standard way to do this, but it can be done with GNU extensions as 2316follows: 2317 2318 #define min(X, Y) \ 2319 ({ typeof (X) x_ = (X); \ 2320 typeof (Y) y_ = (Y); \ 2321 (x_ < y_) ? x_ : y_; }) 2322 2323 The `({ ... })' notation produces a compound statement that acts as 2324an expression. Its value is the value of its last statement. This 2325permits us to define local variables and assign each argument to one. 2326The local variables have underscores after their names to reduce the 2327risk of conflict with an identifier of wider scope (it is impossible to 2328avoid this entirely). Now each argument is evaluated exactly once. 2329 2330 If you do not wish to use GNU C extensions, the only solution is to 2331be careful when _using_ the macro `min'. For example, you can 2332calculate the value of `foo (z)', save it in a variable, and use that 2333variable in `min': 2334 2335 #define min(X, Y) ((X) < (Y) ? (X) : (Y)) 2336 ... 2337 { 2338 int tem = foo (z); 2339 next = min (x + y, tem); 2340 } 2341 2342(where we assume that `foo' returns type `int'). 2343 2344 2345File: cpp.info, Node: Self-Referential Macros, Next: Argument Prescan, Prev: Duplication of Side Effects, Up: Macro Pitfalls 2346 23473.10.5 Self-Referential Macros 2348------------------------------ 2349 2350A "self-referential" macro is one whose name appears in its definition. 2351Recall that all macro definitions are rescanned for more macros to 2352replace. If the self-reference were considered a use of the macro, it 2353would produce an infinitely large expansion. To prevent this, the 2354self-reference is not considered a macro call. It is passed into the 2355preprocessor output unchanged. Consider an example: 2356 2357 #define foo (4 + foo) 2358 2359where `foo' is also a variable in your program. 2360 2361 Following the ordinary rules, each reference to `foo' will expand 2362into `(4 + foo)'; then this will be rescanned and will expand into `(4 2363+ (4 + foo))'; and so on until the computer runs out of memory. 2364 2365 The self-reference rule cuts this process short after one step, at 2366`(4 + foo)'. Therefore, this macro definition has the possibly useful 2367effect of causing the program to add 4 to the value of `foo' wherever 2368`foo' is referred to. 2369 2370 In most cases, it is a bad idea to take advantage of this feature. A 2371person reading the program who sees that `foo' is a variable will not 2372expect that it is a macro as well. The reader will come across the 2373identifier `foo' in the program and think its value should be that of 2374the variable `foo', whereas in fact the value is four greater. 2375 2376 One common, useful use of self-reference is to create a macro which 2377expands to itself. If you write 2378 2379 #define EPERM EPERM 2380 2381then the macro `EPERM' expands to `EPERM'. Effectively, it is left 2382alone by the preprocessor whenever it's used in running text. You can 2383tell that it's a macro with `#ifdef'. You might do this if you want to 2384define numeric constants with an `enum', but have `#ifdef' be true for 2385each constant. 2386 2387 If a macro `x' expands to use a macro `y', and the expansion of `y' 2388refers to the macro `x', that is an "indirect self-reference" of `x'. 2389`x' is not expanded in this case either. Thus, if we have 2390 2391 #define x (4 + y) 2392 #define y (2 * x) 2393 2394then `x' and `y' expand as follows: 2395 2396 x ==> (4 + y) 2397 ==> (4 + (2 * x)) 2398 2399 y ==> (2 * x) 2400 ==> (2 * (4 + y)) 2401 2402Each macro is expanded when it appears in the definition of the other 2403macro, but not when it indirectly appears in its own definition. 2404 2405 2406File: cpp.info, Node: Argument Prescan, Next: Newlines in Arguments, Prev: Self-Referential Macros, Up: Macro Pitfalls 2407 24083.10.6 Argument Prescan 2409----------------------- 2410 2411Macro arguments are completely macro-expanded before they are 2412substituted into a macro body, unless they are stringified or pasted 2413with other tokens. After substitution, the entire macro body, including 2414the substituted arguments, is scanned again for macros to be expanded. 2415The result is that the arguments are scanned _twice_ to expand macro 2416calls in them. 2417 2418 Most of the time, this has no effect. If the argument contained any 2419macro calls, they are expanded during the first scan. The result 2420therefore contains no macro calls, so the second scan does not change 2421it. If the argument were substituted as given, with no prescan, the 2422single remaining scan would find the same macro calls and produce the 2423same results. 2424 2425 You might expect the double scan to change the results when a 2426self-referential macro is used in an argument of another macro (*note 2427Self-Referential Macros::): the self-referential macro would be 2428expanded once in the first scan, and a second time in the second scan. 2429However, this is not what happens. The self-references that do not 2430expand in the first scan are marked so that they will not expand in the 2431second scan either. 2432 2433 You might wonder, "Why mention the prescan, if it makes no 2434difference? And why not skip it and make the preprocessor faster?" 2435The answer is that the prescan does make a difference in three special 2436cases: 2437 2438 * Nested calls to a macro. 2439 2440 We say that "nested" calls to a macro occur when a macro's argument 2441 contains a call to that very macro. For example, if `f' is a macro 2442 that expects one argument, `f (f (1))' is a nested pair of calls to 2443 `f'. The desired expansion is made by expanding `f (1)' and 2444 substituting that into the definition of `f'. The prescan causes 2445 the expected result to happen. Without the prescan, `f (1)' itself 2446 would be substituted as an argument, and the inner use of `f' would 2447 appear during the main scan as an indirect self-reference and 2448 would not be expanded. 2449 2450 * Macros that call other macros that stringify or concatenate. 2451 2452 If an argument is stringified or concatenated, the prescan does not 2453 occur. If you _want_ to expand a macro, then stringify or 2454 concatenate its expansion, you can do that by causing one macro to 2455 call another macro that does the stringification or concatenation. 2456 For instance, if you have 2457 2458 #define AFTERX(x) X_ ## x 2459 #define XAFTERX(x) AFTERX(x) 2460 #define TABLESIZE 1024 2461 #define BUFSIZE TABLESIZE 2462 2463 then `AFTERX(BUFSIZE)' expands to `X_BUFSIZE', and 2464 `XAFTERX(BUFSIZE)' expands to `X_1024'. (Not to `X_TABLESIZE'. 2465 Prescan always does a complete expansion.) 2466 2467 * Macros used in arguments, whose expansions contain unshielded 2468 commas. 2469 2470 This can cause a macro expanded on the second scan to be called 2471 with the wrong number of arguments. Here is an example: 2472 2473 #define foo a,b 2474 #define bar(x) lose(x) 2475 #define lose(x) (1 + (x)) 2476 2477 We would like `bar(foo)' to turn into `(1 + (foo))', which would 2478 then turn into `(1 + (a,b))'. Instead, `bar(foo)' expands into 2479 `lose(a,b)', and you get an error because `lose' requires a single 2480 argument. In this case, the problem is easily solved by the same 2481 parentheses that ought to be used to prevent misnesting of 2482 arithmetic operations: 2483 2484 #define foo (a,b) 2485 or 2486 #define bar(x) lose((x)) 2487 2488 The extra pair of parentheses prevents the comma in `foo''s 2489 definition from being interpreted as an argument separator. 2490 2491 2492 2493File: cpp.info, Node: Newlines in Arguments, Prev: Argument Prescan, Up: Macro Pitfalls 2494 24953.10.7 Newlines in Arguments 2496---------------------------- 2497 2498The invocation of a function-like macro can extend over many logical 2499lines. However, in the present implementation, the entire expansion 2500comes out on one line. Thus line numbers emitted by the compiler or 2501debugger refer to the line the invocation started on, which might be 2502different to the line containing the argument causing the problem. 2503 2504 Here is an example illustrating this: 2505 2506 #define ignore_second_arg(a,b,c) a; c 2507 2508 ignore_second_arg (foo (), 2509 ignored (), 2510 syntax error); 2511 2512The syntax error triggered by the tokens `syntax error' results in an 2513error message citing line three--the line of ignore_second_arg-- even 2514though the problematic code comes from line five. 2515 2516 We consider this a bug, and intend to fix it in the near future. 2517 2518 2519File: cpp.info, Node: Conditionals, Next: Diagnostics, Prev: Macros, Up: Top 2520 25214 Conditionals 2522************** 2523 2524A "conditional" is a directive that instructs the preprocessor to 2525select whether or not to include a chunk of code in the final token 2526stream passed to the compiler. Preprocessor conditionals can test 2527arithmetic expressions, or whether a name is defined as a macro, or both 2528simultaneously using the special `defined' operator. 2529 2530 A conditional in the C preprocessor resembles in some ways an `if' 2531statement in C, but it is important to understand the difference between 2532them. The condition in an `if' statement is tested during the 2533execution of your program. Its purpose is to allow your program to 2534behave differently from run to run, depending on the data it is 2535operating on. The condition in a preprocessing conditional directive is 2536tested when your program is compiled. Its purpose is to allow different 2537code to be included in the program depending on the situation at the 2538time of compilation. 2539 2540 However, the distinction is becoming less clear. Modern compilers 2541often do test `if' statements when a program is compiled, if their 2542conditions are known not to vary at run time, and eliminate code which 2543can never be executed. If you can count on your compiler to do this, 2544you may find that your program is more readable if you use `if' 2545statements with constant conditions (perhaps determined by macros). Of 2546course, you can only use this to exclude code, not type definitions or 2547other preprocessing directives, and you can only do it if the code 2548remains syntactically valid when it is not to be used. 2549 2550 GCC version 3 eliminates this kind of never-executed code even when 2551not optimizing. Older versions did it only when optimizing. 2552 2553* Menu: 2554 2555* Conditional Uses:: 2556* Conditional Syntax:: 2557* Deleted Code:: 2558 2559 2560File: cpp.info, Node: Conditional Uses, Next: Conditional Syntax, Up: Conditionals 2561 25624.1 Conditional Uses 2563==================== 2564 2565There are three general reasons to use a conditional. 2566 2567 * A program may need to use different code depending on the machine 2568 or operating system it is to run on. In some cases the code for 2569 one operating system may be erroneous on another operating system; 2570 for example, it might refer to data types or constants that do not 2571 exist on the other system. When this happens, it is not enough to 2572 avoid executing the invalid code. Its mere presence will cause 2573 the compiler to reject the program. With a preprocessing 2574 conditional, the offending code can be effectively excised from 2575 the program when it is not valid. 2576 2577 * You may want to be able to compile the same source file into two 2578 different programs. One version might make frequent time-consuming 2579 consistency checks on its intermediate data, or print the values of 2580 those data for debugging, and the other not. 2581 2582 * A conditional whose condition is always false is one way to 2583 exclude code from the program but keep it as a sort of comment for 2584 future reference. 2585 2586 Simple programs that do not need system-specific logic or complex 2587debugging hooks generally will not need to use preprocessing 2588conditionals. 2589 2590 2591File: cpp.info, Node: Conditional Syntax, Next: Deleted Code, Prev: Conditional Uses, Up: Conditionals 2592 25934.2 Conditional Syntax 2594====================== 2595 2596A conditional in the C preprocessor begins with a "conditional 2597directive": `#if', `#ifdef' or `#ifndef'. 2598 2599* Menu: 2600 2601* Ifdef:: 2602* If:: 2603* Defined:: 2604* Else:: 2605* Elif:: 2606 2607 2608File: cpp.info, Node: Ifdef, Next: If, Up: Conditional Syntax 2609 26104.2.1 Ifdef 2611----------- 2612 2613The simplest sort of conditional is 2614 2615 #ifdef MACRO 2616 2617 CONTROLLED TEXT 2618 2619 #endif /* MACRO */ 2620 2621 This block is called a "conditional group". CONTROLLED TEXT will be 2622included in the output of the preprocessor if and only if MACRO is 2623defined. We say that the conditional "succeeds" if MACRO is defined, 2624"fails" if it is not. 2625 2626 The CONTROLLED TEXT inside of a conditional can include 2627preprocessing directives. They are executed only if the conditional 2628succeeds. You can nest conditional groups inside other conditional 2629groups, but they must be completely nested. In other words, `#endif' 2630always matches the nearest `#ifdef' (or `#ifndef', or `#if'). Also, 2631you cannot start a conditional group in one file and end it in another. 2632 2633 Even if a conditional fails, the CONTROLLED TEXT inside it is still 2634run through initial transformations and tokenization. Therefore, it 2635must all be lexically valid C. Normally the only way this matters is 2636that all comments and string literals inside a failing conditional group 2637must still be properly ended. 2638 2639 The comment following the `#endif' is not required, but it is a good 2640practice if there is a lot of CONTROLLED TEXT, because it helps people 2641match the `#endif' to the corresponding `#ifdef'. Older programs 2642sometimes put MACRO directly after the `#endif' without enclosing it in 2643a comment. This is invalid code according to the C standard. CPP 2644accepts it with a warning. It never affects which `#ifndef' the 2645`#endif' matches. 2646 2647 Sometimes you wish to use some code if a macro is _not_ defined. 2648You can do this by writing `#ifndef' instead of `#ifdef'. One common 2649use of `#ifndef' is to include code only the first time a header file 2650is included. *Note Once-Only Headers::. 2651 2652 Macro definitions can vary between compilations for several reasons. 2653Here are some samples. 2654 2655 * Some macros are predefined on each kind of machine (*note 2656 System-specific Predefined Macros::). This allows you to provide 2657 code specially tuned for a particular machine. 2658 2659 * System header files define more macros, associated with the 2660 features they implement. You can test these macros with 2661 conditionals to avoid using a system feature on a machine where it 2662 is not implemented. 2663 2664 * Macros can be defined or undefined with the `-D' and `-U' command 2665 line options when you compile the program. You can arrange to 2666 compile the same source file into two different programs by 2667 choosing a macro name to specify which program you want, writing 2668 conditionals to test whether or how this macro is defined, and 2669 then controlling the state of the macro with command line options, 2670 perhaps set in the Makefile. *Note Invocation::. 2671 2672 * Your program might have a special header file (often called 2673 `config.h') that is adjusted when the program is compiled. It can 2674 define or not define macros depending on the features of the 2675 system and the desired capabilities of the program. The 2676 adjustment can be automated by a tool such as `autoconf', or done 2677 by hand. 2678 2679 2680File: cpp.info, Node: If, Next: Defined, Prev: Ifdef, Up: Conditional Syntax 2681 26824.2.2 If 2683-------- 2684 2685The `#if' directive allows you to test the value of an arithmetic 2686expression, rather than the mere existence of one macro. Its syntax is 2687 2688 #if EXPRESSION 2689 2690 CONTROLLED TEXT 2691 2692 #endif /* EXPRESSION */ 2693 2694 EXPRESSION is a C expression of integer type, subject to stringent 2695restrictions. It may contain 2696 2697 * Integer constants. 