1 This is doc/cpp.info, produced by makeinfo version 4.12 from
2 /space/rguenther/gcc-4.7.3/gcc-4.7.3/gcc/doc/cpp.texi.
4 Copyright (C) 1987, 1989, 1991, 1992, 1993, 1994, 1995, 1996, 1997,
5 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
6 2010, 2011 Free Software Foundation, Inc.
8 Permission is granted to copy, distribute and/or modify this document
9 under the terms of the GNU Free Documentation License, Version 1.3 or
10 any later version published by the Free Software Foundation. A copy of
11 the license is included in the section entitled "GNU Free Documentation
14 This manual contains no Invariant Sections. The Front-Cover Texts
15 are (a) (see below), and the Back-Cover Texts are (b) (see below).
17 (a) The FSF's Front-Cover Text is:
21 (b) The FSF's Back-Cover Text is:
23 You have freedom to copy and modify this GNU Manual, like GNU
24 software. Copies published by the Free Software Foundation raise
25 funds for GNU development.
27 INFO-DIR-SECTION Software development
29 * Cpp: (cpp). The GNU C preprocessor.
33 File: cpp.info, Node: Top, Next: Overview, Up: (dir)
38 The C preprocessor implements the macro language used to transform C,
39 C++, and Objective-C programs before they are compiled. It can also be
52 * Preprocessor Output::
54 * Implementation Details::
56 * Environment Variables::
57 * GNU Free Documentation License::
58 * Index of Directives::
62 --- The Detailed Node Listing ---
67 * Initial processing::
69 * The preprocessing language::
77 * Alternatives to Wrapper #ifndef::
84 * Object-like Macros::
85 * Function-like Macros::
91 * Undefining and Redefining Macros::
92 * Directives Within Macro Arguments::
97 * Standard Predefined Macros::
98 * Common Predefined Macros::
99 * System-specific Predefined Macros::
100 * C++ Named Operators::
105 * Operator Precedence Problems::
106 * Swallowing the Semicolon::
107 * Duplication of Side Effects::
108 * Self-Referential Macros::
110 * Newlines in Arguments::
115 * Conditional Syntax::
126 Implementation Details
128 * Implementation-defined behavior::
129 * Implementation limits::
130 * Obsolete Features::
131 * Differences from previous versions::
135 * Obsolete Features::
137 Copyright (C) 1987, 1989, 1991, 1992, 1993, 1994, 1995, 1996, 1997,
138 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009,
139 2010, 2011 Free Software Foundation, Inc.
141 Permission is granted to copy, distribute and/or modify this document
142 under the terms of the GNU Free Documentation License, Version 1.3 or
143 any later version published by the Free Software Foundation. A copy of
144 the license is included in the section entitled "GNU Free Documentation
147 This manual contains no Invariant Sections. The Front-Cover Texts
148 are (a) (see below), and the Back-Cover Texts are (b) (see below).
150 (a) The FSF's Front-Cover Text is:
154 (b) The FSF's Back-Cover Text is:
156 You have freedom to copy and modify this GNU Manual, like GNU
157 software. Copies published by the Free Software Foundation raise
158 funds for GNU development.
161 File: cpp.info, Node: Overview, Next: Header Files, Prev: Top, Up: Top
166 The C preprocessor, often known as "cpp", is a "macro processor" that
167 is used automatically by the C compiler to transform your program
168 before compilation. It is called a macro processor because it allows
169 you to define "macros", which are brief abbreviations for longer
172 The C preprocessor is intended to be used only with C, C++, and
173 Objective-C source code. In the past, it has been abused as a general
174 text processor. It will choke on input which does not obey C's lexical
175 rules. For example, apostrophes will be interpreted as the beginning of
176 character constants, and cause errors. Also, you cannot rely on it
177 preserving characteristics of the input which are not significant to
178 C-family languages. If a Makefile is preprocessed, all the hard tabs
179 will be removed, and the Makefile will not work.
181 Having said that, you can often get away with using cpp on things
182 which are not C. Other Algol-ish programming languages are often safe
183 (Pascal, Ada, etc.) So is assembly, with caution. `-traditional-cpp'
184 mode preserves more white space, and is otherwise more permissive. Many
185 of the problems can be avoided by writing C or C++ style comments
186 instead of native language comments, and keeping macros simple.
188 Wherever possible, you should use a preprocessor geared to the
189 language you are writing in. Modern versions of the GNU assembler have
190 macro facilities. Most high level programming languages have their own
191 conditional compilation and inclusion mechanism. If all else fails,
192 try a true general text processor, such as GNU M4.
194 C preprocessors vary in some details. This manual discusses the GNU
195 C preprocessor, which provides a small superset of the features of ISO
196 Standard C. In its default mode, the GNU C preprocessor does not do a
197 few things required by the standard. These are features which are
198 rarely, if ever, used, and may cause surprising changes to the meaning
199 of a program which does not expect them. To get strict ISO Standard C,
200 you should use the `-std=c90', `-std=c99' or `-std=c11' options,
201 depending on which version of the standard you want. To get all the
202 mandatory diagnostics, you must also use `-pedantic'. *Note
205 This manual describes the behavior of the ISO preprocessor. To
206 minimize gratuitous differences, where the ISO preprocessor's behavior
207 does not conflict with traditional semantics, the traditional
208 preprocessor should behave the same way. The various differences that
209 do exist are detailed in the section *note Traditional Mode::.
211 For clarity, unless noted otherwise, references to `CPP' in this
212 manual refer to GNU CPP.
217 * Initial processing::
219 * The preprocessing language::
222 File: cpp.info, Node: Character sets, Next: Initial processing, Up: Overview
227 Source code character set processing in C and related languages is
228 rather complicated. The C standard discusses two character sets, but
229 there are really at least four.
231 The files input to CPP might be in any character set at all. CPP's
232 very first action, before it even looks for line boundaries, is to
233 convert the file into the character set it uses for internal
234 processing. That set is what the C standard calls the "source"
235 character set. It must be isomorphic with ISO 10646, also known as
236 Unicode. CPP uses the UTF-8 encoding of Unicode.
238 The character sets of the input files are specified using the
239 `-finput-charset=' option.
241 All preprocessing work (the subject of the rest of this manual) is
242 carried out in the source character set. If you request textual output
243 from the preprocessor with the `-E' option, it will be in UTF-8.
245 After preprocessing is complete, string and character constants are
246 converted again, into the "execution" character set. This character
247 set is under control of the user; the default is UTF-8, matching the
248 source character set. Wide string and character constants have their
249 own character set, which is not called out specifically in the
250 standard. Again, it is under control of the user. The default is
251 UTF-16 or UTF-32, whichever fits in the target's `wchar_t' type, in the
252 target machine's byte order.(1) Octal and hexadecimal escape sequences
253 do not undergo conversion; '\x12' has the value 0x12 regardless of the
254 currently selected execution character set. All other escapes are
255 replaced by the character in the source character set that they
256 represent, then converted to the execution character set, just like
257 unescaped characters.
259 Unless the experimental `-fextended-identifiers' option is used, GCC
260 does not permit the use of characters outside the ASCII range, nor `\u'
261 and `\U' escapes, in identifiers. Even with that option, characters
262 outside the ASCII range can only be specified with the `\u' and `\U'
263 escapes, not used directly in identifiers.
265 ---------- Footnotes ----------
267 (1) UTF-16 does not meet the requirements of the C standard for a
268 wide character set, but the choice of 16-bit `wchar_t' is enshrined in
269 some system ABIs so we cannot fix this.
272 File: cpp.info, Node: Initial processing, Next: Tokenization, Prev: Character sets, Up: Overview
274 1.2 Initial processing
275 ======================
277 The preprocessor performs a series of textual transformations on its
278 input. These happen before all other processing. Conceptually, they
279 happen in a rigid order, and the entire file is run through each
280 transformation before the next one begins. CPP actually does them all
281 at once, for performance reasons. These transformations correspond
282 roughly to the first three "phases of translation" described in the C
285 1. The input file is read into memory and broken into lines.
287 Different systems use different conventions to indicate the end of
288 a line. GCC accepts the ASCII control sequences `LF', `CR LF' and
289 `CR' as end-of-line markers. These are the canonical sequences
290 used by Unix, DOS and VMS, and the classic Mac OS (before OSX)
291 respectively. You may therefore safely copy source code written
292 on any of those systems to a different one and use it without
293 conversion. (GCC may lose track of the current line number if a
294 file doesn't consistently use one convention, as sometimes happens
295 when it is edited on computers with different conventions that
296 share a network file system.)
298 If the last line of any input file lacks an end-of-line marker,
299 the end of the file is considered to implicitly supply one. The C
300 standard says that this condition provokes undefined behavior, so
301 GCC will emit a warning message.
303 2. If trigraphs are enabled, they are replaced by their corresponding
304 single characters. By default GCC ignores trigraphs, but if you
305 request a strictly conforming mode with the `-std' option, or you
306 specify the `-trigraphs' option, then it converts them.
308 These are nine three-character sequences, all starting with `??',
309 that are defined by ISO C to stand for single characters. They
310 permit obsolete systems that lack some of C's punctuation to use
311 C. For example, `??/' stands for `\', so '??/n' is a character
312 constant for a newline.
314 Trigraphs are not popular and many compilers implement them
315 incorrectly. Portable code should not rely on trigraphs being
316 either converted or ignored. With `-Wtrigraphs' GCC will warn you
317 when a trigraph may change the meaning of your program if it were
318 converted. *Note Wtrigraphs::.
320 In a string constant, you can prevent a sequence of question marks
321 from being confused with a trigraph by inserting a backslash
322 between the question marks, or by separating the string literal at
323 the trigraph and making use of string literal concatenation.
324 "(??\?)" is the string `(???)', not `(?]'. Traditional C
325 compilers do not recognize these idioms.
327 The nine trigraphs and their replacements are
329 Trigraph: ??( ??) ??< ??> ??= ??/ ??' ??! ??-
330 Replacement: [ ] { } # \ ^ | ~
332 3. Continued lines are merged into one long line.
334 A continued line is a line which ends with a backslash, `\'. The
335 backslash is removed and the following line is joined with the
336 current one. No space is inserted, so you may split a line
337 anywhere, even in the middle of a word. (It is generally more
338 readable to split lines only at white space.)
340 The trailing backslash on a continued line is commonly referred to
341 as a "backslash-newline".
343 If there is white space between a backslash and the end of a line,
344 that is still a continued line. However, as this is usually the
345 result of an editing mistake, and many compilers will not accept
346 it as a continued line, GCC will warn you about it.
348 4. All comments are replaced with single spaces.
350 There are two kinds of comments. "Block comments" begin with `/*'
351 and continue until the next `*/'. Block comments do not nest:
353 /* this is /* one comment */ text outside comment
355 "Line comments" begin with `//' and continue to the end of the
356 current line. Line comments do not nest either, but it does not
357 matter, because they would end in the same place anyway.
359 // this is // one comment
362 It is safe to put line comments inside block comments, or vice versa.
365 // contains line comment
369 // line comment /* contains block comment */
371 But beware of commenting out one end of a block comment with a line
374 // l.c. /* block comment begins
375 oops! this isn't a comment anymore */
377 Comments are not recognized within string literals. "/* blah */" is
378 the string constant `/* blah */', not an empty string.
380 Line comments are not in the 1989 edition of the C standard, but they
381 are recognized by GCC as an extension. In C++ and in the 1999 edition
382 of the C standard, they are an official part of the language.
384 Since these transformations happen before all other processing, you
385 can split a line mechanically with backslash-newline anywhere. You can
386 comment out the end of a line. You can continue a line comment onto the
387 next line with backslash-newline. You can even split `/*', `*/', and
388 `//' onto multiple lines with backslash-newline. For example:
398 is equivalent to `#define FOO 1020'. All these tricks are extremely
399 confusing and should not be used in code intended to be readable.
401 There is no way to prevent a backslash at the end of a line from
402 being interpreted as a backslash-newline. This cannot affect any
403 correct program, however.
406 File: cpp.info, Node: Tokenization, Next: The preprocessing language, Prev: Initial processing, Up: Overview
411 After the textual transformations are finished, the input file is
412 converted into a sequence of "preprocessing tokens". These mostly
413 correspond to the syntactic tokens used by the C compiler, but there are
414 a few differences. White space separates tokens; it is not itself a
415 token of any kind. Tokens do not have to be separated by white space,
416 but it is often necessary to avoid ambiguities.
418 When faced with a sequence of characters that has more than one
419 possible tokenization, the preprocessor is greedy. It always makes
420 each token, starting from the left, as big as possible before moving on
421 to the next token. For instance, `a+++++b' is interpreted as
422 `a ++ ++ + b', not as `a ++ + ++ b', even though the latter
423 tokenization could be part of a valid C program and the former could
426 Once the input file is broken into tokens, the token boundaries never
427 change, except when the `##' preprocessing operator is used to paste
428 tokens together. *Note Concatenation::. For example,
436 The compiler does not re-tokenize the preprocessor's output. Each
437 preprocessing token becomes one compiler token.
439 Preprocessing tokens fall into five broad classes: identifiers,
440 preprocessing numbers, string literals, punctuators, and other. An
441 "identifier" is the same as an identifier in C: any sequence of
442 letters, digits, or underscores, which begins with a letter or
443 underscore. Keywords of C have no significance to the preprocessor;
444 they are ordinary identifiers. You can define a macro whose name is a
445 keyword, for instance. The only identifier which can be considered a
446 preprocessing keyword is `defined'. *Note Defined::.
448 This is mostly true of other languages which use the C preprocessor.
449 However, a few of the keywords of C++ are significant even in the
450 preprocessor. *Note C++ Named Operators::.
452 In the 1999 C standard, identifiers may contain letters which are not
453 part of the "basic source character set", at the implementation's
454 discretion (such as accented Latin letters, Greek letters, or Chinese
455 ideograms). This may be done with an extended character set, or the
456 `\u' and `\U' escape sequences. The implementation of this feature in
457 GCC is experimental; such characters are only accepted in the `\u' and
458 `\U' forms and only if `-fextended-identifiers' is used.
460 As an extension, GCC treats `$' as a letter. This is for
461 compatibility with some systems, such as VMS, where `$' is commonly
462 used in system-defined function and object names. `$' is not a letter
463 in strictly conforming mode, or if you specify the `-$' option. *Note
466 A "preprocessing number" has a rather bizarre definition. The
467 category includes all the normal integer and floating point constants
468 one expects of C, but also a number of other things one might not
469 initially recognize as a number. Formally, preprocessing numbers begin
470 with an optional period, a required decimal digit, and then continue
471 with any sequence of letters, digits, underscores, periods, and
472 exponents. Exponents are the two-character sequences `e+', `e-', `E+',
473 `E-', `p+', `p-', `P+', and `P-'. (The exponents that begin with `p'
474 or `P' are new to C99. They are used for hexadecimal floating-point
477 The purpose of this unusual definition is to isolate the preprocessor
478 from the full complexity of numeric constants. It does not have to
479 distinguish between lexically valid and invalid floating-point numbers,
480 which is complicated. The definition also permits you to split an
481 identifier at any position and get exactly two tokens, which can then be
482 pasted back together with the `##' operator.
484 It's possible for preprocessing numbers to cause programs to be
485 misinterpreted. For example, `0xE+12' is a preprocessing number which
486 does not translate to any valid numeric constant, therefore a syntax
487 error. It does not mean `0xE + 12', which is what you might have
490 "String literals" are string constants, character constants, and
491 header file names (the argument of `#include').(1) String constants
492 and character constants are straightforward: "..." or '...'. In either
493 case embedded quotes should be escaped with a backslash: '\'' is the
494 character constant for `''. There is no limit on the length of a
495 character constant, but the value of a character constant that contains
496 more than one character is implementation-defined. *Note
497 Implementation Details::.
499 Header file names either look like string constants, "...", or are
500 written with angle brackets instead, <...>. In either case, backslash
501 is an ordinary character. There is no way to escape the closing quote
502 or angle bracket. The preprocessor looks for the header file in
503 different places depending on which form you use. *Note Include
506 No string literal may extend past the end of a line. Older versions
507 of GCC accepted multi-line string constants. You may use continued
508 lines instead, or string constant concatenation. *Note Differences
509 from previous versions::.
511 "Punctuators" are all the usual bits of punctuation which are
512 meaningful to C and C++. All but three of the punctuation characters in
513 ASCII are C punctuators. The exceptions are `@', `$', and ``'. In
514 addition, all the two- and three-character operators are punctuators.
515 There are also six "digraphs", which the C++ standard calls
516 "alternative tokens", which are merely alternate ways to spell other
517 punctuators. This is a second attempt to work around missing
518 punctuation in obsolete systems. It has no negative side effects,
519 unlike trigraphs, but does not cover as much ground. The digraphs and
520 their corresponding normal punctuators are:
522 Digraph: <% %> <: :> %: %:%:
523 Punctuator: { } [ ] # ##
525 Any other single character is considered "other". It is passed on to
526 the preprocessor's output unmolested. The C compiler will almost
527 certainly reject source code containing "other" tokens. In ASCII, the
528 only other characters are `@', `$', ``', and control characters other
529 than NUL (all bits zero). (Note that `$' is normally considered a
530 letter.) All characters with the high bit set (numeric range
531 0x7F-0xFF) are also "other" in the present implementation. This will
532 change when proper support for international character sets is added to
535 NUL is a special case because of the high probability that its
536 appearance is accidental, and because it may be invisible to the user
537 (many terminals do not display NUL at all). Within comments, NULs are
538 silently ignored, just as any other character would be. In running
539 text, NUL is considered white space. For example, these two directives
540 have the same meaning.
545 (where `^@' is ASCII NUL). Within string or character constants, NULs
546 are preserved. In the latter two cases the preprocessor emits a
549 ---------- Footnotes ----------
551 (1) The C standard uses the term "string literal" to refer only to
552 what we are calling "string constants".
555 File: cpp.info, Node: The preprocessing language, Prev: Tokenization, Up: Overview
557 1.4 The preprocessing language
558 ==============================
560 After tokenization, the stream of tokens may simply be passed straight
561 to the compiler's parser. However, if it contains any operations in the
562 "preprocessing language", it will be transformed first. This stage
563 corresponds roughly to the standard's "translation phase 4" and is what
564 most people think of as the preprocessor's job.
566 The preprocessing language consists of "directives" to be executed
567 and "macros" to be expanded. Its primary capabilities are:
569 * Inclusion of header files. These are files of declarations that
570 can be substituted into your program.
572 * Macro expansion. You can define "macros", which are abbreviations
573 for arbitrary fragments of C code. The preprocessor will replace
574 the macros with their definitions throughout the program. Some
575 macros are automatically defined for you.
577 * Conditional compilation. You can include or exclude parts of the
578 program according to various conditions.
580 * Line control. If you use a program to combine or rearrange source
581 files into an intermediate file which is then compiled, you can
582 use line control to inform the compiler where each source line
583 originally came from.
585 * Diagnostics. You can detect problems at compile time and issue
588 There are a few more, less useful, features.
590 Except for expansion of predefined macros, all these operations are
591 triggered with "preprocessing directives". Preprocessing directives
592 are lines in your program that start with `#'. Whitespace is allowed
593 before and after the `#'. The `#' is followed by an identifier, the
594 "directive name". It specifies the operation to perform. Directives
595 are commonly referred to as `#NAME' where NAME is the directive name.
596 For example, `#define' is the directive that defines a macro.
598 The `#' which begins a directive cannot come from a macro expansion.
599 Also, the directive name is not macro expanded. Thus, if `foo' is
600 defined as a macro expanding to `define', that does not make `#foo' a
601 valid preprocessing directive.
603 The set of valid directive names is fixed. Programs cannot define
604 new preprocessing directives.
606 Some directives require arguments; these make up the rest of the
607 directive line and must be separated from the directive name by
608 whitespace. For example, `#define' must be followed by a macro name
609 and the intended expansion of the macro.
611 A preprocessing directive cannot cover more than one line. The line
612 may, however, be continued with backslash-newline, or by a block comment
613 which extends past the end of the line. In either case, when the
614 directive is processed, the continuations have already been merged with
615 the first line to make one long line.
618 File: cpp.info, Node: Header Files, Next: Macros, Prev: Overview, Up: Top
623 A header file is a file containing C declarations and macro definitions
624 (*note Macros::) to be shared between several source files. You request
625 the use of a header file in your program by "including" it, with the C
626 preprocessing directive `#include'.
628 Header files serve two purposes.
630 * System header files declare the interfaces to parts of the
631 operating system. You include them in your program to supply the
632 definitions and declarations you need to invoke system calls and
635 * Your own header files contain declarations for interfaces between
636 the source files of your program. Each time you have a group of
637 related declarations and macro definitions all or most of which
638 are needed in several different source files, it is a good idea to
639 create a header file for them.
641 Including a header file produces the same results as copying the
642 header file into each source file that needs it. Such copying would be
643 time-consuming and error-prone. With a header file, the related
644 declarations appear in only one place. If they need to be changed, they
645 can be changed in one place, and programs that include the header file
646 will automatically use the new version when next recompiled. The header
647 file eliminates the labor of finding and changing all the copies as well
648 as the risk that a failure to find one copy will result in
649 inconsistencies within a program.
651 In C, the usual convention is to give header files names that end
652 with `.h'. It is most portable to use only letters, digits, dashes, and
653 underscores in header file names, and at most one dot.
658 * Include Operation::
660 * Once-Only Headers::
661 * Alternatives to Wrapper #ifndef::
662 * Computed Includes::
667 File: cpp.info, Node: Include Syntax, Next: Include Operation, Up: Header Files
672 Both user and system header files are included using the preprocessing
673 directive `#include'. It has two variants:
676 This variant is used for system header files. It searches for a
677 file named FILE in a standard list of system directories. You can
678 prepend directories to this list with the `-I' option (*note
682 This variant is used for header files of your own program. It
683 searches for a file named FILE first in the directory containing
684 the current file, then in the quote directories and then the same
685 directories used for `<FILE>'. You can prepend directories to the
686 list of quote directories with the `-iquote' option.
688 The argument of `#include', whether delimited with quote marks or
689 angle brackets, behaves like a string constant in that comments are not
690 recognized, and macro names are not expanded. Thus, `#include <x/*y>'
691 specifies inclusion of a system header file named `x/*y'.
693 However, if backslashes occur within FILE, they are considered
694 ordinary text characters, not escape characters. None of the character
695 escape sequences appropriate to string constants in C are processed.
696 Thus, `#include "x\n\\y"' specifies a filename containing three
697 backslashes. (Some systems interpret `\' as a pathname separator. All
698 of these also interpret `/' the same way. It is most portable to use
701 It is an error if there is anything (other than comments) on the line
705 File: cpp.info, Node: Include Operation, Next: Search Path, Prev: Include Syntax, Up: Header Files
707 2.2 Include Operation
708 =====================
710 The `#include' directive works by directing the C preprocessor to scan
711 the specified file as input before continuing with the rest of the
712 current file. The output from the preprocessor contains the output
713 already generated, followed by the output resulting from the included
714 file, followed by the output that comes from the text after the
715 `#include' directive. For example, if you have a header file
716 `header.h' as follows,
720 and a main program called `program.c' that uses the header file, like
732 the compiler will see the same token stream as it would if `program.c'
744 Included files are not limited to declarations and macro definitions;
745 those are merely the typical uses. Any fragment of a C program can be
746 included from another file. The include file could even contain the
747 beginning of a statement that is concluded in the containing file, or
748 the end of a statement that was started in the including file. However,
749 an included file must consist of complete tokens. Comments and string
750 literals which have not been closed by the end of an included file are
751 invalid. For error recovery, they are considered to end at the end of
754 To avoid confusion, it is best if header files contain only complete
755 syntactic units--function declarations or definitions, type
758 The line following the `#include' directive is always treated as a
759 separate line by the C preprocessor, even if the included file lacks a
763 File: cpp.info, Node: Search Path, Next: Once-Only Headers, Prev: Include Operation, Up: Header Files
768 GCC looks in several different places for headers. On a normal Unix
769 system, if you do not instruct it otherwise, it will look for headers
770 requested with `#include <FILE>' in:
773 LIBDIR/gcc/TARGET/VERSION/include
777 For C++ programs, it will also look in `/usr/include/g++-v3', first.
778 In the above, TARGET is the canonical name of the system GCC was
779 configured to compile code for; often but not always the same as the
780 canonical name of the system it runs on. VERSION is the version of GCC
783 You can add to this list with the `-IDIR' command line option. All
784 the directories named by `-I' are searched, in left-to-right order,
785 _before_ the default directories. The only exception is when `dir' is
786 already searched by default. In this case, the option is ignored and
787 the search order for system directories remains unchanged.
789 Duplicate directories are removed from the quote and bracket search
790 chains before the two chains are merged to make the final search chain.
791 Thus, it is possible for a directory to occur twice in the final search
792 chain if it was specified in both the quote and bracket chains.
