3 @settitle The C Preprocessor
9 @include gcc-common.texi
12 @c man begin COPYRIGHT
13 Copyright @copyright{} 1987, 1989, 1991, 1992, 1993, 1994, 1995, 1996,
14 1997, 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007,
16 Free Software Foundation, Inc.
18 Permission is granted to copy, distribute and/or modify this document
19 under the terms of the GNU Free Documentation License, Version 1.2 or
20 any later version published by the Free Software Foundation. A copy of
21 the license is included in the
23 section entitled ``GNU Free Documentation License''.
25 @c man begin COPYRIGHT
30 @c man begin COPYRIGHT
31 This manual contains no Invariant Sections. The Front-Cover Texts are
32 (a) (see below), and the Back-Cover Texts are (b) (see below).
34 (a) The FSF's Front-Cover Text is:
38 (b) The FSF's Back-Cover Text is:
40 You have freedom to copy and modify this GNU Manual, like GNU
41 software. Copies published by the Free Software Foundation raise
42 funds for GNU development.
46 @c Create a separate index for command line options.
50 @c Used in cppopts.texi and cppenv.texi.
54 @dircategory Software development
56 * Cpp: (cpp). The GNU C preprocessor.
61 @title The C Preprocessor
63 @author Richard M. Stallman, Zachary Weinberg
65 @c There is a fill at the bottom of the page, so we need a filll to
67 @vskip 0pt plus 1filll
76 The C preprocessor implements the macro language used to transform C,
77 C++, and Objective-C programs before they are compiled. It can also be
89 * Preprocessor Output::
91 * Implementation Details::
93 * Environment Variables::
94 * GNU Free Documentation License::
95 * Index of Directives::
100 --- The Detailed Node Listing ---
105 * Initial processing::
107 * The preprocessing language::
112 * Include Operation::
114 * Once-Only Headers::
115 * Alternatives to Wrapper #ifndef::
116 * Computed Includes::
122 * Object-like Macros::
123 * Function-like Macros::
128 * Predefined Macros::
129 * Undefining and Redefining Macros::
130 * Directives Within Macro Arguments::
135 * Standard Predefined Macros::
136 * Common Predefined Macros::
137 * System-specific Predefined Macros::
138 * C++ Named Operators::
143 * Operator Precedence Problems::
144 * Swallowing the Semicolon::
145 * Duplication of Side Effects::
146 * Self-Referential Macros::
148 * Newlines in Arguments::
153 * Conditional Syntax::
164 Implementation Details
166 * Implementation-defined behavior::
167 * Implementation limits::
168 * Obsolete Features::
169 * Differences from previous versions::
173 * Obsolete Features::
183 @c man begin DESCRIPTION
184 The C preprocessor, often known as @dfn{cpp}, is a @dfn{macro processor}
185 that is used automatically by the C compiler to transform your program
186 before compilation. It is called a macro processor because it allows
187 you to define @dfn{macros}, which are brief abbreviations for longer
190 The C preprocessor is intended to be used only with C, C++, and
191 Objective-C source code. In the past, it has been abused as a general
192 text processor. It will choke on input which does not obey C's lexical
193 rules. For example, apostrophes will be interpreted as the beginning of
194 character constants, and cause errors. Also, you cannot rely on it
195 preserving characteristics of the input which are not significant to
196 C-family languages. If a Makefile is preprocessed, all the hard tabs
197 will be removed, and the Makefile will not work.
199 Having said that, you can often get away with using cpp on things which
200 are not C@. Other Algol-ish programming languages are often safe
201 (Pascal, Ada, etc.) So is assembly, with caution. @option{-traditional-cpp}
202 mode preserves more white space, and is otherwise more permissive. Many
203 of the problems can be avoided by writing C or C++ style comments
204 instead of native language comments, and keeping macros simple.
206 Wherever possible, you should use a preprocessor geared to the language
207 you are writing in. Modern versions of the GNU assembler have macro
208 facilities. Most high level programming languages have their own
209 conditional compilation and inclusion mechanism. If all else fails,
210 try a true general text processor, such as GNU M4.
212 C preprocessors vary in some details. This manual discusses the GNU C
213 preprocessor, which provides a small superset of the features of ISO
214 Standard C@. In its default mode, the GNU C preprocessor does not do a
215 few things required by the standard. These are features which are
216 rarely, if ever, used, and may cause surprising changes to the meaning
217 of a program which does not expect them. To get strict ISO Standard C,
218 you should use the @option{-std=c89} or @option{-std=c99} options, depending
219 on which version of the standard you want. To get all the mandatory
220 diagnostics, you must also use @option{-pedantic}. @xref{Invocation}.
222 This manual describes the behavior of the ISO preprocessor. To
223 minimize gratuitous differences, where the ISO preprocessor's
224 behavior does not conflict with traditional semantics, the
225 traditional preprocessor should behave the same way. The various
226 differences that do exist are detailed in the section @ref{Traditional
229 For clarity, unless noted otherwise, references to @samp{CPP} in this
230 manual refer to GNU CPP@.
235 * Initial processing::
237 * The preprocessing language::
241 @section Character sets
243 Source code character set processing in C and related languages is
244 rather complicated. The C standard discusses two character sets, but
245 there are really at least four.
247 The files input to CPP might be in any character set at all. CPP's
248 very first action, before it even looks for line boundaries, is to
249 convert the file into the character set it uses for internal
250 processing. That set is what the C standard calls the @dfn{source}
251 character set. It must be isomorphic with ISO 10646, also known as
252 Unicode. CPP uses the UTF-8 encoding of Unicode.
254 The character sets of the input files are specified using the
255 @option{-finput-charset=} option.
257 All preprocessing work (the subject of the rest of this manual) is
258 carried out in the source character set. If you request textual
259 output from the preprocessor with the @option{-E} option, it will be
262 After preprocessing is complete, string and character constants are
263 converted again, into the @dfn{execution} character set. This
264 character set is under control of the user; the default is UTF-8,
265 matching the source character set. Wide string and character
266 constants have their own character set, which is not called out
267 specifically in the standard. Again, it is under control of the user.
268 The default is UTF-16 or UTF-32, whichever fits in the target's
269 @code{wchar_t} type, in the target machine's byte
270 order.@footnote{UTF-16 does not meet the requirements of the C
271 standard for a wide character set, but the choice of 16-bit
272 @code{wchar_t} is enshrined in some system ABIs so we cannot fix
273 this.} Octal and hexadecimal escape sequences do not undergo
274 conversion; @t{'\x12'} has the value 0x12 regardless of the currently
275 selected execution character set. All other escapes are replaced by
276 the character in the source character set that they represent, then
277 converted to the execution character set, just like unescaped
280 Unless the experimental @option{-fextended-identifiers} option is used,
281 GCC does not permit the use of characters outside the ASCII range, nor
282 @samp{\u} and @samp{\U} escapes, in identifiers. Even with that
283 option, characters outside the ASCII range can only be specified with
284 the @samp{\u} and @samp{\U} escapes, not used directly in identifiers.
286 @node Initial processing
287 @section Initial processing
289 The preprocessor performs a series of textual transformations on its
290 input. These happen before all other processing. Conceptually, they
291 happen in a rigid order, and the entire file is run through each
292 transformation before the next one begins. CPP actually does them
293 all at once, for performance reasons. These transformations correspond
294 roughly to the first three ``phases of translation'' described in the C
300 The input file is read into memory and broken into lines.
302 Different systems use different conventions to indicate the end of a
303 line. GCC accepts the ASCII control sequences @kbd{LF}, @kbd{@w{CR
304 LF}} and @kbd{CR} as end-of-line markers. These are the canonical
305 sequences used by Unix, DOS and VMS, and the classic Mac OS (before
306 OSX) respectively. You may therefore safely copy source code written
307 on any of those systems to a different one and use it without
308 conversion. (GCC may lose track of the current line number if a file
309 doesn't consistently use one convention, as sometimes happens when it
310 is edited on computers with different conventions that share a network
313 If the last line of any input file lacks an end-of-line marker, the end
314 of the file is considered to implicitly supply one. The C standard says
315 that this condition provokes undefined behavior, so GCC will emit a
320 @anchor{trigraphs}If trigraphs are enabled, they are replaced by their
321 corresponding single characters. By default GCC ignores trigraphs,
322 but if you request a strictly conforming mode with the @option{-std}
323 option, or you specify the @option{-trigraphs} option, then it
326 These are nine three-character sequences, all starting with @samp{??},
327 that are defined by ISO C to stand for single characters. They permit
328 obsolete systems that lack some of C's punctuation to use C@. For
329 example, @samp{??/} stands for @samp{\}, so @t{'??/n'} is a character
330 constant for a newline.