2698 2699 * Character constants, which are interpreted as they would be in 2700 normal code. 2701 2702 * Arithmetic operators for addition, subtraction, multiplication, 2703 division, bitwise operations, shifts, comparisons, and logical 2704 operations (`&&' and `||'). The latter two obey the usual 2705 short-circuiting rules of standard C. 2706 2707 * Macros. All macros in the expression are expanded before actual 2708 computation of the expression's value begins. 2709 2710 * Uses of the `defined' operator, which lets you check whether macros 2711 are defined in the middle of an `#if'. 2712 2713 * Identifiers that are not macros, which are all considered to be the 2714 number zero. This allows you to write `#if MACRO' instead of 2715 `#ifdef MACRO', if you know that MACRO, when defined, will always 2716 have a nonzero value. Function-like macros used without their 2717 function call parentheses are also treated as zero. 2718 2719 In some contexts this shortcut is undesirable. The `-Wundef' 2720 option causes GCC to warn whenever it encounters an identifier 2721 which is not a macro in an `#if'. 2722 2723 The preprocessor does not know anything about types in the language. 2724Therefore, `sizeof' operators are not recognized in `#if', and neither 2725are `enum' constants. They will be taken as identifiers which are not 2726macros, and replaced by zero. In the case of `sizeof', this is likely 2727to cause the expression to be invalid. 2728 2729 The preprocessor calculates the value of EXPRESSION. It carries out 2730all calculations in the widest integer type known to the compiler; on 2731most machines supported by GCC this is 64 bits. This is not the same 2732rule as the compiler uses to calculate the value of a constant 2733expression, and may give different results in some cases. If the value 2734comes out to be nonzero, the `#if' succeeds and the CONTROLLED TEXT is 2735included; otherwise it is skipped. 2736 2737 2738File: cpp.info, Node: Defined, Next: Else, Prev: If, Up: Conditional Syntax 2739 27404.2.3 Defined 2741------------- 2742 2743The special operator `defined' is used in `#if' and `#elif' expressions 2744to test whether a certain name is defined as a macro. `defined NAME' 2745and `defined (NAME)' are both expressions whose value is 1 if NAME is 2746defined as a macro at the current point in the program, and 0 2747otherwise. Thus, `#if defined MACRO' is precisely equivalent to 2748`#ifdef MACRO'. 2749 2750 `defined' is useful when you wish to test more than one macro for 2751existence at once. For example, 2752 2753 #if defined (__vax__) || defined (__ns16000__) 2754 2755would succeed if either of the names `__vax__' or `__ns16000__' is 2756defined as a macro. 2757 2758 Conditionals written like this: 2759 2760 #if defined BUFSIZE && BUFSIZE >= 1024 2761 2762can generally be simplified to just `#if BUFSIZE >= 1024', since if 2763`BUFSIZE' is not defined, it will be interpreted as having the value 2764zero. 2765 2766 If the `defined' operator appears as a result of a macro expansion, 2767the C standard says the behavior is undefined. GNU cpp treats it as a 2768genuine `defined' operator and evaluates it normally. It will warn 2769wherever your code uses this feature if you use the command-line option 2770`-pedantic', since other compilers may handle it differently. 2771 2772 2773File: cpp.info, Node: Else, Next: Elif, Prev: Defined, Up: Conditional Syntax 2774 27754.2.4 Else 2776---------- 2777 2778The `#else' directive can be added to a conditional to provide 2779alternative text to be used if the condition fails. This is what it 2780looks like: 2781 2782 #if EXPRESSION 2783 TEXT-IF-TRUE 2784 #else /* Not EXPRESSION */ 2785 TEXT-IF-FALSE 2786 #endif /* Not EXPRESSION */ 2787 2788If EXPRESSION is nonzero, the TEXT-IF-TRUE is included and the 2789TEXT-IF-FALSE is skipped. If EXPRESSION is zero, the opposite happens. 2790 2791 You can use `#else' with `#ifdef' and `#ifndef', too. 2792 2793 2794File: cpp.info, Node: Elif, Prev: Else, Up: Conditional Syntax 2795 27964.2.5 Elif 2797---------- 2798 2799One common case of nested conditionals is used to check for more than 2800two possible alternatives. For example, you might have 2801 2802 #if X == 1 2803 ... 2804 #else /* X != 1 */ 2805 #if X == 2 2806 ... 2807 #else /* X != 2 */ 2808 ... 2809 #endif /* X != 2 */ 2810 #endif /* X != 1 */ 2811 2812 Another conditional directive, `#elif', allows this to be 2813abbreviated as follows: 2814 2815 #if X == 1 2816 ... 2817 #elif X == 2 2818 ... 2819 #else /* X != 2 and X != 1*/ 2820 ... 2821 #endif /* X != 2 and X != 1*/ 2822 2823 `#elif' stands for "else if". Like `#else', it goes in the middle 2824of a conditional group and subdivides it; it does not require a 2825matching `#endif' of its own. Like `#if', the `#elif' directive 2826includes an expression to be tested. The text following the `#elif' is 2827processed only if the original `#if'-condition failed and the `#elif' 2828condition succeeds. 2829 2830 More than one `#elif' can go in the same conditional group. Then 2831the text after each `#elif' is processed only if the `#elif' condition 2832succeeds after the original `#if' and all previous `#elif' directives 2833within it have failed. 2834 2835 `#else' is allowed after any number of `#elif' directives, but 2836`#elif' may not follow `#else'. 2837 2838 2839File: cpp.info, Node: Deleted Code, Prev: Conditional Syntax, Up: Conditionals 2840 28414.3 Deleted Code 2842================ 2843 2844If you replace or delete a part of the program but want to keep the old 2845code around for future reference, you often cannot simply comment it 2846out. Block comments do not nest, so the first comment inside the old 2847code will end the commenting-out. The probable result is a flood of 2848syntax errors. 2849 2850 One way to avoid this problem is to use an always-false conditional 2851instead. For instance, put `#if 0' before the deleted code and 2852`#endif' after it. This works even if the code being turned off 2853contains conditionals, but they must be entire conditionals (balanced 2854`#if' and `#endif'). 2855 2856 Some people use `#ifdef notdef' instead. This is risky, because 2857`notdef' might be accidentally defined as a macro, and then the 2858conditional would succeed. `#if 0' can be counted on to fail. 2859 2860 Do not use `#if 0' for comments which are not C code. Use a real 2861comment, instead. The interior of `#if 0' must consist of complete 2862tokens; in particular, single-quote characters must balance. Comments 2863often contain unbalanced single-quote characters (known in English as 2864apostrophes). These confuse `#if 0'. They don't confuse `/*'. 2865 2866 2867File: cpp.info, Node: Diagnostics, Next: Line Control, Prev: Conditionals, Up: Top 2868 28695 Diagnostics 2870************* 2871 2872The directive `#error' causes the preprocessor to report a fatal error. 2873The tokens forming the rest of the line following `#error' are used as 2874the error message. 2875 2876 You would use `#error' inside of a conditional that detects a 2877combination of parameters which you know the program does not properly 2878support. For example, if you know that the program will not run 2879properly on a VAX, you might write 2880 2881 #ifdef __vax__ 2882 #error "Won't work on VAXen. See comments at get_last_object." 2883 #endif 2884 2885 If you have several configuration parameters that must be set up by 2886the installation in a consistent way, you can use conditionals to detect 2887an inconsistency and report it with `#error'. For example, 2888 2889 #if !defined(UNALIGNED_INT_ASM_OP) && defined(DWARF2_DEBUGGING_INFO) 2890 #error "DWARF2_DEBUGGING_INFO requires UNALIGNED_INT_ASM_OP." 2891 #endif 2892 2893 The directive `#warning' is like `#error', but causes the 2894preprocessor to issue a warning and continue preprocessing. The tokens 2895following `#warning' are used as the warning message. 2896 2897 You might use `#warning' in obsolete header files, with a message 2898directing the user to the header file which should be used instead. 2899 2900 Neither `#error' nor `#warning' macro-expands its argument. 2901Internal whitespace sequences are each replaced with a single space. 2902The line must consist of complete tokens. It is wisest to make the 2903argument of these directives be a single string constant; this avoids 2904problems with apostrophes and the like. 2905 2906 2907File: cpp.info, Node: Line Control, Next: Pragmas, Prev: Diagnostics, Up: Top 2908 29096 Line Control 2910************** 2911 2912The C preprocessor informs the C compiler of the location in your source 2913code where each token came from. Presently, this is just the file name 2914and line number. All the tokens resulting from macro expansion are 2915reported as having appeared on the line of the source file where the 2916outermost macro was used. We intend to be more accurate in the future. 2917 2918 If you write a program which generates source code, such as the 2919`bison' parser generator, you may want to adjust the preprocessor's 2920notion of the current file name and line number by hand. Parts of the 2921output from `bison' are generated from scratch, other parts come from a 2922standard parser file. The rest are copied verbatim from `bison''s 2923input. You would like compiler error messages and symbolic debuggers 2924to be able to refer to `bison''s input file. 2925 2926 `bison' or any such program can arrange this by writing `#line' 2927directives into the output file. `#line' is a directive that specifies 2928the original line number and source file name for subsequent input in 2929the current preprocessor input file. `#line' has three variants: 2930 2931`#line LINENUM' 2932 LINENUM is a non-negative decimal integer constant. It specifies 2933 the line number which should be reported for the following line of 2934 input. Subsequent lines are counted from LINENUM. 2935 2936`#line LINENUM FILENAME' 2937 LINENUM is the same as for the first form, and has the same 2938 effect. In addition, FILENAME is a string constant. The 2939 following line and all subsequent lines are reported to come from 2940 the file it specifies, until something else happens to change that. 2941 FILENAME is interpreted according to the normal rules for a string 2942 constant: backslash escapes are interpreted. This is different 2943 from `#include'. 2944 2945 Previous versions of CPP did not interpret escapes in `#line'; we 2946 have changed it because the standard requires they be interpreted, 2947 and most other compilers do. 2948 2949`#line ANYTHING ELSE' 2950 ANYTHING ELSE is checked for macro calls, which are expanded. The 2951 result should match one of the above two forms. 2952 2953 `#line' directives alter the results of the `__FILE__' and 2954`__LINE__' predefined macros from that point on. *Note Standard 2955Predefined Macros::. They do not have any effect on `#include''s idea 2956of the directory containing the current file. This is a change from 2957GCC 2.95. Previously, a file reading 2958 2959 #line 1 "../src/gram.y" 2960 #include "gram.h" 2961 2962 would search for `gram.h' in `../src', then the `-I' chain; the 2963directory containing the physical source file would not be searched. 2964In GCC 3.0 and later, the `#include' is not affected by the presence of 2965a `#line' referring to a different directory. 2966 2967 We made this change because the old behavior caused problems when 2968generated source files were transported between machines. For instance, 2969it is common practice to ship generated parsers with a source release, 2970so that people building the distribution do not need to have yacc or 2971Bison installed. These files frequently have `#line' directives 2972referring to the directory tree of the system where the distribution was 2973created. If GCC tries to search for headers in those directories, the 2974build is likely to fail. 2975 2976 The new behavior can cause failures too, if the generated file is not 2977in the same directory as its source and it attempts to include a header 2978which would be visible searching from the directory containing the 2979source file. However, this problem is easily solved with an additional 2980`-I' switch on the command line. The failures caused by the old 2981semantics could sometimes be corrected only by editing the generated 2982files, which is difficult and error-prone. 2983 2984 2985File: cpp.info, Node: Pragmas, Next: Other Directives, Prev: Line Control, Up: Top 2986 29877 Pragmas 2988********* 2989 2990The `#pragma' directive is the method specified by the C standard for 2991providing additional information to the compiler, beyond what is 2992conveyed in the language itself. Three forms of this directive 2993(commonly known as "pragmas") are specified by the 1999 C standard. A 2994C compiler is free to attach any meaning it likes to other pragmas. 2995 2996 GCC has historically preferred to use extensions to the syntax of the 2997language, such as `__attribute__', for this purpose. However, GCC does 2998define a few pragmas of its own. These mostly have effects on the 2999entire translation unit or source file. 3000 3001 In GCC version 3, all GNU-defined, supported pragmas have been given 3002a `GCC' prefix. This is in line with the `STDC' prefix on all pragmas 3003defined by C99. For backward compatibility, pragmas which were 3004recognized by previous versions are still recognized without the `GCC' 3005prefix, but that usage is deprecated. Some older pragmas are 3006deprecated in their entirety. They are not recognized with the `GCC' 3007prefix. *Note Obsolete Features::. 3008 3009 C99 introduces the `_Pragma' operator. This feature addresses a 3010major problem with `#pragma': being a directive, it cannot be produced 3011as the result of macro expansion. `_Pragma' is an operator, much like 3012`sizeof' or `defined', and can be embedded in a macro. 3013 3014 Its syntax is `_Pragma (STRING-LITERAL)', where STRING-LITERAL can 3015be either a normal or wide-character string literal. It is 3016destringized, by replacing all `\\' with a single `\' and all `\"' with 3017a `"'. The result is then processed as if it had appeared as the right 3018hand side of a `#pragma' directive. For example, 3019 3020 _Pragma ("GCC dependency \"parse.y\"") 3021 3022has the same effect as `#pragma GCC dependency "parse.y"'. The same 3023effect could be achieved using macros, for example 3024 3025 #define DO_PRAGMA(x) _Pragma (#x) 3026 DO_PRAGMA (GCC dependency "parse.y") 3027 3028 The standard is unclear on where a `_Pragma' operator can appear. 3029The preprocessor does not accept it within a preprocessing conditional 3030directive like `#if'. To be safe, you are probably best keeping it out 3031of directives other than `#define', and putting it on a line of its own. 3032 3033 This manual documents the pragmas which are meaningful to the 3034preprocessor itself. Other pragmas are meaningful to the C or C++ 3035compilers. They are documented in the GCC manual. 3036 3037`#pragma GCC dependency' 3038 `#pragma GCC dependency' allows you to check the relative dates of 3039 the current file and another file. If the other file is more 3040 recent than the current file, a warning is issued. This is useful 3041 if the current file is derived from the other file, and should be 3042 regenerated. The other file is searched for using the normal 3043 include search path. Optional trailing text can be used to give 3044 more information in the warning message. 3045 3046 #pragma GCC dependency "parse.y" 3047 #pragma GCC dependency "/usr/include/time.h" rerun fixincludes 3048 3049`#pragma GCC poison' 3050 Sometimes, there is an identifier that you want to remove 3051 completely from your program, and make sure that it never creeps 3052 back in. To enforce this, you can "poison" the identifier with 3053 this pragma. `#pragma GCC poison' is followed by a list of 3054 identifiers to poison. If any of those identifiers appears 3055 anywhere in the source after the directive, it is a hard error. 3056 For example, 3057 3058 #pragma GCC poison printf sprintf fprintf 3059 sprintf(some_string, "hello"); 3060 3061 will produce an error. 