794 You can prevent GCC from searching any of the default directories
795 with the `-nostdinc' option. This is useful when you are compiling an
796 operating system kernel or some other program that does not use the
797 standard C library facilities, or the standard C library itself. `-I'
798 options are not ignored as described above when `-nostdinc' is in
801 GCC looks for headers requested with `#include "FILE"' first in the
802 directory containing the current file, then in the directories as
803 specified by `-iquote' options, then in the same places it would have
804 looked for a header requested with angle brackets. For example, if
805 `/usr/include/sys/stat.h' contains `#include "types.h"', GCC looks for
806 `types.h' first in `/usr/include/sys', then in its usual search path.
808 `#line' (*note Line Control::) does not change GCC's idea of the
809 directory containing the current file.
811 You may put `-I-' at any point in your list of `-I' options. This
812 has two effects. First, directories appearing before the `-I-' in the
813 list are searched only for headers requested with quote marks.
814 Directories after `-I-' are searched for all headers. Second, the
815 directory containing the current file is not searched for anything,
816 unless it happens to be one of the directories named by an `-I' switch.
817 `-I-' is deprecated, `-iquote' should be used instead.
819 `-I. -I-' is not the same as no `-I' options at all, and does not
820 cause the same behavior for `<>' includes that `""' includes get with
821 no special options. `-I.' searches the compiler's current working
822 directory for header files. That may or may not be the same as the
823 directory containing the current file.
825 If you need to look for headers in a directory named `-', write
828 There are several more ways to adjust the header search path. They
829 are generally less useful. *Note Invocation::.
832 File: cpp.info, Node: Once-Only Headers, Next: Alternatives to Wrapper #ifndef, Prev: Search Path, Up: Header Files
834 2.4 Once-Only Headers
835 =====================
837 If a header file happens to be included twice, the compiler will process
838 its contents twice. This is very likely to cause an error, e.g. when
839 the compiler sees the same structure definition twice. Even if it does
840 not, it will certainly waste time.
842 The standard way to prevent this is to enclose the entire real
843 contents of the file in a conditional, like this:
846 #ifndef FILE_FOO_SEEN
847 #define FILE_FOO_SEEN
851 #endif /* !FILE_FOO_SEEN */
853 This construct is commonly known as a "wrapper #ifndef". When the
854 header is included again, the conditional will be false, because
855 `FILE_FOO_SEEN' is defined. The preprocessor will skip over the entire
856 contents of the file, and the compiler will not see it twice.
858 CPP optimizes even further. It remembers when a header file has a
859 wrapper `#ifndef'. If a subsequent `#include' specifies that header,
860 and the macro in the `#ifndef' is still defined, it does not bother to
861 rescan the file at all.
863 You can put comments outside the wrapper. They will not interfere
864 with this optimization.
866 The macro `FILE_FOO_SEEN' is called the "controlling macro" or
867 "guard macro". In a user header file, the macro name should not begin
868 with `_'. In a system header file, it should begin with `__' to avoid
869 conflicts with user programs. In any kind of header file, the macro
870 name should contain the name of the file and some additional text, to
871 avoid conflicts with other header files.
874 File: cpp.info, Node: Alternatives to Wrapper #ifndef, Next: Computed Includes, Prev: Once-Only Headers, Up: Header Files
876 2.5 Alternatives to Wrapper #ifndef
877 ===================================
879 CPP supports two more ways of indicating that a header file should be
880 read only once. Neither one is as portable as a wrapper `#ifndef' and
881 we recommend you do not use them in new programs, with the caveat that
882 `#import' is standard practice in Objective-C.
884 CPP supports a variant of `#include' called `#import' which includes
885 a file, but does so at most once. If you use `#import' instead of
886 `#include', then you don't need the conditionals inside the header file
887 to prevent multiple inclusion of the contents. `#import' is standard
888 in Objective-C, but is considered a deprecated extension in C and C++.
890 `#import' is not a well designed feature. It requires the users of
891 a header file to know that it should only be included once. It is much
892 better for the header file's implementor to write the file so that users
893 don't need to know this. Using a wrapper `#ifndef' accomplishes this
896 In the present implementation, a single use of `#import' will
897 prevent the file from ever being read again, by either `#import' or
898 `#include'. You should not rely on this; do not use both `#import' and
899 `#include' to refer to the same header file.
901 Another way to prevent a header file from being included more than
902 once is with the `#pragma once' directive. If `#pragma once' is seen
903 when scanning a header file, that file will never be read again, no
906 `#pragma once' does not have the problems that `#import' does, but
907 it is not recognized by all preprocessors, so you cannot rely on it in
911 File: cpp.info, Node: Computed Includes, Next: Wrapper Headers, Prev: Alternatives to Wrapper #ifndef, Up: Header Files
913 2.6 Computed Includes
914 =====================
916 Sometimes it is necessary to select one of several different header
917 files to be included into your program. They might specify
918 configuration parameters to be used on different sorts of operating
919 systems, for instance. You could do this with a series of conditionals,
922 # include "system_1.h"
924 # include "system_2.h"
929 That rapidly becomes tedious. Instead, the preprocessor offers the
930 ability to use a macro for the header name. This is called a "computed
931 include". Instead of writing a header name as the direct argument of
932 `#include', you simply put a macro name there instead:
934 #define SYSTEM_H "system_1.h"
938 `SYSTEM_H' will be expanded, and the preprocessor will look for
939 `system_1.h' as if the `#include' had been written that way originally.
940 `SYSTEM_H' could be defined by your Makefile with a `-D' option.
942 You must be careful when you define the macro. `#define' saves
943 tokens, not text. The preprocessor has no way of knowing that the macro
944 will be used as the argument of `#include', so it generates ordinary
945 tokens, not a header name. This is unlikely to cause problems if you
946 use double-quote includes, which are close enough to string constants.
947 If you use angle brackets, however, you may have trouble.
949 The syntax of a computed include is actually a bit more general than
950 the above. If the first non-whitespace character after `#include' is
951 not `"' or `<', then the entire line is macro-expanded like running
954 If the line expands to a single string constant, the contents of that
955 string constant are the file to be included. CPP does not re-examine
956 the string for embedded quotes, but neither does it process backslash
957 escapes in the string. Therefore
959 #define HEADER "a\"b"
962 looks for a file named `a\"b'. CPP searches for the file according to
963 the rules for double-quoted includes.
965 If the line expands to a token stream beginning with a `<' token and
966 including a `>' token, then the tokens between the `<' and the first
967 `>' are combined to form the filename to be included. Any whitespace
968 between tokens is reduced to a single space; then any space after the
969 initial `<' is retained, but a trailing space before the closing `>' is
970 ignored. CPP searches for the file according to the rules for
971 angle-bracket includes.
973 In either case, if there are any tokens on the line after the file
974 name, an error occurs and the directive is not processed. It is also
975 an error if the result of expansion does not match either of the two
978 These rules are implementation-defined behavior according to the C
979 standard. To minimize the risk of different compilers interpreting your
980 computed includes differently, we recommend you use only a single
981 object-like macro which expands to a string constant. This will also
982 minimize confusion for people reading your program.
985 File: cpp.info, Node: Wrapper Headers, Next: System Headers, Prev: Computed Includes, Up: Header Files
990 Sometimes it is necessary to adjust the contents of a system-provided
991 header file without editing it directly. GCC's `fixincludes' operation
992 does this, for example. One way to do that would be to create a new
993 header file with the same name and insert it in the search path before
994 the original header. That works fine as long as you're willing to
995 replace the old header entirely. But what if you want to refer to the
996 old header from the new one?
998 You cannot simply include the old header with `#include'. That will
999 start from the beginning, and find your new header again. If your
1000 header is not protected from multiple inclusion (*note Once-Only
1001 Headers::), it will recurse infinitely and cause a fatal error.
1003 You could include the old header with an absolute pathname:
1004 #include "/usr/include/old-header.h"
1005 This works, but is not clean; should the system headers ever move,
1006 you would have to edit the new headers to match.
1008 There is no way to solve this problem within the C standard, but you
1009 can use the GNU extension `#include_next'. It means, "Include the
1010 _next_ file with this name". This directive works like `#include'
1011 except in searching for the specified file: it starts searching the
1012 list of header file directories _after_ the directory in which the
1013 current file was found.
1015 Suppose you specify `-I /usr/local/include', and the list of
1016 directories to search also includes `/usr/include'; and suppose both
1017 directories contain `signal.h'. Ordinary `#include <signal.h>' finds
1018 the file under `/usr/local/include'. If that file contains
1019 `#include_next <signal.h>', it starts searching after that directory,
1020 and finds the file in `/usr/include'.
1022 `#include_next' does not distinguish between `<FILE>' and `"FILE"'
1023 inclusion, nor does it check that the file you specify has the same
1024 name as the current file. It simply looks for the file named, starting
1025 with the directory in the search path after the one where the current
1028 The use of `#include_next' can lead to great confusion. We
1029 recommend it be used only when there is no other alternative. In
1030 particular, it should not be used in the headers belonging to a specific
1031 program; it should be used only to make global corrections along the
1032 lines of `fixincludes'.
1035 File: cpp.info, Node: System Headers, Prev: Wrapper Headers, Up: Header Files
1040 The header files declaring interfaces to the operating system and
1041 runtime libraries often cannot be written in strictly conforming C.
1042 Therefore, GCC gives code found in "system headers" special treatment.
1043 All warnings, other than those generated by `#warning' (*note
1044 Diagnostics::), are suppressed while GCC is processing a system header.
1045 Macros defined in a system header are immune to a few warnings wherever
1046 they are expanded. This immunity is granted on an ad-hoc basis, when
1047 we find that a warning generates lots of false positives because of
1048 code in macros defined in system headers.
1050 Normally, only the headers found in specific directories are
1051 considered system headers. These directories are determined when GCC
1052 is compiled. There are, however, two ways to make normal headers into
1055 The `-isystem' command line option adds its argument to the list of
1056 directories to search for headers, just like `-I'. Any headers found
1057 in that directory will be considered system headers.
1059 All directories named by `-isystem' are searched _after_ all
1060 directories named by `-I', no matter what their order was on the
1061 command line. If the same directory is named by both `-I' and
1062 `-isystem', the `-I' option is ignored. GCC provides an informative
1063 message when this occurs if `-v' is used.
1065 There is also a directive, `#pragma GCC system_header', which tells
1066 GCC to consider the rest of the current include file a system header,
1067 no matter where it was found. Code that comes before the `#pragma' in
1068 the file will not be affected. `#pragma GCC system_header' has no
1069 effect in the primary source file.
1071 On very old systems, some of the pre-defined system header
1072 directories get even more special treatment. GNU C++ considers code in
1073 headers found in those directories to be surrounded by an `extern "C"'
1074 block. There is no way to request this behavior with a `#pragma', or
1075 from the command line.
1078 File: cpp.info, Node: Macros, Next: Conditionals, Prev: Header Files, Up: Top
1083 A "macro" is a fragment of code which has been given a name. Whenever
1084 the name is used, it is replaced by the contents of the macro. There
1085 are two kinds of macros. They differ mostly in what they look like
1086 when they are used. "Object-like" macros resemble data objects when
1087 used, "function-like" macros resemble function calls.
1089 You may define any valid identifier as a macro, even if it is a C
1090 keyword. The preprocessor does not know anything about keywords. This
1091 can be useful if you wish to hide a keyword such as `const' from an
1092 older compiler that does not understand it. However, the preprocessor
1093 operator `defined' (*note Defined::) can never be defined as a macro,
1094 and C++'s named operators (*note C++ Named Operators::) cannot be
1095 macros when you are compiling C++.
1099 * Object-like Macros::
1100 * Function-like Macros::
1105 * Predefined Macros::
1106 * Undefining and Redefining Macros::
1107 * Directives Within Macro Arguments::
1111 File: cpp.info, Node: Object-like Macros, Next: Function-like Macros, Up: Macros
1113 3.1 Object-like Macros
1114 ======================
1116 An "object-like macro" is a simple identifier which will be replaced by
1117 a code fragment. It is called object-like because it looks like a data
1118 object in code that uses it. They are most commonly used to give
1119 symbolic names to numeric constants.
1121 You create macros with the `#define' directive. `#define' is
1122 followed by the name of the macro and then the token sequence it should
1123 be an abbreviation for, which is variously referred to as the macro's
1124 "body", "expansion" or "replacement list". For example,
1126 #define BUFFER_SIZE 1024
1128 defines a macro named `BUFFER_SIZE' as an abbreviation for the token
1129 `1024'. If somewhere after this `#define' directive there comes a C
1130 statement of the form
1132 foo = (char *) malloc (BUFFER_SIZE);
1134 then the C preprocessor will recognize and "expand" the macro
1135 `BUFFER_SIZE'. The C compiler will see the same tokens as it would if
1138 foo = (char *) malloc (1024);
1140 By convention, macro names are written in uppercase. Programs are
1141 easier to read when it is possible to tell at a glance which names are
1144 The macro's body ends at the end of the `#define' line. You may
1145 continue the definition onto multiple lines, if necessary, using
1146 backslash-newline. When the macro is expanded, however, it will all
1147 come out on one line. For example,
1149 #define NUMBERS 1, \
1152 int x[] = { NUMBERS };
1153 ==> int x[] = { 1, 2, 3 };
1155 The most common visible consequence of this is surprising line numbers
1158 There is no restriction on what can go in a macro body provided it
1159 decomposes into valid preprocessing tokens. Parentheses need not
1160 balance, and the body need not resemble valid C code. (If it does not,
1161 you may get error messages from the C compiler when you use the macro.)
1163 The C preprocessor scans your program sequentially. Macro
1164 definitions take effect at the place you write them. Therefore, the
1165 following input to the C preprocessor
1176 When the preprocessor expands a macro name, the macro's expansion
1177 replaces the macro invocation, then the expansion is examined for more
1178 macros to expand. For example,
1180 #define TABLESIZE BUFSIZE
1181 #define BUFSIZE 1024
1186 `TABLESIZE' is expanded first to produce `BUFSIZE', then that macro is
1187 expanded to produce the final result, `1024'.
1189 Notice that `BUFSIZE' was not defined when `TABLESIZE' was defined.
1190 The `#define' for `TABLESIZE' uses exactly the expansion you
1191 specify--in this case, `BUFSIZE'--and does not check to see whether it
1192 too contains macro names. Only when you _use_ `TABLESIZE' is the
1193 result of its expansion scanned for more macro names.
1195 This makes a difference if you change the definition of `BUFSIZE' at
1196 some point in the source file. `TABLESIZE', defined as shown, will
1197 always expand using the definition of `BUFSIZE' that is currently in
1200 #define BUFSIZE 1020
1201 #define TABLESIZE BUFSIZE
1205 Now `TABLESIZE' expands (in two stages) to `37'.
1207 If the expansion of a macro contains its own name, either directly or
1208 via intermediate macros, it is not expanded again when the expansion is
1209 examined for more macros. This prevents infinite recursion. *Note
1210 Self-Referential Macros::, for the precise details.
1213 File: cpp.info, Node: Function-like Macros, Next: Macro Arguments, Prev: Object-like Macros, Up: Macros
1215 3.2 Function-like Macros
1216 ========================
1218 You can also define macros whose use looks like a function call. These
1219 are called "function-like macros". To define a function-like macro,
1220 you use the same `#define' directive, but you put a pair of parentheses
1221 immediately after the macro name. For example,
1223 #define lang_init() c_init()
1227 A function-like macro is only expanded if its name appears with a
1228 pair of parentheses after it. If you write just the name, it is left
1229 alone. This can be useful when you have a function and a macro of the
1230 same name, and you wish to use the function sometimes.
1232 extern void foo(void);
1233 #define foo() /* optimized inline version */
1238 Here the call to `foo()' will use the macro, but the function
1239 pointer will get the address of the real function. If the macro were to
1240 be expanded, it would cause a syntax error.
1242 If you put spaces between the macro name and the parentheses in the
1243 macro definition, that does not define a function-like macro, it defines
1244 an object-like macro whose expansion happens to begin with a pair of
1247 #define lang_init () c_init()
1251 The first two pairs of parentheses in this expansion come from the
1252 macro. The third is the pair that was originally after the macro
1253 invocation. Since `lang_init' is an object-like macro, it does not
1254 consume those parentheses.
1257 File: cpp.info, Node: Macro Arguments, Next: Stringification, Prev: Function-like Macros, Up: Macros
1262 Function-like macros can take "arguments", just like true functions.
1263 To define a macro that uses arguments, you insert "parameters" between
1264 the pair of parentheses in the macro definition that make the macro
1265 function-like. The parameters must be valid C identifiers, separated
1266 by commas and optionally whitespace.
1268 To invoke a macro that takes arguments, you write the name of the
1269 macro followed by a list of "actual arguments" in parentheses, separated
1270 by commas. The invocation of the macro need not be restricted to a
1271 single logical line--it can cross as many lines in the source file as
1272 you wish. The number of arguments you give must match the number of
1273 parameters in the macro definition. When the macro is expanded, each
1274 use of a parameter in its body is replaced by the tokens of the
1275 corresponding argument. (You need not use all of the parameters in the
1278 As an example, here is a macro that computes the minimum of two
1279 numeric values, as it is defined in many C programs, and some uses.
1281 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
1282 x = min(a, b); ==> x = ((a) < (b) ? (a) : (b));
1283 y = min(1, 2); ==> y = ((1) < (2) ? (1) : (2));
1284 z = min(a + 28, *p); ==> z = ((a + 28) < (*p) ? (a + 28) : (*p));
1286 (In this small example you can already see several of the dangers of
1287 macro arguments. *Note Macro Pitfalls::, for detailed explanations.)
1289 Leading and trailing whitespace in each argument is dropped, and all
1290 whitespace between the tokens of an argument is reduced to a single
1291 space. Parentheses within each argument must balance; a comma within
1292 such parentheses does not end the argument. However, there is no
1293 requirement for square brackets or braces to balance, and they do not
1294 prevent a comma from separating arguments. Thus,
1296 macro (array[x = y, x + 1])
1298 passes two arguments to `macro': `array[x = y' and `x + 1]'. If you
1299 want to supply `array[x = y, x + 1]' as an argument, you can write it
1300 as `array[(x = y, x + 1)]', which is equivalent C code.
1302 All arguments to a macro are completely macro-expanded before they
1303 are substituted into the macro body. After substitution, the complete
1304 text is scanned again for macros to expand, including the arguments.
1305 This rule may seem strange, but it is carefully designed so you need
1306 not worry about whether any function call is actually a macro
1307 invocation. You can run into trouble if you try to be too clever,
1308 though. *Note Argument Prescan::, for detailed discussion.
1310 For example, `min (min (a, b), c)' is first expanded to
1312 min (((a) < (b) ? (a) : (b)), (c))
1316 ((((a) < (b) ? (a) : (b))) < (c)
1317 ? (((a) < (b) ? (a) : (b)))
1320 (Line breaks shown here for clarity would not actually be generated.)
1322 You can leave macro arguments empty; this is not an error to the
1323 preprocessor (but many macros will then expand to invalid code). You
1324 cannot leave out arguments entirely; if a macro takes two arguments,
1325 there must be exactly one comma at the top level of its argument list.
1326 Here are some silly examples using `min':
1328 min(, b) ==> (( ) < (b) ? ( ) : (b))
1329 min(a, ) ==> ((a ) < ( ) ? (a ) : ( ))
1330 min(,) ==> (( ) < ( ) ? ( ) : ( ))
1331 min((,),) ==> (((,)) < ( ) ? ((,)) : ( ))
1333 min() error--> macro "min" requires 2 arguments, but only 1 given
1334 min(,,) error--> macro "min" passed 3 arguments, but takes just 2
1336 Whitespace is not a preprocessing token, so if a macro `foo' takes
1337 one argument, `foo ()' and `foo ( )' both supply it an empty argument.
1338 Previous GNU preprocessor implementations and documentation were
1339 incorrect on this point, insisting that a function-like macro that
1340 takes a single argument be passed a space if an empty argument was
1343 Macro parameters appearing inside string literals are not replaced by
1344 their corresponding actual arguments.
1346 #define foo(x) x, "x"
1347 foo(bar) ==> bar, "x"
1350 File: cpp.info, Node: Stringification, Next: Concatenation, Prev: Macro Arguments, Up: Macros
1355 Sometimes you may want to convert a macro argument into a string
1356 constant. Parameters are not replaced inside string constants, but you
1357 can use the `#' preprocessing operator instead. When a macro parameter
1358 is used with a leading `#', the preprocessor replaces it with the
1359 literal text of the actual argument, converted to a string constant.
1360 Unlike normal parameter replacement, the argument is not macro-expanded
1361 first. This is called "stringification".
1363 There is no way to combine an argument with surrounding text and
1364 stringify it all together. Instead, you can write a series of adjacent
1365 string constants and stringified arguments. The preprocessor will
1366 replace the stringified arguments with string constants. The C
1367 compiler will then combine all the adjacent string constants into one
1370 Here is an example of a macro definition that uses stringification:
1372 #define WARN_IF(EXP) \
1374 fprintf (stderr, "Warning: " #EXP "\n"); } \
1377 ==> do { if (x == 0)
1378 fprintf (stderr, "Warning: " "x == 0" "\n"); } while (0);
1380 The argument for `EXP' is substituted once, as-is, into the `if'
1381 statement, and once, stringified, into the argument to `fprintf'. If
1382 `x' were a macro, it would be expanded in the `if' statement, but not
1385 The `do' and `while (0)' are a kludge to make it possible to write
1386 `WARN_IF (ARG);', which the resemblance of `WARN_IF' to a function
1387 would make C programmers want to do; see *note Swallowing the
1390 Stringification in C involves more than putting double-quote
1391 characters around the fragment. The preprocessor backslash-escapes the
1392 quotes surrounding embedded string constants, and all backslashes
1393 within string and character constants, in order to get a valid C string
1394 constant with the proper contents. Thus, stringifying `p = "foo\n";'
1395 results in "p = \"foo\\n\";". However, backslashes that are not inside
1396 string or character constants are not duplicated: `\n' by itself
1397 stringifies to "\n".
1399 All leading and trailing whitespace in text being stringified is
1400 ignored. Any sequence of whitespace in the middle of the text is
1401 converted to a single space in the stringified result. Comments are
1402 replaced by whitespace long before stringification happens, so they
1403 never appear in stringified text.
1405 There is no way to convert a macro argument into a character
1408 If you want to stringify the result of expansion of a macro argument,
1409 you have to use two levels of macros.
1411 #define xstr(s) str(s)
1421 `s' is stringified when it is used in `str', so it is not
1422 macro-expanded first. But `s' is an ordinary argument to `xstr', so it
1423 is completely macro-expanded before `xstr' itself is expanded (*note
1424 Argument Prescan::). Therefore, by the time `str' gets to its
1425 argument, it has already been macro-expanded.
1428 File: cpp.info, Node: Concatenation, Next: Variadic Macros, Prev: Stringification, Up: Macros
1433 It is often useful to merge two tokens into one while expanding macros.
1434 This is called "token pasting" or "token concatenation". The `##'
1435 preprocessing operator performs token pasting. When a macro is
1436 expanded, the two tokens on either side of each `##' operator are
1437 combined into a single token, which then replaces the `##' and the two
1438 original tokens in the macro expansion. Usually both will be
1439 identifiers, or one will be an identifier and the other a preprocessing
1440 number. When pasted, they make a longer identifier. This isn't the
1441 only valid case. It is also possible to concatenate two numbers (or a
1442 number and a name, such as `1.5' and `e3') into a number. Also,
1443 multi-character operators such as `+=' can be formed by token pasting.
1445 However, two tokens that don't together form a valid token cannot be
1446 pasted together. For example, you cannot concatenate `x' with `+' in
1447 either order. If you try, the preprocessor issues a warning and emits
1448 the two tokens. Whether it puts white space between the tokens is
1449 undefined. It is common to find unnecessary uses of `##' in complex
1450 macros. If you get this warning, it is likely that you can simply
1453 Both the tokens combined by `##' could come from the macro body, but
1454 you could just as well write them as one token in the first place.
1455 Token pasting is most useful when one or both of the tokens comes from a
1456 macro argument. If either of the tokens next to an `##' is a parameter
1457 name, it is replaced by its actual argument before `##' executes. As
1458 with stringification, the actual argument is not macro-expanded first.
1459 If the argument is empty, that `##' has no effect.
1461 Keep in mind that the C preprocessor converts comments to whitespace
1462 before macros are even considered. Therefore, you cannot create a
1463 comment by concatenating `/' and `*'. You can put as much whitespace
1464 between `##' and its operands as you like, including comments, and you
1465 can put comments in arguments that will be concatenated. However, it
1466 is an error if `##' appears at either end of a macro body.
1468 Consider a C program that interprets named commands. There probably
1469 needs to be a table of commands, perhaps an array of structures declared
1475 void (*function) (void);
1478 struct command commands[] =
1480 { "quit", quit_command },
1481 { "help", help_command },
1485 It would be cleaner not to have to give each command name twice,
1486 once in the string constant and once in the function name. A macro
1487 which takes the name of a command as an argument can make this
1488 unnecessary. The string constant can be created with stringification,
1489 and the function name by concatenating the argument with `_command'.