332 Trigraphs are not popular and many compilers implement them
333 incorrectly. Portable code should not rely on trigraphs being either
334 converted or ignored. With @option{-Wtrigraphs} GCC will warn you
335 when a trigraph may change the meaning of your program if it were
336 converted. @xref{Wtrigraphs}.
338 In a string constant, you can prevent a sequence of question marks
339 from being confused with a trigraph by inserting a backslash between
340 the question marks, or by separating the string literal at the
341 trigraph and making use of string literal concatenation. @t{"(??\?)"}
342 is the string @samp{(???)}, not @samp{(?]}. Traditional C compilers
343 do not recognize these idioms.
345 The nine trigraphs and their replacements are
348 Trigraph: ??( ??) ??< ??> ??= ??/ ??' ??! ??-
349 Replacement: [ ] @{ @} # \ ^ | ~
353 @cindex continued lines
354 @cindex backslash-newline
355 Continued lines are merged into one long line.
357 A continued line is a line which ends with a backslash, @samp{\}. The
358 backslash is removed and the following line is joined with the current
359 one. No space is inserted, so you may split a line anywhere, even in
360 the middle of a word. (It is generally more readable to split lines
361 only at white space.)
363 The trailing backslash on a continued line is commonly referred to as a
364 @dfn{backslash-newline}.
366 If there is white space between a backslash and the end of a line, that
367 is still a continued line. However, as this is usually the result of an
368 editing mistake, and many compilers will not accept it as a continued
369 line, GCC will warn you about it.
373 @cindex line comments
374 @cindex block comments
375 All comments are replaced with single spaces.
377 There are two kinds of comments. @dfn{Block comments} begin with
378 @samp{/*} and continue until the next @samp{*/}. Block comments do not
382 /* @r{this is} /* @r{one comment} */ @r{text outside comment}
385 @dfn{Line comments} begin with @samp{//} and continue to the end of the
386 current line. Line comments do not nest either, but it does not matter,
387 because they would end in the same place anyway.
390 // @r{this is} // @r{one comment}
391 @r{text outside comment}
395 It is safe to put line comments inside block comments, or vice versa.
400 // @r{contains line comment}
402 */ @r{outside comment}
404 // @r{line comment} /* @r{contains block comment} */
408 But beware of commenting out one end of a block comment with a line
413 // @r{l.c.} /* @r{block comment begins}
414 @r{oops! this isn't a comment anymore} */
418 Comments are not recognized within string literals.
419 @t{@w{"/* blah */"}} is the string constant @samp{@w{/* blah */}}, not
422 Line comments are not in the 1989 edition of the C standard, but they
423 are recognized by GCC as an extension. In C++ and in the 1999 edition
424 of the C standard, they are an official part of the language.
426 Since these transformations happen before all other processing, you can
427 split a line mechanically with backslash-newline anywhere. You can
428 comment out the end of a line. You can continue a line comment onto the
429 next line with backslash-newline. You can even split @samp{/*},
430 @samp{*/}, and @samp{//} onto multiple lines with backslash-newline.
446 is equivalent to @code{@w{#define FOO 1020}}. All these tricks are
447 extremely confusing and should not be used in code intended to be
450 There is no way to prevent a backslash at the end of a line from being
451 interpreted as a backslash-newline. This cannot affect any correct
455 @section Tokenization
458 @cindex preprocessing tokens
459 After the textual transformations are finished, the input file is
460 converted into a sequence of @dfn{preprocessing tokens}. These mostly
461 correspond to the syntactic tokens used by the C compiler, but there are
462 a few differences. White space separates tokens; it is not itself a
463 token of any kind. Tokens do not have to be separated by white space,
464 but it is often necessary to avoid ambiguities.
466 When faced with a sequence of characters that has more than one possible
467 tokenization, the preprocessor is greedy. It always makes each token,
468 starting from the left, as big as possible before moving on to the next
469 token. For instance, @code{a+++++b} is interpreted as
470 @code{@w{a ++ ++ + b}}, not as @code{@w{a ++ + ++ b}}, even though the
471 latter tokenization could be part of a valid C program and the former
474 Once the input file is broken into tokens, the token boundaries never
475 change, except when the @samp{##} preprocessing operator is used to paste
476 tokens together. @xref{Concatenation}. For example,
488 The compiler does not re-tokenize the preprocessor's output. Each
489 preprocessing token becomes one compiler token.
492 Preprocessing tokens fall into five broad classes: identifiers,
493 preprocessing numbers, string literals, punctuators, and other. An
494 @dfn{identifier} is the same as an identifier in C: any sequence of
495 letters, digits, or underscores, which begins with a letter or
496 underscore. Keywords of C have no significance to the preprocessor;
497 they are ordinary identifiers. You can define a macro whose name is a
498 keyword, for instance. The only identifier which can be considered a
499 preprocessing keyword is @code{defined}. @xref{Defined}.
501 This is mostly true of other languages which use the C preprocessor.
502 However, a few of the keywords of C++ are significant even in the
503 preprocessor. @xref{C++ Named Operators}.
505 In the 1999 C standard, identifiers may contain letters which are not
506 part of the ``basic source character set'', at the implementation's
507 discretion (such as accented Latin letters, Greek letters, or Chinese
508 ideograms). This may be done with an extended character set, or the
509 @samp{\u} and @samp{\U} escape sequences. The implementation of this
510 feature in GCC is experimental; such characters are only accepted in
511 the @samp{\u} and @samp{\U} forms and only if
512 @option{-fextended-identifiers} is used.
514 As an extension, GCC treats @samp{$} as a letter. This is for
515 compatibility with some systems, such as VMS, where @samp{$} is commonly
516 used in system-defined function and object names. @samp{$} is not a
517 letter in strictly conforming mode, or if you specify the @option{-$}
518 option. @xref{Invocation}.
521 @cindex preprocessing numbers
522 A @dfn{preprocessing number} has a rather bizarre definition. The
523 category includes all the normal integer and floating point constants
524 one expects of C, but also a number of other things one might not
525 initially recognize as a number. Formally, preprocessing numbers begin
526 with an optional period, a required decimal digit, and then continue
527 with any sequence of letters, digits, underscores, periods, and
528 exponents. Exponents are the two-character sequences @samp{e+},
529 @samp{e-}, @samp{E+}, @samp{E-}, @samp{p+}, @samp{p-}, @samp{P+}, and
530 @samp{P-}. (The exponents that begin with @samp{p} or @samp{P} are new
531 to C99. They are used for hexadecimal floating-point constants.)
533 The purpose of this unusual definition is to isolate the preprocessor
534 from the full complexity of numeric constants. It does not have to
535 distinguish between lexically valid and invalid floating-point numbers,
536 which is complicated. The definition also permits you to split an
537 identifier at any position and get exactly two tokens, which can then be
538 pasted back together with the @samp{##} operator.
540 It's possible for preprocessing numbers to cause programs to be
541 misinterpreted. For example, @code{0xE+12} is a preprocessing number
542 which does not translate to any valid numeric constant, therefore a
543 syntax error. It does not mean @code{@w{0xE + 12}}, which is what you
546 @cindex string literals
547 @cindex string constants
548 @cindex character constants
549 @cindex header file names
550 @c the @: prevents makeinfo from turning '' into ".
551 @dfn{String literals} are string constants, character constants, and
552 header file names (the argument of @samp{#include}).@footnote{The C
553 standard uses the term @dfn{string literal} to refer only to what we are
554 calling @dfn{string constants}.} String constants and character
555 constants are straightforward: @t{"@dots{}"} or @t{'@dots{}'}. In
556 either case embedded quotes should be escaped with a backslash:
557 @t{'\'@:'} is the character constant for @samp{'}. There is no limit on
558 the length of a character constant, but the value of a character
559 constant that contains more than one character is
560 implementation-defined. @xref{Implementation Details}.