3062 3063 If a poisoned identifier appears as part of the expansion of a 3064 macro which was defined before the identifier was poisoned, it 3065 will _not_ cause an error. This lets you poison an identifier 3066 without worrying about system headers defining macros that use it. 3067 3068 For example, 3069 3070 #define strrchr rindex 3071 #pragma GCC poison rindex 3072 strrchr(some_string, 'h'); 3073 3074 will not produce an error. 3075 3076`#pragma GCC system_header' 3077 This pragma takes no arguments. It causes the rest of the code in 3078 the current file to be treated as if it came from a system header. 3079 *Note System Headers::. 3080 3081 3082 3083File: cpp.info, Node: Other Directives, Next: Preprocessor Output, Prev: Pragmas, Up: Top 3084 30858 Other Directives 3086****************** 3087 3088The `#ident' directive takes one argument, a string constant. On some 3089systems, that string constant is copied into a special segment of the 3090object file. On other systems, the directive is ignored. The `#sccs' 3091directive is a synonym for `#ident'. 3092 3093 These directives are not part of the C standard, but they are not 3094official GNU extensions either. What historical information we have 3095been able to find, suggests they originated with System V. 3096 3097 The "null directive" consists of a `#' followed by a newline, with 3098only whitespace (including comments) in between. A null directive is 3099understood as a preprocessing directive but has no effect on the 3100preprocessor output. The primary significance of the existence of the 3101null directive is that an input line consisting of just a `#' will 3102produce no output, rather than a line of output containing just a `#'. 3103Supposedly some old C programs contain such lines. 3104 3105 3106File: cpp.info, Node: Preprocessor Output, Next: Traditional Mode, Prev: Other Directives, Up: Top 3107 31089 Preprocessor Output 3109********************* 3110 3111When the C preprocessor is used with the C, C++, or Objective-C 3112compilers, it is integrated into the compiler and communicates a stream 3113of binary tokens directly to the compiler's parser. However, it can 3114also be used in the more conventional standalone mode, where it produces 3115textual output. 3116 3117 The output from the C preprocessor looks much like the input, except 3118that all preprocessing directive lines have been replaced with blank 3119lines and all comments with spaces. Long runs of blank lines are 3120discarded. 3121 3122 The ISO standard specifies that it is implementation defined whether 3123a preprocessor preserves whitespace between tokens, or replaces it with 3124e.g. a single space. In GNU CPP, whitespace between tokens is collapsed 3125to become a single space, with the exception that the first token on a 3126non-directive line is preceded with sufficient spaces that it appears in 3127the same column in the preprocessed output that it appeared in the 3128original source file. This is so the output is easy to read. *Note 3129Differences from previous versions::. CPP does not insert any 3130whitespace where there was none in the original source, except where 3131necessary to prevent an accidental token paste. 3132 3133 Source file name and line number information is conveyed by lines of 3134the form 3135 3136 # LINENUM FILENAME FLAGS 3137 3138These are called "linemarkers". They are inserted as needed into the 3139output (but never within a string or character constant). They mean 3140that the following line originated in file FILENAME at line LINENUM. 3141FILENAME will never contain any non-printing characters; they are 3142replaced with octal escape sequences. 3143 3144 After the file name comes zero or more flags, which are `1', `2', 3145`3', or `4'. If there are multiple flags, spaces separate them. Here 3146is what the flags mean: 3147 3148`1' 3149 This indicates the start of a new file. 3150 3151`2' 3152 This indicates returning to a file (after having included another 3153 file). 3154 3155`3' 3156 This indicates that the following text comes from a system header 3157 file, so certain warnings should be suppressed. 3158 3159`4' 3160 This indicates that the following text should be treated as being 3161 wrapped in an implicit `extern "C"' block. 3162 3163 As an extension, the preprocessor accepts linemarkers in 3164non-assembler input files. They are treated like the corresponding 3165`#line' directive, (*note Line Control::), except that trailing flags 3166are permitted, and are interpreted with the meanings described above. 3167If multiple flags are given, they must be in ascending order. 3168 3169 Some directives may be duplicated in the output of the preprocessor. 3170These are `#ident' (always), `#pragma' (only if the preprocessor does 3171not handle the pragma itself), and `#define' and `#undef' (with certain 3172debugging options). If this happens, the `#' of the directive will 3173always be in the first column, and there will be no space between the 3174`#' and the directive name. If macro expansion happens to generate 3175tokens which might be mistaken for a duplicated directive, a space will 3176be inserted between the `#' and the directive name. 3177 3178 3179File: cpp.info, Node: Traditional Mode, Next: Implementation Details, Prev: Preprocessor Output, Up: Top 3180 318110 Traditional Mode 3182******************* 3183 3184Traditional (pre-standard) C preprocessing is rather different from the 3185preprocessing specified by the standard. When GCC is given the 3186`-traditional-cpp' option, it attempts to emulate a traditional 3187preprocessor. 3188 3189 GCC versions 3.2 and later only support traditional mode semantics in 3190the preprocessor, and not in the compiler front ends. This chapter 3191outlines the traditional preprocessor semantics we implemented. 3192 3193 The implementation does not correspond precisely to the behavior of 3194earlier versions of GCC, nor to any true traditional preprocessor. 3195After all, inconsistencies among traditional implementations were a 3196major motivation for C standardization. However, we intend that it 3197should be compatible with true traditional preprocessors in all ways 3198that actually matter. 3199 3200* Menu: 3201 3202* Traditional lexical analysis:: 3203* Traditional macros:: 3204* Traditional miscellany:: 3205* Traditional warnings:: 3206 3207 3208File: cpp.info, Node: Traditional lexical analysis, Next: Traditional macros, Up: Traditional Mode 3209 321010.1 Traditional lexical analysis 3211================================= 3212 3213The traditional preprocessor does not decompose its input into tokens 3214the same way a standards-conforming preprocessor does. The input is 3215simply treated as a stream of text with minimal internal form. 3216 3217 This implementation does not treat trigraphs (*note trigraphs::) 3218specially since they were an invention of the standards committee. It 3219handles arbitrarily-positioned escaped newlines properly and splices 3220the lines as you would expect; many traditional preprocessors did not 3221do this. 3222 3223 The form of horizontal whitespace in the input file is preserved in 3224the output. In particular, hard tabs remain hard tabs. This can be 3225useful if, for example, you are preprocessing a Makefile. 3226 3227 Traditional CPP only recognizes C-style block comments, and treats 3228the `/*' sequence as introducing a comment only if it lies outside 3229quoted text. Quoted text is introduced by the usual single and double 3230quotes, and also by an initial `<' in a `#include' directive. 3231 3232 Traditionally, comments are completely removed and are not replaced 3233with a space. Since a traditional compiler does its own tokenization 3234of the output of the preprocessor, this means that comments can 3235effectively be used as token paste operators. However, comments behave 3236like separators for text handled by the preprocessor itself, since it 3237doesn't re-lex its input. For example, in 3238 3239 #if foo/**/bar 3240 3241`foo' and `bar' are distinct identifiers and expanded separately if 3242they happen to be macros. In other words, this directive is equivalent 3243to 3244 3245 #if foo bar 3246 3247rather than 3248 3249 #if foobar 3250 3251 Generally speaking, in traditional mode an opening quote need not 3252have a matching closing quote. In particular, a macro may be defined 3253with replacement text that contains an unmatched quote. Of course, if 3254you attempt to compile preprocessed output containing an unmatched quote 3255you will get a syntax error. 3256 3257 However, all preprocessing directives other than `#define' require 3258matching quotes. For example: 3259 3260 #define m This macro's fine and has an unmatched quote 3261 "/* This is not a comment. */ 3262 /* This is a comment. The following #include directive 3263 is ill-formed. */ 3264 #include <stdio.h 3265 3266 Just as for the ISO preprocessor, what would be a closing quote can 3267be escaped with a backslash to prevent the quoted text from closing. 3268 3269 3270File: cpp.info, Node: Traditional macros, Next: Traditional miscellany, Prev: Traditional lexical analysis, Up: Traditional Mode 3271 327210.2 Traditional macros 3273======================= 3274 3275The major difference between traditional and ISO macros is that the 3276former expand to text rather than to a token sequence. CPP removes all 3277leading and trailing horizontal whitespace from a macro's replacement 3278text before storing it, but preserves the form of internal whitespace. 3279 3280 One consequence is that it is legitimate for the replacement text to 3281contain an unmatched quote (*note Traditional lexical analysis::). An 3282unclosed string or character constant continues into the text following 3283the macro call. Similarly, the text at the end of a macro's expansion 3284can run together with the text after the macro invocation to produce a 3285single token. 3286 3287 Normally comments are removed from the replacement text after the 3288macro is expanded, but if the `-CC' option is passed on the command 3289line comments are preserved. (In fact, the current implementation 3290removes comments even before saving the macro replacement text, but it 3291careful to do it in such a way that the observed effect is identical 3292even in the function-like macro case.) 3293 3294 The ISO stringification operator `#' and token paste operator `##' 3295have no special meaning. As explained later, an effect similar to 3296these operators can be obtained in a different way. Macro names that 3297are embedded in quotes, either from the main file or after macro 3298replacement, do not expand. 3299 3300 CPP replaces an unquoted object-like macro name with its replacement 3301text, and then rescans it for further macros to replace. Unlike 3302standard macro expansion, traditional macro expansion has no provision 3303to prevent recursion. If an object-like macro appears unquoted in its 3304replacement text, it will be replaced again during the rescan pass, and 3305so on _ad infinitum_. GCC detects when it is expanding recursive 3306macros, emits an error message, and continues after the offending macro 3307invocation. 3308 3309 #define PLUS + 3310 #define INC(x) PLUS+x 3311 INC(foo); 3312 ==> ++foo; 3313 3314 Function-like macros are similar in form but quite different in 3315behavior to their ISO counterparts. Their arguments are contained 3316within parentheses, are comma-separated, and can cross physical lines. 3317Commas within nested parentheses are not treated as argument 3318separators. Similarly, a quote in an argument cannot be left unclosed; 3319a following comma or parenthesis that comes before the closing quote is 3320treated like any other character. There is no facility for handling 3321variadic macros. 3322 3323 This implementation removes all comments from macro arguments, unless 3324the `-C' option is given. The form of all other horizontal whitespace 3325in arguments is preserved, including leading and trailing whitespace. 3326In particular 3327 3328 f( ) 3329 3330is treated as an invocation of the macro `f' with a single argument 3331consisting of a single space. If you want to invoke a function-like 3332macro that takes no arguments, you must not leave any whitespace 3333between the parentheses. 3334 3335 If a macro argument crosses a new line, the new line is replaced with 3336a space when forming the argument. If the previous line contained an 3337unterminated quote, the following line inherits the quoted state. 3338 3339 Traditional preprocessors replace parameters in the replacement text 3340with their arguments regardless of whether the parameters are within 3341quotes or not. This provides a way to stringize arguments. For example 3342 3343 #define str(x) "x" 3344 str(/* A comment */some text ) 3345 ==> "some text " 3346 3347Note that the comment is removed, but that the trailing space is 3348preserved. Here is an example of using a comment to effect token 3349pasting. 3350 3351 #define suffix(x) foo_/**/x 3352 suffix(bar) 3353 ==> foo_bar 3354 3355 3356File: cpp.info, Node: Traditional miscellany, Next: Traditional warnings, Prev: Traditional macros, Up: Traditional Mode 3357 335810.3 Traditional miscellany 3359=========================== 3360 3361Here are some things to be aware of when using the traditional 3362preprocessor. 3363 3364 * Preprocessing directives are recognized only when their leading 3365 `#' appears in the first column. There can be no whitespace 3366 between the beginning of the line and the `#', but whitespace can 3367 follow the `#'. 3368 3369 * A true traditional C preprocessor does not recognize `#error' or 3370 `#pragma', and may not recognize `#elif'. CPP supports all the 3371 directives in traditional mode that it supports in ISO mode, 3372 including extensions, with the exception that the effects of 3373 `#pragma GCC poison' are undefined. 3374 3375 * __STDC__ is not defined. 3376 3377 * If you use digraphs the behavior is undefined. 3378 3379 * If a line that looks like a directive appears within macro 3380 arguments, the behavior is undefined. 3381 3382 3383 3384File: cpp.info, Node: Traditional warnings, Prev: Traditional miscellany, Up: Traditional Mode 3385 338610.4 Traditional warnings 3387========================= 3388 3389You can request warnings about features that did not exist, or worked 3390differently, in traditional C with the `-Wtraditional' option. GCC 3391does not warn about features of ISO C which you must use when you are 3392using a conforming compiler, such as the `#' and `##' operators. 3393 3394 Presently `-Wtraditional' warns about: 3395 3396 * Macro parameters that appear within string literals in the macro 3397 body. In traditional C macro replacement takes place within 3398 string literals, but does not in ISO C. 3399 3400 * In traditional C, some preprocessor directives did not exist. 3401 Traditional preprocessors would only consider a line to be a 3402 directive if the `#' appeared in column 1 on the line. Therefore 3403 `-Wtraditional' warns about directives that traditional C 3404 understands but would ignore because the `#' does not appear as the 3405 first character on the line. It also suggests you hide directives 3406 like `#pragma' not understood by traditional C by indenting them. 3407 Some traditional implementations would not recognize `#elif', so it 3408 suggests avoiding it altogether. 3409 3410 * A function-like macro that appears without an argument list. In 3411 some traditional preprocessors this was an error. In ISO C it 3412 merely means that the macro is not expanded. 3413 3414 * The unary plus operator. This did not exist in traditional C. 3415 3416 * The `U' and `LL' integer constant suffixes, which were not 3417 available in traditional C. (Traditional C does support the `L' 3418 suffix for simple long integer constants.) You are not warned 3419 about uses of these suffixes in macros defined in system headers. 3420 For instance, `UINT_MAX' may well be defined as `4294967295U', but 3421 you will not be warned if you use `UINT_MAX'. 3422 3423 You can usually avoid the warning, and the related warning about 3424 constants which are so large that they are unsigned, by writing the 3425 integer constant in question in hexadecimal, with no U suffix. 