1490 Here is how it is done:
1492 #define COMMAND(NAME) { #NAME, NAME ## _command }
1494 struct command commands[] =
1502 File: cpp.info, Node: Variadic Macros, Next: Predefined Macros, Prev: Concatenation, Up: Macros
1507 A macro can be declared to accept a variable number of arguments much as
1508 a function can. The syntax for defining the macro is similar to that of
1509 a function. Here is an example:
1511 #define eprintf(...) fprintf (stderr, __VA_ARGS__)
1513 This kind of macro is called "variadic". When the macro is invoked,
1514 all the tokens in its argument list after the last named argument (this
1515 macro has none), including any commas, become the "variable argument".
1516 This sequence of tokens replaces the identifier `__VA_ARGS__' in the
1517 macro body wherever it appears. Thus, we have this expansion:
1519 eprintf ("%s:%d: ", input_file, lineno)
1520 ==> fprintf (stderr, "%s:%d: ", input_file, lineno)
1522 The variable argument is completely macro-expanded before it is
1523 inserted into the macro expansion, just like an ordinary argument. You
1524 may use the `#' and `##' operators to stringify the variable argument
1525 or to paste its leading or trailing token with another token. (But see
1526 below for an important special case for `##'.)
1528 If your macro is complicated, you may want a more descriptive name
1529 for the variable argument than `__VA_ARGS__'. CPP permits this, as an
1530 extension. You may write an argument name immediately before the
1531 `...'; that name is used for the variable argument. The `eprintf'
1532 macro above could be written
1534 #define eprintf(args...) fprintf (stderr, args)
1536 using this extension. You cannot use `__VA_ARGS__' and this extension
1539 You can have named arguments as well as variable arguments in a
1540 variadic macro. We could define `eprintf' like this, instead:
1542 #define eprintf(format, ...) fprintf (stderr, format, __VA_ARGS__)
1544 This formulation looks more descriptive, but unfortunately it is less
1545 flexible: you must now supply at least one argument after the format
1546 string. In standard C, you cannot omit the comma separating the named
1547 argument from the variable arguments. Furthermore, if you leave the
1548 variable argument empty, you will get a syntax error, because there
1549 will be an extra comma after the format string.
1551 eprintf("success!\n", );
1552 ==> fprintf(stderr, "success!\n", );
1554 GNU CPP has a pair of extensions which deal with this problem.
1555 First, you are allowed to leave the variable argument out entirely:
1557 eprintf ("success!\n")
1558 ==> fprintf(stderr, "success!\n", );
1560 Second, the `##' token paste operator has a special meaning when placed
1561 between a comma and a variable argument. If you write
1563 #define eprintf(format, ...) fprintf (stderr, format, ##__VA_ARGS__)
1565 and the variable argument is left out when the `eprintf' macro is used,
1566 then the comma before the `##' will be deleted. This does _not_ happen
1567 if you pass an empty argument, nor does it happen if the token
1568 preceding `##' is anything other than a comma.
1570 eprintf ("success!\n")
1571 ==> fprintf(stderr, "success!\n");
1573 The above explanation is ambiguous about the case where the only macro
1574 parameter is a variable arguments parameter, as it is meaningless to
1575 try to distinguish whether no argument at all is an empty argument or a
1576 missing argument. In this case the C99 standard is clear that the
1577 comma must remain, however the existing GCC extension used to swallow
1578 the comma. So CPP retains the comma when conforming to a specific C
1579 standard, and drops it otherwise.
1581 C99 mandates that the only place the identifier `__VA_ARGS__' can
1582 appear is in the replacement list of a variadic macro. It may not be
1583 used as a macro name, macro argument name, or within a different type
1584 of macro. It may also be forbidden in open text; the standard is
1585 ambiguous. We recommend you avoid using it except for its defined
1588 Variadic macros are a new feature in C99. GNU CPP has supported them
1589 for a long time, but only with a named variable argument (`args...',
1590 not `...' and `__VA_ARGS__'). If you are concerned with portability to
1591 previous versions of GCC, you should use only named variable arguments.
1592 On the other hand, if you are concerned with portability to other
1593 conforming implementations of C99, you should use only `__VA_ARGS__'.
1595 Previous versions of CPP implemented the comma-deletion extension
1596 much more generally. We have restricted it in this release to minimize
1597 the differences from C99. To get the same effect with both this and
1598 previous versions of GCC, the token preceding the special `##' must be
1599 a comma, and there must be white space between that comma and whatever
1600 comes immediately before it:
1602 #define eprintf(format, args...) fprintf (stderr, format , ##args)
1604 *Note Differences from previous versions::, for the gory details.
1607 File: cpp.info, Node: Predefined Macros, Next: Undefining and Redefining Macros, Prev: Variadic Macros, Up: Macros
1609 3.7 Predefined Macros
1610 =====================
1612 Several object-like macros are predefined; you use them without
1613 supplying their definitions. They fall into three classes: standard,
1614 common, and system-specific.
1616 In C++, there is a fourth category, the named operators. They act
1617 like predefined macros, but you cannot undefine them.
1621 * Standard Predefined Macros::
1622 * Common Predefined Macros::
1623 * System-specific Predefined Macros::
1624 * C++ Named Operators::
1627 File: cpp.info, Node: Standard Predefined Macros, Next: Common Predefined Macros, Up: Predefined Macros
1629 3.7.1 Standard Predefined Macros
1630 --------------------------------
1632 The standard predefined macros are specified by the relevant language
1633 standards, so they are available with all compilers that implement
1634 those standards. Older compilers may not provide all of them. Their
1635 names all start with double underscores.
1638 This macro expands to the name of the current input file, in the
1639 form of a C string constant. This is the path by which the
1640 preprocessor opened the file, not the short name specified in
1641 `#include' or as the input file name argument. For example,
1642 `"/usr/local/include/myheader.h"' is a possible expansion of this
1646 This macro expands to the current input line number, in the form
1647 of a decimal integer constant. While we call it a predefined
1648 macro, it's a pretty strange macro, since its "definition" changes
1649 with each new line of source code.
1651 `__FILE__' and `__LINE__' are useful in generating an error message
1652 to report an inconsistency detected by the program; the message can
1653 state the source line at which the inconsistency was detected. For
1656 fprintf (stderr, "Internal error: "
1657 "negative string length "
1658 "%d at %s, line %d.",
1659 length, __FILE__, __LINE__);
1661 An `#include' directive changes the expansions of `__FILE__' and
1662 `__LINE__' to correspond to the included file. At the end of that
1663 file, when processing resumes on the input file that contained the
1664 `#include' directive, the expansions of `__FILE__' and `__LINE__'
1665 revert to the values they had before the `#include' (but `__LINE__' is
1666 then incremented by one as processing moves to the line after the
1669 A `#line' directive changes `__LINE__', and may change `__FILE__' as
1670 well. *Note Line Control::.
1672 C99 introduces `__func__', and GCC has provided `__FUNCTION__' for a
1673 long time. Both of these are strings containing the name of the
1674 current function (there are slight semantic differences; see the GCC
1675 manual). Neither of them is a macro; the preprocessor does not know the
1676 name of the current function. They tend to be useful in conjunction
1677 with `__FILE__' and `__LINE__', though.
1680 This macro expands to a string constant that describes the date on
1681 which the preprocessor is being run. The string constant contains
1682 eleven characters and looks like `"Feb 12 1996"'. If the day of
1683 the month is less than 10, it is padded with a space on the left.
1685 If GCC cannot determine the current date, it will emit a warning
1686 message (once per compilation) and `__DATE__' will expand to
1690 This macro expands to a string constant that describes the time at
1691 which the preprocessor is being run. The string constant contains
1692 eight characters and looks like `"23:59:01"'.
1694 If GCC cannot determine the current time, it will emit a warning
1695 message (once per compilation) and `__TIME__' will expand to
1699 In normal operation, this macro expands to the constant 1, to
1700 signify that this compiler conforms to ISO Standard C. If GNU CPP
1701 is used with a compiler other than GCC, this is not necessarily
1702 true; however, the preprocessor always conforms to the standard
1703 unless the `-traditional-cpp' option is used.
1705 This macro is not defined if the `-traditional-cpp' option is used.
1707 On some hosts, the system compiler uses a different convention,
1708 where `__STDC__' is normally 0, but is 1 if the user specifies
1709 strict conformance to the C Standard. CPP follows the host
1710 convention when processing system header files, but when
1711 processing user files `__STDC__' is always 1. This has been
1712 reported to cause problems; for instance, some versions of Solaris
1713 provide X Windows headers that expect `__STDC__' to be either
1714 undefined or 1. *Note Invocation::.
1717 This macro expands to the C Standard's version number, a long
1718 integer constant of the form `YYYYMML' where YYYY and MM are the
1719 year and month of the Standard version. This signifies which
1720 version of the C Standard the compiler conforms to. Like
1721 `__STDC__', this is not necessarily accurate for the entire
1722 implementation, unless GNU CPP is being used with GCC.
1724 The value `199409L' signifies the 1989 C standard as amended in
1725 1994, which is the current default; the value `199901L' signifies
1726 the 1999 revision of the C standard. Support for the 1999
1727 revision is not yet complete.
1729 This macro is not defined if the `-traditional-cpp' option is
1730 used, nor when compiling C++ or Objective-C.
1733 This macro is defined, with value 1, if the compiler's target is a
1734 "hosted environment". A hosted environment has the complete
1735 facilities of the standard C library available.
1738 This macro is defined when the C++ compiler is in use. You can use
1739 `__cplusplus' to test whether a header is compiled by a C compiler
1740 or a C++ compiler. This macro is similar to `__STDC_VERSION__', in
1741 that it expands to a version number. A fully conforming
1742 implementation of the 1998 C++ standard will define this macro to
1743 `199711L'. The GNU C++ compiler is not yet fully conforming, so
1744 it uses `1' instead. It is hoped to complete the implementation
1745 of standard C++ in the near future.
1748 This macro is defined, with value 1, when the Objective-C compiler
1749 is in use. You can use `__OBJC__' to test whether a header is
1750 compiled by a C compiler or an Objective-C compiler.
1753 This macro is defined with value 1 when preprocessing assembly
1758 File: cpp.info, Node: Common Predefined Macros, Next: System-specific Predefined Macros, Prev: Standard Predefined Macros, Up: Predefined Macros
1760 3.7.2 Common Predefined Macros
1761 ------------------------------
1763 The common predefined macros are GNU C extensions. They are available
1764 with the same meanings regardless of the machine or operating system on
1765 which you are using GNU C or GNU Fortran. Their names all start with
1769 This macro expands to sequential integral values starting from 0.
1770 In conjunction with the `##' operator, this provides a convenient
1771 means to generate unique identifiers. Care must be taken to
1772 ensure that `__COUNTER__' is not expanded prior to inclusion of
1773 precompiled headers which use it. Otherwise, the precompiled
1774 headers will not be used.
1777 The GNU Fortran compiler defines this.
1781 `__GNUC_PATCHLEVEL__'
1782 These macros are defined by all GNU compilers that use the C
1783 preprocessor: C, C++, Objective-C and Fortran. Their values are
1784 the major version, minor version, and patch level of the compiler,
1785 as integer constants. For example, GCC 3.2.1 will define
1786 `__GNUC__' to 3, `__GNUC_MINOR__' to 2, and `__GNUC_PATCHLEVEL__'
1787 to 1. These macros are also defined if you invoke the
1788 preprocessor directly.
1790 `__GNUC_PATCHLEVEL__' is new to GCC 3.0; it is also present in the
1791 widely-used development snapshots leading up to 3.0 (which identify
1792 themselves as GCC 2.96 or 2.97, depending on which snapshot you
1795 If all you need to know is whether or not your program is being
1796 compiled by GCC, or a non-GCC compiler that claims to accept the
1797 GNU C dialects, you can simply test `__GNUC__'. If you need to
1798 write code which depends on a specific version, you must be more
1799 careful. Each time the minor version is increased, the patch
1800 level is reset to zero; each time the major version is increased
1801 (which happens rarely), the minor version and patch level are
1802 reset. If you wish to use the predefined macros directly in the
1803 conditional, you will need to write it like this:
1805 /* Test for GCC > 3.2.0 */
1806 #if __GNUC__ > 3 || \
1807 (__GNUC__ == 3 && (__GNUC_MINOR__ > 2 || \
1808 (__GNUC_MINOR__ == 2 && \
1809 __GNUC_PATCHLEVEL__ > 0))
1811 Another approach is to use the predefined macros to calculate a
1812 single number, then compare that against a threshold:
1814 #define GCC_VERSION (__GNUC__ * 10000 \
1815 + __GNUC_MINOR__ * 100 \
1816 + __GNUC_PATCHLEVEL__)
1818 /* Test for GCC > 3.2.0 */
1819 #if GCC_VERSION > 30200
1821 Many people find this form easier to understand.
1824 The GNU C++ compiler defines this. Testing it is equivalent to
1825 testing `(__GNUC__ && __cplusplus)'.
1828 GCC defines this macro if and only if the `-ansi' switch, or a
1829 `-std' switch specifying strict conformance to some version of ISO
1830 C or ISO C++, was specified when GCC was invoked. It is defined
1831 to `1'. This macro exists primarily to direct GNU libc's header
1832 files to restrict their definitions to the minimal set found in
1833 the 1989 C standard.
1836 This macro expands to the name of the main input file, in the form
1837 of a C string constant. This is the source file that was specified
1838 on the command line of the preprocessor or C compiler.
1841 This macro expands to a decimal integer constant that represents
1842 the depth of nesting in include files. The value of this macro is
1843 incremented on every `#include' directive and decremented at the
1844 end of every included file. It starts out at 0, its value within
1845 the base file specified on the command line.
1848 This macro is defined if the target uses the ELF object format.
1851 This macro expands to a string constant which describes the
1852 version of the compiler in use. You should not rely on its
1853 contents having any particular form, but it can be counted on to
1854 contain at least the release number.
1859 These macros describe the compilation mode. `__OPTIMIZE__' is
1860 defined in all optimizing compilations. `__OPTIMIZE_SIZE__' is
1861 defined if the compiler is optimizing for size, not speed.
1862 `__NO_INLINE__' is defined if no functions will be inlined into
1863 their callers (when not optimizing, or when inlining has been
1864 specifically disabled by `-fno-inline').
1866 These macros cause certain GNU header files to provide optimized
1867 definitions, using macros or inline functions, of system library
1868 functions. You should not use these macros in any way unless you
1869 make sure that programs will execute with the same effect whether
1870 or not they are defined. If they are defined, their value is 1.
1872 `__GNUC_GNU_INLINE__'
1873 GCC defines this macro if functions declared `inline' will be
1874 handled in GCC's traditional gnu90 mode. Object files will contain
1875 externally visible definitions of all functions declared `inline'
1876 without `extern' or `static'. They will not contain any
1877 definitions of any functions declared `extern inline'.
1879 `__GNUC_STDC_INLINE__'
1880 GCC defines this macro if functions declared `inline' will be
1881 handled according to the ISO C99 standard. Object files will
1882 contain externally visible definitions of all functions declared
1883 `extern inline'. They will not contain definitions of any
1884 functions declared `inline' without `extern'.
1886 If this macro is defined, GCC supports the `gnu_inline' function
1887 attribute as a way to always get the gnu90 behavior. Support for
1888 this and `__GNUC_GNU_INLINE__' was added in GCC 4.1.3. If neither
1889 macro is defined, an older version of GCC is being used: `inline'
1890 functions will be compiled in gnu90 mode, and the `gnu_inline'
1891 function attribute will not be recognized.
1894 GCC defines this macro if and only if the data type `char' is
1895 unsigned on the target machine. It exists to cause the standard
1896 header file `limits.h' to work correctly. You should not use this
1897 macro yourself; instead, refer to the standard macros defined in
1900 `__WCHAR_UNSIGNED__'
1901 Like `__CHAR_UNSIGNED__', this macro is defined if and only if the
1902 data type `wchar_t' is unsigned and the front-end is in C++ mode.
1904 `__REGISTER_PREFIX__'
1905 This macro expands to a single token (not a string constant) which
1906 is the prefix applied to CPU register names in assembly language
1907 for this target. You can use it to write assembly that is usable
1908 in multiple environments. For example, in the `m68k-aout'
1909 environment it expands to nothing, but in the `m68k-coff'
1910 environment it expands to a single `%'.
1912 `__USER_LABEL_PREFIX__'
1913 This macro expands to a single token which is the prefix applied to
1914 user labels (symbols visible to C code) in assembly. For example,
1915 in the `m68k-aout' environment it expands to an `_', but in the
1916 `m68k-coff' environment it expands to nothing.
1918 This macro will have the correct definition even if
1919 `-f(no-)underscores' is in use, but it will not be correct if
1920 target-specific options that adjust this prefix are used (e.g. the
1921 OSF/rose `-mno-underscores' option).
1929 `__SIG_ATOMIC_TYPE__'
1938 `__INT_LEAST8_TYPE__'
1939 `__INT_LEAST16_TYPE__'
1940 `__INT_LEAST32_TYPE__'
1941 `__INT_LEAST64_TYPE__'
1942 `__UINT_LEAST8_TYPE__'
1943 `__UINT_LEAST16_TYPE__'
1944 `__UINT_LEAST32_TYPE__'
1945 `__UINT_LEAST64_TYPE__'
1946 `__INT_FAST8_TYPE__'
1947 `__INT_FAST16_TYPE__'
1948 `__INT_FAST32_TYPE__'
1949 `__INT_FAST64_TYPE__'
1950 `__UINT_FAST8_TYPE__'
1951 `__UINT_FAST16_TYPE__'
1952 `__UINT_FAST32_TYPE__'
1953 `__UINT_FAST64_TYPE__'
1956 These macros are defined to the correct underlying types for the
1957 `size_t', `ptrdiff_t', `wchar_t', `wint_t', `intmax_t',
1958 `uintmax_t', `sig_atomic_t', `int8_t', `int16_t', `int32_t',
1959 `int64_t', `uint8_t', `uint16_t', `uint32_t', `uint64_t',
1960 `int_least8_t', `int_least16_t', `int_least32_t', `int_least64_t',
1961 `uint_least8_t', `uint_least16_t', `uint_least32_t',
1962 `uint_least64_t', `int_fast8_t', `int_fast16_t', `int_fast32_t',
1963 `int_fast64_t', `uint_fast8_t', `uint_fast16_t', `uint_fast32_t',
1964 `uint_fast64_t', `intptr_t', and `uintptr_t' typedefs,
1965 respectively. They exist to make the standard header files
1966 `stddef.h', `stdint.h', and `wchar.h' work correctly. You should
1967 not use these macros directly; instead, include the appropriate
1968 headers and use the typedefs. Some of these macros may not be
1969 defined on particular systems if GCC does not provide a `stdint.h'
1970 header on those systems.
1973 Defined to the number of bits used in the representation of the
1974 `char' data type. It exists to make the standard header given
1975 numerical limits work correctly. You should not use this macro
1976 directly; instead, include the appropriate headers.
1989 `__SIG_ATOMIC_MAX__'
1998 `__INT_LEAST8_MAX__'
1999 `__INT_LEAST16_MAX__'
2000 `__INT_LEAST32_MAX__'
2001 `__INT_LEAST64_MAX__'
2002 `__UINT_LEAST8_MAX__'
2003 `__UINT_LEAST16_MAX__'
2004 `__UINT_LEAST32_MAX__'
2005 `__UINT_LEAST64_MAX__'
2007 `__INT_FAST16_MAX__'
2008 `__INT_FAST32_MAX__'
2009 `__INT_FAST64_MAX__'
2010 `__UINT_FAST8_MAX__'
2011 `__UINT_FAST16_MAX__'
2012 `__UINT_FAST32_MAX__'
2013 `__UINT_FAST64_MAX__'
2018 `__SIG_ATOMIC_MIN__'
2019 Defined to the maximum value of the `signed char', `wchar_t',
2020 `signed short', `signed int', `signed long', `signed long long',
2021 `wint_t', `size_t', `ptrdiff_t', `intmax_t', `uintmax_t',
2022 `sig_atomic_t', `int8_t', `int16_t', `int32_t', `int64_t',
2023 `uint8_t', `uint16_t', `uint32_t', `uint64_t', `int_least8_t',
2024 `int_least16_t', `int_least32_t', `int_least64_t',
2025 `uint_least8_t', `uint_least16_t', `uint_least32_t',
2026 `uint_least64_t', `int_fast8_t', `int_fast16_t', `int_fast32_t',
2027 `int_fast64_t', `uint_fast8_t', `uint_fast16_t', `uint_fast32_t',
2028 `uint_fast64_t', `intptr_t', and `uintptr_t' types and to the
2029 minimum value of the `wchar_t', `wint_t', and `sig_atomic_t' types
2030 respectively. They exist to make the standard header given
2031 numerical limits work correctly. You should not use these macros
2032 directly; instead, include the appropriate headers. Some of these
2033 macros may not be defined on particular systems if GCC does not
2034 provide a `stdint.h' header on those systems.
2046 Defined to implementations of the standard `stdint.h' macros with
2047 the same names without the leading `__'. They exist the make the
2048 implementation of that header work correctly. You should not use
2049 these macros directly; instead, include the appropriate headers.
2050 Some of these macros may not be defined on particular systems if
2051 GCC does not provide a `stdint.h' header on those systems.
2055 `__SIZEOF_LONG_LONG__'
2057 `__SIZEOF_POINTER__'
2060 `__SIZEOF_LONG_DOUBLE__'
2062 `__SIZEOF_WCHAR_T__'
2064 `__SIZEOF_PTRDIFF_T__'
2065 Defined to the number of bytes of the C standard data types: `int',
2066 `long', `long long', `short', `void *', `float', `double', `long
2067 double', `size_t', `wchar_t', `wint_t' and `ptrdiff_t'.
2070 `__ORDER_LITTLE_ENDIAN__'
2071 `__ORDER_BIG_ENDIAN__'
2072 `__ORDER_PDP_ENDIAN__'
2073 `__BYTE_ORDER__' is defined to one of the values
2074 `__ORDER_LITTLE_ENDIAN__', `__ORDER_BIG_ENDIAN__', or
2075 `__ORDER_PDP_ENDIAN__' to reflect the layout of multi-byte and
2076 multi-word quantities in memory. If `__BYTE_ORDER__' is equal to
2077 `__ORDER_LITTLE_ENDIAN__' or `__ORDER_BIG_ENDIAN__', then
2078 multi-byte and multi-word quantities are laid out identically: the
2079 byte (word) at the lowest address is the least significant or most
2080 significant byte (word) of the quantity, respectively. If
2081 `__BYTE_ORDER__' is equal to `__ORDER_PDP_ENDIAN__', then bytes in
2082 16-bit words are laid out in a little-endian fashion, whereas the
2083 16-bit subwords of a 32-bit quantity are laid out in big-endian
2086 You should use these macros for testing like this:
2088 /* Test for a little-endian machine */
2089 #if __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__
2091 `__FLOAT_WORD_ORDER__'
2092 `__FLOAT_WORD_ORDER__' is defined to one of the values
2093 `__ORDER_LITTLE_ENDIAN__' or `__ORDER_BIG_ENDIAN__' to reflect the
2094 layout of the words of multi-word floating-point quantities.
2097 This macro is defined, with value 1, when compiling a C++ source
2098 file with warnings about deprecated constructs enabled. These
2099 warnings are enabled by default, but can be disabled with
2103 This macro is defined, with value 1, when compiling a C++ source
2104 file with exceptions enabled. If `-fno-exceptions' is used when
2105 compiling the file, then this macro is not defined.
2108 This macro is defined, with value 1, when compiling a C++ source
2109 file with runtime type identification enabled. If `-fno-rtti' is
2110 used when compiling the file, then this macro is not defined.
2112 `__USING_SJLJ_EXCEPTIONS__'
2113 This macro is defined, with value 1, if the compiler uses the old
2114 mechanism based on `setjmp' and `longjmp' for exception handling.
2116 `__GXX_EXPERIMENTAL_CXX0X__'
2117 This macro is defined when compiling a C++ source file with the
2118 option `-std=c++0x' or `-std=gnu++0x'. It indicates that some
2119 features likely to be included in C++0x are available. Note that
2120 these features are experimental, and may change or be removed in
2121 future versions of GCC.
2124 This macro is defined when compiling a C++ source file. It has the
2125 value 1 if the compiler will use weak symbols, COMDAT sections, or
2126 other similar techniques to collapse symbols with "vague linkage"
2127 that are defined in multiple translation units. If the compiler
2128 will not collapse such symbols, this macro is defined with value
2129 0. In general, user code should not need to make use of this
2130 macro; the purpose of this macro is to ease implementation of the
2131 C++ runtime library provided with G++.
2134 This macro is defined, with value 1, if (and only if) the NeXT
2135 runtime (as in `-fnext-runtime') is in use for Objective-C. If
2136 the GNU runtime is used, this macro is not defined, so that you
2137 can use this macro to determine which runtime (NeXT or GNU) is
2142 These macros are defined, with value 1, if (and only if) the
2143 compilation is for a target where `long int' and pointer both use
2144 64-bits and `int' uses 32-bit.