562 Header file names either look like string constants, @t{"@dots{}"}, or are
563 written with angle brackets instead, @t{<@dots{}>}. In either case,
564 backslash is an ordinary character. There is no way to escape the
565 closing quote or angle bracket. The preprocessor looks for the header
566 file in different places depending on which form you use. @xref{Include
569 No string literal may extend past the end of a line. Older versions
570 of GCC accepted multi-line string constants. You may use continued
571 lines instead, or string constant concatenation. @xref{Differences
572 from previous versions}.
576 @cindex alternative tokens
577 @dfn{Punctuators} are all the usual bits of punctuation which are
578 meaningful to C and C++. All but three of the punctuation characters in
579 ASCII are C punctuators. The exceptions are @samp{@@}, @samp{$}, and
580 @samp{`}. In addition, all the two- and three-character operators are
581 punctuators. There are also six @dfn{digraphs}, which the C++ standard
582 calls @dfn{alternative tokens}, which are merely alternate ways to spell
583 other punctuators. This is a second attempt to work around missing
584 punctuation in obsolete systems. It has no negative side effects,
585 unlike trigraphs, but does not cover as much ground. The digraphs and
586 their corresponding normal punctuators are:
589 Digraph: <% %> <: :> %: %:%:
590 Punctuator: @{ @} [ ] # ##
594 Any other single character is considered ``other''. It is passed on to
595 the preprocessor's output unmolested. The C compiler will almost
596 certainly reject source code containing ``other'' tokens. In ASCII, the
597 only other characters are @samp{@@}, @samp{$}, @samp{`}, and control
598 characters other than NUL (all bits zero). (Note that @samp{$} is
599 normally considered a letter.) All characters with the high bit set
600 (numeric range 0x7F--0xFF) are also ``other'' in the present
601 implementation. This will change when proper support for international
602 character sets is added to GCC@.
604 NUL is a special case because of the high probability that its
605 appearance is accidental, and because it may be invisible to the user
606 (many terminals do not display NUL at all). Within comments, NULs are
607 silently ignored, just as any other character would be. In running
608 text, NUL is considered white space. For example, these two directives
609 have the same meaning.
617 (where @samp{^@@} is ASCII NUL)@. Within string or character constants,
618 NULs are preserved. In the latter two cases the preprocessor emits a
621 @node The preprocessing language
622 @section The preprocessing language
624 @cindex preprocessing directives
625 @cindex directive line
626 @cindex directive name
628 After tokenization, the stream of tokens may simply be passed straight
629 to the compiler's parser. However, if it contains any operations in the
630 @dfn{preprocessing language}, it will be transformed first. This stage
631 corresponds roughly to the standard's ``translation phase 4'' and is
632 what most people think of as the preprocessor's job.
634 The preprocessing language consists of @dfn{directives} to be executed
635 and @dfn{macros} to be expanded. Its primary capabilities are:
639 Inclusion of header files. These are files of declarations that can be
640 substituted into your program.
643 Macro expansion. You can define @dfn{macros}, which are abbreviations
644 for arbitrary fragments of C code. The preprocessor will replace the
645 macros with their definitions throughout the program. Some macros are
646 automatically defined for you.
649 Conditional compilation. You can include or exclude parts of the
650 program according to various conditions.
653 Line control. If you use a program to combine or rearrange source files
654 into an intermediate file which is then compiled, you can use line
655 control to inform the compiler where each source line originally came
659 Diagnostics. You can detect problems at compile time and issue errors
663 There are a few more, less useful, features.
665 Except for expansion of predefined macros, all these operations are
666 triggered with @dfn{preprocessing directives}. Preprocessing directives
667 are lines in your program that start with @samp{#}. Whitespace is
668 allowed before and after the @samp{#}. The @samp{#} is followed by an
669 identifier, the @dfn{directive name}. It specifies the operation to
670 perform. Directives are commonly referred to as @samp{#@var{name}}
671 where @var{name} is the directive name. For example, @samp{#define} is
672 the directive that defines a macro.
674 The @samp{#} which begins a directive cannot come from a macro
675 expansion. Also, the directive name is not macro expanded. Thus, if
676 @code{foo} is defined as a macro expanding to @code{define}, that does
677 not make @samp{#foo} a valid preprocessing directive.
679 The set of valid directive names is fixed. Programs cannot define new
680 preprocessing directives.
682 Some directives require arguments; these make up the rest of the
683 directive line and must be separated from the directive name by
684 whitespace. For example, @samp{#define} must be followed by a macro
685 name and the intended expansion of the macro.
687 A preprocessing directive cannot cover more than one line. The line
688 may, however, be continued with backslash-newline, or by a block comment
689 which extends past the end of the line. In either case, when the
690 directive is processed, the continuations have already been merged with
691 the first line to make one long line.
694 @chapter Header Files
697 A header file is a file containing C declarations and macro definitions
698 (@pxref{Macros}) to be shared between several source files. You request
699 the use of a header file in your program by @dfn{including} it, with the
700 C preprocessing directive @samp{#include}.
702 Header files serve two purposes.
706 @cindex system header files
707 System header files declare the interfaces to parts of the operating
708 system. You include them in your program to supply the definitions and
709 declarations you need to invoke system calls and libraries.
712 Your own header files contain declarations for interfaces between the
713 source files of your program. Each time you have a group of related
714 declarations and macro definitions all or most of which are needed in
715 several different source files, it is a good idea to create a header
719 Including a header file produces the same results as copying the header
720 file into each source file that needs it. Such copying would be
721 time-consuming and error-prone. With a header file, the related
722 declarations appear in only one place. If they need to be changed, they
723 can be changed in one place, and programs that include the header file
724 will automatically use the new version when next recompiled. The header
725 file eliminates the labor of finding and changing all the copies as well
726 as the risk that a failure to find one copy will result in
727 inconsistencies within a program.
729 In C, the usual convention is to give header files names that end with
730 @file{.h}. It is most portable to use only letters, digits, dashes, and
731 underscores in header file names, and at most one dot.
735 * Include Operation::
737 * Once-Only Headers::
738 * Alternatives to Wrapper #ifndef::
739 * Computed Includes::
745 @section Include Syntax
748 Both user and system header files are included using the preprocessing
749 directive @samp{#include}. It has two variants:
752 @item #include <@var{file}>
753 This variant is used for system header files. It searches for a file
754 named @var{file} in a standard list of system directories. You can prepend
755 directories to this list with the @option{-I} option (@pxref{Invocation}).
757 @item #include "@var{file}"
758 This variant is used for header files of your own program. It
759 searches for a file named @var{file} first in the directory containing
760 the current file, then in the quote directories and then the same
761 directories used for @code{<@var{file}>}. You can prepend directories
762 to the list of quote directories with the @option{-iquote} option.
765 The argument of @samp{#include}, whether delimited with quote marks or
766 angle brackets, behaves like a string constant in that comments are not
767 recognized, and macro names are not expanded. Thus, @code{@w{#include
768 <x/*y>}} specifies inclusion of a system header file named @file{x/*y}.
770 However, if backslashes occur within @var{file}, they are considered
771 ordinary text characters, not escape characters. None of the character
772 escape sequences appropriate to string constants in C are processed.
773 Thus, @code{@w{#include "x\n\\y"}} specifies a filename containing three
774 backslashes. (Some systems interpret @samp{\} as a pathname separator.
775 All of these also interpret @samp{/} the same way. It is most portable
776 to use only @samp{/}.)