3426 Take care, though, because this gives the wrong result in exotic 3427 cases. 3428 3429 3430File: cpp.info, Node: Implementation Details, Next: Invocation, Prev: Traditional Mode, Up: Top 3431 343211 Implementation Details 3433************************* 3434 3435Here we document details of how the preprocessor's implementation 3436affects its user-visible behavior. You should try to avoid undue 3437reliance on behavior described here, as it is possible that it will 3438change subtly in future implementations. 3439 3440 Also documented here are obsolete features and changes from previous 3441versions of CPP. 3442 3443* Menu: 3444 3445* Implementation-defined behavior:: 3446* Implementation limits:: 3447* Obsolete Features:: 3448* Differences from previous versions:: 3449 3450 3451File: cpp.info, Node: Implementation-defined behavior, Next: Implementation limits, Up: Implementation Details 3452 345311.1 Implementation-defined behavior 3454==================================== 3455 3456This is how CPP behaves in all the cases which the C standard describes 3457as "implementation-defined". This term means that the implementation 3458is free to do what it likes, but must document its choice and stick to 3459it. 3460 3461 * The mapping of physical source file multi-byte characters to the 3462 execution character set. 3463 3464 Currently, CPP requires its input to be ASCII or UTF-8. The 3465 execution character set may be controlled by the user, with the 3466 `-fexec-charset' and `-fwide-exec-charset' options. 3467 3468 * Identifier characters. 3469 3470 The C and C++ standards allow identifiers to be composed of `_' 3471 and the alphanumeric characters. C++ and C99 also allow universal 3472 character names, and C99 further permits implementation-defined 3473 characters. GCC currently only permits universal character names 3474 if `-fextended-identifiers' is used, because the implementation of 3475 universal character names in identifiers is experimental. 3476 3477 GCC allows the `$' character in identifiers as an extension for 3478 most targets. This is true regardless of the `std=' switch, since 3479 this extension cannot conflict with standards-conforming programs. 3480 When preprocessing assembler, however, dollars are not identifier 3481 characters by default. 3482 3483 Currently the targets that by default do not permit `$' are AVR, 3484 IP2K, MMIX, MIPS Irix 3, ARM aout, and PowerPC targets for the AIX 3485 and BeOS operating systems. 3486 3487 You can override the default with `-fdollars-in-identifiers' or 3488 `fno-dollars-in-identifiers'. *Note fdollars-in-identifiers::. 3489 3490 * Non-empty sequences of whitespace characters. 3491 3492 In textual output, each whitespace sequence is collapsed to a 3493 single space. For aesthetic reasons, the first token on each 3494 non-directive line of output is preceded with sufficient spaces 3495 that it appears in the same column as it did in the original 3496 source file. 3497 3498 * The numeric value of character constants in preprocessor 3499 expressions. 3500 3501 The preprocessor and compiler interpret character constants in the 3502 same way; i.e. escape sequences such as `\a' are given the values 3503 they would have on the target machine. 3504 3505 The compiler values a multi-character character constant a 3506 character at a time, shifting the previous value left by the 3507 number of bits per target character, and then or-ing in the 3508 bit-pattern of the new character truncated to the width of a 3509 target character. The final bit-pattern is given type `int', and 3510 is therefore signed, regardless of whether single characters are 3511 signed or not (a slight change from versions 3.1 and earlier of 3512 GCC). If there are more characters in the constant than would fit 3513 in the target `int' the compiler issues a warning, and the excess 3514 leading characters are ignored. 3515 3516 For example, `'ab'' for a target with an 8-bit `char' would be 3517 interpreted as 3518 `(int) ((unsigned char) 'a' * 256 + (unsigned char) 'b')', and 3519 `'\234a'' as 3520 `(int) ((unsigned char) '\234' * 256 + (unsigned char) 'a')'. 3521 3522 * Source file inclusion. 3523 3524 For a discussion on how the preprocessor locates header files, 3525 *Note Include Operation::. 3526 3527 * Interpretation of the filename resulting from a macro-expanded 3528 `#include' directive. 3529 3530 *Note Computed Includes::. 3531 3532 * Treatment of a `#pragma' directive that after macro-expansion 3533 results in a standard pragma. 3534 3535 No macro expansion occurs on any `#pragma' directive line, so the 3536 question does not arise. 3537 3538 Note that GCC does not yet implement any of the standard pragmas. 3539 3540 3541 3542File: cpp.info, Node: Implementation limits, Next: Obsolete Features, Prev: Implementation-defined behavior, Up: Implementation Details 3543 354411.2 Implementation limits 3545========================== 3546 3547CPP has a small number of internal limits. This section lists the 3548limits which the C standard requires to be no lower than some minimum, 3549and all the others known. It is intended that there should be as few 3550limits as possible. If you encounter an undocumented or inconvenient 3551limit, please report that as a bug. *Note Reporting Bugs: (gcc)Bugs. 3552 3553 Where we say something is limited "only by available memory", that 3554means that internal data structures impose no intrinsic limit, and space 3555is allocated with `malloc' or equivalent. The actual limit will 3556therefore depend on many things, such as the size of other things 3557allocated by the compiler at the same time, the amount of memory 3558consumed by other processes on the same computer, etc. 3559 3560 * Nesting levels of `#include' files. 3561 3562 We impose an arbitrary limit of 200 levels, to avoid runaway 3563 recursion. The standard requires at least 15 levels. 3564 3565 * Nesting levels of conditional inclusion. 3566 3567 The C standard mandates this be at least 63. CPP is limited only 3568 by available memory. 3569 3570 * Levels of parenthesized expressions within a full expression. 3571 3572 The C standard requires this to be at least 63. In preprocessor 3573 conditional expressions, it is limited only by available memory. 3574 3575 * Significant initial characters in an identifier or macro name. 3576 3577 The preprocessor treats all characters as significant. The C 3578 standard requires only that the first 63 be significant. 3579 3580 * Number of macros simultaneously defined in a single translation 3581 unit. 3582 3583 The standard requires at least 4095 be possible. CPP is limited 3584 only by available memory. 3585 3586 * Number of parameters in a macro definition and arguments in a 3587 macro call. 3588 3589 We allow `USHRT_MAX', which is no smaller than 65,535. The minimum 3590 required by the standard is 127. 3591 3592 * Number of characters on a logical source line. 3593 3594 The C standard requires a minimum of 4096 be permitted. CPP places 3595 no limits on this, but you may get incorrect column numbers 3596 reported in diagnostics for lines longer than 65,535 characters. 3597 3598 * Maximum size of a source file. 3599 3600 The standard does not specify any lower limit on the maximum size 3601 of a source file. GNU cpp maps files into memory, so it is 3602 limited by the available address space. This is generally at 3603 least two gigabytes. Depending on the operating system, the size 3604 of physical memory may or may not be a limitation. 3605 3606 3607 3608File: cpp.info, Node: Obsolete Features, Next: Differences from previous versions, Prev: Implementation limits, Up: Implementation Details 3609 361011.3 Obsolete Features 3611====================== 3612 3613CPP has a number of features which are present mainly for compatibility 3614with older programs. We discourage their use in new code. In some 3615cases, we plan to remove the feature in a future version of GCC. 3616 3617* Menu: 3618 3619* Assertions:: 3620* Obsolete once-only headers:: 3621 3622 3623File: cpp.info, Node: Assertions, Next: Obsolete once-only headers, Up: Obsolete Features 3624 362511.3.1 Assertions 3626----------------- 3627 3628"Assertions" are a deprecated alternative to macros in writing 3629conditionals to test what sort of computer or system the compiled 3630program will run on. Assertions are usually predefined, but you can 3631define them with preprocessing directives or command-line options. 3632 3633 Assertions were intended to provide a more systematic way to describe 3634the compiler's target system. However, in practice they are just as 3635unpredictable as the system-specific predefined macros. In addition, 3636they are not part of any standard, and only a few compilers support 3637them. Therefore, the use of assertions is *less* portable than the use 3638of system-specific predefined macros. We recommend you do not use them 3639at all. 3640 3641 An assertion looks like this: 3642 3643 #PREDICATE (ANSWER) 3644 3645PREDICATE must be a single identifier. ANSWER can be any sequence of 3646tokens; all characters are significant except for leading and trailing 3647whitespace, and differences in internal whitespace sequences are 3648ignored. (This is similar to the rules governing macro redefinition.) 3649Thus, `(x + y)' is different from `(x+y)' but equivalent to 3650`( x + y )'. Parentheses do not nest inside an answer. 3651 3652 To test an assertion, you write it in an `#if'. For example, this 3653conditional succeeds if either `vax' or `ns16000' has been asserted as 3654an answer for `machine'. 3655 3656 #if #machine (vax) || #machine (ns16000) 3657 3658You can test whether _any_ answer is asserted for a predicate by 3659omitting the answer in the conditional: 3660 3661 #if #machine 3662 3663 Assertions are made with the `#assert' directive. Its sole argument 3664is the assertion to make, without the leading `#' that identifies 3665assertions in conditionals. 3666 3667 #assert PREDICATE (ANSWER) 3668 3669You may make several assertions with the same predicate and different 3670answers. Subsequent assertions do not override previous ones for the 3671same predicate. All the answers for any given predicate are 3672simultaneously true. 3673 3674 Assertions can be canceled with the `#unassert' directive. It has 3675the same syntax as `#assert'. In that form it cancels only the answer 3676which was specified on the `#unassert' line; other answers for that 3677predicate remain true. You can cancel an entire predicate by leaving 3678out the answer: 3679 3680 #unassert PREDICATE 3681 3682In either form, if no such assertion has been made, `#unassert' has no 3683effect. 3684 3685 You can also make or cancel assertions using command line options. 3686*Note Invocation::. 3687 3688 3689File: cpp.info, Node: Obsolete once-only headers, Prev: Assertions, Up: Obsolete Features 3690 369111.3.2 Obsolete once-only headers 3692--------------------------------- 3693 3694CPP supports two more ways of indicating that a header file should be 3695read only once. Neither one is as portable as a wrapper `#ifndef', and 3696we recommend you do not use them in new programs. 3697 3698 In the Objective-C language, there is a variant of `#include' called 3699`#import' which includes a file, but does so at most once. If you use 3700`#import' instead of `#include', then you don't need the conditionals 3701inside the header file to prevent multiple inclusion of the contents. 3702GCC permits the use of `#import' in C and C++ as well as Objective-C. 3703However, it is not in standard C or C++ and should therefore not be 3704used by portable programs. 3705 3706 `#import' is not a well designed feature. It requires the users of 3707a header file to know that it should only be included once. It is much 3708better for the header file's implementor to write the file so that users 3709don't need to know this. Using a wrapper `#ifndef' accomplishes this 3710goal. 3711 3712 In the present implementation, a single use of `#import' will 3713prevent the file from ever being read again, by either `#import' or 3714`#include'. You should not rely on this; do not use both `#import' and 3715`#include' to refer to the same header file. 3716 3717 Another way to prevent a header file from being included more than 3718once is with the `#pragma once' directive. If `#pragma once' is seen 3719when scanning a header file, that file will never be read again, no 3720matter what. 3721 3722 `#pragma once' does not have the problems that `#import' does, but 3723it is not recognized by all preprocessors, so you cannot rely on it in 3724a portable program. 3725 3726 3727File: cpp.info, Node: Differences from previous versions, Prev: Obsolete Features, Up: Implementation Details 3728 372911.4 Differences from previous versions 3730======================================= 3731 3732This section details behavior which has changed from previous versions 3733of CPP. We do not plan to change it again in the near future, but we 3734do not promise not to, either. 3735 3736 The "previous versions" discussed here are 2.95 and before. The 3737behavior of GCC 3.0 is mostly the same as the behavior of the widely 3738used 2.96 and 2.97 development snapshots. Where there are differences, 3739they generally represent bugs in the snapshots. 3740 3741 * -I- deprecated 3742 3743 This option has been deprecated in 4.0. `-iquote' is meant to 3744 replace the need for this option. 3745 3746 * Order of evaluation of `#' and `##' operators 3747 3748 The standard does not specify the order of evaluation of a chain of 3749 `##' operators, nor whether `#' is evaluated before, after, or at 3750 the same time as `##'. You should therefore not write any code 3751 which depends on any specific ordering. It is possible to 3752 guarantee an ordering, if you need one, by suitable use of nested 3753 macros. 3754 3755 An example of where this might matter is pasting the arguments `1', 3756 `e' and `-2'. This would be fine for left-to-right pasting, but 3757 right-to-left pasting would produce an invalid token `e-2'. 3758 3759 GCC 3.0 evaluates `#' and `##' at the same time and strictly left 3760 to right. Older versions evaluated all `#' operators first, then 3761 all `##' operators, in an unreliable order. 3762 3763 * The form of whitespace between tokens in preprocessor output 3764 3765 *Note Preprocessor Output::, for the current textual format. This 3766 is also the format used by stringification. Normally, the 3767 preprocessor communicates tokens directly to the compiler's 3768 parser, and whitespace does not come up at all. 3769 3770 Older versions of GCC preserved all whitespace provided by the 3771 user and inserted lots more whitespace of their own, because they 3772 could not accurately predict when extra spaces were needed to 3773 prevent accidental token pasting. 3774 3775 * Optional argument when invoking rest argument macros 3776 3777 As an extension, GCC permits you to omit the variable arguments 3778 entirely when you use a variable argument macro. This is 3779 forbidden by the 1999 C standard, and will provoke a pedantic 3780 warning with GCC 3.0. Previous versions accepted it silently. 3781 3782 * `##' swallowing preceding text in rest argument macros 3783 3784 Formerly, in a macro expansion, if `##' appeared before a variable 3785 arguments parameter, and the set of tokens specified for that 3786 argument in the macro invocation was empty, previous versions of 3787 CPP would back up and remove the preceding sequence of 3788 non-whitespace characters (*not* the preceding token). This 3789 extension is in direct conflict with the 1999 C standard and has 3790 been drastically pared back. 3791 3792 In the current version of the preprocessor, if `##' appears between 3793 a comma and a variable arguments parameter, and the variable 3794 argument is omitted entirely, the comma will be removed from the 3795 expansion. If the variable argument is empty, or the token before 3796 `##' is not a comma, then `##' behaves as a normal token paste. 