2147 This macro is defined, with value 1, when `-fstack-protector' is in
2151 This macro is defined, with value 2, when `-fstack-protector-all'
2155 This macro expands to a string constant that describes the date
2156 and time of the last modification of the current source file. The
2157 string constant contains abbreviated day of the week, month, day
2158 of the month, time in hh:mm:ss form, year and looks like
2159 `"Sun Sep 16 01:03:52 1973"'. If the day of the month is less
2160 than 10, it is padded with a space on the left.
2162 If GCC cannot determine the current date, it will emit a warning
2163 message (once per compilation) and `__TIMESTAMP__' will expand to
2164 `"??? ??? ?? ??:??:?? ????"'.
2166 `__GCC_HAVE_SYNC_COMPARE_AND_SWAP_1'
2167 `__GCC_HAVE_SYNC_COMPARE_AND_SWAP_2'
2168 `__GCC_HAVE_SYNC_COMPARE_AND_SWAP_4'
2169 `__GCC_HAVE_SYNC_COMPARE_AND_SWAP_8'
2170 `__GCC_HAVE_SYNC_COMPARE_AND_SWAP_16'
2171 These macros are defined when the target processor supports atomic
2172 compare and swap operations on operands 1, 2, 4, 8 or 16 bytes in
2173 length, respectively.
2175 `__GCC_HAVE_DWARF2_CFI_ASM'
2176 This macro is defined when the compiler is emitting Dwarf2 CFI
2177 directives to the assembler. When this is defined, it is possible
2178 to emit those same directives in inline assembly.
2183 These macros are defined with value 1 if the backend supports the
2184 `fma', `fmaf', and `fmal' builtin functions, so that the include
2185 file `math.h' can define the macros `FP_FAST_FMA', `FP_FAST_FMAF',
2186 and `FP_FAST_FMAL' for compatibility with the 1999 C standard.
2189 File: cpp.info, Node: System-specific Predefined Macros, Next: C++ Named Operators, Prev: Common Predefined Macros, Up: Predefined Macros
2191 3.7.3 System-specific Predefined Macros
2192 ---------------------------------------
2194 The C preprocessor normally predefines several macros that indicate what
2195 type of system and machine is in use. They are obviously different on
2196 each target supported by GCC. This manual, being for all systems and
2197 machines, cannot tell you what their names are, but you can use `cpp
2198 -dM' to see them all. *Note Invocation::. All system-specific
2199 predefined macros expand to the constant 1, so you can test them with
2200 either `#ifdef' or `#if'.
2202 The C standard requires that all system-specific macros be part of
2203 the "reserved namespace". All names which begin with two underscores,
2204 or an underscore and a capital letter, are reserved for the compiler and
2205 library to use as they wish. However, historically system-specific
2206 macros have had names with no special prefix; for instance, it is common
2207 to find `unix' defined on Unix systems. For all such macros, GCC
2208 provides a parallel macro with two underscores added at the beginning
2209 and the end. If `unix' is defined, `__unix__' will be defined too.
2210 There will never be more than two underscores; the parallel of `_mips'
2213 When the `-ansi' option, or any `-std' option that requests strict
2214 conformance, is given to the compiler, all the system-specific
2215 predefined macros outside the reserved namespace are suppressed. The
2216 parallel macros, inside the reserved namespace, remain defined.
2218 We are slowly phasing out all predefined macros which are outside the
2219 reserved namespace. You should never use them in new programs, and we
2220 encourage you to correct older code to use the parallel macros whenever
2221 you find it. We don't recommend you use the system-specific macros that
2222 are in the reserved namespace, either. It is better in the long run to
2223 check specifically for features you need, using a tool such as
2227 File: cpp.info, Node: C++ Named Operators, Prev: System-specific Predefined Macros, Up: Predefined Macros
2229 3.7.4 C++ Named Operators
2230 -------------------------
2232 In C++, there are eleven keywords which are simply alternate spellings
2233 of operators normally written with punctuation. These keywords are
2234 treated as such even in the preprocessor. They function as operators in
2235 `#if', and they cannot be defined as macros or poisoned. In C, you can
2236 request that those keywords take their C++ meaning by including
2237 `iso646.h'. That header defines each one as a normal object-like macro
2238 expanding to the appropriate punctuator.
2240 These are the named operators and their corresponding punctuators:
2242 Named Operator Punctuator
2256 File: cpp.info, Node: Undefining and Redefining Macros, Next: Directives Within Macro Arguments, Prev: Predefined Macros, Up: Macros
2258 3.8 Undefining and Redefining Macros
2259 ====================================
2261 If a macro ceases to be useful, it may be "undefined" with the `#undef'
2262 directive. `#undef' takes a single argument, the name of the macro to
2263 undefine. You use the bare macro name, even if the macro is
2264 function-like. It is an error if anything appears on the line after
2265 the macro name. `#undef' has no effect if the name is not a macro.
2270 x = FOO; ==> x = FOO;
2272 Once a macro has been undefined, that identifier may be "redefined"
2273 as a macro by a subsequent `#define' directive. The new definition
2274 need not have any resemblance to the old definition.
2276 However, if an identifier which is currently a macro is redefined,
2277 then the new definition must be "effectively the same" as the old one.
2278 Two macro definitions are effectively the same if:
2279 * Both are the same type of macro (object- or function-like).
2281 * All the tokens of the replacement list are the same.
2283 * If there are any parameters, they are the same.
2285 * Whitespace appears in the same places in both. It need not be
2286 exactly the same amount of whitespace, though. Remember that
2287 comments count as whitespace.
2289 These definitions are effectively the same:
2290 #define FOUR (2 + 2)
2291 #define FOUR (2 + 2)
2292 #define FOUR (2 /* two */ + 2)
2294 #define FOUR (2 + 2)
2295 #define FOUR ( 2+2 )
2296 #define FOUR (2 * 2)
2297 #define FOUR(score,and,seven,years,ago) (2 + 2)
2299 If a macro is redefined with a definition that is not effectively the
2300 same as the old one, the preprocessor issues a warning and changes the
2301 macro to use the new definition. If the new definition is effectively
2302 the same, the redefinition is silently ignored. This allows, for
2303 instance, two different headers to define a common macro. The
2304 preprocessor will only complain if the definitions do not match.
2307 File: cpp.info, Node: Directives Within Macro Arguments, Next: Macro Pitfalls, Prev: Undefining and Redefining Macros, Up: Macros
2309 3.9 Directives Within Macro Arguments
2310 =====================================
2312 Occasionally it is convenient to use preprocessor directives within the
2313 arguments of a macro. The C and C++ standards declare that behavior in
2314 these cases is undefined.
2316 Versions of CPP prior to 3.2 would reject such constructs with an
2317 error message. This was the only syntactic difference between normal
2318 functions and function-like macros, so it seemed attractive to remove
2319 this limitation, and people would often be surprised that they could
2320 not use macros in this way. Moreover, sometimes people would use
2321 conditional compilation in the argument list to a normal library
2322 function like `printf', only to find that after a library upgrade
2323 `printf' had changed to be a function-like macro, and their code would
2324 no longer compile. So from version 3.2 we changed CPP to successfully
2325 process arbitrary directives within macro arguments in exactly the same
2326 way as it would have processed the directive were the function-like
2327 macro invocation not present.
2329 If, within a macro invocation, that macro is redefined, then the new
2330 definition takes effect in time for argument pre-expansion, but the
2331 original definition is still used for argument replacement. Here is a
2332 pathological example:
2344 with the semantics described above.
2347 File: cpp.info, Node: Macro Pitfalls, Prev: Directives Within Macro Arguments, Up: Macros
2352 In this section we describe some special rules that apply to macros and
2353 macro expansion, and point out certain cases in which the rules have
2354 counter-intuitive consequences that you must watch out for.
2359 * Operator Precedence Problems::
2360 * Swallowing the Semicolon::
2361 * Duplication of Side Effects::
2362 * Self-Referential Macros::
2363 * Argument Prescan::
2364 * Newlines in Arguments::
2367 File: cpp.info, Node: Misnesting, Next: Operator Precedence Problems, Up: Macro Pitfalls
2372 When a macro is called with arguments, the arguments are substituted
2373 into the macro body and the result is checked, together with the rest of
2374 the input file, for more macro calls. It is possible to piece together
2375 a macro call coming partially from the macro body and partially from the
2376 arguments. For example,
2378 #define twice(x) (2*(x))
2379 #define call_with_1(x) x(1)
2384 Macro definitions do not have to have balanced parentheses. By
2385 writing an unbalanced open parenthesis in a macro body, it is possible
2386 to create a macro call that begins inside the macro body but ends
2387 outside of it. For example,
2389 #define strange(file) fprintf (file, "%s %d",
2391 strange(stderr) p, 35)
2392 ==> fprintf (stderr, "%s %d", p, 35)
2394 The ability to piece together a macro call can be useful, but the
2395 use of unbalanced open parentheses in a macro body is just confusing,
2396 and should be avoided.
2399 File: cpp.info, Node: Operator Precedence Problems, Next: Swallowing the Semicolon, Prev: Misnesting, Up: Macro Pitfalls
2401 3.10.2 Operator Precedence Problems
2402 -----------------------------------
2404 You may have noticed that in most of the macro definition examples shown
2405 above, each occurrence of a macro argument name had parentheses around
2406 it. In addition, another pair of parentheses usually surround the
2407 entire macro definition. Here is why it is best to write macros that
2410 Suppose you define a macro as follows,
2412 #define ceil_div(x, y) (x + y - 1) / y
2414 whose purpose is to divide, rounding up. (One use for this operation is
2415 to compute how many `int' objects are needed to hold a certain number
2416 of `char' objects.) Then suppose it is used as follows:
2418 a = ceil_div (b & c, sizeof (int));
2419 ==> a = (b & c + sizeof (int) - 1) / sizeof (int);
2421 This does not do what is intended. The operator-precedence rules of C
2422 make it equivalent to this:
2424 a = (b & (c + sizeof (int) - 1)) / sizeof (int);
2426 What we want is this:
2428 a = ((b & c) + sizeof (int) - 1)) / sizeof (int);
2430 Defining the macro as
2432 #define ceil_div(x, y) ((x) + (y) - 1) / (y)
2434 provides the desired result.
2436 Unintended grouping can result in another way. Consider `sizeof
2437 ceil_div(1, 2)'. That has the appearance of a C expression that would
2438 compute the size of the type of `ceil_div (1, 2)', but in fact it means
2439 something very different. Here is what it expands to:
2441 sizeof ((1) + (2) - 1) / (2)
2443 This would take the size of an integer and divide it by two. The
2444 precedence rules have put the division outside the `sizeof' when it was
2445 intended to be inside.
2447 Parentheses around the entire macro definition prevent such problems.
2448 Here, then, is the recommended way to define `ceil_div':
2450 #define ceil_div(x, y) (((x) + (y) - 1) / (y))
2453 File: cpp.info, Node: Swallowing the Semicolon, Next: Duplication of Side Effects, Prev: Operator Precedence Problems, Up: Macro Pitfalls
2455 3.10.3 Swallowing the Semicolon
2456 -------------------------------
2458 Often it is desirable to define a macro that expands into a compound
2459 statement. Consider, for example, the following macro, that advances a
2460 pointer (the argument `p' says where to find it) across whitespace
2463 #define SKIP_SPACES(p, limit) \
2464 { char *lim = (limit); \
2466 if (*p++ != ' ') { \
2469 Here backslash-newline is used to split the macro definition, which must
2470 be a single logical line, so that it resembles the way such code would
2471 be laid out if not part of a macro definition.
2473 A call to this macro might be `SKIP_SPACES (p, lim)'. Strictly
2474 speaking, the call expands to a compound statement, which is a complete
2475 statement with no need for a semicolon to end it. However, since it
2476 looks like a function call, it minimizes confusion if you can use it
2477 like a function call, writing a semicolon afterward, as in `SKIP_SPACES
2480 This can cause trouble before `else' statements, because the
2481 semicolon is actually a null statement. Suppose you write
2484 SKIP_SPACES (p, lim);
2487 The presence of two statements--the compound statement and a null
2488 statement--in between the `if' condition and the `else' makes invalid C
2491 The definition of the macro `SKIP_SPACES' can be altered to solve
2492 this problem, using a `do ... while' statement. Here is how:
2494 #define SKIP_SPACES(p, limit) \
2495 do { char *lim = (limit); \
2497 if (*p++ != ' ') { \
2501 Now `SKIP_SPACES (p, lim);' expands into
2505 which is one statement. The loop executes exactly once; most compilers
2506 generate no extra code for it.
2509 File: cpp.info, Node: Duplication of Side Effects, Next: Self-Referential Macros, Prev: Swallowing the Semicolon, Up: Macro Pitfalls
2511 3.10.4 Duplication of Side Effects
2512 ----------------------------------
2514 Many C programs define a macro `min', for "minimum", like this:
2516 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
2518 When you use this macro with an argument containing a side effect,
2521 next = min (x + y, foo (z));
2523 it expands as follows:
2525 next = ((x + y) < (foo (z)) ? (x + y) : (foo (z)));
2527 where `x + y' has been substituted for `X' and `foo (z)' for `Y'.
2529 The function `foo' is used only once in the statement as it appears
2530 in the program, but the expression `foo (z)' has been substituted twice
2531 into the macro expansion. As a result, `foo' might be called two times
2532 when the statement is executed. If it has side effects or if it takes
2533 a long time to compute, the results might not be what you intended. We
2534 say that `min' is an "unsafe" macro.
2536 The best solution to this problem is to define `min' in a way that
2537 computes the value of `foo (z)' only once. The C language offers no
2538 standard way to do this, but it can be done with GNU extensions as
2542 ({ typeof (X) x_ = (X); \
2543 typeof (Y) y_ = (Y); \
2544 (x_ < y_) ? x_ : y_; })
2546 The `({ ... })' notation produces a compound statement that acts as
2547 an expression. Its value is the value of its last statement. This
2548 permits us to define local variables and assign each argument to one.
2549 The local variables have underscores after their names to reduce the
2550 risk of conflict with an identifier of wider scope (it is impossible to
2551 avoid this entirely). Now each argument is evaluated exactly once.
2553 If you do not wish to use GNU C extensions, the only solution is to
2554 be careful when _using_ the macro `min'. For example, you can
2555 calculate the value of `foo (z)', save it in a variable, and use that
2558 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
2562 next = min (x + y, tem);
2565 (where we assume that `foo' returns type `int').
2568 File: cpp.info, Node: Self-Referential Macros, Next: Argument Prescan, Prev: Duplication of Side Effects, Up: Macro Pitfalls
2570 3.10.5 Self-Referential Macros
2571 ------------------------------
2573 A "self-referential" macro is one whose name appears in its definition.
2574 Recall that all macro definitions are rescanned for more macros to
2575 replace. If the self-reference were considered a use of the macro, it
2576 would produce an infinitely large expansion. To prevent this, the
2577 self-reference is not considered a macro call. It is passed into the
2578 preprocessor output unchanged. Consider an example:
2580 #define foo (4 + foo)
2582 where `foo' is also a variable in your program.
2584 Following the ordinary rules, each reference to `foo' will expand
2585 into `(4 + foo)'; then this will be rescanned and will expand into `(4
2586 + (4 + foo))'; and so on until the computer runs out of memory.
2588 The self-reference rule cuts this process short after one step, at
2589 `(4 + foo)'. Therefore, this macro definition has the possibly useful
2590 effect of causing the program to add 4 to the value of `foo' wherever
2591 `foo' is referred to.
2593 In most cases, it is a bad idea to take advantage of this feature. A
2594 person reading the program who sees that `foo' is a variable will not
2595 expect that it is a macro as well. The reader will come across the
2596 identifier `foo' in the program and think its value should be that of
2597 the variable `foo', whereas in fact the value is four greater.
2599 One common, useful use of self-reference is to create a macro which
2600 expands to itself. If you write
2604 then the macro `EPERM' expands to `EPERM'. Effectively, it is left
2605 alone by the preprocessor whenever it's used in running text. You can
2606 tell that it's a macro with `#ifdef'. You might do this if you want to
2607 define numeric constants with an `enum', but have `#ifdef' be true for
2610 If a macro `x' expands to use a macro `y', and the expansion of `y'
2611 refers to the macro `x', that is an "indirect self-reference" of `x'.
2612 `x' is not expanded in this case either. Thus, if we have
2617 then `x' and `y' expand as follows:
2625 Each macro is expanded when it appears in the definition of the other
2626 macro, but not when it indirectly appears in its own definition.
2629 File: cpp.info, Node: Argument Prescan, Next: Newlines in Arguments, Prev: Self-Referential Macros, Up: Macro Pitfalls
2631 3.10.6 Argument Prescan
2632 -----------------------
2634 Macro arguments are completely macro-expanded before they are
2635 substituted into a macro body, unless they are stringified or pasted
2636 with other tokens. After substitution, the entire macro body, including
2637 the substituted arguments, is scanned again for macros to be expanded.
2638 The result is that the arguments are scanned _twice_ to expand macro
2641 Most of the time, this has no effect. If the argument contained any
2642 macro calls, they are expanded during the first scan. The result
2643 therefore contains no macro calls, so the second scan does not change
2644 it. If the argument were substituted as given, with no prescan, the
2645 single remaining scan would find the same macro calls and produce the
2648 You might expect the double scan to change the results when a
2649 self-referential macro is used in an argument of another macro (*note
2650 Self-Referential Macros::): the self-referential macro would be
2651 expanded once in the first scan, and a second time in the second scan.
2652 However, this is not what happens. The self-references that do not
2653 expand in the first scan are marked so that they will not expand in the
2656 You might wonder, "Why mention the prescan, if it makes no
2657 difference? And why not skip it and make the preprocessor faster?"
2658 The answer is that the prescan does make a difference in three special
2661 * Nested calls to a macro.
2663 We say that "nested" calls to a macro occur when a macro's argument
2664 contains a call to that very macro. For example, if `f' is a macro
2665 that expects one argument, `f (f (1))' is a nested pair of calls to
2666 `f'. The desired expansion is made by expanding `f (1)' and
2667 substituting that into the definition of `f'. The prescan causes
2668 the expected result to happen. Without the prescan, `f (1)' itself
2669 would be substituted as an argument, and the inner use of `f' would
2670 appear during the main scan as an indirect self-reference and
2671 would not be expanded.
2673 * Macros that call other macros that stringify or concatenate.
2675 If an argument is stringified or concatenated, the prescan does not
2676 occur. If you _want_ to expand a macro, then stringify or
2677 concatenate its expansion, you can do that by causing one macro to
2678 call another macro that does the stringification or concatenation.
2679 For instance, if you have
2681 #define AFTERX(x) X_ ## x
2682 #define XAFTERX(x) AFTERX(x)
2683 #define TABLESIZE 1024
2684 #define BUFSIZE TABLESIZE
2686 then `AFTERX(BUFSIZE)' expands to `X_BUFSIZE', and
2687 `XAFTERX(BUFSIZE)' expands to `X_1024'. (Not to `X_TABLESIZE'.
2688 Prescan always does a complete expansion.)
2690 * Macros used in arguments, whose expansions contain unshielded
2693 This can cause a macro expanded on the second scan to be called
2694 with the wrong number of arguments. Here is an example:
2697 #define bar(x) lose(x)
2698 #define lose(x) (1 + (x))
2700 We would like `bar(foo)' to turn into `(1 + (foo))', which would
2701 then turn into `(1 + (a,b))'. Instead, `bar(foo)' expands into
2702 `lose(a,b)', and you get an error because `lose' requires a single
2703 argument. In this case, the problem is easily solved by the same
2704 parentheses that ought to be used to prevent misnesting of
2705 arithmetic operations:
2709 #define bar(x) lose((x))
2711 The extra pair of parentheses prevents the comma in `foo''s
2712 definition from being interpreted as an argument separator.
2716 File: cpp.info, Node: Newlines in Arguments, Prev: Argument Prescan, Up: Macro Pitfalls
2718 3.10.7 Newlines in Arguments
2719 ----------------------------
2721 The invocation of a function-like macro can extend over many logical
2722 lines. However, in the present implementation, the entire expansion
2723 comes out on one line. Thus line numbers emitted by the compiler or
2724 debugger refer to the line the invocation started on, which might be
2725 different to the line containing the argument causing the problem.
2727 Here is an example illustrating this:
2729 #define ignore_second_arg(a,b,c) a; c
2731 ignore_second_arg (foo (),
2735 The syntax error triggered by the tokens `syntax error' results in an
2736 error message citing line three--the line of ignore_second_arg-- even
2737 though the problematic code comes from line five.
2739 We consider this a bug, and intend to fix it in the near future.
2742 File: cpp.info, Node: Conditionals, Next: Diagnostics, Prev: Macros, Up: Top
2747 A "conditional" is a directive that instructs the preprocessor to
2748 select whether or not to include a chunk of code in the final token
2749 stream passed to the compiler. Preprocessor conditionals can test
2750 arithmetic expressions, or whether a name is defined as a macro, or both
2751 simultaneously using the special `defined' operator.
2753 A conditional in the C preprocessor resembles in some ways an `if'
2754 statement in C, but it is important to understand the difference between
2755 them. The condition in an `if' statement is tested during the
2756 execution of your program. Its purpose is to allow your program to
2757 behave differently from run to run, depending on the data it is
2758 operating on. The condition in a preprocessing conditional directive is
2759 tested when your program is compiled. Its purpose is to allow different
2760 code to be included in the program depending on the situation at the
2761 time of compilation.
2763 However, the distinction is becoming less clear. Modern compilers
2764 often do test `if' statements when a program is compiled, if their
2765 conditions are known not to vary at run time, and eliminate code which
2766 can never be executed. If you can count on your compiler to do this,
2767 you may find that your program is more readable if you use `if'
2768 statements with constant conditions (perhaps determined by macros). Of
2769 course, you can only use this to exclude code, not type definitions or
2770 other preprocessing directives, and you can only do it if the code
2771 remains syntactically valid when it is not to be used.
2773 GCC version 3 eliminates this kind of never-executed code even when
2774 not optimizing. Older versions did it only when optimizing.
2778 * Conditional Uses::
2779 * Conditional Syntax::
2783 File: cpp.info, Node: Conditional Uses, Next: Conditional Syntax, Up: Conditionals
2785 4.1 Conditional Uses
2786 ====================
2788 There are three general reasons to use a conditional.
2790 * A program may need to use different code depending on the machine
2791 or operating system it is to run on. In some cases the code for
2792 one operating system may be erroneous on another operating system;
2793 for example, it might refer to data types or constants that do not
2794 exist on the other system. When this happens, it is not enough to
2795 avoid executing the invalid code. Its mere presence will cause
2796 the compiler to reject the program. With a preprocessing
2797 conditional, the offending code can be effectively excised from
2798 the program when it is not valid.
2800 * You may want to be able to compile the same source file into two
2801 different programs. One version might make frequent time-consuming
2802 consistency checks on its intermediate data, or print the values of
2803 those data for debugging, and the other not.
2805 * A conditional whose condition is always false is one way to
2806 exclude code from the program but keep it as a sort of comment for
2809 Simple programs that do not need system-specific logic or complex
2810 debugging hooks generally will not need to use preprocessing
2814 File: cpp.info, Node: Conditional Syntax, Next: Deleted Code, Prev: Conditional Uses, Up: Conditionals
2816 4.2 Conditional Syntax
2817 ======================
2819 A conditional in the C preprocessor begins with a "conditional
2820 directive": `#if', `#ifdef' or `#ifndef'.
2831 File: cpp.info, Node: Ifdef, Next: If, Up: Conditional Syntax
2836 The simplest sort of conditional is
2844 This block is called a "conditional group". CONTROLLED TEXT will be
2845 included in the output of the preprocessor if and only if MACRO is
2846 defined. We say that the conditional "succeeds" if MACRO is defined,
2847 "fails" if it is not.
2849 The CONTROLLED TEXT inside of a conditional can include
2850 preprocessing directives. They are executed only if the conditional
2851 succeeds. You can nest conditional groups inside other conditional
2852 groups, but they must be completely nested. In other words, `#endif'
2853 always matches the nearest `#ifdef' (or `#ifndef', or `#if'). Also,
2854 you cannot start a conditional group in one file and end it in another.
2856 Even if a conditional fails, the CONTROLLED TEXT inside it is still
2857 run through initial transformations and tokenization. Therefore, it
2858 must all be lexically valid C. Normally the only way this matters is
2859 that all comments and string literals inside a failing conditional group
2860 must still be properly ended.
2862 The comment following the `#endif' is not required, but it is a good
2863 practice if there is a lot of CONTROLLED TEXT, because it helps people
2864 match the `#endif' to the corresponding `#ifdef'. Older programs
2865 sometimes put MACRO directly after the `#endif' without enclosing it in
2866 a comment. This is invalid code according to the C standard. CPP
2867 accepts it with a warning. It never affects which `#ifndef' the
2870 Sometimes you wish to use some code if a macro is _not_ defined.
2871 You can do this by writing `#ifndef' instead of `#ifdef'. One common
2872 use of `#ifndef' is to include code only the first time a header file
2873 is included. *Note Once-Only Headers::.