778 It is an error if there is anything (other than comments) on the line
781 @node Include Operation
782 @section Include Operation
784 The @samp{#include} directive works by directing the C preprocessor to
785 scan the specified file as input before continuing with the rest of the
786 current file. The output from the preprocessor contains the output
787 already generated, followed by the output resulting from the included
788 file, followed by the output that comes from the text after the
789 @samp{#include} directive. For example, if you have a header file
790 @file{header.h} as follows,
797 and a main program called @file{program.c} that uses the header file,
812 the compiler will see the same token stream as it would if
813 @file{program.c} read
826 Included files are not limited to declarations and macro definitions;
827 those are merely the typical uses. Any fragment of a C program can be
828 included from another file. The include file could even contain the
829 beginning of a statement that is concluded in the containing file, or
830 the end of a statement that was started in the including file. However,
831 an included file must consist of complete tokens. Comments and string
832 literals which have not been closed by the end of an included file are
833 invalid. For error recovery, they are considered to end at the end of
836 To avoid confusion, it is best if header files contain only complete
837 syntactic units---function declarations or definitions, type
840 The line following the @samp{#include} directive is always treated as a
841 separate line by the C preprocessor, even if the included file lacks a
847 GCC looks in several different places for headers. On a normal Unix
848 system, if you do not instruct it otherwise, it will look for headers
849 requested with @code{@w{#include <@var{file}>}} in:
853 @var{libdir}/gcc/@var{target}/@var{version}/include
854 /usr/@var{target}/include
858 For C++ programs, it will also look in @file{/usr/include/g++-v3},
859 first. In the above, @var{target} is the canonical name of the system
860 GCC was configured to compile code for; often but not always the same as
861 the canonical name of the system it runs on. @var{version} is the
862 version of GCC in use.
864 You can add to this list with the @option{-I@var{dir}} command line
865 option. All the directories named by @option{-I} are searched, in
866 left-to-right order, @emph{before} the default directories. The only
867 exception is when @file{dir} is already searched by default. In
868 this case, the option is ignored and the search order for system
869 directories remains unchanged.
871 Duplicate directories are removed from the quote and bracket search
872 chains before the two chains are merged to make the final search chain.
873 Thus, it is possible for a directory to occur twice in the final search
874 chain if it was specified in both the quote and bracket chains.
876 You can prevent GCC from searching any of the default directories with
877 the @option{-nostdinc} option. This is useful when you are compiling an
878 operating system kernel or some other program that does not use the
879 standard C library facilities, or the standard C library itself.
880 @option{-I} options are not ignored as described above when
881 @option{-nostdinc} is in effect.
883 GCC looks for headers requested with @code{@w{#include "@var{file}"}}
884 first in the directory containing the current file, then in the
885 directories as specified by @option{-iquote} options, then in the same
886 places it would have looked for a header requested with angle
887 brackets. For example, if @file{/usr/include/sys/stat.h} contains
888 @code{@w{#include "types.h"}}, GCC looks for @file{types.h} first in
889 @file{/usr/include/sys}, then in its usual search path.
891 @samp{#line} (@pxref{Line Control}) does not change GCC's idea of the
892 directory containing the current file.
894 You may put @option{-I-} at any point in your list of @option{-I} options.
895 This has two effects. First, directories appearing before the
896 @option{-I-} in the list are searched only for headers requested with
897 quote marks. Directories after @option{-I-} are searched for all
898 headers. Second, the directory containing the current file is not
899 searched for anything, unless it happens to be one of the directories
900 named by an @option{-I} switch. @option{-I-} is deprecated, @option{-iquote}
901 should be used instead.
903 @option{-I. -I-} is not the same as no @option{-I} options at all, and does
904 not cause the same behavior for @samp{<>} includes that @samp{""}
905 includes get with no special options. @option{-I.} searches the
906 compiler's current working directory for header files. That may or may
907 not be the same as the directory containing the current file.
909 If you need to look for headers in a directory named @file{-}, write
912 There are several more ways to adjust the header search path. They are
913 generally less useful. @xref{Invocation}.
915 @node Once-Only Headers
916 @section Once-Only Headers
917 @cindex repeated inclusion
918 @cindex including just once
919 @cindex wrapper @code{#ifndef}
921 If a header file happens to be included twice, the compiler will process
922 its contents twice. This is very likely to cause an error, e.g.@: when the
923 compiler sees the same structure definition twice. Even if it does not,
924 it will certainly waste time.
926 The standard way to prevent this is to enclose the entire real contents
927 of the file in a conditional, like this:
932 #ifndef FILE_FOO_SEEN
933 #define FILE_FOO_SEEN
935 @var{the entire file}
937 #endif /* !FILE_FOO_SEEN */
941 This construct is commonly known as a @dfn{wrapper #ifndef}.
942 When the header is included again, the conditional will be false,
943 because @code{FILE_FOO_SEEN} is defined. The preprocessor will skip
944 over the entire contents of the file, and the compiler will not see it
947 CPP optimizes even further. It remembers when a header file has a
948 wrapper @samp{#ifndef}. If a subsequent @samp{#include} specifies that
949 header, and the macro in the @samp{#ifndef} is still defined, it does
950 not bother to rescan the file at all.
952 You can put comments outside the wrapper. They will not interfere with
955 @cindex controlling macro
957 The macro @code{FILE_FOO_SEEN} is called the @dfn{controlling macro} or
958 @dfn{guard macro}. In a user header file, the macro name should not
959 begin with @samp{_}. In a system header file, it should begin with
960 @samp{__} to avoid conflicts with user programs. In any kind of header
961 file, the macro name should contain the name of the file and some
962 additional text, to avoid conflicts with other header files.
964 @node Alternatives to Wrapper #ifndef
965 @section Alternatives to Wrapper #ifndef
967 CPP supports two more ways of indicating that a header file should be
968 read only once. Neither one is as portable as a wrapper @samp{#ifndef}
969 and we recommend you do not use them in new programs, with the caveat
970 that @samp{#import} is standard practice in Objective-C.
973 CPP supports a variant of @samp{#include} called @samp{#import} which
974 includes a file, but does so at most once. If you use @samp{#import}
975 instead of @samp{#include}, then you don't need the conditionals
976 inside the header file to prevent multiple inclusion of the contents.
977 @samp{#import} is standard in Objective-C, but is considered a
978 deprecated extension in C and C++.
980 @samp{#import} is not a well designed feature. It requires the users of
981 a header file to know that it should only be included once. It is much
982 better for the header file's implementor to write the file so that users
983 don't need to know this. Using a wrapper @samp{#ifndef} accomplishes
986 In the present implementation, a single use of @samp{#import} will
987 prevent the file from ever being read again, by either @samp{#import} or
988 @samp{#include}. You should not rely on this; do not use both
989 @samp{#import} and @samp{#include} to refer to the same header file.
991 Another way to prevent a header file from being included more than once
992 is with the @samp{#pragma once} directive. If @samp{#pragma once} is
993 seen when scanning a header file, that file will never be read again, no
996 @samp{#pragma once} does not have the problems that @samp{#import} does,
997 but it is not recognized by all preprocessors, so you cannot rely on it
998 in a portable program.
1000 @node Computed Includes
1001 @section Computed Includes
1002 @cindex computed includes
1003 @cindex macros in include
1005 Sometimes it is necessary to select one of several different header
1006 files to be included into your program. They might specify
1007 configuration parameters to be used on different sorts of operating
1008 systems, for instance. You could do this with a series of conditionals,
1012 # include "system_1.h"
1014 # include "system_2.h"
1020 That rapidly becomes tedious. Instead, the preprocessor offers the
1021 ability to use a macro for the header name. This is called a
1022 @dfn{computed include}. Instead of writing a header name as the direct
1023 argument of @samp{#include}, you simply put a macro name there instead:
1026 #define SYSTEM_H "system_1.h"
1032 @code{SYSTEM_H} will be expanded, and the preprocessor will look for
1033 @file{system_1.h} as if the @samp{#include} had been written that way
1034 originally. @code{SYSTEM_H} could be defined by your Makefile with a
1037 You must be careful when you define the macro. @samp{#define} saves
1038 tokens, not text. The preprocessor has no way of knowing that the macro
1039 will be used as the argument of @samp{#include}, so it generates
1040 ordinary tokens, not a header name. This is unlikely to cause problems
1041 if you use double-quote includes, which are close enough to string
1042 constants. If you use angle brackets, however, you may have trouble.
1044 The syntax of a computed include is actually a bit more general than the
1045 above. If the first non-whitespace character after @samp{#include} is
1046 not @samp{"} or @samp{<}, then the entire line is macro-expanded
1047 like running text would be.
1049 If the line expands to a single string constant, the contents of that
1050 string constant are the file to be included. CPP does not re-examine the
1051 string for embedded quotes, but neither does it process backslash
1052 escapes in the string. Therefore
1055 #define HEADER "a\"b"
1060 looks for a file named @file{a\"b}. CPP searches for the file according
1061 to the rules for double-quoted includes.