3797 3798 * `#line' and `#include' 3799 3800 The `#line' directive used to change GCC's notion of the 3801 "directory containing the current file", used by `#include' with a 3802 double-quoted header file name. In 3.0 and later, it does not. 3803 *Note Line Control::, for further explanation. 3804 3805 * Syntax of `#line' 3806 3807 In GCC 2.95 and previous, the string constant argument to `#line' 3808 was treated the same way as the argument to `#include': backslash 3809 escapes were not honored, and the string ended at the second `"'. 3810 This is not compliant with the C standard. In GCC 3.0, an attempt 3811 was made to correct the behavior, so that the string was treated 3812 as a real string constant, but it turned out to be buggy. In 3.1, 3813 the bugs have been fixed. (We are not fixing the bugs in 3.0 3814 because they affect relatively few people and the fix is quite 3815 invasive.) 3816 3817 3818 3819File: cpp.info, Node: Invocation, Next: Environment Variables, Prev: Implementation Details, Up: Top 3820 382112 Invocation 3822************* 3823 3824Most often when you use the C preprocessor you will not have to invoke 3825it explicitly: the C compiler will do so automatically. However, the 3826preprocessor is sometimes useful on its own. All the options listed 3827here are also acceptable to the C compiler and have the same meaning, 3828except that the C compiler has different rules for specifying the output 3829file. 3830 3831 _Note:_ Whether you use the preprocessor by way of `gcc' or `cpp', 3832the "compiler driver" is run first. This program's purpose is to 3833translate your command into invocations of the programs that do the 3834actual work. Their command line interfaces are similar but not 3835identical to the documented interface, and may change without notice. 3836 3837 The C preprocessor expects two file names as arguments, INFILE and 3838OUTFILE. The preprocessor reads INFILE together with any other files 3839it specifies with `#include'. All the output generated by the combined 3840input files is written in OUTFILE. 3841 3842 Either INFILE or OUTFILE may be `-', which as INFILE means to read 3843from standard input and as OUTFILE means to write to standard output. 3844Also, if either file is omitted, it means the same as if `-' had been 3845specified for that file. 3846 3847 Unless otherwise noted, or the option ends in `=', all options which 3848take an argument may have that argument appear either immediately after 3849the option, or with a space between option and argument: `-Ifoo' and 3850`-I foo' have the same effect. 3851 3852 Many options have multi-letter names; therefore multiple 3853single-letter options may _not_ be grouped: `-dM' is very different from 3854`-d -M'. 3855 3856`-D NAME' 3857 Predefine NAME as a macro, with definition `1'. 3858 3859`-D NAME=DEFINITION' 3860 The contents of DEFINITION are tokenized and processed as if they 3861 appeared during translation phase three in a `#define' directive. 3862 In particular, the definition will be truncated by embedded 3863 newline characters. 3864 3865 If you are invoking the preprocessor from a shell or shell-like 3866 program you may need to use the shell's quoting syntax to protect 3867 characters such as spaces that have a meaning in the shell syntax. 3868 3869 If you wish to define a function-like macro on the command line, 3870 write its argument list with surrounding parentheses before the 3871 equals sign (if any). Parentheses are meaningful to most shells, 3872 so you will need to quote the option. With `sh' and `csh', 3873 `-D'NAME(ARGS...)=DEFINITION'' works. 3874 3875 `-D' and `-U' options are processed in the order they are given on 3876 the command line. All `-imacros FILE' and `-include FILE' options 3877 are processed after all `-D' and `-U' options. 3878 3879`-U NAME' 3880 Cancel any previous definition of NAME, either built in or 3881 provided with a `-D' option. 3882 3883`-undef' 3884 Do not predefine any system-specific or GCC-specific macros. The 3885 standard predefined macros remain defined. *Note Standard 3886 Predefined Macros::. 3887 3888`-I DIR' 3889 Add the directory DIR to the list of directories to be searched 3890 for header files. *Note Search Path::. Directories named by `-I' 3891 are searched before the standard system include directories. If 3892 the directory DIR is a standard system include directory, the 3893 option is ignored to ensure that the default search order for 3894 system directories and the special treatment of system headers are 3895 not defeated (*note System Headers::) . 3896 3897`-o FILE' 3898 Write output to FILE. This is the same as specifying FILE as the 3899 second non-option argument to `cpp'. `gcc' has a different 3900 interpretation of a second non-option argument, so you must use 3901 `-o' to specify the output file. 3902 3903`-Wall' 3904 Turns on all optional warnings which are desirable for normal code. 3905 At present this is `-Wcomment', `-Wtrigraphs', `-Wmultichar' and a 3906 warning about integer promotion causing a change of sign in `#if' 3907 expressions. Note that many of the preprocessor's warnings are on 3908 by default and have no options to control them. 3909 3910`-Wcomment' 3911`-Wcomments' 3912 Warn whenever a comment-start sequence `/*' appears in a `/*' 3913 comment, or whenever a backslash-newline appears in a `//' comment. 3914 (Both forms have the same effect.) 3915 3916`-Wtrigraphs' 3917 Most trigraphs in comments cannot affect the meaning of the 3918 program. However, a trigraph that would form an escaped newline 3919 (`??/' at the end of a line) can, by changing where the comment 3920 begins or ends. Therefore, only trigraphs that would form escaped 3921 newlines produce warnings inside a comment. 3922 3923 This option is implied by `-Wall'. If `-Wall' is not given, this 3924 option is still enabled unless trigraphs are enabled. To get 3925 trigraph conversion without warnings, but get the other `-Wall' 3926 warnings, use `-trigraphs -Wall -Wno-trigraphs'. 3927 3928`-Wtraditional' 3929 Warn about certain constructs that behave differently in 3930 traditional and ISO C. Also warn about ISO C constructs that have 3931 no traditional C equivalent, and problematic constructs which 3932 should be avoided. *Note Traditional Mode::. 3933 3934`-Wimport' 3935 Warn the first time `#import' is used. 3936 3937`-Wundef' 3938 Warn whenever an identifier which is not a macro is encountered in 3939 an `#if' directive, outside of `defined'. Such identifiers are 3940 replaced with zero. 3941 3942`-Wunused-macros' 3943 Warn about macros defined in the main file that are unused. A 3944 macro is "used" if it is expanded or tested for existence at least 3945 once. The preprocessor will also warn if the macro has not been 3946 used at the time it is redefined or undefined. 3947 3948 Built-in macros, macros defined on the command line, and macros 3949 defined in include files are not warned about. 3950 3951 _Note:_ If a macro is actually used, but only used in skipped 3952 conditional blocks, then CPP will report it as unused. To avoid 3953 the warning in such a case, you might improve the scope of the 3954 macro's definition by, for example, moving it into the first 3955 skipped block. Alternatively, you could provide a dummy use with 3956 something like: 3957 3958 #if defined the_macro_causing_the_warning 3959 #endif 3960 3961`-Wendif-labels' 3962 Warn whenever an `#else' or an `#endif' are followed by text. 3963 This usually happens in code of the form 3964 3965 #if FOO 3966 ... 3967 #else FOO 3968 ... 3969 #endif FOO 3970 3971 The second and third `FOO' should be in comments, but often are not 3972 in older programs. This warning is on by default. 3973 3974`-Werror' 3975 Make all warnings into hard errors. Source code which triggers 3976 warnings will be rejected. 3977 3978`-Wsystem-headers' 3979 Issue warnings for code in system headers. These are normally 3980 unhelpful in finding bugs in your own code, therefore suppressed. 3981 If you are responsible for the system library, you may want to see 3982 them. 3983 3984`-w' 3985 Suppress all warnings, including those which GNU CPP issues by 3986 default. 3987 3988`-pedantic' 3989 Issue all the mandatory diagnostics listed in the C standard. 3990 Some of them are left out by default, since they trigger 3991 frequently on harmless code. 3992 3993`-pedantic-errors' 3994 Issue all the mandatory diagnostics, and make all mandatory 3995 diagnostics into errors. This includes mandatory diagnostics that 3996 GCC issues without `-pedantic' but treats as warnings. 3997 3998`-M' 3999 Instead of outputting the result of preprocessing, output a rule 4000 suitable for `make' describing the dependencies of the main source 4001 file. The preprocessor outputs one `make' rule containing the 4002 object file name for that source file, a colon, and the names of 4003 all the included files, including those coming from `-include' or 4004 `-imacros' command line options. 4005 4006 Unless specified explicitly (with `-MT' or `-MQ'), the object file 4007 name consists of the basename of the source file with any suffix 4008 replaced with object file suffix. If there are many included 4009 files then the rule is split into several lines using `\'-newline. 4010 The rule has no commands. 4011 4012 This option does not suppress the preprocessor's debug output, 4013 such as `-dM'. To avoid mixing such debug output with the 4014 dependency rules you should explicitly specify the dependency 4015 output file with `-MF', or use an environment variable like 4016 `DEPENDENCIES_OUTPUT' (*note Environment Variables::). Debug 4017 output will still be sent to the regular output stream as normal. 4018 4019 Passing `-M' to the driver implies `-E', and suppresses warnings 4020 with an implicit `-w'. 4021 4022`-MM' 4023 Like `-M' but do not mention header files that are found in system 4024 header directories, nor header files that are included, directly 4025 or indirectly, from such a header. 4026 4027 This implies that the choice of angle brackets or double quotes in 4028 an `#include' directive does not in itself determine whether that 4029 header will appear in `-MM' dependency output. This is a slight 4030 change in semantics from GCC versions 3.0 and earlier. 4031 4032`-MF FILE' 4033 When used with `-M' or `-MM', specifies a file to write the 4034 dependencies to. If no `-MF' switch is given the preprocessor 4035 sends the rules to the same place it would have sent preprocessed 4036 output. 4037 4038 When used with the driver options `-MD' or `-MMD', `-MF' overrides 4039 the default dependency output file. 4040 4041`-MG' 4042 In conjunction with an option such as `-M' requesting dependency 4043 generation, `-MG' assumes missing header files are generated files 4044 and adds them to the dependency list without raising an error. 4045 The dependency filename is taken directly from the `#include' 4046 directive without prepending any path. `-MG' also suppresses 4047 preprocessed output, as a missing header file renders this useless. 4048 4049 This feature is used in automatic updating of makefiles. 4050 4051`-MP' 4052 This option instructs CPP to add a phony target for each dependency 4053 other than the main file, causing each to depend on nothing. These 4054 dummy rules work around errors `make' gives if you remove header 4055 files without updating the `Makefile' to match. 4056 4057 This is typical output: 4058 4059 test.o: test.c test.h 4060 4061 test.h: 4062 4063`-MT TARGET' 4064 Change the target of the rule emitted by dependency generation. By 4065 default CPP takes the name of the main input file, including any 4066 path, deletes any file suffix such as `.c', and appends the 4067 platform's usual object suffix. The result is the target. 4068 4069 An `-MT' option will set the target to be exactly the string you 4070 specify. If you want multiple targets, you can specify them as a 4071 single argument to `-MT', or use multiple `-MT' options. 4072 4073 For example, `-MT '$(objpfx)foo.o'' might give 4074 4075 $(objpfx)foo.o: foo.c 4076 4077`-MQ TARGET' 4078 Same as `-MT', but it quotes any characters which are special to 4079 Make. `-MQ '$(objpfx)foo.o'' gives 4080 4081 $$(objpfx)foo.o: foo.c 4082 4083 The default target is automatically quoted, as if it were given 4084 with `-MQ'. 4085 4086`-MD' 4087 `-MD' is equivalent to `-M -MF FILE', except that `-E' is not 4088 implied. The driver determines FILE based on whether an `-o' 4089 option is given. If it is, the driver uses its argument but with 4090 a suffix of `.d', otherwise it take the basename of the input file 4091 and applies a `.d' suffix. 4092 4093 If `-MD' is used in conjunction with `-E', any `-o' switch is 4094 understood to specify the dependency output file (*note -MF: 4095 dashMF.), but if used without `-E', each `-o' is understood to 4096 specify a target object file. 4097 4098 Since `-E' is not implied, `-MD' can be used to generate a 4099 dependency output file as a side-effect of the compilation process. 4100 4101`-MMD' 4102 Like `-MD' except mention only user header files, not system 4103 header files. 4104 4105`-x c' 4106`-x c++' 4107`-x objective-c' 4108`-x assembler-with-cpp' 4109 Specify the source language: C, C++, Objective-C, or assembly. 4110 This has nothing to do with standards conformance or extensions; 4111 it merely selects which base syntax to expect. If you give none 4112 of these options, cpp will deduce the language from the extension 4113 of the source file: `.c', `.cc', `.m', or `.S'. Some other common 4114 extensions for C++ and assembly are also recognized. If cpp does 4115 not recognize the extension, it will treat the file as C; this is 4116 the most generic mode. 4117 4118 _Note:_ Previous versions of cpp accepted a `-lang' option which 4119 selected both the language and the standards conformance level. 4120 This option has been removed, because it conflicts with the `-l' 4121 option. 4122 4123`-std=STANDARD' 4124`-ansi' 4125 Specify the standard to which the code should conform. Currently 4126 CPP knows about C and C++ standards; others may be added in the 4127 future. 4128 4129 STANDARD may be one of: 4130 `iso9899:1990' 4131 `c89' 4132 The ISO C standard from 1990. `c89' is the customary 4133 shorthand for this version of the standard. 4134 4135 The `-ansi' option is equivalent to `-std=c89'. 4136 4137 `iso9899:199409' 4138 The 1990 C standard, as amended in 1994. 4139 4140 `iso9899:1999' 4141 `c99' 4142 `iso9899:199x' 4143 `c9x' 4144 The revised ISO C standard, published in December 1999. 4145 Before publication, this was known as C9X. 4146 4147 `gnu89' 4148 The 1990 C standard plus GNU extensions. This is the default. 4149 4150 `gnu99' 4151 `gnu9x' 4152 The 1999 C standard plus GNU extensions. 4153 4154 `c++98' 4155 The 1998 ISO C++ standard plus amendments. 4156 4157 `gnu++98' 4158 The same as `-std=c++98' plus GNU extensions. This is the 4159 default for C++ code. 4160 4161`-I-' 4162 Split the include path. Any directories specified with `-I' 4163 options before `-I-' are searched only for headers requested with 4164 `#include "FILE"'; they are not searched for `#include <FILE>'. 4165 If additional directories are specified with `-I' options after 4166 the `-I-', those directories are searched for all `#include' 4167 directives. 4168 4169 In addition, `-I-' inhibits the use of the directory of the current 4170 file directory as the first search directory for `#include "FILE"'. 4171 *Note Search Path::. This option has been deprecated. 4172 4173`-nostdinc' 4174 Do not search the standard system directories for header files. 4175 Only the directories you have specified with `-I' options (and the 4176 directory of the current file, if appropriate) are searched. 4177 4178`-nostdinc++' 4179 Do not search for header files in the C++-specific standard 4180 directories, but do still search the other standard directories. 4181 (This option is used when building the C++ library.) 4182 4183`-include FILE' 4184 Process FILE as if `#include "file"' appeared as the first line of 4185 the primary source file. However, the first directory searched 4186 for FILE is the preprocessor's working directory _instead of_ the 4187 directory containing the main source file. If not found there, it 4188 is searched for in the remainder of the `#include "..."' search 4189 chain as normal. 