2875 Macro definitions can vary between compilations for several reasons.
2876 Here are some samples.
2878 * Some macros are predefined on each kind of machine (*note
2879 System-specific Predefined Macros::). This allows you to provide
2880 code specially tuned for a particular machine.
2882 * System header files define more macros, associated with the
2883 features they implement. You can test these macros with
2884 conditionals to avoid using a system feature on a machine where it
2887 * Macros can be defined or undefined with the `-D' and `-U' command
2888 line options when you compile the program. You can arrange to
2889 compile the same source file into two different programs by
2890 choosing a macro name to specify which program you want, writing
2891 conditionals to test whether or how this macro is defined, and
2892 then controlling the state of the macro with command line options,
2893 perhaps set in the Makefile. *Note Invocation::.
2895 * Your program might have a special header file (often called
2896 `config.h') that is adjusted when the program is compiled. It can
2897 define or not define macros depending on the features of the
2898 system and the desired capabilities of the program. The
2899 adjustment can be automated by a tool such as `autoconf', or done
2903 File: cpp.info, Node: If, Next: Defined, Prev: Ifdef, Up: Conditional Syntax
2908 The `#if' directive allows you to test the value of an arithmetic
2909 expression, rather than the mere existence of one macro. Its syntax is
2915 #endif /* EXPRESSION */
2917 EXPRESSION is a C expression of integer type, subject to stringent
2918 restrictions. It may contain
2920 * Integer constants.
2922 * Character constants, which are interpreted as they would be in
2925 * Arithmetic operators for addition, subtraction, multiplication,
2926 division, bitwise operations, shifts, comparisons, and logical
2927 operations (`&&' and `||'). The latter two obey the usual
2928 short-circuiting rules of standard C.
2930 * Macros. All macros in the expression are expanded before actual
2931 computation of the expression's value begins.
2933 * Uses of the `defined' operator, which lets you check whether macros
2934 are defined in the middle of an `#if'.
2936 * Identifiers that are not macros, which are all considered to be the
2937 number zero. This allows you to write `#if MACRO' instead of
2938 `#ifdef MACRO', if you know that MACRO, when defined, will always
2939 have a nonzero value. Function-like macros used without their
2940 function call parentheses are also treated as zero.
2942 In some contexts this shortcut is undesirable. The `-Wundef'
2943 option causes GCC to warn whenever it encounters an identifier
2944 which is not a macro in an `#if'.
2946 The preprocessor does not know anything about types in the language.
2947 Therefore, `sizeof' operators are not recognized in `#if', and neither
2948 are `enum' constants. They will be taken as identifiers which are not
2949 macros, and replaced by zero. In the case of `sizeof', this is likely
2950 to cause the expression to be invalid.
2952 The preprocessor calculates the value of EXPRESSION. It carries out
2953 all calculations in the widest integer type known to the compiler; on
2954 most machines supported by GCC this is 64 bits. This is not the same
2955 rule as the compiler uses to calculate the value of a constant
2956 expression, and may give different results in some cases. If the value
2957 comes out to be nonzero, the `#if' succeeds and the CONTROLLED TEXT is
2958 included; otherwise it is skipped.
2961 File: cpp.info, Node: Defined, Next: Else, Prev: If, Up: Conditional Syntax
2966 The special operator `defined' is used in `#if' and `#elif' expressions
2967 to test whether a certain name is defined as a macro. `defined NAME'
2968 and `defined (NAME)' are both expressions whose value is 1 if NAME is
2969 defined as a macro at the current point in the program, and 0
2970 otherwise. Thus, `#if defined MACRO' is precisely equivalent to
2973 `defined' is useful when you wish to test more than one macro for
2974 existence at once. For example,
2976 #if defined (__vax__) || defined (__ns16000__)
2978 would succeed if either of the names `__vax__' or `__ns16000__' is
2981 Conditionals written like this:
2983 #if defined BUFSIZE && BUFSIZE >= 1024
2985 can generally be simplified to just `#if BUFSIZE >= 1024', since if
2986 `BUFSIZE' is not defined, it will be interpreted as having the value
2989 If the `defined' operator appears as a result of a macro expansion,
2990 the C standard says the behavior is undefined. GNU cpp treats it as a
2991 genuine `defined' operator and evaluates it normally. It will warn
2992 wherever your code uses this feature if you use the command-line option
2993 `-pedantic', since other compilers may handle it differently.
2996 File: cpp.info, Node: Else, Next: Elif, Prev: Defined, Up: Conditional Syntax
3001 The `#else' directive can be added to a conditional to provide
3002 alternative text to be used if the condition fails. This is what it
3007 #else /* Not EXPRESSION */
3009 #endif /* Not EXPRESSION */
3011 If EXPRESSION is nonzero, the TEXT-IF-TRUE is included and the
3012 TEXT-IF-FALSE is skipped. If EXPRESSION is zero, the opposite happens.
3014 You can use `#else' with `#ifdef' and `#ifndef', too.
3017 File: cpp.info, Node: Elif, Prev: Else, Up: Conditional Syntax
3022 One common case of nested conditionals is used to check for more than
3023 two possible alternatives. For example, you might have
3035 Another conditional directive, `#elif', allows this to be
3036 abbreviated as follows:
3042 #else /* X != 2 and X != 1*/
3044 #endif /* X != 2 and X != 1*/
3046 `#elif' stands for "else if". Like `#else', it goes in the middle
3047 of a conditional group and subdivides it; it does not require a
3048 matching `#endif' of its own. Like `#if', the `#elif' directive
3049 includes an expression to be tested. The text following the `#elif' is
3050 processed only if the original `#if'-condition failed and the `#elif'
3053 More than one `#elif' can go in the same conditional group. Then
3054 the text after each `#elif' is processed only if the `#elif' condition
3055 succeeds after the original `#if' and all previous `#elif' directives
3056 within it have failed.
3058 `#else' is allowed after any number of `#elif' directives, but
3059 `#elif' may not follow `#else'.
3062 File: cpp.info, Node: Deleted Code, Prev: Conditional Syntax, Up: Conditionals
3067 If you replace or delete a part of the program but want to keep the old
3068 code around for future reference, you often cannot simply comment it
3069 out. Block comments do not nest, so the first comment inside the old
3070 code will end the commenting-out. The probable result is a flood of
3073 One way to avoid this problem is to use an always-false conditional
3074 instead. For instance, put `#if 0' before the deleted code and
3075 `#endif' after it. This works even if the code being turned off
3076 contains conditionals, but they must be entire conditionals (balanced
3077 `#if' and `#endif').
3079 Some people use `#ifdef notdef' instead. This is risky, because
3080 `notdef' might be accidentally defined as a macro, and then the
3081 conditional would succeed. `#if 0' can be counted on to fail.
3083 Do not use `#if 0' for comments which are not C code. Use a real
3084 comment, instead. The interior of `#if 0' must consist of complete
3085 tokens; in particular, single-quote characters must balance. Comments
3086 often contain unbalanced single-quote characters (known in English as
3087 apostrophes). These confuse `#if 0'. They don't confuse `/*'.
3090 File: cpp.info, Node: Diagnostics, Next: Line Control, Prev: Conditionals, Up: Top
3095 The directive `#error' causes the preprocessor to report a fatal error.
3096 The tokens forming the rest of the line following `#error' are used as
3099 You would use `#error' inside of a conditional that detects a
3100 combination of parameters which you know the program does not properly
3101 support. For example, if you know that the program will not run
3102 properly on a VAX, you might write
3105 #error "Won't work on VAXen. See comments at get_last_object."
3108 If you have several configuration parameters that must be set up by
3109 the installation in a consistent way, you can use conditionals to detect
3110 an inconsistency and report it with `#error'. For example,
3112 #if !defined(UNALIGNED_INT_ASM_OP) && defined(DWARF2_DEBUGGING_INFO)
3113 #error "DWARF2_DEBUGGING_INFO requires UNALIGNED_INT_ASM_OP."
3116 The directive `#warning' is like `#error', but causes the
3117 preprocessor to issue a warning and continue preprocessing. The tokens
3118 following `#warning' are used as the warning message.
3120 You might use `#warning' in obsolete header files, with a message
3121 directing the user to the header file which should be used instead.
3123 Neither `#error' nor `#warning' macro-expands its argument.
3124 Internal whitespace sequences are each replaced with a single space.
3125 The line must consist of complete tokens. It is wisest to make the
3126 argument of these directives be a single string constant; this avoids
3127 problems with apostrophes and the like.
3130 File: cpp.info, Node: Line Control, Next: Pragmas, Prev: Diagnostics, Up: Top
3135 The C preprocessor informs the C compiler of the location in your source
3136 code where each token came from. Presently, this is just the file name
3137 and line number. All the tokens resulting from macro expansion are
3138 reported as having appeared on the line of the source file where the
3139 outermost macro was used. We intend to be more accurate in the future.
3141 If you write a program which generates source code, such as the
3142 `bison' parser generator, you may want to adjust the preprocessor's
3143 notion of the current file name and line number by hand. Parts of the
3144 output from `bison' are generated from scratch, other parts come from a
3145 standard parser file. The rest are copied verbatim from `bison''s
3146 input. You would like compiler error messages and symbolic debuggers
3147 to be able to refer to `bison''s input file.
3149 `bison' or any such program can arrange this by writing `#line'
3150 directives into the output file. `#line' is a directive that specifies
3151 the original line number and source file name for subsequent input in
3152 the current preprocessor input file. `#line' has three variants:
3155 LINENUM is a non-negative decimal integer constant. It specifies
3156 the line number which should be reported for the following line of
3157 input. Subsequent lines are counted from LINENUM.
3159 `#line LINENUM FILENAME'
3160 LINENUM is the same as for the first form, and has the same
3161 effect. In addition, FILENAME is a string constant. The
3162 following line and all subsequent lines are reported to come from
3163 the file it specifies, until something else happens to change that.
3164 FILENAME is interpreted according to the normal rules for a string
3165 constant: backslash escapes are interpreted. This is different
3168 Previous versions of CPP did not interpret escapes in `#line'; we
3169 have changed it because the standard requires they be interpreted,
3170 and most other compilers do.
3172 `#line ANYTHING ELSE'
3173 ANYTHING ELSE is checked for macro calls, which are expanded. The
3174 result should match one of the above two forms.
3176 `#line' directives alter the results of the `__FILE__' and
3177 `__LINE__' predefined macros from that point on. *Note Standard
3178 Predefined Macros::. They do not have any effect on `#include''s idea
3179 of the directory containing the current file. This is a change from
3180 GCC 2.95. Previously, a file reading
3182 #line 1 "../src/gram.y"
3185 would search for `gram.h' in `../src', then the `-I' chain; the
3186 directory containing the physical source file would not be searched.
3187 In GCC 3.0 and later, the `#include' is not affected by the presence of
3188 a `#line' referring to a different directory.
3190 We made this change because the old behavior caused problems when
3191 generated source files were transported between machines. For instance,
3192 it is common practice to ship generated parsers with a source release,
3193 so that people building the distribution do not need to have yacc or
3194 Bison installed. These files frequently have `#line' directives
3195 referring to the directory tree of the system where the distribution was
3196 created. If GCC tries to search for headers in those directories, the
3197 build is likely to fail.
3199 The new behavior can cause failures too, if the generated file is not
3200 in the same directory as its source and it attempts to include a header
3201 which would be visible searching from the directory containing the
3202 source file. However, this problem is easily solved with an additional
3203 `-I' switch on the command line. The failures caused by the old
3204 semantics could sometimes be corrected only by editing the generated
3205 files, which is difficult and error-prone.
3208 File: cpp.info, Node: Pragmas, Next: Other Directives, Prev: Line Control, Up: Top
3213 The `#pragma' directive is the method specified by the C standard for
3214 providing additional information to the compiler, beyond what is
3215 conveyed in the language itself. Three forms of this directive
3216 (commonly known as "pragmas") are specified by the 1999 C standard. A
3217 C compiler is free to attach any meaning it likes to other pragmas.
3219 GCC has historically preferred to use extensions to the syntax of the
3220 language, such as `__attribute__', for this purpose. However, GCC does
3221 define a few pragmas of its own. These mostly have effects on the
3222 entire translation unit or source file.
3224 In GCC version 3, all GNU-defined, supported pragmas have been given
3225 a `GCC' prefix. This is in line with the `STDC' prefix on all pragmas
3226 defined by C99. For backward compatibility, pragmas which were
3227 recognized by previous versions are still recognized without the `GCC'
3228 prefix, but that usage is deprecated. Some older pragmas are
3229 deprecated in their entirety. They are not recognized with the `GCC'
3230 prefix. *Note Obsolete Features::.
3232 C99 introduces the `_Pragma' operator. This feature addresses a
3233 major problem with `#pragma': being a directive, it cannot be produced
3234 as the result of macro expansion. `_Pragma' is an operator, much like
3235 `sizeof' or `defined', and can be embedded in a macro.
3237 Its syntax is `_Pragma (STRING-LITERAL)', where STRING-LITERAL can
3238 be either a normal or wide-character string literal. It is
3239 destringized, by replacing all `\\' with a single `\' and all `\"' with
3240 a `"'. The result is then processed as if it had appeared as the right
3241 hand side of a `#pragma' directive. For example,
3243 _Pragma ("GCC dependency \"parse.y\"")
3245 has the same effect as `#pragma GCC dependency "parse.y"'. The same
3246 effect could be achieved using macros, for example
3248 #define DO_PRAGMA(x) _Pragma (#x)
3249 DO_PRAGMA (GCC dependency "parse.y")
3251 The standard is unclear on where a `_Pragma' operator can appear.
3252 The preprocessor does not accept it within a preprocessing conditional
3253 directive like `#if'. To be safe, you are probably best keeping it out
3254 of directives other than `#define', and putting it on a line of its own.
3256 This manual documents the pragmas which are meaningful to the
3257 preprocessor itself. Other pragmas are meaningful to the C or C++
3258 compilers. They are documented in the GCC manual.
3260 GCC plugins may provide their own pragmas.
3262 `#pragma GCC dependency'
3263 `#pragma GCC dependency' allows you to check the relative dates of
3264 the current file and another file. If the other file is more
3265 recent than the current file, a warning is issued. This is useful
3266 if the current file is derived from the other file, and should be
3267 regenerated. The other file is searched for using the normal
3268 include search path. Optional trailing text can be used to give
3269 more information in the warning message.
3271 #pragma GCC dependency "parse.y"
3272 #pragma GCC dependency "/usr/include/time.h" rerun fixincludes
3274 `#pragma GCC poison'
3275 Sometimes, there is an identifier that you want to remove
3276 completely from your program, and make sure that it never creeps
3277 back in. To enforce this, you can "poison" the identifier with
3278 this pragma. `#pragma GCC poison' is followed by a list of
3279 identifiers to poison. If any of those identifiers appears
3280 anywhere in the source after the directive, it is a hard error.
3283 #pragma GCC poison printf sprintf fprintf
3284 sprintf(some_string, "hello");
3286 will produce an error.
3288 If a poisoned identifier appears as part of the expansion of a
3289 macro which was defined before the identifier was poisoned, it
3290 will _not_ cause an error. This lets you poison an identifier
3291 without worrying about system headers defining macros that use it.
3295 #define strrchr rindex
3296 #pragma GCC poison rindex
3297 strrchr(some_string, 'h');
3299 will not produce an error.
3301 `#pragma GCC system_header'
3302 This pragma takes no arguments. It causes the rest of the code in
3303 the current file to be treated as if it came from a system header.
3304 *Note System Headers::.
3308 File: cpp.info, Node: Other Directives, Next: Preprocessor Output, Prev: Pragmas, Up: Top
3313 The `#ident' directive takes one argument, a string constant. On some
3314 systems, that string constant is copied into a special segment of the
3315 object file. On other systems, the directive is ignored. The `#sccs'
3316 directive is a synonym for `#ident'.
3318 These directives are not part of the C standard, but they are not
3319 official GNU extensions either. What historical information we have
3320 been able to find, suggests they originated with System V.
3322 The "null directive" consists of a `#' followed by a newline, with
3323 only whitespace (including comments) in between. A null directive is
3324 understood as a preprocessing directive but has no effect on the
3325 preprocessor output. The primary significance of the existence of the
3326 null directive is that an input line consisting of just a `#' will
3327 produce no output, rather than a line of output containing just a `#'.
3328 Supposedly some old C programs contain such lines.
3331 File: cpp.info, Node: Preprocessor Output, Next: Traditional Mode, Prev: Other Directives, Up: Top
3333 9 Preprocessor Output
3334 *********************
3336 When the C preprocessor is used with the C, C++, or Objective-C
3337 compilers, it is integrated into the compiler and communicates a stream
3338 of binary tokens directly to the compiler's parser. However, it can
3339 also be used in the more conventional standalone mode, where it produces
3342 The output from the C preprocessor looks much like the input, except
3343 that all preprocessing directive lines have been replaced with blank
3344 lines and all comments with spaces. Long runs of blank lines are
3347 The ISO standard specifies that it is implementation defined whether
3348 a preprocessor preserves whitespace between tokens, or replaces it with
3349 e.g. a single space. In GNU CPP, whitespace between tokens is collapsed
3350 to become a single space, with the exception that the first token on a
3351 non-directive line is preceded with sufficient spaces that it appears in
3352 the same column in the preprocessed output that it appeared in the
3353 original source file. This is so the output is easy to read. *Note
3354 Differences from previous versions::. CPP does not insert any
3355 whitespace where there was none in the original source, except where
3356 necessary to prevent an accidental token paste.
3358 Source file name and line number information is conveyed by lines of
3361 # LINENUM FILENAME FLAGS
3363 These are called "linemarkers". They are inserted as needed into the
3364 output (but never within a string or character constant). They mean
3365 that the following line originated in file FILENAME at line LINENUM.
3366 FILENAME will never contain any non-printing characters; they are
3367 replaced with octal escape sequences.
3369 After the file name comes zero or more flags, which are `1', `2',
3370 `3', or `4'. If there are multiple flags, spaces separate them. Here
3371 is what the flags mean:
3374 This indicates the start of a new file.
3377 This indicates returning to a file (after having included another
3381 This indicates that the following text comes from a system header
3382 file, so certain warnings should be suppressed.
3385 This indicates that the following text should be treated as being
3386 wrapped in an implicit `extern "C"' block.
3388 As an extension, the preprocessor accepts linemarkers in
3389 non-assembler input files. They are treated like the corresponding
3390 `#line' directive, (*note Line Control::), except that trailing flags
3391 are permitted, and are interpreted with the meanings described above.
3392 If multiple flags are given, they must be in ascending order.
3394 Some directives may be duplicated in the output of the preprocessor.
3395 These are `#ident' (always), `#pragma' (only if the preprocessor does
3396 not handle the pragma itself), and `#define' and `#undef' (with certain
3397 debugging options). If this happens, the `#' of the directive will
3398 always be in the first column, and there will be no space between the
3399 `#' and the directive name. If macro expansion happens to generate
3400 tokens which might be mistaken for a duplicated directive, a space will
3401 be inserted between the `#' and the directive name.
3404 File: cpp.info, Node: Traditional Mode, Next: Implementation Details, Prev: Preprocessor Output, Up: Top
3409 Traditional (pre-standard) C preprocessing is rather different from the
3410 preprocessing specified by the standard. When GCC is given the
3411 `-traditional-cpp' option, it attempts to emulate a traditional
3414 GCC versions 3.2 and later only support traditional mode semantics in
3415 the preprocessor, and not in the compiler front ends. This chapter
3416 outlines the traditional preprocessor semantics we implemented.
3418 The implementation does not correspond precisely to the behavior of
3419 earlier versions of GCC, nor to any true traditional preprocessor.
3420 After all, inconsistencies among traditional implementations were a
3421 major motivation for C standardization. However, we intend that it
3422 should be compatible with true traditional preprocessors in all ways
3423 that actually matter.
3427 * Traditional lexical analysis::
3428 * Traditional macros::
3429 * Traditional miscellany::
3430 * Traditional warnings::
3433 File: cpp.info, Node: Traditional lexical analysis, Next: Traditional macros, Up: Traditional Mode
3435 10.1 Traditional lexical analysis
3436 =================================
3438 The traditional preprocessor does not decompose its input into tokens
3439 the same way a standards-conforming preprocessor does. The input is
3440 simply treated as a stream of text with minimal internal form.
3442 This implementation does not treat trigraphs (*note trigraphs::)
3443 specially since they were an invention of the standards committee. It
3444 handles arbitrarily-positioned escaped newlines properly and splices
3445 the lines as you would expect; many traditional preprocessors did not
3448 The form of horizontal whitespace in the input file is preserved in
3449 the output. In particular, hard tabs remain hard tabs. This can be
3450 useful if, for example, you are preprocessing a Makefile.
3452 Traditional CPP only recognizes C-style block comments, and treats
3453 the `/*' sequence as introducing a comment only if it lies outside
3454 quoted text. Quoted text is introduced by the usual single and double
3455 quotes, and also by an initial `<' in a `#include' directive.
3457 Traditionally, comments are completely removed and are not replaced
3458 with a space. Since a traditional compiler does its own tokenization
3459 of the output of the preprocessor, this means that comments can
3460 effectively be used as token paste operators. However, comments behave
3461 like separators for text handled by the preprocessor itself, since it
3462 doesn't re-lex its input. For example, in
3466 `foo' and `bar' are distinct identifiers and expanded separately if
3467 they happen to be macros. In other words, this directive is equivalent
3476 Generally speaking, in traditional mode an opening quote need not
3477 have a matching closing quote. In particular, a macro may be defined
3478 with replacement text that contains an unmatched quote. Of course, if
3479 you attempt to compile preprocessed output containing an unmatched quote
3480 you will get a syntax error.
3482 However, all preprocessing directives other than `#define' require
3483 matching quotes. For example:
3485 #define m This macro's fine and has an unmatched quote
3486 "/* This is not a comment. */
3487 /* This is a comment. The following #include directive
3491 Just as for the ISO preprocessor, what would be a closing quote can
3492 be escaped with a backslash to prevent the quoted text from closing.
3495 File: cpp.info, Node: Traditional macros, Next: Traditional miscellany, Prev: Traditional lexical analysis, Up: Traditional Mode
3497 10.2 Traditional macros
3498 =======================
3500 The major difference between traditional and ISO macros is that the
3501 former expand to text rather than to a token sequence. CPP removes all
3502 leading and trailing horizontal whitespace from a macro's replacement
3503 text before storing it, but preserves the form of internal whitespace.
3505 One consequence is that it is legitimate for the replacement text to
3506 contain an unmatched quote (*note Traditional lexical analysis::). An
3507 unclosed string or character constant continues into the text following
3508 the macro call. Similarly, the text at the end of a macro's expansion
3509 can run together with the text after the macro invocation to produce a
3512 Normally comments are removed from the replacement text after the
3513 macro is expanded, but if the `-CC' option is passed on the command
3514 line comments are preserved. (In fact, the current implementation
3515 removes comments even before saving the macro replacement text, but it
3516 careful to do it in such a way that the observed effect is identical
3517 even in the function-like macro case.)
3519 The ISO stringification operator `#' and token paste operator `##'
3520 have no special meaning. As explained later, an effect similar to
3521 these operators can be obtained in a different way. Macro names that
3522 are embedded in quotes, either from the main file or after macro
3523 replacement, do not expand.
3525 CPP replaces an unquoted object-like macro name with its replacement
3526 text, and then rescans it for further macros to replace. Unlike
3527 standard macro expansion, traditional macro expansion has no provision
3528 to prevent recursion. If an object-like macro appears unquoted in its
3529 replacement text, it will be replaced again during the rescan pass, and
3530 so on _ad infinitum_. GCC detects when it is expanding recursive
3531 macros, emits an error message, and continues after the offending macro
3535 #define INC(x) PLUS+x
3539 Function-like macros are similar in form but quite different in
3540 behavior to their ISO counterparts. Their arguments are contained
3541 within parentheses, are comma-separated, and can cross physical lines.
3542 Commas within nested parentheses are not treated as argument
3543 separators. Similarly, a quote in an argument cannot be left unclosed;
3544 a following comma or parenthesis that comes before the closing quote is
3545 treated like any other character. There is no facility for handling
3548 This implementation removes all comments from macro arguments, unless
3549 the `-C' option is given. The form of all other horizontal whitespace
3550 in arguments is preserved, including leading and trailing whitespace.
3555 is treated as an invocation of the macro `f' with a single argument
3556 consisting of a single space. If you want to invoke a function-like
3557 macro that takes no arguments, you must not leave any whitespace
3558 between the parentheses.
3560 If a macro argument crosses a new line, the new line is replaced with
3561 a space when forming the argument. If the previous line contained an
3562 unterminated quote, the following line inherits the quoted state.