1063 If the line expands to a token stream beginning with a @samp{<} token
1064 and including a @samp{>} token, then the tokens between the @samp{<} and
1065 the first @samp{>} are combined to form the filename to be included.
1066 Any whitespace between tokens is reduced to a single space; then any
1067 space after the initial @samp{<} is retained, but a trailing space
1068 before the closing @samp{>} is ignored. CPP searches for the file
1069 according to the rules for angle-bracket includes.
1071 In either case, if there are any tokens on the line after the file name,
1072 an error occurs and the directive is not processed. It is also an error
1073 if the result of expansion does not match either of the two expected
1076 These rules are implementation-defined behavior according to the C
1077 standard. To minimize the risk of different compilers interpreting your
1078 computed includes differently, we recommend you use only a single
1079 object-like macro which expands to a string constant. This will also
1080 minimize confusion for people reading your program.
1082 @node Wrapper Headers
1083 @section Wrapper Headers
1084 @cindex wrapper headers
1085 @cindex overriding a header file
1086 @findex #include_next
1088 Sometimes it is necessary to adjust the contents of a system-provided
1089 header file without editing it directly. GCC's @command{fixincludes}
1090 operation does this, for example. One way to do that would be to create
1091 a new header file with the same name and insert it in the search path
1092 before the original header. That works fine as long as you're willing
1093 to replace the old header entirely. But what if you want to refer to
1094 the old header from the new one?
1096 You cannot simply include the old header with @samp{#include}. That
1097 will start from the beginning, and find your new header again. If your
1098 header is not protected from multiple inclusion (@pxref{Once-Only
1099 Headers}), it will recurse infinitely and cause a fatal error.
1101 You could include the old header with an absolute pathname:
1103 #include "/usr/include/old-header.h"
1106 This works, but is not clean; should the system headers ever move, you
1107 would have to edit the new headers to match.
1109 There is no way to solve this problem within the C standard, but you can
1110 use the GNU extension @samp{#include_next}. It means, ``Include the
1111 @emph{next} file with this name''. This directive works like
1112 @samp{#include} except in searching for the specified file: it starts
1113 searching the list of header file directories @emph{after} the directory
1114 in which the current file was found.
1116 Suppose you specify @option{-I /usr/local/include}, and the list of
1117 directories to search also includes @file{/usr/include}; and suppose
1118 both directories contain @file{signal.h}. Ordinary @code{@w{#include
1119 <signal.h>}} finds the file under @file{/usr/local/include}. If that
1120 file contains @code{@w{#include_next <signal.h>}}, it starts searching
1121 after that directory, and finds the file in @file{/usr/include}.
1123 @samp{#include_next} does not distinguish between @code{<@var{file}>}
1124 and @code{"@var{file}"} inclusion, nor does it check that the file you
1125 specify has the same name as the current file. It simply looks for the
1126 file named, starting with the directory in the search path after the one
1127 where the current file was found.
1129 The use of @samp{#include_next} can lead to great confusion. We
1130 recommend it be used only when there is no other alternative. In
1131 particular, it should not be used in the headers belonging to a specific
1132 program; it should be used only to make global corrections along the
1133 lines of @command{fixincludes}.
1135 @node System Headers
1136 @section System Headers
1137 @cindex system header files
1139 The header files declaring interfaces to the operating system and
1140 runtime libraries often cannot be written in strictly conforming C@.
1141 Therefore, GCC gives code found in @dfn{system headers} special
1142 treatment. All warnings, other than those generated by @samp{#warning}
1143 (@pxref{Diagnostics}), are suppressed while GCC is processing a system
1144 header. Macros defined in a system header are immune to a few warnings
1145 wherever they are expanded. This immunity is granted on an ad-hoc
1146 basis, when we find that a warning generates lots of false positives
1147 because of code in macros defined in system headers.
1149 Normally, only the headers found in specific directories are considered
1150 system headers. These directories are determined when GCC is compiled.
1151 There are, however, two ways to make normal headers into system headers.
1153 The @option{-isystem} command line option adds its argument to the list of
1154 directories to search for headers, just like @option{-I}. Any headers
1155 found in that directory will be considered system headers.
1157 All directories named by @option{-isystem} are searched @emph{after} all
1158 directories named by @option{-I}, no matter what their order was on the
1159 command line. If the same directory is named by both @option{-I} and
1160 @option{-isystem}, the @option{-I} option is ignored. GCC provides an
1161 informative message when this occurs if @option{-v} is used.
1163 @findex #pragma GCC system_header
1164 There is also a directive, @code{@w{#pragma GCC system_header}}, which
1165 tells GCC to consider the rest of the current include file a system
1166 header, no matter where it was found. Code that comes before the
1167 @samp{#pragma} in the file will not be affected. @code{@w{#pragma GCC
1168 system_header}} has no effect in the primary source file.
1170 On very old systems, some of the pre-defined system header directories
1171 get even more special treatment. GNU C++ considers code in headers
1172 found in those directories to be surrounded by an @code{@w{extern "C"}}
1173 block. There is no way to request this behavior with a @samp{#pragma},
1174 or from the command line.
1179 A @dfn{macro} is a fragment of code which has been given a name.
1180 Whenever the name is used, it is replaced by the contents of the macro.
1181 There are two kinds of macros. They differ mostly in what they look
1182 like when they are used. @dfn{Object-like} macros resemble data objects
1183 when used, @dfn{function-like} macros resemble function calls.
1185 You may define any valid identifier as a macro, even if it is a C
1186 keyword. The preprocessor does not know anything about keywords. This
1187 can be useful if you wish to hide a keyword such as @code{const} from an
1188 older compiler that does not understand it. However, the preprocessor
1189 operator @code{defined} (@pxref{Defined}) can never be defined as a
1190 macro, and C++'s named operators (@pxref{C++ Named Operators}) cannot be
1191 macros when you are compiling C++.
1194 * Object-like Macros::
1195 * Function-like Macros::
1200 * Predefined Macros::
1201 * Undefining and Redefining Macros::
1202 * Directives Within Macro Arguments::
1206 @node Object-like Macros
1207 @section Object-like Macros
1208 @cindex object-like macro
1209 @cindex symbolic constants
1210 @cindex manifest constants
1212 An @dfn{object-like macro} is a simple identifier which will be replaced
1213 by a code fragment. It is called object-like because it looks like a
1214 data object in code that uses it. They are most commonly used to give
1215 symbolic names to numeric constants.
1218 You create macros with the @samp{#define} directive. @samp{#define} is
1219 followed by the name of the macro and then the token sequence it should
1220 be an abbreviation for, which is variously referred to as the macro's
1221 @dfn{body}, @dfn{expansion} or @dfn{replacement list}. For example,
1224 #define BUFFER_SIZE 1024
1228 defines a macro named @code{BUFFER_SIZE} as an abbreviation for the
1229 token @code{1024}. If somewhere after this @samp{#define} directive
1230 there comes a C statement of the form
1233 foo = (char *) malloc (BUFFER_SIZE);
1237 then the C preprocessor will recognize and @dfn{expand} the macro
1238 @code{BUFFER_SIZE}. The C compiler will see the same tokens as it would
1242 foo = (char *) malloc (1024);
1245 By convention, macro names are written in uppercase. Programs are
1246 easier to read when it is possible to tell at a glance which names are
1249 The macro's body ends at the end of the @samp{#define} line. You may
1250 continue the definition onto multiple lines, if necessary, using
1251 backslash-newline. When the macro is expanded, however, it will all
1252 come out on one line. For example,
1255 #define NUMBERS 1, \
1258 int x[] = @{ NUMBERS @};
1259 @expansion{} int x[] = @{ 1, 2, 3 @};
1263 The most common visible consequence of this is surprising line numbers
1266 There is no restriction on what can go in a macro body provided it
1267 decomposes into valid preprocessing tokens. Parentheses need not
1268 balance, and the body need not resemble valid C code. (If it does not,
1269 you may get error messages from the C compiler when you use the macro.)