4190 4191 If multiple `-include' options are given, the files are included 4192 in the order they appear on the command line. 4193 4194`-imacros FILE' 4195 Exactly like `-include', except that any output produced by 4196 scanning FILE is thrown away. Macros it defines remain defined. 4197 This allows you to acquire all the macros from a header without 4198 also processing its declarations. 4199 4200 All files specified by `-imacros' are processed before all files 4201 specified by `-include'. 4202 4203`-idirafter DIR' 4204 Search DIR for header files, but do it _after_ all directories 4205 specified with `-I' and the standard system directories have been 4206 exhausted. DIR is treated as a system include directory. 4207 4208`-iprefix PREFIX' 4209 Specify PREFIX as the prefix for subsequent `-iwithprefix' 4210 options. If the prefix represents a directory, you should include 4211 the final `/'. 4212 4213`-iwithprefix DIR' 4214`-iwithprefixbefore DIR' 4215 Append DIR to the prefix specified previously with `-iprefix', and 4216 add the resulting directory to the include search path. 4217 `-iwithprefixbefore' puts it in the same place `-I' would; 4218 `-iwithprefix' puts it where `-idirafter' would. 4219 4220`-isysroot DIR' 4221 This option is like the `--sysroot' option, but applies only to 4222 header files. See the `--sysroot' option for more information. 4223 4224`-imultilib DIR' 4225 Use DIR as a subdirectory of the directory containing 4226 target-specific C++ headers. 4227 4228`-isystem DIR' 4229 Search DIR for header files, after all directories specified by 4230 `-I' but before the standard system directories. Mark it as a 4231 system directory, so that it gets the same special treatment as is 4232 applied to the standard system directories. *Note System 4233 Headers::. 4234 4235`-iquote DIR' 4236 Search DIR only for header files requested with `#include "FILE"'; 4237 they are not searched for `#include <FILE>', before all 4238 directories specified by `-I' and before the standard system 4239 directories. *Note Search Path::. 4240 4241`-fdollars-in-identifiers' 4242 Accept `$' in identifiers. *Note Identifier characters::. 4243 4244`-fextended-identifiers' 4245 Accept universal character names in identifiers. This option is 4246 experimental; in a future version of GCC, it will be enabled by 4247 default for C99 and C++. 4248 4249`-fpreprocessed' 4250 Indicate to the preprocessor that the input file has already been 4251 preprocessed. This suppresses things like macro expansion, 4252 trigraph conversion, escaped newline splicing, and processing of 4253 most directives. The preprocessor still recognizes and removes 4254 comments, so that you can pass a file preprocessed with `-C' to 4255 the compiler without problems. In this mode the integrated 4256 preprocessor is little more than a tokenizer for the front ends. 4257 4258 `-fpreprocessed' is implicit if the input file has one of the 4259 extensions `.i', `.ii' or `.mi'. These are the extensions that 4260 GCC uses for preprocessed files created by `-save-temps'. 4261 4262`-ftabstop=WIDTH' 4263 Set the distance between tab stops. This helps the preprocessor 4264 report correct column numbers in warnings or errors, even if tabs 4265 appear on the line. If the value is less than 1 or greater than 4266 100, the option is ignored. The default is 8. 4267 4268`-fexec-charset=CHARSET' 4269 Set the execution character set, used for string and character 4270 constants. The default is UTF-8. CHARSET can be any encoding 4271 supported by the system's `iconv' library routine. 4272 4273`-fwide-exec-charset=CHARSET' 4274 Set the wide execution character set, used for wide string and 4275 character constants. The default is UTF-32 or UTF-16, whichever 4276 corresponds to the width of `wchar_t'. As with `-fexec-charset', 4277 CHARSET can be any encoding supported by the system's `iconv' 4278 library routine; however, you will have problems with encodings 4279 that do not fit exactly in `wchar_t'. 4280 4281`-finput-charset=CHARSET' 4282 Set the input character set, used for translation from the 4283 character set of the input file to the source character set used 4284 by GCC. If the locale does not specify, or GCC cannot get this 4285 information from the locale, the default is UTF-8. This can be 4286 overridden by either the locale or this command line option. 4287 Currently the command line option takes precedence if there's a 4288 conflict. CHARSET can be any encoding supported by the system's 4289 `iconv' library routine. 4290 4291`-fworking-directory' 4292 Enable generation of linemarkers in the preprocessor output that 4293 will let the compiler know the current working directory at the 4294 time of preprocessing. When this option is enabled, the 4295 preprocessor will emit, after the initial linemarker, a second 4296 linemarker with the current working directory followed by two 4297 slashes. GCC will use this directory, when it's present in the 4298 preprocessed input, as the directory emitted as the current 4299 working directory in some debugging information formats. This 4300 option is implicitly enabled if debugging information is enabled, 4301 but this can be inhibited with the negated form 4302 `-fno-working-directory'. If the `-P' flag is present in the 4303 command line, this option has no effect, since no `#line' 4304 directives are emitted whatsoever. 4305 4306`-fno-show-column' 4307 Do not print column numbers in diagnostics. This may be necessary 4308 if diagnostics are being scanned by a program that does not 4309 understand the column numbers, such as `dejagnu'. 4310 4311`-A PREDICATE=ANSWER' 4312 Make an assertion with the predicate PREDICATE and answer ANSWER. 4313 This form is preferred to the older form `-A PREDICATE(ANSWER)', 4314 which is still supported, because it does not use shell special 4315 characters. *Note Assertions::. 4316 4317`-A -PREDICATE=ANSWER' 4318 Cancel an assertion with the predicate PREDICATE and answer ANSWER. 4319 4320`-dCHARS' 4321 CHARS is a sequence of one or more of the following characters, 4322 and must not be preceded by a space. Other characters are 4323 interpreted by the compiler proper, or reserved for future 4324 versions of GCC, and so are silently ignored. If you specify 4325 characters whose behavior conflicts, the result is undefined. 4326 4327 `M' 4328 Instead of the normal output, generate a list of `#define' 4329 directives for all the macros defined during the execution of 4330 the preprocessor, including predefined macros. This gives 4331 you a way of finding out what is predefined in your version 4332 of the preprocessor. Assuming you have no file `foo.h', the 4333 command 4334 4335 touch foo.h; cpp -dM foo.h 4336 4337 will show all the predefined macros. 4338 4339 `D' 4340 Like `M' except in two respects: it does _not_ include the 4341 predefined macros, and it outputs _both_ the `#define' 4342 directives and the result of preprocessing. Both kinds of 4343 output go to the standard output file. 4344 4345 `N' 4346 Like `D', but emit only the macro names, not their expansions. 4347 4348 `I' 4349 Output `#include' directives in addition to the result of 4350 preprocessing. 4351 4352`-P' 4353 Inhibit generation of linemarkers in the output from the 4354 preprocessor. This might be useful when running the preprocessor 4355 on something that is not C code, and will be sent to a program 4356 which might be confused by the linemarkers. *Note Preprocessor 4357 Output::. 4358 4359`-C' 4360 Do not discard comments. All comments are passed through to the 4361 output file, except for comments in processed directives, which 4362 are deleted along with the directive. 4363 4364 You should be prepared for side effects when using `-C'; it causes 4365 the preprocessor to treat comments as tokens in their own right. 4366 For example, comments appearing at the start of what would be a 4367 directive line have the effect of turning that line into an 4368 ordinary source line, since the first token on the line is no 4369 longer a `#'. 4370 4371`-CC' 4372 Do not discard comments, including during macro expansion. This is 4373 like `-C', except that comments contained within macros are also 4374 passed through to the output file where the macro is expanded. 4375 4376 In addition to the side-effects of the `-C' option, the `-CC' 4377 option causes all C++-style comments inside a macro to be 4378 converted to C-style comments. This is to prevent later use of 4379 that macro from inadvertently commenting out the remainder of the 4380 source line. 4381 4382 The `-CC' option is generally used to support lint comments. 4383 4384`-traditional-cpp' 4385 Try to imitate the behavior of old-fashioned C preprocessors, as 4386 opposed to ISO C preprocessors. *Note Traditional Mode::. 4387 4388`-trigraphs' 4389 Process trigraph sequences. *Note Initial processing::. 4390 4391`-remap' 4392 Enable special code to work around file systems which only permit 4393 very short file names, such as MS-DOS. 4394 4395`--help' 4396`--target-help' 4397 Print text describing all the command line options instead of 4398 preprocessing anything. 4399 4400`-v' 4401 Verbose mode. Print out GNU CPP's version number at the beginning 4402 of execution, and report the final form of the include path. 4403 4404`-H' 4405 Print the name of each header file used, in addition to other 4406 normal activities. Each name is indented to show how deep in the 4407 `#include' stack it is. Precompiled header files are also 4408 printed, even if they are found to be invalid; an invalid 4409 precompiled header file is printed with `...x' and a valid one 4410 with `...!' . 4411 4412`-version' 4413`--version' 4414 Print out GNU CPP's version number. With one dash, proceed to 4415 preprocess as normal. With two dashes, exit immediately. 4416 4417 4418File: cpp.info, Node: Environment Variables, Next: GNU Free Documentation License, Prev: Invocation, Up: Top 4419 442013 Environment Variables 4421************************ 4422 4423This section describes the environment variables that affect how CPP 4424operates. You can use them to specify directories or prefixes to use 4425when searching for include files, or to control dependency output. 4426 4427 Note that you can also specify places to search using options such as 4428`-I', and control dependency output with options like `-M' (*note 4429Invocation::). These take precedence over environment variables, which 4430in turn take precedence over the configuration of GCC. 4431 4432`CPATH' 4433`C_INCLUDE_PATH' 4434`CPLUS_INCLUDE_PATH' 4435`OBJC_INCLUDE_PATH' 4436 Each variable's value is a list of directories separated by a 4437 special character, much like `PATH', in which to look for header 4438 files. The special character, `PATH_SEPARATOR', is 4439 target-dependent and determined at GCC build time. For Microsoft 4440 Windows-based targets it is a semicolon, and for almost all other 4441 targets it is a colon. 4442 4443 `CPATH' specifies a list of directories to be searched as if 4444 specified with `-I', but after any paths given with `-I' options 4445 on the command line. This environment variable is used regardless 4446 of which language is being preprocessed. 4447 4448 The remaining environment variables apply only when preprocessing 4449 the particular language indicated. Each specifies a list of 4450 directories to be searched as if specified with `-isystem', but 4451 after any paths given with `-isystem' options on the command line. 4452 4453 In all these variables, an empty element instructs the compiler to 4454 search its current working directory. Empty elements can appear 4455 at the beginning or end of a path. For instance, if the value of 4456 `CPATH' is `:/special/include', that has the same effect as 4457 `-I. -I/special/include'. 4458 4459 See also *Note Search Path::. 4460 4461`DEPENDENCIES_OUTPUT' 4462 If this variable is set, its value specifies how to output 4463 dependencies for Make based on the non-system header files 4464 processed by the compiler. System header files are ignored in the 4465 dependency output. 4466 4467 The value of `DEPENDENCIES_OUTPUT' can be just a file name, in 4468 which case the Make rules are written to that file, guessing the 4469 target name from the source file name. Or the value can have the 4470 form `FILE TARGET', in which case the rules are written to file 4471 FILE using TARGET as the target name. 4472 4473 In other words, this environment variable is equivalent to 4474 combining the options `-MM' and `-MF' (*note Invocation::), with 4475 an optional `-MT' switch too. 4476 4477`SUNPRO_DEPENDENCIES' 4478 This variable is the same as `DEPENDENCIES_OUTPUT' (see above), 4479 except that system header files are not ignored, so it implies 4480 `-M' rather than `-MM'. However, the dependence on the main input 4481 file is omitted. *Note Invocation::. 4482 4483 4484File: cpp.info, Node: GNU Free Documentation License, Next: Index of Directives, Prev: Environment Variables, Up: Top 4485 4486GNU Free Documentation License 4487****************************** 4488 4489 Version 1.2, November 2002 4490 4491 Copyright (C) 2000,2001,2002 Free Software Foundation, Inc. 4492 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA 4493 4494 Everyone is permitted to copy and distribute verbatim copies 4495 of this license document, but changing it is not allowed. 4496 4497 0. PREAMBLE 4498 4499 The purpose of this License is to make a manual, textbook, or other 4500 functional and useful document "free" in the sense of freedom: to 4501 assure everyone the effective freedom to copy and redistribute it, 4502 with or without modifying it, either commercially or 4503 noncommercially. Secondarily, this License preserves for the 4504 author and publisher a way to get credit for their work, while not 4505 being considered responsible for modifications made by others. 4506 4507 This License is a kind of "copyleft", which means that derivative 4508 works of the document must themselves be free in the same sense. 4509 It complements the GNU General Public License, which is a copyleft 4510 license designed for free software. 4511 4512 We have designed this License in order to use it for manuals for 4513 free software, because free software needs free documentation: a 4514 free program should come with manuals providing the same freedoms 4515 that the software does. But this License is not limited to 4516 software manuals; it can be used for any textual work, regardless 4517 of subject matter or whether it is published as a printed book. 4518 We recommend this License principally for works whose purpose is 4519 instruction or reference. 4520 4521 1. APPLICABILITY AND DEFINITIONS 4522 4523 This License applies to any manual or other work, in any medium, 4524 that contains a notice placed by the copyright holder saying it 4525 can be distributed under the terms of this License. Such a notice 4526 grants a world-wide, royalty-free license, unlimited in duration, 4527 to use that work under the conditions stated herein. The 4528 "Document", below, refers to any such manual or work. Any member 4529 of the public is a licensee, and is addressed as "you". You 4530 accept the license if you copy, modify or distribute the work in a 4531 way requiring permission under copyright law. 4532 4533 A "Modified Version" of the Document means any work containing the 4534 Document or a portion of it, either copied verbatim, or with 4535 modifications and/or translated into another language. 