3564 Traditional preprocessors replace parameters in the replacement text
3565 with their arguments regardless of whether the parameters are within
3566 quotes or not. This provides a way to stringize arguments. For example
3569 str(/* A comment */some text )
3572 Note that the comment is removed, but that the trailing space is
3573 preserved. Here is an example of using a comment to effect token
3576 #define suffix(x) foo_/**/x
3581 File: cpp.info, Node: Traditional miscellany, Next: Traditional warnings, Prev: Traditional macros, Up: Traditional Mode
3583 10.3 Traditional miscellany
3584 ===========================
3586 Here are some things to be aware of when using the traditional
3589 * Preprocessing directives are recognized only when their leading
3590 `#' appears in the first column. There can be no whitespace
3591 between the beginning of the line and the `#', but whitespace can
3594 * A true traditional C preprocessor does not recognize `#error' or
3595 `#pragma', and may not recognize `#elif'. CPP supports all the
3596 directives in traditional mode that it supports in ISO mode,
3597 including extensions, with the exception that the effects of
3598 `#pragma GCC poison' are undefined.
3600 * __STDC__ is not defined.
3602 * If you use digraphs the behavior is undefined.
3604 * If a line that looks like a directive appears within macro
3605 arguments, the behavior is undefined.
3609 File: cpp.info, Node: Traditional warnings, Prev: Traditional miscellany, Up: Traditional Mode
3611 10.4 Traditional warnings
3612 =========================
3614 You can request warnings about features that did not exist, or worked
3615 differently, in traditional C with the `-Wtraditional' option. GCC
3616 does not warn about features of ISO C which you must use when you are
3617 using a conforming compiler, such as the `#' and `##' operators.
3619 Presently `-Wtraditional' warns about:
3621 * Macro parameters that appear within string literals in the macro
3622 body. In traditional C macro replacement takes place within
3623 string literals, but does not in ISO C.
3625 * In traditional C, some preprocessor directives did not exist.
3626 Traditional preprocessors would only consider a line to be a
3627 directive if the `#' appeared in column 1 on the line. Therefore
3628 `-Wtraditional' warns about directives that traditional C
3629 understands but would ignore because the `#' does not appear as the
3630 first character on the line. It also suggests you hide directives
3631 like `#pragma' not understood by traditional C by indenting them.
3632 Some traditional implementations would not recognize `#elif', so it
3633 suggests avoiding it altogether.
3635 * A function-like macro that appears without an argument list. In
3636 some traditional preprocessors this was an error. In ISO C it
3637 merely means that the macro is not expanded.
3639 * The unary plus operator. This did not exist in traditional C.
3641 * The `U' and `LL' integer constant suffixes, which were not
3642 available in traditional C. (Traditional C does support the `L'
3643 suffix for simple long integer constants.) You are not warned
3644 about uses of these suffixes in macros defined in system headers.
3645 For instance, `UINT_MAX' may well be defined as `4294967295U', but
3646 you will not be warned if you use `UINT_MAX'.
3648 You can usually avoid the warning, and the related warning about
3649 constants which are so large that they are unsigned, by writing the
3650 integer constant in question in hexadecimal, with no U suffix.
3651 Take care, though, because this gives the wrong result in exotic
3655 File: cpp.info, Node: Implementation Details, Next: Invocation, Prev: Traditional Mode, Up: Top
3657 11 Implementation Details
3658 *************************
3660 Here we document details of how the preprocessor's implementation
3661 affects its user-visible behavior. You should try to avoid undue
3662 reliance on behavior described here, as it is possible that it will
3663 change subtly in future implementations.
3665 Also documented here are obsolete features and changes from previous
3670 * Implementation-defined behavior::
3671 * Implementation limits::
3672 * Obsolete Features::
3673 * Differences from previous versions::
3676 File: cpp.info, Node: Implementation-defined behavior, Next: Implementation limits, Up: Implementation Details
3678 11.1 Implementation-defined behavior
3679 ====================================
3681 This is how CPP behaves in all the cases which the C standard describes
3682 as "implementation-defined". This term means that the implementation
3683 is free to do what it likes, but must document its choice and stick to
3686 * The mapping of physical source file multi-byte characters to the
3687 execution character set.
3689 The input character set can be specified using the
3690 `-finput-charset' option, while the execution character set may be
3691 controlled using the `-fexec-charset' and `-fwide-exec-charset'
3694 * Identifier characters. The C and C++ standards allow identifiers
3695 to be composed of `_' and the alphanumeric characters. C++ and
3696 C99 also allow universal character names, and C99 further permits
3697 implementation-defined characters. GCC currently only permits
3698 universal character names if `-fextended-identifiers' is used,
3699 because the implementation of universal character names in
3700 identifiers is experimental.
3702 GCC allows the `$' character in identifiers as an extension for
3703 most targets. This is true regardless of the `std=' switch, since
3704 this extension cannot conflict with standards-conforming programs.
3705 When preprocessing assembler, however, dollars are not identifier
3706 characters by default.
3708 Currently the targets that by default do not permit `$' are AVR,
3709 IP2K, MMIX, MIPS Irix 3, ARM aout, and PowerPC targets for the AIX
3712 You can override the default with `-fdollars-in-identifiers' or
3713 `fno-dollars-in-identifiers'. *Note fdollars-in-identifiers::.
3715 * Non-empty sequences of whitespace characters.
3717 In textual output, each whitespace sequence is collapsed to a
3718 single space. For aesthetic reasons, the first token on each
3719 non-directive line of output is preceded with sufficient spaces
3720 that it appears in the same column as it did in the original
3723 * The numeric value of character constants in preprocessor
3726 The preprocessor and compiler interpret character constants in the
3727 same way; i.e. escape sequences such as `\a' are given the values
3728 they would have on the target machine.
3730 The compiler evaluates a multi-character character constant a
3731 character at a time, shifting the previous value left by the
3732 number of bits per target character, and then or-ing in the
3733 bit-pattern of the new character truncated to the width of a
3734 target character. The final bit-pattern is given type `int', and
3735 is therefore signed, regardless of whether single characters are
3736 signed or not (a slight change from versions 3.1 and earlier of
3737 GCC). If there are more characters in the constant than would fit
3738 in the target `int' the compiler issues a warning, and the excess
3739 leading characters are ignored.
3741 For example, `'ab'' for a target with an 8-bit `char' would be
3743 `(int) ((unsigned char) 'a' * 256 + (unsigned char) 'b')', and
3745 `(int) ((unsigned char) '\234' * 256 + (unsigned char) 'a')'.
3747 * Source file inclusion.
3749 For a discussion on how the preprocessor locates header files,
3750 *note Include Operation::.
3752 * Interpretation of the filename resulting from a macro-expanded
3753 `#include' directive.
3755 *Note Computed Includes::.
3757 * Treatment of a `#pragma' directive that after macro-expansion
3758 results in a standard pragma.
3760 No macro expansion occurs on any `#pragma' directive line, so the
3761 question does not arise.
3763 Note that GCC does not yet implement any of the standard pragmas.
3767 File: cpp.info, Node: Implementation limits, Next: Obsolete Features, Prev: Implementation-defined behavior, Up: Implementation Details
3769 11.2 Implementation limits
3770 ==========================
3772 CPP has a small number of internal limits. This section lists the
3773 limits which the C standard requires to be no lower than some minimum,
3774 and all the others known. It is intended that there should be as few
3775 limits as possible. If you encounter an undocumented or inconvenient
3776 limit, please report that as a bug. *Note Reporting Bugs: (gcc)Bugs.
3778 Where we say something is limited "only by available memory", that
3779 means that internal data structures impose no intrinsic limit, and space
3780 is allocated with `malloc' or equivalent. The actual limit will
3781 therefore depend on many things, such as the size of other things
3782 allocated by the compiler at the same time, the amount of memory
3783 consumed by other processes on the same computer, etc.
3785 * Nesting levels of `#include' files.
3787 We impose an arbitrary limit of 200 levels, to avoid runaway
3788 recursion. The standard requires at least 15 levels.
3790 * Nesting levels of conditional inclusion.
3792 The C standard mandates this be at least 63. CPP is limited only
3793 by available memory.
3795 * Levels of parenthesized expressions within a full expression.
3797 The C standard requires this to be at least 63. In preprocessor
3798 conditional expressions, it is limited only by available memory.
3800 * Significant initial characters in an identifier or macro name.
3802 The preprocessor treats all characters as significant. The C
3803 standard requires only that the first 63 be significant.
3805 * Number of macros simultaneously defined in a single translation
3808 The standard requires at least 4095 be possible. CPP is limited
3809 only by available memory.
3811 * Number of parameters in a macro definition and arguments in a
3814 We allow `USHRT_MAX', which is no smaller than 65,535. The minimum
3815 required by the standard is 127.
3817 * Number of characters on a logical source line.
3819 The C standard requires a minimum of 4096 be permitted. CPP places
3820 no limits on this, but you may get incorrect column numbers
3821 reported in diagnostics for lines longer than 65,535 characters.
3823 * Maximum size of a source file.
3825 The standard does not specify any lower limit on the maximum size
3826 of a source file. GNU cpp maps files into memory, so it is
3827 limited by the available address space. This is generally at
3828 least two gigabytes. Depending on the operating system, the size
3829 of physical memory may or may not be a limitation.
3833 File: cpp.info, Node: Obsolete Features, Next: Differences from previous versions, Prev: Implementation limits, Up: Implementation Details
3835 11.3 Obsolete Features
3836 ======================
3838 CPP has some features which are present mainly for compatibility with
3839 older programs. We discourage their use in new code. In some cases,
3840 we plan to remove the feature in a future version of GCC.
3845 "Assertions" are a deprecated alternative to macros in writing
3846 conditionals to test what sort of computer or system the compiled
3847 program will run on. Assertions are usually predefined, but you can
3848 define them with preprocessing directives or command-line options.
3850 Assertions were intended to provide a more systematic way to describe
3851 the compiler's target system and we added them for compatibility with
3852 existing compilers. In practice they are just as unpredictable as the
3853 system-specific predefined macros. In addition, they are not part of
3854 any standard, and only a few compilers support them. Therefore, the
3855 use of assertions is *less* portable than the use of system-specific
3856 predefined macros. We recommend you do not use them at all.
3858 An assertion looks like this:
3862 PREDICATE must be a single identifier. ANSWER can be any sequence of
3863 tokens; all characters are significant except for leading and trailing
3864 whitespace, and differences in internal whitespace sequences are
3865 ignored. (This is similar to the rules governing macro redefinition.)
3866 Thus, `(x + y)' is different from `(x+y)' but equivalent to
3867 `( x + y )'. Parentheses do not nest inside an answer.
3869 To test an assertion, you write it in an `#if'. For example, this
3870 conditional succeeds if either `vax' or `ns16000' has been asserted as
3871 an answer for `machine'.
3873 #if #machine (vax) || #machine (ns16000)
3875 You can test whether _any_ answer is asserted for a predicate by
3876 omitting the answer in the conditional:
3880 Assertions are made with the `#assert' directive. Its sole argument
3881 is the assertion to make, without the leading `#' that identifies
3882 assertions in conditionals.
3884 #assert PREDICATE (ANSWER)
3886 You may make several assertions with the same predicate and different
3887 answers. Subsequent assertions do not override previous ones for the
3888 same predicate. All the answers for any given predicate are
3889 simultaneously true.
3891 Assertions can be canceled with the `#unassert' directive. It has
3892 the same syntax as `#assert'. In that form it cancels only the answer
3893 which was specified on the `#unassert' line; other answers for that
3894 predicate remain true. You can cancel an entire predicate by leaving
3899 In either form, if no such assertion has been made, `#unassert' has no
3902 You can also make or cancel assertions using command line options.
3906 File: cpp.info, Node: Differences from previous versions, Prev: Obsolete Features, Up: Implementation Details
3908 11.4 Differences from previous versions
3909 =======================================
3911 This section details behavior which has changed from previous versions
3912 of CPP. We do not plan to change it again in the near future, but we
3913 do not promise not to, either.
3915 The "previous versions" discussed here are 2.95 and before. The
3916 behavior of GCC 3.0 is mostly the same as the behavior of the widely
3917 used 2.96 and 2.97 development snapshots. Where there are differences,
3918 they generally represent bugs in the snapshots.
3922 This option has been deprecated in 4.0. `-iquote' is meant to
3923 replace the need for this option.
3925 * Order of evaluation of `#' and `##' operators
3927 The standard does not specify the order of evaluation of a chain of
3928 `##' operators, nor whether `#' is evaluated before, after, or at
3929 the same time as `##'. You should therefore not write any code
3930 which depends on any specific ordering. It is possible to
3931 guarantee an ordering, if you need one, by suitable use of nested
3934 An example of where this might matter is pasting the arguments `1',
3935 `e' and `-2'. This would be fine for left-to-right pasting, but
3936 right-to-left pasting would produce an invalid token `e-2'.
3938 GCC 3.0 evaluates `#' and `##' at the same time and strictly left
3939 to right. Older versions evaluated all `#' operators first, then
3940 all `##' operators, in an unreliable order.
3942 * The form of whitespace between tokens in preprocessor output
3944 *Note Preprocessor Output::, for the current textual format. This
3945 is also the format used by stringification. Normally, the
3946 preprocessor communicates tokens directly to the compiler's
3947 parser, and whitespace does not come up at all.
3949 Older versions of GCC preserved all whitespace provided by the
3950 user and inserted lots more whitespace of their own, because they
3951 could not accurately predict when extra spaces were needed to
3952 prevent accidental token pasting.
3954 * Optional argument when invoking rest argument macros
3956 As an extension, GCC permits you to omit the variable arguments
3957 entirely when you use a variable argument macro. This is
3958 forbidden by the 1999 C standard, and will provoke a pedantic
3959 warning with GCC 3.0. Previous versions accepted it silently.
3961 * `##' swallowing preceding text in rest argument macros
3963 Formerly, in a macro expansion, if `##' appeared before a variable
3964 arguments parameter, and the set of tokens specified for that
3965 argument in the macro invocation was empty, previous versions of
3966 CPP would back up and remove the preceding sequence of
3967 non-whitespace characters (*not* the preceding token). This
3968 extension is in direct conflict with the 1999 C standard and has
3969 been drastically pared back.
3971 In the current version of the preprocessor, if `##' appears between
3972 a comma and a variable arguments parameter, and the variable
3973 argument is omitted entirely, the comma will be removed from the
3974 expansion. If the variable argument is empty, or the token before
3975 `##' is not a comma, then `##' behaves as a normal token paste.
3977 * `#line' and `#include'
3979 The `#line' directive used to change GCC's notion of the
3980 "directory containing the current file", used by `#include' with a
3981 double-quoted header file name. In 3.0 and later, it does not.
3982 *Note Line Control::, for further explanation.
3986 In GCC 2.95 and previous, the string constant argument to `#line'
3987 was treated the same way as the argument to `#include': backslash
3988 escapes were not honored, and the string ended at the second `"'.
3989 This is not compliant with the C standard. In GCC 3.0, an attempt
3990 was made to correct the behavior, so that the string was treated
3991 as a real string constant, but it turned out to be buggy. In 3.1,
3992 the bugs have been fixed. (We are not fixing the bugs in 3.0
3993 because they affect relatively few people and the fix is quite
3998 File: cpp.info, Node: Invocation, Next: Environment Variables, Prev: Implementation Details, Up: Top
4003 Most often when you use the C preprocessor you will not have to invoke
4004 it explicitly: the C compiler will do so automatically. However, the
4005 preprocessor is sometimes useful on its own. All the options listed
4006 here are also acceptable to the C compiler and have the same meaning,
4007 except that the C compiler has different rules for specifying the output
4010 _Note:_ Whether you use the preprocessor by way of `gcc' or `cpp',
4011 the "compiler driver" is run first. This program's purpose is to
4012 translate your command into invocations of the programs that do the
4013 actual work. Their command line interfaces are similar but not
4014 identical to the documented interface, and may change without notice.
4016 The C preprocessor expects two file names as arguments, INFILE and
4017 OUTFILE. The preprocessor reads INFILE together with any other files
4018 it specifies with `#include'. All the output generated by the combined
4019 input files is written in OUTFILE.
4021 Either INFILE or OUTFILE may be `-', which as INFILE means to read
4022 from standard input and as OUTFILE means to write to standard output.
4023 Also, if either file is omitted, it means the same as if `-' had been
4024 specified for that file.
4026 Unless otherwise noted, or the option ends in `=', all options which
4027 take an argument may have that argument appear either immediately after
4028 the option, or with a space between option and argument: `-Ifoo' and
4029 `-I foo' have the same effect.
4031 Many options have multi-letter names; therefore multiple
4032 single-letter options may _not_ be grouped: `-dM' is very different from
4036 Predefine NAME as a macro, with definition `1'.
4038 `-D NAME=DEFINITION'
4039 The contents of DEFINITION are tokenized and processed as if they
4040 appeared during translation phase three in a `#define' directive.
4041 In particular, the definition will be truncated by embedded
4044 If you are invoking the preprocessor from a shell or shell-like
4045 program you may need to use the shell's quoting syntax to protect
4046 characters such as spaces that have a meaning in the shell syntax.
4048 If you wish to define a function-like macro on the command line,
4049 write its argument list with surrounding parentheses before the
4050 equals sign (if any). Parentheses are meaningful to most shells,
4051 so you will need to quote the option. With `sh' and `csh',
4052 `-D'NAME(ARGS...)=DEFINITION'' works.
4054 `-D' and `-U' options are processed in the order they are given on
4055 the command line. All `-imacros FILE' and `-include FILE' options
4056 are processed after all `-D' and `-U' options.
4059 Cancel any previous definition of NAME, either built in or
4060 provided with a `-D' option.
4063 Do not predefine any system-specific or GCC-specific macros. The
4064 standard predefined macros remain defined. *Note Standard
4065 Predefined Macros::.
4068 Add the directory DIR to the list of directories to be searched
4069 for header files. *Note Search Path::. Directories named by `-I'
4070 are searched before the standard system include directories. If
4071 the directory DIR is a standard system include directory, the
4072 option is ignored to ensure that the default search order for
4073 system directories and the special treatment of system headers are
4074 not defeated (*note System Headers::) . If DIR begins with `=',
4075 then the `=' will be replaced by the sysroot prefix; see
4076 `--sysroot' and `-isysroot'.
4079 Write output to FILE. This is the same as specifying FILE as the
4080 second non-option argument to `cpp'. `gcc' has a different
4081 interpretation of a second non-option argument, so you must use
4082 `-o' to specify the output file.
4085 Turns on all optional warnings which are desirable for normal code.
4086 At present this is `-Wcomment', `-Wtrigraphs', `-Wmultichar' and a
4087 warning about integer promotion causing a change of sign in `#if'
4088 expressions. Note that many of the preprocessor's warnings are on
4089 by default and have no options to control them.
4093 Warn whenever a comment-start sequence `/*' appears in a `/*'
4094 comment, or whenever a backslash-newline appears in a `//' comment.
4095 (Both forms have the same effect.)
4098 Most trigraphs in comments cannot affect the meaning of the
4099 program. However, a trigraph that would form an escaped newline
4100 (`??/' at the end of a line) can, by changing where the comment
4101 begins or ends. Therefore, only trigraphs that would form escaped
4102 newlines produce warnings inside a comment.
4104 This option is implied by `-Wall'. If `-Wall' is not given, this
4105 option is still enabled unless trigraphs are enabled. To get
4106 trigraph conversion without warnings, but get the other `-Wall'
4107 warnings, use `-trigraphs -Wall -Wno-trigraphs'.
4110 Warn about certain constructs that behave differently in
4111 traditional and ISO C. Also warn about ISO C constructs that have
4112 no traditional C equivalent, and problematic constructs which
4113 should be avoided. *Note Traditional Mode::.
4116 Warn whenever an identifier which is not a macro is encountered in
4117 an `#if' directive, outside of `defined'. Such identifiers are
4121 Warn about macros defined in the main file that are unused. A
4122 macro is "used" if it is expanded or tested for existence at least
4123 once. The preprocessor will also warn if the macro has not been
4124 used at the time it is redefined or undefined.
4126 Built-in macros, macros defined on the command line, and macros
4127 defined in include files are not warned about.
4129 _Note:_ If a macro is actually used, but only used in skipped
4130 conditional blocks, then CPP will report it as unused. To avoid
4131 the warning in such a case, you might improve the scope of the
4132 macro's definition by, for example, moving it into the first
4133 skipped block. Alternatively, you could provide a dummy use with
4136 #if defined the_macro_causing_the_warning
4140 Warn whenever an `#else' or an `#endif' are followed by text.
4141 This usually happens in code of the form
4149 The second and third `FOO' should be in comments, but often are not
4150 in older programs. This warning is on by default.
4153 Make all warnings into hard errors. Source code which triggers
4154 warnings will be rejected.
4157 Issue warnings for code in system headers. These are normally
4158 unhelpful in finding bugs in your own code, therefore suppressed.
4159 If you are responsible for the system library, you may want to see
4163 Suppress all warnings, including those which GNU CPP issues by
4167 Issue all the mandatory diagnostics listed in the C standard.
4168 Some of them are left out by default, since they trigger
4169 frequently on harmless code.
4172 Issue all the mandatory diagnostics, and make all mandatory
4173 diagnostics into errors. This includes mandatory diagnostics that
4174 GCC issues without `-pedantic' but treats as warnings.
4177 Instead of outputting the result of preprocessing, output a rule
4178 suitable for `make' describing the dependencies of the main source
4179 file. The preprocessor outputs one `make' rule containing the
4180 object file name for that source file, a colon, and the names of
4181 all the included files, including those coming from `-include' or
4182 `-imacros' command line options.
4184 Unless specified explicitly (with `-MT' or `-MQ'), the object file
4185 name consists of the name of the source file with any suffix
4186 replaced with object file suffix and with any leading directory
4187 parts removed. If there are many included files then the rule is
4188 split into several lines using `\'-newline. The rule has no
4191 This option does not suppress the preprocessor's debug output,
4192 such as `-dM'. To avoid mixing such debug output with the
4193 dependency rules you should explicitly specify the dependency
4194 output file with `-MF', or use an environment variable like
4195 `DEPENDENCIES_OUTPUT' (*note Environment Variables::). Debug
4196 output will still be sent to the regular output stream as normal.
4198 Passing `-M' to the driver implies `-E', and suppresses warnings
4199 with an implicit `-w'.
4202 Like `-M' but do not mention header files that are found in system
4203 header directories, nor header files that are included, directly
4204 or indirectly, from such a header.
4206 This implies that the choice of angle brackets or double quotes in
4207 an `#include' directive does not in itself determine whether that
4208 header will appear in `-MM' dependency output. This is a slight
4209 change in semantics from GCC versions 3.0 and earlier.
4212 When used with `-M' or `-MM', specifies a file to write the
4213 dependencies to. If no `-MF' switch is given the preprocessor
4214 sends the rules to the same place it would have sent preprocessed
4217 When used with the driver options `-MD' or `-MMD', `-MF' overrides
4218 the default dependency output file.
4221 In conjunction with an option such as `-M' requesting dependency
4222 generation, `-MG' assumes missing header files are generated files
4223 and adds them to the dependency list without raising an error.
4224 The dependency filename is taken directly from the `#include'
4225 directive without prepending any path. `-MG' also suppresses
4226 preprocessed output, as a missing header file renders this useless.
4228 This feature is used in automatic updating of makefiles.
4231 This option instructs CPP to add a phony target for each dependency
4232 other than the main file, causing each to depend on nothing. These
4233 dummy rules work around errors `make' gives if you remove header
4234 files without updating the `Makefile' to match.
4236 This is typical output:
4238 test.o: test.c test.h
4243 Change the target of the rule emitted by dependency generation. By
4244 default CPP takes the name of the main input file, deletes any
4245 directory components and any file suffix such as `.c', and appends
4246 the platform's usual object suffix. The result is the target.
4248 An `-MT' option will set the target to be exactly the string you
4249 specify. If you want multiple targets, you can specify them as a
4250 single argument to `-MT', or use multiple `-MT' options.
4252 For example, `-MT '$(objpfx)foo.o'' might give
4254 $(objpfx)foo.o: foo.c
4257 Same as `-MT', but it quotes any characters which are special to
4258 Make. `-MQ '$(objpfx)foo.o'' gives
4260 $$(objpfx)foo.o: foo.c
4262 The default target is automatically quoted, as if it were given
4266 `-MD' is equivalent to `-M -MF FILE', except that `-E' is not
4267 implied. The driver determines FILE based on whether an `-o'
4268 option is given. If it is, the driver uses its argument but with
4269 a suffix of `.d', otherwise it takes the name of the input file,
4270 removes any directory components and suffix, and applies a `.d'
4273 If `-MD' is used in conjunction with `-E', any `-o' switch is
4274 understood to specify the dependency output file (*note -MF:
4275 dashMF.), but if used without `-E', each `-o' is understood to
4276 specify a target object file.
4278 Since `-E' is not implied, `-MD' can be used to generate a
4279 dependency output file as a side-effect of the compilation process.