1271 The C preprocessor scans your program sequentially. Macro definitions
1272 take effect at the place you write them. Therefore, the following input
1273 to the C preprocessor
1289 When the preprocessor expands a macro name, the macro's expansion
1290 replaces the macro invocation, then the expansion is examined for more
1291 macros to expand. For example,
1295 #define TABLESIZE BUFSIZE
1296 #define BUFSIZE 1024
1298 @expansion{} BUFSIZE
1304 @code{TABLESIZE} is expanded first to produce @code{BUFSIZE}, then that
1305 macro is expanded to produce the final result, @code{1024}.
1307 Notice that @code{BUFSIZE} was not defined when @code{TABLESIZE} was
1308 defined. The @samp{#define} for @code{TABLESIZE} uses exactly the
1309 expansion you specify---in this case, @code{BUFSIZE}---and does not
1310 check to see whether it too contains macro names. Only when you
1311 @emph{use} @code{TABLESIZE} is the result of its expansion scanned for
1314 This makes a difference if you change the definition of @code{BUFSIZE}
1315 at some point in the source file. @code{TABLESIZE}, defined as shown,
1316 will always expand using the definition of @code{BUFSIZE} that is
1317 currently in effect:
1320 #define BUFSIZE 1020
1321 #define TABLESIZE BUFSIZE
1327 Now @code{TABLESIZE} expands (in two stages) to @code{37}.
1329 If the expansion of a macro contains its own name, either directly or
1330 via intermediate macros, it is not expanded again when the expansion is
1331 examined for more macros. This prevents infinite recursion.
1332 @xref{Self-Referential Macros}, for the precise details.
1334 @node Function-like Macros
1335 @section Function-like Macros
1336 @cindex function-like macros
1338 You can also define macros whose use looks like a function call. These
1339 are called @dfn{function-like macros}. To define a function-like macro,
1340 you use the same @samp{#define} directive, but you put a pair of
1341 parentheses immediately after the macro name. For example,
1344 #define lang_init() c_init()
1346 @expansion{} c_init()
1349 A function-like macro is only expanded if its name appears with a pair
1350 of parentheses after it. If you write just the name, it is left alone.
1351 This can be useful when you have a function and a macro of the same
1352 name, and you wish to use the function sometimes.
1355 extern void foo(void);
1356 #define foo() /* @r{optimized inline version} */
1362 Here the call to @code{foo()} will use the macro, but the function
1363 pointer will get the address of the real function. If the macro were to
1364 be expanded, it would cause a syntax error.
1366 If you put spaces between the macro name and the parentheses in the
1367 macro definition, that does not define a function-like macro, it defines
1368 an object-like macro whose expansion happens to begin with a pair of
1372 #define lang_init () c_init()
1374 @expansion{} () c_init()()
1377 The first two pairs of parentheses in this expansion come from the
1378 macro. The third is the pair that was originally after the macro
1379 invocation. Since @code{lang_init} is an object-like macro, it does not
1380 consume those parentheses.
1382 @node Macro Arguments
1383 @section Macro Arguments
1385 @cindex macros with arguments
1386 @cindex arguments in macro definitions
1388 Function-like macros can take @dfn{arguments}, just like true functions.
1389 To define a macro that uses arguments, you insert @dfn{parameters}
1390 between the pair of parentheses in the macro definition that make the
1391 macro function-like. The parameters must be valid C identifiers,
1392 separated by commas and optionally whitespace.
1394 To invoke a macro that takes arguments, you write the name of the macro
1395 followed by a list of @dfn{actual arguments} in parentheses, separated
1396 by commas. The invocation of the macro need not be restricted to a
1397 single logical line---it can cross as many lines in the source file as
1398 you wish. The number of arguments you give must match the number of
1399 parameters in the macro definition. When the macro is expanded, each
1400 use of a parameter in its body is replaced by the tokens of the
1401 corresponding argument. (You need not use all of the parameters in the
1404 As an example, here is a macro that computes the minimum of two numeric
1405 values, as it is defined in many C programs, and some uses.
1408 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
1409 x = min(a, b); @expansion{} x = ((a) < (b) ? (a) : (b));
1410 y = min(1, 2); @expansion{} y = ((1) < (2) ? (1) : (2));
1411 z = min(a + 28, *p); @expansion{} z = ((a + 28) < (*p) ? (a + 28) : (*p));
1415 (In this small example you can already see several of the dangers of
1416 macro arguments. @xref{Macro Pitfalls}, for detailed explanations.)
1418 Leading and trailing whitespace in each argument is dropped, and all
1419 whitespace between the tokens of an argument is reduced to a single
1420 space. Parentheses within each argument must balance; a comma within
1421 such parentheses does not end the argument. However, there is no
1422 requirement for square brackets or braces to balance, and they do not
1423 prevent a comma from separating arguments. Thus,
1426 macro (array[x = y, x + 1])
1430 passes two arguments to @code{macro}: @code{array[x = y} and @code{x +
1431 1]}. If you want to supply @code{array[x = y, x + 1]} as an argument,
1432 you can write it as @code{array[(x = y, x + 1)]}, which is equivalent C
1435 All arguments to a macro are completely macro-expanded before they are
1436 substituted into the macro body. After substitution, the complete text
1437 is scanned again for macros to expand, including the arguments. This rule
1438 may seem strange, but it is carefully designed so you need not worry
1439 about whether any function call is actually a macro invocation. You can
1440 run into trouble if you try to be too clever, though. @xref{Argument
1441 Prescan}, for detailed discussion.
1443 For example, @code{min (min (a, b), c)} is first expanded to
1446 min (((a) < (b) ? (a) : (b)), (c))
1454 ((((a) < (b) ? (a) : (b))) < (c)
1455 ? (((a) < (b) ? (a) : (b)))
1461 (Line breaks shown here for clarity would not actually be generated.)
1463 @cindex empty macro arguments
1464 You can leave macro arguments empty; this is not an error to the
1465 preprocessor (but many macros will then expand to invalid code).
1466 You cannot leave out arguments entirely; if a macro takes two arguments,
1467 there must be exactly one comma at the top level of its argument list.
1468 Here are some silly examples using @code{min}:
1471 min(, b) @expansion{} (( ) < (b) ? ( ) : (b))
1472 min(a, ) @expansion{} ((a ) < ( ) ? (a ) : ( ))
1473 min(,) @expansion{} (( ) < ( ) ? ( ) : ( ))
1474 min((,),) @expansion{} (((,)) < ( ) ? ((,)) : ( ))
1476 min() @error{} macro "min" requires 2 arguments, but only 1 given
1477 min(,,) @error{} macro "min" passed 3 arguments, but takes just 2
1480 Whitespace is not a preprocessing token, so if a macro @code{foo} takes
1481 one argument, @code{@w{foo ()}} and @code{@w{foo ( )}} both supply it an
1482 empty argument. Previous GNU preprocessor implementations and
1483 documentation were incorrect on this point, insisting that a
1484 function-like macro that takes a single argument be passed a space if an
1485 empty argument was required.
1487 Macro parameters appearing inside string literals are not replaced by
1488 their corresponding actual arguments.
1491 #define foo(x) x, "x"
1492 foo(bar) @expansion{} bar, "x"
1495 @node Stringification
1496 @section Stringification
1497 @cindex stringification
1498 @cindex @samp{#} operator
1500 Sometimes you may want to convert a macro argument into a string
1501 constant. Parameters are not replaced inside string constants, but you
1502 can use the @samp{#} preprocessing operator instead. When a macro
1503 parameter is used with a leading @samp{#}, the preprocessor replaces it
1504 with the literal text of the actual argument, converted to a string
1505 constant. Unlike normal parameter replacement, the argument is not
1506 macro-expanded first. This is called @dfn{stringification}.
1508 There is no way to combine an argument with surrounding text and
1509 stringify it all together. Instead, you can write a series of adjacent
1510 string constants and stringified arguments. The preprocessor will
1511 replace the stringified arguments with string constants. The C
1512 compiler will then combine all the adjacent string constants into one
1515 Here is an example of a macro definition that uses stringification:
1519 #define WARN_IF(EXP) \
1521 fprintf (stderr, "Warning: " #EXP "\n"); @} \
1524 @expansion{} do @{ if (x == 0)
1525 fprintf (stderr, "Warning: " "x == 0" "\n"); @} while (0);
1530 The argument for @code{EXP} is substituted once, as-is, into the
1531 @code{if} statement, and once, stringified, into the argument to
1532 @code{fprintf}. If @code{x} were a macro, it would be expanded in the
1533 @code{if} statement, but not in the string.