4536 4537 A "Secondary Section" is a named appendix or a front-matter section 4538 of the Document that deals exclusively with the relationship of the 4539 publishers or authors of the Document to the Document's overall 4540 subject (or to related matters) and contains nothing that could 4541 fall directly within that overall subject. (Thus, if the Document 4542 is in part a textbook of mathematics, a Secondary Section may not 4543 explain any mathematics.) The relationship could be a matter of 4544 historical connection with the subject or with related matters, or 4545 of legal, commercial, philosophical, ethical or political position 4546 regarding them. 4547 4548 The "Invariant Sections" are certain Secondary Sections whose 4549 titles are designated, as being those of Invariant Sections, in 4550 the notice that says that the Document is released under this 4551 License. If a section does not fit the above definition of 4552 Secondary then it is not allowed to be designated as Invariant. 4553 The Document may contain zero Invariant Sections. If the Document 4554 does not identify any Invariant Sections then there are none. 4555 4556 The "Cover Texts" are certain short passages of text that are 4557 listed, as Front-Cover Texts or Back-Cover Texts, in the notice 4558 that says that the Document is released under this License. A 4559 Front-Cover Text may be at most 5 words, and a Back-Cover Text may 4560 be at most 25 words. 4561 4562 A "Transparent" copy of the Document means a machine-readable copy, 4563 represented in a format whose specification is available to the 4564 general public, that is suitable for revising the document 4565 straightforwardly with generic text editors or (for images 4566 composed of pixels) generic paint programs or (for drawings) some 4567 widely available drawing editor, and that is suitable for input to 4568 text formatters or for automatic translation to a variety of 4569 formats suitable for input to text formatters. A copy made in an 4570 otherwise Transparent file format whose markup, or absence of 4571 markup, has been arranged to thwart or discourage subsequent 4572 modification by readers is not Transparent. An image format is 4573 not Transparent if used for any substantial amount of text. A 4574 copy that is not "Transparent" is called "Opaque". 4575 4576 Examples of suitable formats for Transparent copies include plain 4577 ASCII without markup, Texinfo input format, LaTeX input format, 4578 SGML or XML using a publicly available DTD, and 4579 standard-conforming simple HTML, PostScript or PDF designed for 4580 human modification. Examples of transparent image formats include 4581 PNG, XCF and JPG. Opaque formats include proprietary formats that 4582 can be read and edited only by proprietary word processors, SGML or 4583 XML for which the DTD and/or processing tools are not generally 4584 available, and the machine-generated HTML, PostScript or PDF 4585 produced by some word processors for output purposes only. 4586 4587 The "Title Page" means, for a printed book, the title page itself, 4588 plus such following pages as are needed to hold, legibly, the 4589 material this License requires to appear in the title page. For 4590 works in formats which do not have any title page as such, "Title 4591 Page" means the text near the most prominent appearance of the 4592 work's title, preceding the beginning of the body of the text. 4593 4594 A section "Entitled XYZ" means a named subunit of the Document 4595 whose title either is precisely XYZ or contains XYZ in parentheses 4596 following text that translates XYZ in another language. (Here XYZ 4597 stands for a specific section name mentioned below, such as 4598 "Acknowledgements", "Dedications", "Endorsements", or "History".) 4599 To "Preserve the Title" of such a section when you modify the 4600 Document means that it remains a section "Entitled XYZ" according 4601 to this definition. 4602 4603 The Document may include Warranty Disclaimers next to the notice 4604 which states that this License applies to the Document. These 4605 Warranty Disclaimers are considered to be included by reference in 4606 this License, but only as regards disclaiming warranties: any other 4607 implication that these Warranty Disclaimers may have is void and 4608 has no effect on the meaning of this License. 4609 4610 2. VERBATIM COPYING 4611 4612 You may copy and distribute the Document in any medium, either 4613 commercially or noncommercially, provided that this License, the 4614 copyright notices, and the license notice saying this License 4615 applies to the Document are reproduced in all copies, and that you 4616 add no other conditions whatsoever to those of this License. You 4617 may not use technical measures to obstruct or control the reading 4618 or further copying of the copies you make or distribute. However, 4619 you may accept compensation in exchange for copies. If you 4620 distribute a large enough number of copies you must also follow 4621 the conditions in section 3. 4622 4623 You may also lend copies, under the same conditions stated above, 4624 and you may publicly display copies. 4625 4626 3. COPYING IN QUANTITY 4627 4628 If you publish printed copies (or copies in media that commonly 4629 have printed covers) of the Document, numbering more than 100, and 4630 the Document's license notice requires Cover Texts, you must 4631 enclose the copies in covers that carry, clearly and legibly, all 4632 these Cover Texts: Front-Cover Texts on the front cover, and 4633 Back-Cover Texts on the back cover. Both covers must also clearly 4634 and legibly identify you as the publisher of these copies. The 4635 front cover must present the full title with all words of the 4636 title equally prominent and visible. You may add other material 4637 on the covers in addition. Copying with changes limited to the 4638 covers, as long as they preserve the title of the Document and 4639 satisfy these conditions, can be treated as verbatim copying in 4640 other respects. 4641 4642 If the required texts for either cover are too voluminous to fit 4643 legibly, you should put the first ones listed (as many as fit 4644 reasonably) on the actual cover, and continue the rest onto 4645 adjacent pages. 4646 4647 If you publish or distribute Opaque copies of the Document 4648 numbering more than 100, you must either include a 4649 machine-readable Transparent copy along with each Opaque copy, or 4650 state in or with each Opaque copy a computer-network location from 4651 which the general network-using public has access to download 4652 using public-standard network protocols a complete Transparent 4653 copy of the Document, free of added material. If you use the 4654 latter option, you must take reasonably prudent steps, when you 4655 begin distribution of Opaque copies in quantity, to ensure that 4656 this Transparent copy will remain thus accessible at the stated 4657 location until at least one year after the last time you 4658 distribute an Opaque copy (directly or through your agents or 4659 retailers) of that edition to the public. 4660 4661 It is requested, but not required, that you contact the authors of 4662 the Document well before redistributing any large number of 4663 copies, to give them a chance to provide you with an updated 4664 version of the Document. 4665 4666 4. MODIFICATIONS 4667 4668 You may copy and distribute a Modified Version of the Document 4669 under the conditions of sections 2 and 3 above, provided that you 4670 release the Modified Version under precisely this License, with 4671 the Modified Version filling the role of the Document, thus 4672 licensing distribution and modification of the Modified Version to 4673 whoever possesses a copy of it. In addition, you must do these 4674 things in the Modified Version: 4675 4676 A. Use in the Title Page (and on the covers, if any) a title 4677 distinct from that of the Document, and from those of 4678 previous versions (which should, if there were any, be listed 4679 in the History section of the Document). You may use the 4680 same title as a previous version if the original publisher of 4681 that version gives permission. 4682 4683 B. List on the Title Page, as authors, one or more persons or 4684 entities responsible for authorship of the modifications in 4685 the Modified Version, together with at least five of the 4686 principal authors of the Document (all of its principal 4687 authors, if it has fewer than five), unless they release you 4688 from this requirement. 4689 4690 C. State on the Title page the name of the publisher of the 4691 Modified Version, as the publisher. 4692 4693 D. Preserve all the copyright notices of the Document. 4694 4695 E. Add an appropriate copyright notice for your modifications 4696 adjacent to the other copyright notices. 4697 4698 F. Include, immediately after the copyright notices, a license 4699 notice giving the public permission to use the Modified 4700 Version under the terms of this License, in the form shown in 4701 the Addendum below. 4702 4703 G. Preserve in that license notice the full lists of Invariant 4704 Sections and required Cover Texts given in the Document's 4705 license notice. 4706 4707 H. Include an unaltered copy of this License. 4708 4709 I. Preserve the section Entitled "History", Preserve its Title, 4710 and add to it an item stating at least the title, year, new 4711 authors, and publisher of the Modified Version as given on 4712 the Title Page. If there is no section Entitled "History" in 4713 the Document, create one stating the title, year, authors, 4714 and publisher of the Document as given on its Title Page, 4715 then add an item describing the Modified Version as stated in 4716 the previous sentence. 4717 4718 J. Preserve the network location, if any, given in the Document 4719 for public access to a Transparent copy of the Document, and 4720 likewise the network locations given in the Document for 4721 previous versions it was based on. These may be placed in 4722 the "History" section. You may omit a network location for a 4723 work that was published at least four years before the 4724 Document itself, or if the original publisher of the version 4725 it refers to gives permission. 4726 4727 K. For any section Entitled "Acknowledgements" or "Dedications", 4728 Preserve the Title of the section, and preserve in the 4729 section all the substance and tone of each of the contributor 4730 acknowledgements and/or dedications given therein. 4731 4732 L. Preserve all the Invariant Sections of the Document, 4733 unaltered in their text and in their titles. Section numbers 4734 or the equivalent are not considered part of the section 4735 titles. 4736 4737 M. Delete any section Entitled "Endorsements". Such a section 4738 may not be included in the Modified Version. 4739 4740 N. Do not retitle any existing section to be Entitled 4741 "Endorsements" or to conflict in title with any Invariant 4742 Section. 4743 4744 O. Preserve any Warranty Disclaimers. 4745 4746 If the Modified Version includes new front-matter sections or 4747 appendices that qualify as Secondary Sections and contain no 4748 material copied from the Document, you may at your option 4749 designate some or all of these sections as invariant. To do this, 4750 add their titles to the list of Invariant Sections in the Modified 4751 Version's license notice. These titles must be distinct from any 4752 other section titles. 4753 4754 You may add a section Entitled "Endorsements", provided it contains 4755 nothing but endorsements of your Modified Version by various 4756 parties--for example, statements of peer review or that the text 4757 has been approved by an organization as the authoritative 4758 definition of a standard. 4759 4760 You may add a passage of up to five words as a Front-Cover Text, 4761 and a passage of up to 25 words as a Back-Cover Text, to the end 4762 of the list of Cover Texts in the Modified Version. Only one 4763 passage of Front-Cover Text and one of Back-Cover Text may be 4764 added by (or through arrangements made by) any one entity. If the 4765 Document already includes a cover text for the same cover, 4766 previously added by you or by arrangement made by the same entity 4767 you are acting on behalf of, you may not add another; but you may 4768 replace the old one, on explicit permission from the previous 4769 publisher that added the old one. 4770 4771 The author(s) and publisher(s) of the Document do not by this 4772 License give permission to use their names for publicity for or to 4773 assert or imply endorsement of any Modified Version. 4774 4775 5. COMBINING DOCUMENTS 4776 4777 You may combine the Document with other documents released under 4778 this License, under the terms defined in section 4 above for 4779 modified versions, provided that you include in the combination 4780 all of the Invariant Sections of all of the original documents, 4781 unmodified, and list them all as Invariant Sections of your 4782 combined work in its license notice, and that you preserve all 4783 their Warranty Disclaimers. 4784 4785 The combined work need only contain one copy of this License, and 4786 multiple identical Invariant Sections may be replaced with a single 4787 copy. If there are multiple Invariant Sections with the same name 4788 but different contents, make the title of each such section unique 4789 by adding at the end of it, in parentheses, the name of the 4790 original author or publisher of that section if known, or else a 4791 unique number. Make the same adjustment to the section titles in 4792 the list of Invariant Sections in the license notice of the 4793 combined work. 4794 4795 In the combination, you must combine any sections Entitled 4796 "History" in the various original documents, forming one section 4797 Entitled "History"; likewise combine any sections Entitled 4798 "Acknowledgements", and any sections Entitled "Dedications". You 4799 must delete all sections Entitled "Endorsements." 4800 4801 6. COLLECTIONS OF DOCUMENTS 4802 4803 You may make a collection consisting of the Document and other 4804 documents released under this License, and replace the individual 4805 copies of this License in the various documents with a single copy 4806 that is included in the collection, provided that you follow the 4807 rules of this License for verbatim copying of each of the 4808 documents in all other respects. 4809 4810 You may extract a single document from such a collection, and 4811 distribute it individually under this License, provided you insert 4812 a copy of this License into the extracted document, and follow 4813 this License in all other respects regarding verbatim copying of 4814 that document. 4815 4816 7. AGGREGATION WITH INDEPENDENT WORKS 4817 4818 A compilation of the Document or its derivatives with other 4819 separate and independent documents or works, in or on a volume of 4820 a storage or distribution medium, is called an "aggregate" if the 4821 copyright resulting from the compilation is not used to limit the 4822 legal rights of the compilation's users beyond what the individual 4823 works permit. When the Document is included in an aggregate, this 4824 License does not apply to the other works in the aggregate which 4825 are not themselves derivative works of the Document. 4826 4827 If the Cover Text requirement of section 3 is applicable to these 4828 copies of the Document, then if the Document is less than one half 4829 of the entire aggregate, the Document's Cover Texts may be placed 4830 on covers that bracket the Document within the aggregate, or the 4831 electronic equivalent of covers if the Document is in electronic 4832 form. Otherwise they must appear on printed covers that bracket 4833 the whole aggregate. 4834 4835 8. TRANSLATION 4836 4837 Translation is considered a kind of modification, so you may 4838 distribute translations of the Document under the terms of section 4839 4. Replacing Invariant Sections with translations requires special 4840 permission from their copyright holders, but you may include 4841 translations of some or all Invariant Sections in addition to the 4842 original versions of these Invariant Sections. You may include a 4843 translation of this License, and all the license notices in the 4844 Document, and any Warranty Disclaimers, provided that you also 4845 include the original English version of this License and the 4846 original versions of those notices and disclaimers. In case of a 4847 disagreement between the translation and the original version of 4848 this License or a notice or disclaimer, the original version will 4849 prevail. 