4282 Like `-MD' except mention only user header files, not system
4288 `-x assembler-with-cpp'
4289 Specify the source language: C, C++, Objective-C, or assembly.
4290 This has nothing to do with standards conformance or extensions;
4291 it merely selects which base syntax to expect. If you give none
4292 of these options, cpp will deduce the language from the extension
4293 of the source file: `.c', `.cc', `.m', or `.S'. Some other common
4294 extensions for C++ and assembly are also recognized. If cpp does
4295 not recognize the extension, it will treat the file as C; this is
4296 the most generic mode.
4298 _Note:_ Previous versions of cpp accepted a `-lang' option which
4299 selected both the language and the standards conformance level.
4300 This option has been removed, because it conflicts with the `-l'
4305 Specify the standard to which the code should conform. Currently
4306 CPP knows about C and C++ standards; others may be added in the
4309 STANDARD may be one of:
4313 The ISO C standard from 1990. `c90' is the customary
4314 shorthand for this version of the standard.
4316 The `-ansi' option is equivalent to `-std=c90'.
4319 The 1990 C standard, as amended in 1994.
4325 The revised ISO C standard, published in December 1999.
4326 Before publication, this was known as C9X.
4331 The revised ISO C standard, published in December 2011.
4332 Before publication, this was known as C1X.
4336 The 1990 C standard plus GNU extensions. This is the default.
4340 The 1999 C standard plus GNU extensions.
4344 The 2011 C standard plus GNU extensions.
4347 The 1998 ISO C++ standard plus amendments.
4350 The same as `-std=c++98' plus GNU extensions. This is the
4351 default for C++ code.
4354 Split the include path. Any directories specified with `-I'
4355 options before `-I-' are searched only for headers requested with
4356 `#include "FILE"'; they are not searched for `#include <FILE>'.
4357 If additional directories are specified with `-I' options after
4358 the `-I-', those directories are searched for all `#include'
4361 In addition, `-I-' inhibits the use of the directory of the current
4362 file directory as the first search directory for `#include "FILE"'.
4363 *Note Search Path::. This option has been deprecated.
4366 Do not search the standard system directories for header files.
4367 Only the directories you have specified with `-I' options (and the
4368 directory of the current file, if appropriate) are searched.
4371 Do not search for header files in the C++-specific standard
4372 directories, but do still search the other standard directories.
4373 (This option is used when building the C++ library.)
4376 Process FILE as if `#include "file"' appeared as the first line of
4377 the primary source file. However, the first directory searched
4378 for FILE is the preprocessor's working directory _instead of_ the
4379 directory containing the main source file. If not found there, it
4380 is searched for in the remainder of the `#include "..."' search
4383 If multiple `-include' options are given, the files are included
4384 in the order they appear on the command line.
4387 Exactly like `-include', except that any output produced by
4388 scanning FILE is thrown away. Macros it defines remain defined.
4389 This allows you to acquire all the macros from a header without
4390 also processing its declarations.
4392 All files specified by `-imacros' are processed before all files
4393 specified by `-include'.
4396 Search DIR for header files, but do it _after_ all directories
4397 specified with `-I' and the standard system directories have been
4398 exhausted. DIR is treated as a system include directory. If DIR
4399 begins with `=', then the `=' will be replaced by the sysroot
4400 prefix; see `--sysroot' and `-isysroot'.
4403 Specify PREFIX as the prefix for subsequent `-iwithprefix'
4404 options. If the prefix represents a directory, you should include
4408 `-iwithprefixbefore DIR'
4409 Append DIR to the prefix specified previously with `-iprefix', and
4410 add the resulting directory to the include search path.
4411 `-iwithprefixbefore' puts it in the same place `-I' would;
4412 `-iwithprefix' puts it where `-idirafter' would.
4415 This option is like the `--sysroot' option, but applies only to
4416 header files (except for Darwin targets, where it applies to both
4417 header files and libraries). See the `--sysroot' option for more
4421 Use DIR as a subdirectory of the directory containing
4422 target-specific C++ headers.
4425 Search DIR for header files, after all directories specified by
4426 `-I' but before the standard system directories. Mark it as a
4427 system directory, so that it gets the same special treatment as is
4428 applied to the standard system directories. *Note System
4429 Headers::. If DIR begins with `=', then the `=' will be replaced
4430 by the sysroot prefix; see `--sysroot' and `-isysroot'.
4433 Search DIR only for header files requested with `#include "FILE"';
4434 they are not searched for `#include <FILE>', before all
4435 directories specified by `-I' and before the standard system
4436 directories. *Note Search Path::. If DIR begins with `=', then
4437 the `=' will be replaced by the sysroot prefix; see `--sysroot'
4441 When preprocessing, handle directives, but do not expand macros.
4443 The option's behavior depends on the `-E' and `-fpreprocessed'
4446 With `-E', preprocessing is limited to the handling of directives
4447 such as `#define', `#ifdef', and `#error'. Other preprocessor
4448 operations, such as macro expansion and trigraph conversion are
4449 not performed. In addition, the `-dD' option is implicitly
4452 With `-fpreprocessed', predefinition of command line and most
4453 builtin macros is disabled. Macros such as `__LINE__', which are
4454 contextually dependent, are handled normally. This enables
4455 compilation of files previously preprocessed with `-E
4458 With both `-E' and `-fpreprocessed', the rules for
4459 `-fpreprocessed' take precedence. This enables full preprocessing
4460 of files previously preprocessed with `-E -fdirectives-only'.
4462 `-fdollars-in-identifiers'
4463 Accept `$' in identifiers. *Note Identifier characters::.
4465 `-fextended-identifiers'
4466 Accept universal character names in identifiers. This option is
4467 experimental; in a future version of GCC, it will be enabled by
4468 default for C99 and C++.
4471 Indicate to the preprocessor that the input file has already been
4472 preprocessed. This suppresses things like macro expansion,
4473 trigraph conversion, escaped newline splicing, and processing of
4474 most directives. The preprocessor still recognizes and removes
4475 comments, so that you can pass a file preprocessed with `-C' to
4476 the compiler without problems. In this mode the integrated
4477 preprocessor is little more than a tokenizer for the front ends.
4479 `-fpreprocessed' is implicit if the input file has one of the
4480 extensions `.i', `.ii' or `.mi'. These are the extensions that
4481 GCC uses for preprocessed files created by `-save-temps'.
4484 Set the distance between tab stops. This helps the preprocessor
4485 report correct column numbers in warnings or errors, even if tabs
4486 appear on the line. If the value is less than 1 or greater than
4487 100, the option is ignored. The default is 8.
4490 This option is only useful for debugging GCC. When used with
4491 `-E', dumps debugging information about location maps. Every
4492 token in the output is preceded by the dump of the map its location
4493 belongs to. The dump of the map holding the location of a token
4495 {`P':`/file/path';`F':`/includer/path';`L':LINE_NUM;`C':COL_NUM;`S':SYSTEM_HEADER_P;`M':MAP_ADDRESS;`E':MACRO_EXPANSION_P,`loc':LOCATION}
4497 When used without `-E', this option has no effect.
4499 `-ftrack-macro-expansion[=LEVEL]'
4500 Track locations of tokens across macro expansions. This allows the
4501 compiler to emit diagnostic about the current macro expansion stack
4502 when a compilation error occurs in a macro expansion. Using this
4503 option makes the preprocessor and the compiler consume more
4504 memory. The LEVEL parameter can be used to choose the level of
4505 precision of token location tracking thus decreasing the memory
4506 consumption if necessary. Value `0' of LEVEL de-activates this
4507 option just as if no `-ftrack-macro-expansion' was present on the
4508 command line. Value `1' tracks tokens locations in a degraded mode
4509 for the sake of minimal memory overhead. In this mode all tokens
4510 resulting from the expansion of an argument of a function-like
4511 macro have the same location. Value `2' tracks tokens locations
4512 completely. This value is the most memory hungry. When this
4513 option is given no argument, the default parameter value is `2'.
4515 `-fexec-charset=CHARSET'
4516 Set the execution character set, used for string and character
4517 constants. The default is UTF-8. CHARSET can be any encoding
4518 supported by the system's `iconv' library routine.
4520 `-fwide-exec-charset=CHARSET'
4521 Set the wide execution character set, used for wide string and
4522 character constants. The default is UTF-32 or UTF-16, whichever
4523 corresponds to the width of `wchar_t'. As with `-fexec-charset',
4524 CHARSET can be any encoding supported by the system's `iconv'
4525 library routine; however, you will have problems with encodings
4526 that do not fit exactly in `wchar_t'.
4528 `-finput-charset=CHARSET'
4529 Set the input character set, used for translation from the
4530 character set of the input file to the source character set used
4531 by GCC. If the locale does not specify, or GCC cannot get this
4532 information from the locale, the default is UTF-8. This can be
4533 overridden by either the locale or this command line option.
4534 Currently the command line option takes precedence if there's a
4535 conflict. CHARSET can be any encoding supported by the system's
4536 `iconv' library routine.
4538 `-fworking-directory'
4539 Enable generation of linemarkers in the preprocessor output that
4540 will let the compiler know the current working directory at the
4541 time of preprocessing. When this option is enabled, the
4542 preprocessor will emit, after the initial linemarker, a second
4543 linemarker with the current working directory followed by two
4544 slashes. GCC will use this directory, when it's present in the
4545 preprocessed input, as the directory emitted as the current
4546 working directory in some debugging information formats. This
4547 option is implicitly enabled if debugging information is enabled,
4548 but this can be inhibited with the negated form
4549 `-fno-working-directory'. If the `-P' flag is present in the
4550 command line, this option has no effect, since no `#line'
4551 directives are emitted whatsoever.
4554 Do not print column numbers in diagnostics. This may be necessary
4555 if diagnostics are being scanned by a program that does not
4556 understand the column numbers, such as `dejagnu'.
4558 `-A PREDICATE=ANSWER'
4559 Make an assertion with the predicate PREDICATE and answer ANSWER.
4560 This form is preferred to the older form `-A PREDICATE(ANSWER)',
4561 which is still supported, because it does not use shell special
4562 characters. *Note Obsolete Features::.
4564 `-A -PREDICATE=ANSWER'
4565 Cancel an assertion with the predicate PREDICATE and answer ANSWER.
4568 CHARS is a sequence of one or more of the following characters,
4569 and must not be preceded by a space. Other characters are
4570 interpreted by the compiler proper, or reserved for future
4571 versions of GCC, and so are silently ignored. If you specify
4572 characters whose behavior conflicts, the result is undefined.
4575 Instead of the normal output, generate a list of `#define'
4576 directives for all the macros defined during the execution of
4577 the preprocessor, including predefined macros. This gives
4578 you a way of finding out what is predefined in your version
4579 of the preprocessor. Assuming you have no file `foo.h', the
4582 touch foo.h; cpp -dM foo.h
4584 will show all the predefined macros.
4586 If you use `-dM' without the `-E' option, `-dM' is
4587 interpreted as a synonym for `-fdump-rtl-mach'. *Note
4588 Debugging Options: (gcc)Debugging Options.
4591 Like `M' except in two respects: it does _not_ include the
4592 predefined macros, and it outputs _both_ the `#define'
4593 directives and the result of preprocessing. Both kinds of
4594 output go to the standard output file.
4597 Like `D', but emit only the macro names, not their expansions.
4600 Output `#include' directives in addition to the result of
4604 Like `D' except that only macros that are expanded, or whose
4605 definedness is tested in preprocessor directives, are output;
4606 the output is delayed until the use or test of the macro; and
4607 `#undef' directives are also output for macros tested but
4608 undefined at the time.
4611 Inhibit generation of linemarkers in the output from the
4612 preprocessor. This might be useful when running the preprocessor
4613 on something that is not C code, and will be sent to a program
4614 which might be confused by the linemarkers. *Note Preprocessor
4618 Do not discard comments. All comments are passed through to the
4619 output file, except for comments in processed directives, which
4620 are deleted along with the directive.
4622 You should be prepared for side effects when using `-C'; it causes
4623 the preprocessor to treat comments as tokens in their own right.
4624 For example, comments appearing at the start of what would be a
4625 directive line have the effect of turning that line into an
4626 ordinary source line, since the first token on the line is no
4630 Do not discard comments, including during macro expansion. This is
4631 like `-C', except that comments contained within macros are also
4632 passed through to the output file where the macro is expanded.
4634 In addition to the side-effects of the `-C' option, the `-CC'
4635 option causes all C++-style comments inside a macro to be
4636 converted to C-style comments. This is to prevent later use of
4637 that macro from inadvertently commenting out the remainder of the
4640 The `-CC' option is generally used to support lint comments.
4643 Try to imitate the behavior of old-fashioned C preprocessors, as
4644 opposed to ISO C preprocessors. *Note Traditional Mode::.
4647 Process trigraph sequences. *Note Initial processing::.
4650 Enable special code to work around file systems which only permit
4651 very short file names, such as MS-DOS.
4655 Print text describing all the command line options instead of
4656 preprocessing anything.
4659 Verbose mode. Print out GNU CPP's version number at the beginning
4660 of execution, and report the final form of the include path.
4663 Print the name of each header file used, in addition to other
4664 normal activities. Each name is indented to show how deep in the
4665 `#include' stack it is. Precompiled header files are also
4666 printed, even if they are found to be invalid; an invalid
4667 precompiled header file is printed with `...x' and a valid one
4672 Print out GNU CPP's version number. With one dash, proceed to
4673 preprocess as normal. With two dashes, exit immediately.
4676 File: cpp.info, Node: Environment Variables, Next: GNU Free Documentation License, Prev: Invocation, Up: Top
4678 13 Environment Variables
4679 ************************
4681 This section describes the environment variables that affect how CPP
4682 operates. You can use them to specify directories or prefixes to use
4683 when searching for include files, or to control dependency output.
4685 Note that you can also specify places to search using options such as
4686 `-I', and control dependency output with options like `-M' (*note
4687 Invocation::). These take precedence over environment variables, which
4688 in turn take precedence over the configuration of GCC.
4692 `CPLUS_INCLUDE_PATH'
4694 Each variable's value is a list of directories separated by a
4695 special character, much like `PATH', in which to look for header
4696 files. The special character, `PATH_SEPARATOR', is
4697 target-dependent and determined at GCC build time. For Microsoft
4698 Windows-based targets it is a semicolon, and for almost all other
4699 targets it is a colon.
4701 `CPATH' specifies a list of directories to be searched as if
4702 specified with `-I', but after any paths given with `-I' options
4703 on the command line. This environment variable is used regardless
4704 of which language is being preprocessed.
4706 The remaining environment variables apply only when preprocessing
4707 the particular language indicated. Each specifies a list of
4708 directories to be searched as if specified with `-isystem', but
4709 after any paths given with `-isystem' options on the command line.
4711 In all these variables, an empty element instructs the compiler to
4712 search its current working directory. Empty elements can appear
4713 at the beginning or end of a path. For instance, if the value of
4714 `CPATH' is `:/special/include', that has the same effect as
4715 `-I. -I/special/include'.
4717 See also *note Search Path::.
4719 `DEPENDENCIES_OUTPUT'
4720 If this variable is set, its value specifies how to output
4721 dependencies for Make based on the non-system header files
4722 processed by the compiler. System header files are ignored in the
4725 The value of `DEPENDENCIES_OUTPUT' can be just a file name, in
4726 which case the Make rules are written to that file, guessing the
4727 target name from the source file name. Or the value can have the
4728 form `FILE TARGET', in which case the rules are written to file
4729 FILE using TARGET as the target name.
4731 In other words, this environment variable is equivalent to
4732 combining the options `-MM' and `-MF' (*note Invocation::), with
4733 an optional `-MT' switch too.
4735 `SUNPRO_DEPENDENCIES'
4736 This variable is the same as `DEPENDENCIES_OUTPUT' (see above),
4737 except that system header files are not ignored, so it implies
4738 `-M' rather than `-MM'. However, the dependence on the main input
4739 file is omitted. *Note Invocation::.
4742 File: cpp.info, Node: GNU Free Documentation License, Next: Index of Directives, Prev: Environment Variables, Up: Top
4744 GNU Free Documentation License
4745 ******************************
4747 Version 1.3, 3 November 2008
4749 Copyright (C) 2000, 2001, 2002, 2007, 2008 Free Software Foundation, Inc.
4752 Everyone is permitted to copy and distribute verbatim copies
4753 of this license document, but changing it is not allowed.
4757 The purpose of this License is to make a manual, textbook, or other
4758 functional and useful document "free" in the sense of freedom: to
4759 assure everyone the effective freedom to copy and redistribute it,
4760 with or without modifying it, either commercially or
4761 noncommercially. Secondarily, this License preserves for the
4762 author and publisher a way to get credit for their work, while not
4763 being considered responsible for modifications made by others.
4765 This License is a kind of "copyleft", which means that derivative
4766 works of the document must themselves be free in the same sense.
4767 It complements the GNU General Public License, which is a copyleft
4768 license designed for free software.
4770 We have designed this License in order to use it for manuals for
4771 free software, because free software needs free documentation: a
4772 free program should come with manuals providing the same freedoms
4773 that the software does. But this License is not limited to
4774 software manuals; it can be used for any textual work, regardless
4775 of subject matter or whether it is published as a printed book.
4776 We recommend this License principally for works whose purpose is
4777 instruction or reference.
4779 1. APPLICABILITY AND DEFINITIONS
4781 This License applies to any manual or other work, in any medium,
4782 that contains a notice placed by the copyright holder saying it
4783 can be distributed under the terms of this License. Such a notice
4784 grants a world-wide, royalty-free license, unlimited in duration,
4785 to use that work under the conditions stated herein. The
4786 "Document", below, refers to any such manual or work. Any member
4787 of the public is a licensee, and is addressed as "you". You
4788 accept the license if you copy, modify or distribute the work in a
4789 way requiring permission under copyright law.
4791 A "Modified Version" of the Document means any work containing the
4792 Document or a portion of it, either copied verbatim, or with
4793 modifications and/or translated into another language.
4795 A "Secondary Section" is a named appendix or a front-matter section
4796 of the Document that deals exclusively with the relationship of the
4797 publishers or authors of the Document to the Document's overall
4798 subject (or to related matters) and contains nothing that could
4799 fall directly within that overall subject. (Thus, if the Document
4800 is in part a textbook of mathematics, a Secondary Section may not
4801 explain any mathematics.) The relationship could be a matter of
4802 historical connection with the subject or with related matters, or
4803 of legal, commercial, philosophical, ethical or political position
4806 The "Invariant Sections" are certain Secondary Sections whose
4807 titles are designated, as being those of Invariant Sections, in
4808 the notice that says that the Document is released under this
4809 License. If a section does not fit the above definition of
4810 Secondary then it is not allowed to be designated as Invariant.
4811 The Document may contain zero Invariant Sections. If the Document
4812 does not identify any Invariant Sections then there are none.
4814 The "Cover Texts" are certain short passages of text that are
4815 listed, as Front-Cover Texts or Back-Cover Texts, in the notice
4816 that says that the Document is released under this License. A
4817 Front-Cover Text may be at most 5 words, and a Back-Cover Text may
4818 be at most 25 words.
4820 A "Transparent" copy of the Document means a machine-readable copy,
4821 represented in a format whose specification is available to the
4822 general public, that is suitable for revising the document
4823 straightforwardly with generic text editors or (for images
4824 composed of pixels) generic paint programs or (for drawings) some
4825 widely available drawing editor, and that is suitable for input to
4826 text formatters or for automatic translation to a variety of
4827 formats suitable for input to text formatters. A copy made in an
4828 otherwise Transparent file format whose markup, or absence of
4829 markup, has been arranged to thwart or discourage subsequent
4830 modification by readers is not Transparent. An image format is
4831 not Transparent if used for any substantial amount of text. A
4832 copy that is not "Transparent" is called "Opaque".
4834 Examples of suitable formats for Transparent copies include plain
4835 ASCII without markup, Texinfo input format, LaTeX input format,
4836 SGML or XML using a publicly available DTD, and
4837 standard-conforming simple HTML, PostScript or PDF designed for
4838 human modification. Examples of transparent image formats include
4839 PNG, XCF and JPG. Opaque formats include proprietary formats that
4840 can be read and edited only by proprietary word processors, SGML or
4841 XML for which the DTD and/or processing tools are not generally
4842 available, and the machine-generated HTML, PostScript or PDF
4843 produced by some word processors for output purposes only.
4845 The "Title Page" means, for a printed book, the title page itself,
4846 plus such following pages as are needed to hold, legibly, the
4847 material this License requires to appear in the title page. For
4848 works in formats which do not have any title page as such, "Title
4849 Page" means the text near the most prominent appearance of the
4850 work's title, preceding the beginning of the body of the text.
4852 The "publisher" means any person or entity that distributes copies
4853 of the Document to the public.
4855 A section "Entitled XYZ" means a named subunit of the Document
4856 whose title either is precisely XYZ or contains XYZ in parentheses
4857 following text that translates XYZ in another language. (Here XYZ
4858 stands for a specific section name mentioned below, such as
4859 "Acknowledgements", "Dedications", "Endorsements", or "History".)
4860 To "Preserve the Title" of such a section when you modify the
4861 Document means that it remains a section "Entitled XYZ" according
4864 The Document may include Warranty Disclaimers next to the notice
4865 which states that this License applies to the Document. These
4866 Warranty Disclaimers are considered to be included by reference in
4867 this License, but only as regards disclaiming warranties: any other
4868 implication that these Warranty Disclaimers may have is void and
4869 has no effect on the meaning of this License.
4873 You may copy and distribute the Document in any medium, either
4874 commercially or noncommercially, provided that this License, the
4875 copyright notices, and the license notice saying this License
4876 applies to the Document are reproduced in all copies, and that you
4877 add no other conditions whatsoever to those of this License. You
4878 may not use technical measures to obstruct or control the reading
4879 or further copying of the copies you make or distribute. However,
4880 you may accept compensation in exchange for copies. If you
4881 distribute a large enough number of copies you must also follow
4882 the conditions in section 3.
4884 You may also lend copies, under the same conditions stated above,
4885 and you may publicly display copies.
4887 3. COPYING IN QUANTITY
4889 If you publish printed copies (or copies in media that commonly
4890 have printed covers) of the Document, numbering more than 100, and
4891 the Document's license notice requires Cover Texts, you must
4892 enclose the copies in covers that carry, clearly and legibly, all
4893 these Cover Texts: Front-Cover Texts on the front cover, and
4894 Back-Cover Texts on the back cover. Both covers must also clearly
4895 and legibly identify you as the publisher of these copies. The
4896 front cover must present the full title with all words of the
4897 title equally prominent and visible. You may add other material
4898 on the covers in addition. Copying with changes limited to the
4899 covers, as long as they preserve the title of the Document and
4900 satisfy these conditions, can be treated as verbatim copying in
4903 If the required texts for either cover are too voluminous to fit
4904 legibly, you should put the first ones listed (as many as fit
4905 reasonably) on the actual cover, and continue the rest onto
4908 If you publish or distribute Opaque copies of the Document
4909 numbering more than 100, you must either include a
4910 machine-readable Transparent copy along with each Opaque copy, or
4911 state in or with each Opaque copy a computer-network location from
4912 which the general network-using public has access to download
4913 using public-standard network protocols a complete Transparent
4914 copy of the Document, free of added material. If you use the
4915 latter option, you must take reasonably prudent steps, when you
4916 begin distribution of Opaque copies in quantity, to ensure that
4917 this Transparent copy will remain thus accessible at the stated
4918 location until at least one year after the last time you
4919 distribute an Opaque copy (directly or through your agents or
4920 retailers) of that edition to the public.
4922 It is requested, but not required, that you contact the authors of
4923 the Document well before redistributing any large number of
4924 copies, to give them a chance to provide you with an updated
4925 version of the Document.
4929 You may copy and distribute a Modified Version of the Document
4930 under the conditions of sections 2 and 3 above, provided that you
4931 release the Modified Version under precisely this License, with
4932 the Modified Version filling the role of the Document, thus
4933 licensing distribution and modification of the Modified Version to
4934 whoever possesses a copy of it. In addition, you must do these
4935 things in the Modified Version:
4937 A. Use in the Title Page (and on the covers, if any) a title
4938 distinct from that of the Document, and from those of
4939 previous versions (which should, if there were any, be listed
4940 in the History section of the Document). You may use the
4941 same title as a previous version if the original publisher of
4942 that version gives permission.
4944 B. List on the Title Page, as authors, one or more persons or
4945 entities responsible for authorship of the modifications in
4946 the Modified Version, together with at least five of the
4947 principal authors of the Document (all of its principal
4948 authors, if it has fewer than five), unless they release you
4949 from this requirement.
4951 C. State on the Title page the name of the publisher of the
4952 Modified Version, as the publisher.
4954 D. Preserve all the copyright notices of the Document.
4956 E. Add an appropriate copyright notice for your modifications
4957 adjacent to the other copyright notices.