1535 The @code{do} and @code{while (0)} are a kludge to make it possible to
1536 write @code{WARN_IF (@var{arg});}, which the resemblance of
1537 @code{WARN_IF} to a function would make C programmers want to do; see
1538 @ref{Swallowing the Semicolon}.
1540 Stringification in C involves more than putting double-quote characters
1541 around the fragment. The preprocessor backslash-escapes the quotes
1542 surrounding embedded string constants, and all backslashes within string and
1543 character constants, in order to get a valid C string constant with the
1544 proper contents. Thus, stringifying @code{@w{p = "foo\n";}} results in
1545 @t{@w{"p = \"foo\\n\";"}}. However, backslashes that are not inside string
1546 or character constants are not duplicated: @samp{\n} by itself
1547 stringifies to @t{"\n"}.
1549 All leading and trailing whitespace in text being stringified is
1550 ignored. Any sequence of whitespace in the middle of the text is
1551 converted to a single space in the stringified result. Comments are
1552 replaced by whitespace long before stringification happens, so they
1553 never appear in stringified text.
1555 There is no way to convert a macro argument into a character constant.
1557 If you want to stringify the result of expansion of a macro argument,
1558 you have to use two levels of macros.
1561 #define xstr(s) str(s)
1567 @expansion{} xstr (4)
1568 @expansion{} str (4)
1572 @code{s} is stringified when it is used in @code{str}, so it is not
1573 macro-expanded first. But @code{s} is an ordinary argument to
1574 @code{xstr}, so it is completely macro-expanded before @code{xstr}
1575 itself is expanded (@pxref{Argument Prescan}). Therefore, by the time
1576 @code{str} gets to its argument, it has already been macro-expanded.
1579 @section Concatenation
1580 @cindex concatenation
1581 @cindex token pasting
1582 @cindex token concatenation
1583 @cindex @samp{##} operator
1585 It is often useful to merge two tokens into one while expanding macros.
1586 This is called @dfn{token pasting} or @dfn{token concatenation}. The
1587 @samp{##} preprocessing operator performs token pasting. When a macro
1588 is expanded, the two tokens on either side of each @samp{##} operator
1589 are combined into a single token, which then replaces the @samp{##} and
1590 the two original tokens in the macro expansion. Usually both will be
1591 identifiers, or one will be an identifier and the other a preprocessing
1592 number. When pasted, they make a longer identifier. This isn't the
1593 only valid case. It is also possible to concatenate two numbers (or a
1594 number and a name, such as @code{1.5} and @code{e3}) into a number.
1595 Also, multi-character operators such as @code{+=} can be formed by
1598 However, two tokens that don't together form a valid token cannot be
1599 pasted together. For example, you cannot concatenate @code{x} with
1600 @code{+} in either order. If you try, the preprocessor issues a warning
1601 and emits the two tokens. Whether it puts white space between the
1602 tokens is undefined. It is common to find unnecessary uses of @samp{##}
1603 in complex macros. If you get this warning, it is likely that you can
1604 simply remove the @samp{##}.
1606 Both the tokens combined by @samp{##} could come from the macro body,
1607 but you could just as well write them as one token in the first place.
1608 Token pasting is most useful when one or both of the tokens comes from a
1609 macro argument. If either of the tokens next to an @samp{##} is a
1610 parameter name, it is replaced by its actual argument before @samp{##}
1611 executes. As with stringification, the actual argument is not
1612 macro-expanded first. If the argument is empty, that @samp{##} has no
1615 Keep in mind that the C preprocessor converts comments to whitespace
1616 before macros are even considered. Therefore, you cannot create a
1617 comment by concatenating @samp{/} and @samp{*}. You can put as much
1618 whitespace between @samp{##} and its operands as you like, including
1619 comments, and you can put comments in arguments that will be
1620 concatenated. However, it is an error if @samp{##} appears at either
1621 end of a macro body.
1623 Consider a C program that interprets named commands. There probably
1624 needs to be a table of commands, perhaps an array of structures declared
1632 void (*function) (void);
1637 struct command commands[] =
1639 @{ "quit", quit_command @},
1640 @{ "help", help_command @},
1646 It would be cleaner not to have to give each command name twice, once in
1647 the string constant and once in the function name. A macro which takes the
1648 name of a command as an argument can make this unnecessary. The string
1649 constant can be created with stringification, and the function name by
1650 concatenating the argument with @samp{_command}. Here is how it is done:
1653 #define COMMAND(NAME) @{ #NAME, NAME ## _command @}
1655 struct command commands[] =
1663 @node Variadic Macros
1664 @section Variadic Macros
1665 @cindex variable number of arguments
1666 @cindex macros with variable arguments
1667 @cindex variadic macros
1669 A macro can be declared to accept a variable number of arguments much as
1670 a function can. The syntax for defining the macro is similar to that of
1671 a function. Here is an example:
1674 #define eprintf(@dots{}) fprintf (stderr, __VA_ARGS__)
1677 This kind of macro is called @dfn{variadic}. When the macro is invoked,
1678 all the tokens in its argument list after the last named argument (this
1679 macro has none), including any commas, become the @dfn{variable
1680 argument}. This sequence of tokens replaces the identifier
1681 @code{@w{__VA_ARGS__}} in the macro body wherever it appears. Thus, we
1682 have this expansion:
1685 eprintf ("%s:%d: ", input_file, lineno)
1686 @expansion{} fprintf (stderr, "%s:%d: ", input_file, lineno)
1689 The variable argument is completely macro-expanded before it is inserted
1690 into the macro expansion, just like an ordinary argument. You may use
1691 the @samp{#} and @samp{##} operators to stringify the variable argument
1692 or to paste its leading or trailing token with another token. (But see
1693 below for an important special case for @samp{##}.)
1695 If your macro is complicated, you may want a more descriptive name for
1696 the variable argument than @code{@w{__VA_ARGS__}}. CPP permits
1697 this, as an extension. You may write an argument name immediately
1698 before the @samp{@dots{}}; that name is used for the variable argument.
1699 The @code{eprintf} macro above could be written
1702 #define eprintf(args@dots{}) fprintf (stderr, args)
1706 using this extension. You cannot use @code{@w{__VA_ARGS__}} and this
1707 extension in the same macro.
1709 You can have named arguments as well as variable arguments in a variadic
1710 macro. We could define @code{eprintf} like this, instead:
1713 #define eprintf(format, @dots{}) fprintf (stderr, format, __VA_ARGS__)
1717 This formulation looks more descriptive, but unfortunately it is less
1718 flexible: you must now supply at least one argument after the format
1719 string. In standard C, you cannot omit the comma separating the named
1720 argument from the variable arguments. Furthermore, if you leave the
1721 variable argument empty, you will get a syntax error, because
1722 there will be an extra comma after the format string.
1725 eprintf("success!\n", );
1726 @expansion{} fprintf(stderr, "success!\n", );
1729 GNU CPP has a pair of extensions which deal with this problem. First,
1730 you are allowed to leave the variable argument out entirely:
1733 eprintf ("success!\n")
1734 @expansion{} fprintf(stderr, "success!\n", );
1738 Second, the @samp{##} token paste operator has a special meaning when
1739 placed between a comma and a variable argument. If you write
1742 #define eprintf(format, @dots{}) fprintf (stderr, format, ##__VA_ARGS__)
1746 and the variable argument is left out when the @code{eprintf} macro is
1747 used, then the comma before the @samp{##} will be deleted. This does
1748 @emph{not} happen if you pass an empty argument, nor does it happen if
1749 the token preceding @samp{##} is anything other than a comma.
1752 eprintf ("success!\n")
1753 @expansion{} fprintf(stderr, "success!\n");
1757 The above explanation is ambiguous about the case where the only macro
1758 parameter is a variable arguments parameter, as it is meaningless to
1759 try to distinguish whether no argument at all is an empty argument or
1760 a missing argument. In this case the C99 standard is clear that the
1761 comma must remain, however the existing GCC extension used to swallow
1762 the comma. So CPP retains the comma when conforming to a specific C
1763 standard, and drops it otherwise.