4850 4851 If a section in the Document is Entitled "Acknowledgements", 4852 "Dedications", or "History", the requirement (section 4) to 4853 Preserve its Title (section 1) will typically require changing the 4854 actual title. 4855 4856 9. TERMINATION 4857 4858 You may not copy, modify, sublicense, or distribute the Document 4859 except as expressly provided for under this License. Any other 4860 attempt to copy, modify, sublicense or distribute the Document is 4861 void, and will automatically terminate your rights under this 4862 License. However, parties who have received copies, or rights, 4863 from you under this License will not have their licenses 4864 terminated so long as such parties remain in full compliance. 4865 4866 10. FUTURE REVISIONS OF THIS LICENSE 4867 4868 The Free Software Foundation may publish new, revised versions of 4869 the GNU Free Documentation License from time to time. Such new 4870 versions will be similar in spirit to the present version, but may 4871 differ in detail to address new problems or concerns. See 4872 `http://www.gnu.org/copyleft/'. 4873 4874 Each version of the License is given a distinguishing version 4875 number. If the Document specifies that a particular numbered 4876 version of this License "or any later version" applies to it, you 4877 have the option of following the terms and conditions either of 4878 that specified version or of any later version that has been 4879 published (not as a draft) by the Free Software Foundation. If 4880 the Document does not specify a version number of this License, 4881 you may choose any version ever published (not as a draft) by the 4882 Free Software Foundation. 4883 4884ADDENDUM: How to use this License for your documents 4885==================================================== 4886 4887To use this License in a document you have written, include a copy of 4888the License in the document and put the following copyright and license 4889notices just after the title page: 4890 4891 Copyright (C) YEAR YOUR NAME. 4892 Permission is granted to copy, distribute and/or modify this document 4893 under the terms of the GNU Free Documentation License, Version 1.2 4894 or any later version published by the Free Software Foundation; 4895 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover 4896 Texts. A copy of the license is included in the section entitled ``GNU 4897 Free Documentation License''. 4898 4899 If you have Invariant Sections, Front-Cover Texts and Back-Cover 4900Texts, replace the "with...Texts." line with this: 4901 4902 with the Invariant Sections being LIST THEIR TITLES, with 4903 the Front-Cover Texts being LIST, and with the Back-Cover Texts 4904 being LIST. 4905 4906 If you have Invariant Sections without Cover Texts, or some other 4907combination of the three, merge those two alternatives to suit the 4908situation. 4909 4910 If your document contains nontrivial examples of program code, we 4911recommend releasing these examples in parallel under your choice of 4912free software license, such as the GNU General Public License, to 4913permit their use in free software. 4914 4915 4916File: cpp.info, Node: Index of Directives, Next: Option Index, Prev: GNU Free Documentation License, Up: Top 4917 4918Index of Directives 4919******************* 4920 4921�[index�] 4922* Menu: 4923 4924* #assert: Assertions. (line 41) 4925* #define: Object-like Macros. (line 11) 4926* #elif: Elif. (line 6) 4927* #else: Else. (line 6) 4928* #endif: Ifdef. (line 6) 4929* #error: Diagnostics. (line 6) 4930* #ident: Other Directives. (line 6) 4931* #if: Conditional Syntax. (line 6) 4932* #ifdef: Ifdef. (line 6) 4933* #ifndef: Ifdef. (line 40) 4934* #import: Obsolete once-only headers. 4935 (line 10) 4936* #include: Include Syntax. (line 6) 4937* #include_next: Wrapper Headers. (line 6) 4938* #line: Line Control. (line 20) 4939* #pragma GCC dependency: Pragmas. (line 53) 4940* #pragma GCC poison: Pragmas. (line 65) 4941* #pragma GCC system_header <1>: Pragmas. (line 92) 4942* #pragma GCC system_header: System Headers. (line 31) 4943* #sccs: Other Directives. (line 6) 4944* #unassert: Assertions. (line 52) 4945* #undef: Undefining and Redefining Macros. 4946 (line 6) 4947* #warning: Diagnostics. (line 27) 4948 4949 4950File: cpp.info, Node: Option Index, Next: Concept Index, Prev: Index of Directives, Up: Top 4951 4952Option Index 4953************ 4954 4955CPP's command line options and environment variables are indexed here 4956without any initial `-' or `--'. 4957 4958�[index�] 4959* Menu: 4960 4961* A: Invocation. (line 494) 4962* ansi: Invocation. (line 307) 4963* C: Invocation. (line 542) 4964* C_INCLUDE_PATH: Environment Variables. 4965 (line 16) 4966* CPATH: Environment Variables. 4967 (line 15) 4968* CPLUS_INCLUDE_PATH: Environment Variables. 4969 (line 17) 4970* D: Invocation. (line 39) 4971* dD: Invocation. (line 522) 4972* DEPENDENCIES_OUTPUT: Environment Variables. 4973 (line 44) 4974* dI: Invocation. (line 531) 4975* dM: Invocation. (line 510) 4976* dN: Invocation. (line 528) 4977* fdollars-in-identifiers: Invocation. (line 424) 4978* fexec-charset: Invocation. (line 451) 4979* fextended-identifiers: Invocation. (line 427) 4980* finput-charset: Invocation. (line 464) 4981* fno-show-column: Invocation. (line 489) 4982* fno-working-directory: Invocation. (line 474) 4983* fpreprocessed: Invocation. (line 432) 4984* ftabstop: Invocation. (line 445) 4985* fwide-exec-charset: Invocation. (line 456) 4986* fworking-directory: Invocation. (line 474) 4987* H: Invocation. (line 587) 4988* help: Invocation. (line 579) 4989* I: Invocation. (line 71) 4990* I-: Invocation. (line 344) 4991* idirafter: Invocation. (line 386) 4992* imacros: Invocation. (line 377) 4993* imultilib: Invocation. (line 407) 4994* include: Invocation. (line 366) 4995* iprefix: Invocation. (line 391) 4996* iquote: Invocation. (line 418) 4997* isysroot: Invocation. (line 403) 4998* isystem: Invocation. (line 411) 4999* iwithprefix: Invocation. (line 397) 5000* iwithprefixbefore: Invocation. (line 397) 5001* M: Invocation. (line 181) 5002* MD: Invocation. (line 269) 5003* MF: Invocation. (line 215) 5004* MG: Invocation. (line 224) 5005* MM: Invocation. (line 205) 5006* MMD: Invocation. (line 284) 5007* MP: Invocation. (line 234) 5008* MQ: Invocation. (line 260) 5009* MT: Invocation. (line 246) 5010* nostdinc: Invocation. (line 356) 5011* nostdinc++: Invocation. (line 361) 5012* o: Invocation. (line 80) 5013* OBJC_INCLUDE_PATH: Environment Variables. 5014 (line 18) 5015* P: Invocation. (line 535) 5016* pedantic: Invocation. (line 171) 5017* pedantic-errors: Invocation. (line 176) 5018* remap: Invocation. (line 574) 5019* std=: Invocation. (line 307) 5020* SUNPRO_DEPENDENCIES: Environment Variables. 5021 (line 60) 5022* target-help: Invocation. (line 579) 5023* traditional-cpp: Invocation. (line 567) 5024* trigraphs: Invocation. (line 571) 5025* U: Invocation. (line 62) 5026* undef: Invocation. (line 66) 5027* v: Invocation. (line 583) 5028* version: Invocation. (line 596) 5029* w: Invocation. (line 167) 5030* Wall: Invocation. (line 86) 5031* Wcomment: Invocation. (line 94) 5032* Wcomments: Invocation. (line 94) 5033* Wendif-labels: Invocation. (line 144) 5034* Werror: Invocation. (line 157) 5035* Wimport: Invocation. (line 117) 5036* Wsystem-headers: Invocation. (line 161) 5037* Wtraditional: Invocation. (line 111) 5038* Wtrigraphs: Invocation. (line 99) 5039* Wundef: Invocation. (line 120) 5040* Wunused-macros: Invocation. (line 125) 5041* x: Invocation. (line 291) 5042 5043 5044File: cpp.info, Node: Concept Index, Prev: Option Index, Up: Top 5045 5046Concept Index 5047************* 5048 5049�[index�] 5050* Menu: 5051 5052* # operator: Stringification. (line 6) 5053* ## operator: Concatenation. (line 6) 5054* _Pragma: Pragmas. (line 25) 5055* alternative tokens: Tokenization. (line 106) 5056* arguments: Macro Arguments. (line 6) 5057* arguments in macro definitions: Macro Arguments. (line 6) 5058* assertions: Assertions. (line 6) 5059* assertions, canceling: Assertions. (line 52) 5060* backslash-newline: Initial processing. (line 61) 5061* block comments: Initial processing. (line 77) 5062* C++ named operators: C++ Named Operators. (line 6) 5063* character constants: Tokenization. (line 85) 5064* character set, execution: Invocation. (line 451) 5065* character set, input: Invocation. (line 464) 5066* character set, wide execution: Invocation. (line 456) 5067* command line: Invocation. (line 6) 5068* commenting out code: Deleted Code. (line 6) 5069* comments: Initial processing. (line 77) 5070* common predefined macros: Common Predefined Macros. 5071 (line 6) 5072* computed includes: Computed Includes. (line 6) 5073* concatenation: Concatenation. (line 6) 5074* conditional group: Ifdef. (line 14) 5075* conditionals: Conditionals. (line 6) 5076* continued lines: Initial processing. (line 61) 5077* controlling macro: Once-Only Headers. (line 35) 5078* defined: Defined. (line 6) 5079* dependencies for make as output: Environment Variables. 5080 (line 45) 5081* dependencies, make: Invocation. (line 181) 5082* diagnostic: Diagnostics. (line 6) 5083* differences from previous versions: Differences from previous versions. 5084 (line 6) 5085* digraphs: Tokenization. (line 106) 5086* directive line: The preprocessing language. 5087 (line 6) 5088* directive name: The preprocessing language. 5089 (line 6) 5090* directives: The preprocessing language. 5091 (line 6) 5092* empty macro arguments: Macro Arguments. (line 66) 5093* environment variables: Environment Variables. 5094 (line 6) 5095* expansion of arguments: Argument Prescan. (line 6) 5096* FDL, GNU Free Documentation License: GNU Free Documentation License. 5097 (line 6) 5098* function-like macros: Function-like Macros. 5099 (line 6) 5100* grouping options: Invocation. (line 34) 5101* guard macro: Once-Only Headers. (line 35) 5102* header file: Header Files. (line 6) 5103* header file names: Tokenization. (line 85) 5104* identifiers: Tokenization. (line 34) 5105* implementation limits: Implementation limits. 5106 (line 6) 5107* implementation-defined behavior: Implementation-defined behavior. 5108 (line 6) 5109* including just once: Once-Only Headers. (line 6) 5110* invocation: Invocation. (line 6) 5111* iso646.h: C++ Named Operators. (line 6) 5112* line comments: Initial processing. (line 77) 5113* line control: Line Control. (line 6) 5114* line endings: Initial processing. (line 14) 5115* linemarkers: Preprocessor Output. (line 28) 5116* macro argument expansion: Argument Prescan. (line 6) 5117* macro arguments and directives: Directives Within Macro Arguments. 5118 (line 6) 5119* macros in include: Computed Includes. (line 6) 5120* macros with arguments: Macro Arguments. (line 6) 5121* macros with variable arguments: Variadic Macros. (line 6) 5122* make: Invocation. (line 181) 5123* manifest constants: Object-like Macros. (line 6) 5124* named operators: C++ Named Operators. (line 6) 5125* newlines in macro arguments: Newlines in Arguments. 5126 (line 6) 5127* null directive: Other Directives. (line 15) 5128* numbers: Tokenization. (line 61) 5129* object-like macro: Object-like Macros. (line 6) 5130* options: Invocation. (line 38) 5131* options, grouping: Invocation. (line 34) 5132* other tokens: Tokenization. (line 120) 5133* output format: Preprocessor Output. (line 12) 5134* overriding a header file: Wrapper Headers. (line 6) 5135* parentheses in macro bodies: Operator Precedence Problems. 5136 (line 6) 5137* pitfalls of macros: Macro Pitfalls. (line 6) 5138* predefined macros: Predefined Macros. (line 6) 5139* predefined macros, system-specific: System-specific Predefined Macros. 5140 (line 6) 5141* predicates: Assertions. (line 19) 5142* preprocessing directives: The preprocessing language. 5143 (line 6) 5144* preprocessing numbers: Tokenization. (line 61) 5145* preprocessing tokens: Tokenization. (line 6) 5146* prescan of macro arguments: Argument Prescan. (line 6) 5147* problems with macros: Macro Pitfalls. (line 6) 5148* punctuators: Tokenization. (line 106) 5149* redefining macros: Undefining and Redefining Macros. 5150 (line 6) 5151* repeated inclusion: Once-Only Headers. (line 6) 5152* reporting errors: Diagnostics. (line 6) 5153* reporting warnings: Diagnostics. (line 6) 5154* reserved namespace: System-specific Predefined Macros. 5155 (line 6) 5156* self-reference: Self-Referential Macros. 5157 (line 6) 5158* semicolons (after macro calls): Swallowing the Semicolon. 5159 (line 6) 5160* side effects (in macro arguments): Duplication of Side Effects. 5161 (line 6) 5162* standard predefined macros.: Standard Predefined Macros. 5163 (line 6) 5164* string constants: Tokenization. (line 85) 5165* string literals: Tokenization. (line 85) 5166* stringification: Stringification. (line 6) 5167* symbolic constants: Object-like Macros. (line 6) 5168* system header files <1>: System Headers. (line 6) 5169* system header files: Header Files. (line 13) 5170* system-specific predefined macros: System-specific Predefined Macros. 5171 (line 6) 5172* testing predicates: Assertions. (line 30) 5173* token concatenation: Concatenation. (line 6) 5174* token pasting: Concatenation. (line 6) 5175* tokens: Tokenization. (line 6) 5176* trigraphs: Initial processing. (line 32) 5177* undefining macros: Undefining and Redefining Macros. 5178 (line 6) 5179* unsafe macros: Duplication of Side Effects. 5180 (line 6) 5181* variable number of arguments: Variadic Macros. (line 6) 5182* variadic macros: Variadic Macros. (line 6) 5183* wrapper #ifndef: Once-Only Headers. (line 6) 5184* wrapper headers: Wrapper Headers. (line 6) 5185 5186 5187 5188Tag Table: 5189Node: Top1094 5190Node: Overview3778 5191Node: Character sets6599 5192Ref: Character sets-Footnote-18782 5193Node: Initial processing8963 5194Ref: trigraphs10522 5195Node: Tokenization14724 5196Ref: Tokenization-Footnote-121860 5197Node: The preprocessing language21971 5198Node: Header Files24849 5199Node: Include Syntax26729 5200Node: Include Operation28366 5201Node: Search Path30214 5202Node: Once-Only Headers33404 5203Node: Computed Includes35049 5204Node: Wrapper Headers38193 5205Node: System Headers40619 5206Node: Macros42669 5207Node: Object-like Macros43810 5208Node: Function-like Macros47400 5209Node: Macro Arguments49016 5210Node: Stringification53161 5211Node: Concatenation56367 5212Node: Variadic Macros59475 5213Node: Predefined Macros64262 5214Node: Standard Predefined Macros64850 5215Node: Common Predefined Macros70786 5216Node: System-specific Predefined Macros81678 5217Node: C++ Named Operators83699 5218Node: Undefining and Redefining Macros84663 5219Node: Directives Within Macro Arguments86767 5220Node: Macro Pitfalls88315 5221Node: Misnesting88848 5222Node: Operator Precedence Problems89960 5223Node: Swallowing the Semicolon91826 5224Node: Duplication of Side Effects93849 5225Node: Self-Referential Macros96032 5226Node: Argument Prescan98441 5227Node: Newlines in Arguments102195 5228Node: Conditionals103146 5229Node: Conditional Uses104976 5230Node: Conditional Syntax106334 5231Node: Ifdef106654 5232Node: If109815 5233Node: Defined112119 5234Node: Else113402 5235Node: Elif113972 5236Node: Deleted Code115261 5237Node: Diagnostics116508 5238Node: Line Control118125 5239Node: Pragmas121929 5240Node: Other Directives126199 5241Node: Preprocessor Output127249 5242Node: Traditional Mode130450 5243Node: Traditional lexical analysis131508 5244Node: Traditional macros134011 5245Node: Traditional miscellany137813 5246Node: Traditional warnings138810 5247Node: Implementation Details141007 5248Node: Implementation-defined behavior141628 5249Ref: Identifier characters142361 5250Node: Implementation limits145446 5251Node: Obsolete Features148120 5252Node: Assertions148575 5253Node: Obsolete once-only headers151116 5254Node: Differences from previous versions152851 5255Node: Invocation157059 5256Ref: Wtrigraphs161390 5257Ref: dashMF166174 5258Ref: fdollars-in-identifiers174196 5259Node: Environment Variables181849 5260Node: GNU Free Documentation License184815 5261Node: Index of Directives207248 5262Node: Option Index209177 5263Node: Concept Index215288 5264 5265End Tag Table 5266