4959 F. Include, immediately after the copyright notices, a license
4960 notice giving the public permission to use the Modified
4961 Version under the terms of this License, in the form shown in
4964 G. Preserve in that license notice the full lists of Invariant
4965 Sections and required Cover Texts given in the Document's
4968 H. Include an unaltered copy of this License.
4970 I. Preserve the section Entitled "History", Preserve its Title,
4971 and add to it an item stating at least the title, year, new
4972 authors, and publisher of the Modified Version as given on
4973 the Title Page. If there is no section Entitled "History" in
4974 the Document, create one stating the title, year, authors,
4975 and publisher of the Document as given on its Title Page,
4976 then add an item describing the Modified Version as stated in
4977 the previous sentence.
4979 J. Preserve the network location, if any, given in the Document
4980 for public access to a Transparent copy of the Document, and
4981 likewise the network locations given in the Document for
4982 previous versions it was based on. These may be placed in
4983 the "History" section. You may omit a network location for a
4984 work that was published at least four years before the
4985 Document itself, or if the original publisher of the version
4986 it refers to gives permission.
4988 K. For any section Entitled "Acknowledgements" or "Dedications",
4989 Preserve the Title of the section, and preserve in the
4990 section all the substance and tone of each of the contributor
4991 acknowledgements and/or dedications given therein.
4993 L. Preserve all the Invariant Sections of the Document,
4994 unaltered in their text and in their titles. Section numbers
4995 or the equivalent are not considered part of the section
4998 M. Delete any section Entitled "Endorsements". Such a section
4999 may not be included in the Modified Version.
5001 N. Do not retitle any existing section to be Entitled
5002 "Endorsements" or to conflict in title with any Invariant
5005 O. Preserve any Warranty Disclaimers.
5007 If the Modified Version includes new front-matter sections or
5008 appendices that qualify as Secondary Sections and contain no
5009 material copied from the Document, you may at your option
5010 designate some or all of these sections as invariant. To do this,
5011 add their titles to the list of Invariant Sections in the Modified
5012 Version's license notice. These titles must be distinct from any
5013 other section titles.
5015 You may add a section Entitled "Endorsements", provided it contains
5016 nothing but endorsements of your Modified Version by various
5017 parties--for example, statements of peer review or that the text
5018 has been approved by an organization as the authoritative
5019 definition of a standard.
5021 You may add a passage of up to five words as a Front-Cover Text,
5022 and a passage of up to 25 words as a Back-Cover Text, to the end
5023 of the list of Cover Texts in the Modified Version. Only one
5024 passage of Front-Cover Text and one of Back-Cover Text may be
5025 added by (or through arrangements made by) any one entity. If the
5026 Document already includes a cover text for the same cover,
5027 previously added by you or by arrangement made by the same entity
5028 you are acting on behalf of, you may not add another; but you may
5029 replace the old one, on explicit permission from the previous
5030 publisher that added the old one.
5032 The author(s) and publisher(s) of the Document do not by this
5033 License give permission to use their names for publicity for or to
5034 assert or imply endorsement of any Modified Version.
5036 5. COMBINING DOCUMENTS
5038 You may combine the Document with other documents released under
5039 this License, under the terms defined in section 4 above for
5040 modified versions, provided that you include in the combination
5041 all of the Invariant Sections of all of the original documents,
5042 unmodified, and list them all as Invariant Sections of your
5043 combined work in its license notice, and that you preserve all
5044 their Warranty Disclaimers.
5046 The combined work need only contain one copy of this License, and
5047 multiple identical Invariant Sections may be replaced with a single
5048 copy. If there are multiple Invariant Sections with the same name
5049 but different contents, make the title of each such section unique
5050 by adding at the end of it, in parentheses, the name of the
5051 original author or publisher of that section if known, or else a
5052 unique number. Make the same adjustment to the section titles in
5053 the list of Invariant Sections in the license notice of the
5056 In the combination, you must combine any sections Entitled
5057 "History" in the various original documents, forming one section
5058 Entitled "History"; likewise combine any sections Entitled
5059 "Acknowledgements", and any sections Entitled "Dedications". You
5060 must delete all sections Entitled "Endorsements."
5062 6. COLLECTIONS OF DOCUMENTS
5064 You may make a collection consisting of the Document and other
5065 documents released under this License, and replace the individual
5066 copies of this License in the various documents with a single copy
5067 that is included in the collection, provided that you follow the
5068 rules of this License for verbatim copying of each of the
5069 documents in all other respects.
5071 You may extract a single document from such a collection, and
5072 distribute it individually under this License, provided you insert
5073 a copy of this License into the extracted document, and follow
5074 this License in all other respects regarding verbatim copying of
5077 7. AGGREGATION WITH INDEPENDENT WORKS
5079 A compilation of the Document or its derivatives with other
5080 separate and independent documents or works, in or on a volume of
5081 a storage or distribution medium, is called an "aggregate" if the
5082 copyright resulting from the compilation is not used to limit the
5083 legal rights of the compilation's users beyond what the individual
5084 works permit. When the Document is included in an aggregate, this
5085 License does not apply to the other works in the aggregate which
5086 are not themselves derivative works of the Document.
5088 If the Cover Text requirement of section 3 is applicable to these
5089 copies of the Document, then if the Document is less than one half
5090 of the entire aggregate, the Document's Cover Texts may be placed
5091 on covers that bracket the Document within the aggregate, or the
5092 electronic equivalent of covers if the Document is in electronic
5093 form. Otherwise they must appear on printed covers that bracket
5094 the whole aggregate.
5098 Translation is considered a kind of modification, so you may
5099 distribute translations of the Document under the terms of section
5100 4. Replacing Invariant Sections with translations requires special
5101 permission from their copyright holders, but you may include
5102 translations of some or all Invariant Sections in addition to the
5103 original versions of these Invariant Sections. You may include a
5104 translation of this License, and all the license notices in the
5105 Document, and any Warranty Disclaimers, provided that you also
5106 include the original English version of this License and the
5107 original versions of those notices and disclaimers. In case of a
5108 disagreement between the translation and the original version of
5109 this License or a notice or disclaimer, the original version will
5112 If a section in the Document is Entitled "Acknowledgements",
5113 "Dedications", or "History", the requirement (section 4) to
5114 Preserve its Title (section 1) will typically require changing the
5119 You may not copy, modify, sublicense, or distribute the Document
5120 except as expressly provided under this License. Any attempt
5121 otherwise to copy, modify, sublicense, or distribute it is void,
5122 and will automatically terminate your rights under this License.
5124 However, if you cease all violation of this License, then your
5125 license from a particular copyright holder is reinstated (a)
5126 provisionally, unless and until the copyright holder explicitly
5127 and finally terminates your license, and (b) permanently, if the
5128 copyright holder fails to notify you of the violation by some
5129 reasonable means prior to 60 days after the cessation.
5131 Moreover, your license from a particular copyright holder is
5132 reinstated permanently if the copyright holder notifies you of the
5133 violation by some reasonable means, this is the first time you have
5134 received notice of violation of this License (for any work) from
5135 that copyright holder, and you cure the violation prior to 30 days
5136 after your receipt of the notice.
5138 Termination of your rights under this section does not terminate
5139 the licenses of parties who have received copies or rights from
5140 you under this License. If your rights have been terminated and
5141 not permanently reinstated, receipt of a copy of some or all of
5142 the same material does not give you any rights to use it.
5144 10. FUTURE REVISIONS OF THIS LICENSE
5146 The Free Software Foundation may publish new, revised versions of
5147 the GNU Free Documentation License from time to time. Such new
5148 versions will be similar in spirit to the present version, but may
5149 differ in detail to address new problems or concerns. See
5150 `http://www.gnu.org/copyleft/'.
5152 Each version of the License is given a distinguishing version
5153 number. If the Document specifies that a particular numbered
5154 version of this License "or any later version" applies to it, you
5155 have the option of following the terms and conditions either of
5156 that specified version or of any later version that has been
5157 published (not as a draft) by the Free Software Foundation. If
5158 the Document does not specify a version number of this License,
5159 you may choose any version ever published (not as a draft) by the
5160 Free Software Foundation. If the Document specifies that a proxy
5161 can decide which future versions of this License can be used, that
5162 proxy's public statement of acceptance of a version permanently
5163 authorizes you to choose that version for the Document.
5167 "Massive Multiauthor Collaboration Site" (or "MMC Site") means any
5168 World Wide Web server that publishes copyrightable works and also
5169 provides prominent facilities for anybody to edit those works. A
5170 public wiki that anybody can edit is an example of such a server.
5171 A "Massive Multiauthor Collaboration" (or "MMC") contained in the
5172 site means any set of copyrightable works thus published on the MMC
5175 "CC-BY-SA" means the Creative Commons Attribution-Share Alike 3.0
5176 license published by Creative Commons Corporation, a not-for-profit
5177 corporation with a principal place of business in San Francisco,
5178 California, as well as future copyleft versions of that license
5179 published by that same organization.
5181 "Incorporate" means to publish or republish a Document, in whole or
5182 in part, as part of another Document.
5184 An MMC is "eligible for relicensing" if it is licensed under this
5185 License, and if all works that were first published under this
5186 License somewhere other than this MMC, and subsequently
5187 incorporated in whole or in part into the MMC, (1) had no cover
5188 texts or invariant sections, and (2) were thus incorporated prior
5189 to November 1, 2008.
5191 The operator of an MMC Site may republish an MMC contained in the
5192 site under CC-BY-SA on the same site at any time before August 1,
5193 2009, provided the MMC is eligible for relicensing.
5196 ADDENDUM: How to use this License for your documents
5197 ====================================================
5199 To use this License in a document you have written, include a copy of
5200 the License in the document and put the following copyright and license
5201 notices just after the title page:
5203 Copyright (C) YEAR YOUR NAME.
5204 Permission is granted to copy, distribute and/or modify this document
5205 under the terms of the GNU Free Documentation License, Version 1.3
5206 or any later version published by the Free Software Foundation;
5207 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
5208 Texts. A copy of the license is included in the section entitled ``GNU
5209 Free Documentation License''.
5211 If you have Invariant Sections, Front-Cover Texts and Back-Cover
5212 Texts, replace the "with...Texts." line with this:
5214 with the Invariant Sections being LIST THEIR TITLES, with
5215 the Front-Cover Texts being LIST, and with the Back-Cover Texts
5218 If you have Invariant Sections without Cover Texts, or some other
5219 combination of the three, merge those two alternatives to suit the
5222 If your document contains nontrivial examples of program code, we
5223 recommend releasing these examples in parallel under your choice of
5224 free software license, such as the GNU General Public License, to
5225 permit their use in free software.
5228 File: cpp.info, Node: Index of Directives, Next: Option Index, Prev: GNU Free Documentation License, Up: Top
5236 * #assert: Obsolete Features. (line 48)
5237 * #define: Object-like Macros. (line 11)
5238 * #elif: Elif. (line 6)
5239 * #else: Else. (line 6)
5240 * #endif: Ifdef. (line 6)
5241 * #error: Diagnostics. (line 6)
5242 * #ident: Other Directives. (line 6)
5243 * #if: Conditional Syntax. (line 6)
5244 * #ifdef: Ifdef. (line 6)
5245 * #ifndef: Ifdef. (line 40)
5246 * #import: Alternatives to Wrapper #ifndef.
5248 * #include: Include Syntax. (line 6)
5249 * #include_next: Wrapper Headers. (line 6)
5250 * #line: Line Control. (line 20)
5251 * #pragma GCC dependency: Pragmas. (line 55)
5252 * #pragma GCC poison: Pragmas. (line 67)
5253 * #pragma GCC system_header <1>: Pragmas. (line 94)
5254 * #pragma GCC system_header: System Headers. (line 31)
5255 * #sccs: Other Directives. (line 6)
5256 * #unassert: Obsolete Features. (line 59)
5257 * #undef: Undefining and Redefining Macros.
5259 * #warning: Diagnostics. (line 27)
5262 File: cpp.info, Node: Option Index, Next: Concept Index, Prev: Index of Directives, Up: Top
5267 CPP's command line options and environment variables are indexed here
5268 without any initial `-' or `--'.
5273 * A: Invocation. (line 562)
5274 * ansi: Invocation. (line 308)
5275 * C: Invocation. (line 621)
5276 * C_INCLUDE_PATH: Environment Variables.
5278 * CPATH: Environment Variables.
5280 * CPLUS_INCLUDE_PATH: Environment Variables.
5282 * D: Invocation. (line 39)
5283 * dD: Invocation. (line 594)
5284 * DEPENDENCIES_OUTPUT: Environment Variables.
5286 * dI: Invocation. (line 603)
5287 * dM: Invocation. (line 578)
5288 * dN: Invocation. (line 600)
5289 * dU: Invocation. (line 607)
5290 * fdebug-cpp: Invocation. (line 493)
5291 * fdirectives-only: Invocation. (line 444)
5292 * fdollars-in-identifiers: Invocation. (line 466)
5293 * fexec-charset: Invocation. (line 519)
5294 * fextended-identifiers: Invocation. (line 469)
5295 * finput-charset: Invocation. (line 532)
5296 * fno-show-column: Invocation. (line 557)
5297 * fno-working-directory: Invocation. (line 542)
5298 * fpreprocessed: Invocation. (line 474)
5299 * ftabstop: Invocation. (line 487)
5300 * ftrack-macro-expansion: Invocation. (line 503)
5301 * fwide-exec-charset: Invocation. (line 524)
5302 * fworking-directory: Invocation. (line 542)
5303 * H: Invocation. (line 666)
5304 * help: Invocation. (line 658)
5305 * I: Invocation. (line 71)
5306 * I-: Invocation. (line 357)
5307 * idirafter: Invocation. (line 399)
5308 * imacros: Invocation. (line 390)
5309 * imultilib: Invocation. (line 424)
5310 * include: Invocation. (line 379)
5311 * iprefix: Invocation. (line 406)
5312 * iquote: Invocation. (line 436)
5313 * isysroot: Invocation. (line 418)
5314 * isystem: Invocation. (line 428)
5315 * iwithprefix: Invocation. (line 412)
5316 * iwithprefixbefore: Invocation. (line 412)
5317 * M: Invocation. (line 180)
5318 * MD: Invocation. (line 269)
5319 * MF: Invocation. (line 215)
5320 * MG: Invocation. (line 224)
5321 * MM: Invocation. (line 205)
5322 * MMD: Invocation. (line 285)
5323 * MP: Invocation. (line 234)
5324 * MQ: Invocation. (line 260)
5325 * MT: Invocation. (line 246)
5326 * nostdinc: Invocation. (line 369)
5327 * nostdinc++: Invocation. (line 374)
5328 * o: Invocation. (line 82)
5329 * OBJC_INCLUDE_PATH: Environment Variables.
5331 * P: Invocation. (line 614)
5332 * pedantic: Invocation. (line 170)
5333 * pedantic-errors: Invocation. (line 175)
5334 * remap: Invocation. (line 653)
5335 * std=: Invocation. (line 308)
5336 * SUNPRO_DEPENDENCIES: Environment Variables.
5338 * target-help: Invocation. (line 658)
5339 * traditional-cpp: Invocation. (line 646)
5340 * trigraphs: Invocation. (line 650)
5341 * U: Invocation. (line 62)
5342 * undef: Invocation. (line 66)
5343 * v: Invocation. (line 662)
5344 * version: Invocation. (line 675)
5345 * w: Invocation. (line 166)
5346 * Wall: Invocation. (line 88)
5347 * Wcomment: Invocation. (line 96)
5348 * Wcomments: Invocation. (line 96)
5349 * Wendif-labels: Invocation. (line 143)
5350 * Werror: Invocation. (line 156)
5351 * Wsystem-headers: Invocation. (line 160)
5352 * Wtraditional: Invocation. (line 113)
5353 * Wtrigraphs: Invocation. (line 101)
5354 * Wundef: Invocation. (line 119)
5355 * Wunused-macros: Invocation. (line 124)
5356 * x: Invocation. (line 292)
5359 File: cpp.info, Node: Concept Index, Prev: Option Index, Up: Top
5367 * # operator: Stringification. (line 6)
5368 * ## operator: Concatenation. (line 6)
5369 * _Pragma: Pragmas. (line 25)
5370 * alternative tokens: Tokenization. (line 106)
5371 * arguments: Macro Arguments. (line 6)
5372 * arguments in macro definitions: Macro Arguments. (line 6)
5373 * assertions: Obsolete Features. (line 13)
5374 * assertions, canceling: Obsolete Features. (line 59)
5375 * backslash-newline: Initial processing. (line 61)
5376 * block comments: Initial processing. (line 77)
5377 * C++ named operators: C++ Named Operators. (line 6)
5378 * character constants: Tokenization. (line 85)
5379 * character set, execution: Invocation. (line 519)
5380 * character set, input: Invocation. (line 532)
5381 * character set, wide execution: Invocation. (line 524)
5382 * command line: Invocation. (line 6)
5383 * commenting out code: Deleted Code. (line 6)
5384 * comments: Initial processing. (line 77)
5385 * common predefined macros: Common Predefined Macros.
5387 * computed includes: Computed Includes. (line 6)
5388 * concatenation: Concatenation. (line 6)
5389 * conditional group: Ifdef. (line 14)
5390 * conditionals: Conditionals. (line 6)
5391 * continued lines: Initial processing. (line 61)
5392 * controlling macro: Once-Only Headers. (line 35)
5393 * defined: Defined. (line 6)
5394 * dependencies for make as output: Environment Variables.
5396 * dependencies, make: Invocation. (line 180)
5397 * diagnostic: Diagnostics. (line 6)
5398 * differences from previous versions: Differences from previous versions.
5400 * digraphs: Tokenization. (line 106)
5401 * directive line: The preprocessing language.
5403 * directive name: The preprocessing language.
5405 * directives: The preprocessing language.
5407 * empty macro arguments: Macro Arguments. (line 66)
5408 * environment variables: Environment Variables.
5410 * expansion of arguments: Argument Prescan. (line 6)
5411 * FDL, GNU Free Documentation License: GNU Free Documentation License.
5413 * function-like macros: Function-like Macros.
5415 * grouping options: Invocation. (line 34)
5416 * guard macro: Once-Only Headers. (line 35)
5417 * header file: Header Files. (line 6)
5418 * header file names: Tokenization. (line 85)
5419 * identifiers: Tokenization. (line 34)
5420 * implementation limits: Implementation limits.
5422 * implementation-defined behavior: Implementation-defined behavior.
5424 * including just once: Once-Only Headers. (line 6)
5425 * invocation: Invocation. (line 6)
5426 * iso646.h: C++ Named Operators. (line 6)
5427 * line comments: Initial processing. (line 77)
5428 * line control: Line Control. (line 6)
5429 * line endings: Initial processing. (line 14)
5430 * linemarkers: Preprocessor Output. (line 28)
5431 * macro argument expansion: Argument Prescan. (line 6)
5432 * macro arguments and directives: Directives Within Macro Arguments.
5434 * macros in include: Computed Includes. (line 6)
5435 * macros with arguments: Macro Arguments. (line 6)
5436 * macros with variable arguments: Variadic Macros. (line 6)
5437 * make: Invocation. (line 180)
5438 * manifest constants: Object-like Macros. (line 6)
5439 * named operators: C++ Named Operators. (line 6)
5440 * newlines in macro arguments: Newlines in Arguments.
5442 * null directive: Other Directives. (line 15)
5443 * numbers: Tokenization. (line 61)
5444 * object-like macro: Object-like Macros. (line 6)
5445 * options: Invocation. (line 38)
5446 * options, grouping: Invocation. (line 34)
5447 * other tokens: Tokenization. (line 120)
5448 * output format: Preprocessor Output. (line 12)
5449 * overriding a header file: Wrapper Headers. (line 6)
5450 * parentheses in macro bodies: Operator Precedence Problems.
5452 * pitfalls of macros: Macro Pitfalls. (line 6)
5453 * predefined macros: Predefined Macros. (line 6)
5454 * predefined macros, system-specific: System-specific Predefined Macros.
5456 * predicates: Obsolete Features. (line 26)
5457 * preprocessing directives: The preprocessing language.
5459 * preprocessing numbers: Tokenization. (line 61)
5460 * preprocessing tokens: Tokenization. (line 6)
5461 * prescan of macro arguments: Argument Prescan. (line 6)
5462 * problems with macros: Macro Pitfalls. (line 6)
5463 * punctuators: Tokenization. (line 106)
5464 * redefining macros: Undefining and Redefining Macros.
5466 * repeated inclusion: Once-Only Headers. (line 6)
5467 * reporting errors: Diagnostics. (line 6)
5468 * reporting warnings: Diagnostics. (line 6)
5469 * reserved namespace: System-specific Predefined Macros.
5471 * self-reference: Self-Referential Macros.
5473 * semicolons (after macro calls): Swallowing the Semicolon.
5475 * side effects (in macro arguments): Duplication of Side Effects.
5477 * standard predefined macros.: Standard Predefined Macros.
5479 * string constants: Tokenization. (line 85)
5480 * string literals: Tokenization. (line 85)
5481 * stringification: Stringification. (line 6)
5482 * symbolic constants: Object-like Macros. (line 6)
5483 * system header files <1>: System Headers. (line 6)
5484 * system header files: Header Files. (line 13)
5485 * system-specific predefined macros: System-specific Predefined Macros.
5487 * testing predicates: Obsolete Features. (line 37)
5488 * token concatenation: Concatenation. (line 6)
5489 * token pasting: Concatenation. (line 6)
5490 * tokens: Tokenization. (line 6)
5491 * trigraphs: Initial processing. (line 32)
5492 * undefining macros: Undefining and Redefining Macros.
5494 * unsafe macros: Duplication of Side Effects.
5496 * variable number of arguments: Variadic Macros. (line 6)
5497 * variadic macros: Variadic Macros. (line 6)
5498 * wrapper #ifndef: Once-Only Headers. (line 6)
5499 * wrapper headers: Wrapper Headers. (line 6)
5505 Node: Overview
\7f3855
5506 Node: Character sets
\7f6688
5507 Ref: Character sets-Footnote-1
\7f8871
5508 Node: Initial processing
\7f9052
5509 Ref: trigraphs
\7f10611
5510 Node: Tokenization
\7f14813
5511 Ref: Tokenization-Footnote-1
\7f21949
5512 Node: The preprocessing language
\7f22060
5513 Node: Header Files
\7f24938
5514 Node: Include Syntax
\7f26854
5515 Node: Include Operation
\7f28491
5516 Node: Search Path
\7f30339
5517 Node: Once-Only Headers
\7f33529
5518 Node: Alternatives to Wrapper #ifndef
\7f35188
5519 Node: Computed Includes
\7f36931
5520 Node: Wrapper Headers
\7f40089
5521 Node: System Headers
\7f42515
5522 Node: Macros
\7f44565
5523 Node: Object-like Macros
\7f45706
5524 Node: Function-like Macros
\7f49296
5525 Node: Macro Arguments
\7f50912
5526 Node: Stringification
\7f55057
5527 Node: Concatenation
\7f58263
5528 Node: Variadic Macros
\7f61371
5529 Node: Predefined Macros
\7f66158
5530 Node: Standard Predefined Macros
\7f66746
5531 Node: Common Predefined Macros
\7f72683
5532 Node: System-specific Predefined Macros
\7f90198
5533 Node: C++ Named Operators
\7f92219
5534 Node: Undefining and Redefining Macros
\7f93183
5535 Node: Directives Within Macro Arguments
\7f95287
5536 Node: Macro Pitfalls
\7f96835
5537 Node: Misnesting
\7f97368
5538 Node: Operator Precedence Problems
\7f98480
5539 Node: Swallowing the Semicolon
\7f100346
5540 Node: Duplication of Side Effects
\7f102369
5541 Node: Self-Referential Macros
\7f104552
5542 Node: Argument Prescan
\7f106961
5543 Node: Newlines in Arguments
\7f110715
5544 Node: Conditionals
\7f111666
5545 Node: Conditional Uses
\7f113496
5546 Node: Conditional Syntax
\7f114854
5547 Node: Ifdef
\7f115174
5549 Node: Defined
\7f120639
5552 Node: Deleted Code
\7f123781
5553 Node: Diagnostics
\7f125028
5554 Node: Line Control
\7f126645
5555 Node: Pragmas
\7f130449
5556 Node: Other Directives
\7f134766
5557 Node: Preprocessor Output
\7f135816
5558 Node: Traditional Mode
\7f139017
5559 Node: Traditional lexical analysis
\7f140075
5560 Node: Traditional macros
\7f142578
5561 Node: Traditional miscellany
\7f146380
5562 Node: Traditional warnings
\7f147377
5563 Node: Implementation Details
\7f149574
5564 Node: Implementation-defined behavior
\7f150195
5565 Ref: Identifier characters
\7f150947
5566 Node: Implementation limits
\7f154025
5567 Node: Obsolete Features
\7f156699
5568 Node: Differences from previous versions
\7f159587
5569 Node: Invocation
\7f163795
5570 Ref: Wtrigraphs
\7f168247
5571 Ref: dashMF
\7f173022
5572 Ref: fdollars-in-identifiers
\7f182753
5573 Node: Environment Variables
\7f192434
5574 Node: GNU Free Documentation License
\7f195400
5575 Node: Index of Directives
\7f220564
5576 Node: Option Index
\7f222498
5577 Node: Concept Index
\7f228828