1765 C99 mandates that the only place the identifier @code{@w{__VA_ARGS__}}
1766 can appear is in the replacement list of a variadic macro. It may not
1767 be used as a macro name, macro argument name, or within a different type
1768 of macro. It may also be forbidden in open text; the standard is
1769 ambiguous. We recommend you avoid using it except for its defined
1772 Variadic macros are a new feature in C99. GNU CPP has supported them
1773 for a long time, but only with a named variable argument
1774 (@samp{args@dots{}}, not @samp{@dots{}} and @code{@w{__VA_ARGS__}}). If you are
1775 concerned with portability to previous versions of GCC, you should use
1776 only named variable arguments. On the other hand, if you are concerned
1777 with portability to other conforming implementations of C99, you should
1778 use only @code{@w{__VA_ARGS__}}.
1780 Previous versions of CPP implemented the comma-deletion extension
1781 much more generally. We have restricted it in this release to minimize
1782 the differences from C99. To get the same effect with both this and
1783 previous versions of GCC, the token preceding the special @samp{##} must
1784 be a comma, and there must be white space between that comma and
1785 whatever comes immediately before it:
1788 #define eprintf(format, args@dots{}) fprintf (stderr, format , ##args)
1792 @xref{Differences from previous versions}, for the gory details.
1794 @node Predefined Macros
1795 @section Predefined Macros
1797 @cindex predefined macros
1798 Several object-like macros are predefined; you use them without
1799 supplying their definitions. They fall into three classes: standard,
1800 common, and system-specific.
1802 In C++, there is a fourth category, the named operators. They act like
1803 predefined macros, but you cannot undefine them.
1806 * Standard Predefined Macros::
1807 * Common Predefined Macros::
1808 * System-specific Predefined Macros::
1809 * C++ Named Operators::
1812 @node Standard Predefined Macros
1813 @subsection Standard Predefined Macros
1814 @cindex standard predefined macros.
1816 The standard predefined macros are specified by the relevant
1817 language standards, so they are available with all compilers that
1818 implement those standards. Older compilers may not provide all of
1819 them. Their names all start with double underscores.
1823 This macro expands to the name of the current input file, in the form of
1824 a C string constant. This is the path by which the preprocessor opened
1825 the file, not the short name specified in @samp{#include} or as the
1826 input file name argument. For example,
1827 @code{"/usr/local/include/myheader.h"} is a possible expansion of this
1831 This macro expands to the current input line number, in the form of a
1832 decimal integer constant. While we call it a predefined macro, it's
1833 a pretty strange macro, since its ``definition'' changes with each
1834 new line of source code.
1837 @code{__FILE__} and @code{__LINE__} are useful in generating an error
1838 message to report an inconsistency detected by the program; the message
1839 can state the source line at which the inconsistency was detected. For
1843 fprintf (stderr, "Internal error: "
1844 "negative string length "
1845 "%d at %s, line %d.",
1846 length, __FILE__, __LINE__);
1849 An @samp{#include} directive changes the expansions of @code{__FILE__}
1850 and @code{__LINE__} to correspond to the included file. At the end of
1851 that file, when processing resumes on the input file that contained
1852 the @samp{#include} directive, the expansions of @code{__FILE__} and
1853 @code{__LINE__} revert to the values they had before the
1854 @samp{#include} (but @code{__LINE__} is then incremented by one as
1855 processing moves to the line after the @samp{#include}).
1857 A @samp{#line} directive changes @code{__LINE__}, and may change
1858 @code{__FILE__} as well. @xref{Line Control}.
1860 C99 introduces @code{__func__}, and GCC has provided @code{__FUNCTION__}
1861 for a long time. Both of these are strings containing the name of the
1862 current function (there are slight semantic differences; see the GCC
1863 manual). Neither of them is a macro; the preprocessor does not know the
1864 name of the current function. They tend to be useful in conjunction
1865 with @code{__FILE__} and @code{__LINE__}, though.
1870 This macro expands to a string constant that describes the date on which
1871 the preprocessor is being run. The string constant contains eleven
1872 characters and looks like @code{@w{"Feb 12 1996"}}. If the day of the
1873 month is less than 10, it is padded with a space on the left.
1875 If GCC cannot determine the current date, it will emit a warning message
1876 (once per compilation) and @code{__DATE__} will expand to
1877 @code{@w{"??? ?? ????"}}.
1880 This macro expands to a string constant that describes the time at
1881 which the preprocessor is being run. The string constant contains
1882 eight characters and looks like @code{"23:59:01"}.
1884 If GCC cannot determine the current time, it will emit a warning message
1885 (once per compilation) and @code{__TIME__} will expand to
1889 In normal operation, this macro expands to the constant 1, to signify
1890 that this compiler conforms to ISO Standard C@. If GNU CPP is used with
1891 a compiler other than GCC, this is not necessarily true; however, the
1892 preprocessor always conforms to the standard unless the
1893 @option{-traditional-cpp} option is used.
1895 This macro is not defined if the @option{-traditional-cpp} option is used.
1897 On some hosts, the system compiler uses a different convention, where
1898 @code{__STDC__} is normally 0, but is 1 if the user specifies strict
1899 conformance to the C Standard. CPP follows the host convention when
1900 processing system header files, but when processing user files
1901 @code{__STDC__} is always 1. This has been reported to cause problems;
1902 for instance, some versions of Solaris provide X Windows headers that
1903 expect @code{__STDC__} to be either undefined or 1. @xref{Invocation}.
1905 @item __STDC_VERSION__
1906 This macro expands to the C Standard's version number, a long integer
1907 constant of the form @code{@var{yyyy}@var{mm}L} where @var{yyyy} and
1908 @var{mm} are the year and month of the Standard version. This signifies
1909 which version of the C Standard the compiler conforms to. Like
1910 @code{__STDC__}, this is not necessarily accurate for the entire
1911 implementation, unless GNU CPP is being used with GCC@.
1913 The value @code{199409L} signifies the 1989 C standard as amended in
1914 1994, which is the current default; the value @code{199901L} signifies
1915 the 1999 revision of the C standard. Support for the 1999 revision is
1918 This macro is not defined if the @option{-traditional-cpp} option is
1919 used, nor when compiling C++ or Objective-C@.
1921 @item __STDC_HOSTED__
1922 This macro is defined, with value 1, if the compiler's target is a
1923 @dfn{hosted environment}. A hosted environment has the complete
1924 facilities of the standard C library available.
1927 This macro is defined when the C++ compiler is in use. You can use
1928 @code{__cplusplus} to test whether a header is compiled by a C compiler
1929 or a C++ compiler. This macro is similar to @code{__STDC_VERSION__}, in
1930 that it expands to a version number. A fully conforming implementation
1931 of the 1998 C++ standard will define this macro to @code{199711L}. The
1932 GNU C++ compiler is not yet fully conforming, so it uses @code{1}
1933 instead. It is hoped to complete the implementation of standard C++
1937 This macro is defined, with value 1, when the Objective-C compiler is in
1938 use. You can use @code{__OBJC__} to test whether a header is compiled
1939 by a C compiler or an Objective-C compiler.
1942 This macro is defined with value 1 when preprocessing assembly
1947 @node Common Predefined Macros
1948 @subsection Common Predefined Macros
1949 @cindex common predefined macros
1951 The common predefined macros are GNU C extensions. They are available
1952 with the same meanings regardless of the machine or operating system on
1953 which you are using GNU C or GNU Fortran. Their names all start with
1959 This macro expands to sequential integral values starting from 0. In
1960 conjunction with the @code{##} operator, this provides a convenient means to
1961 generate unique identifiers. Care must be taken to ensure that
1962 @code{__COUNTER__} is not expanded prior to inclusion of precompiled headers
1963 which use it. Otherwise, the precompiled headers will not be used.
1966 The GNU Fortran compiler defines this.
1969 @itemx __GNUC_MINOR__
1970 @itemx __GNUC_PATCHLEVEL__
1971 These macros are defined by all GNU compilers that use the C
1972 preprocessor: C, C++, Objective-C and Fortran. Their values are the major
1973 version, minor version, and patch level of the compiler, as integer
1974 constants. For example, GCC 3.2.1 will define @code{__GNUC__} to 3,
1975 @code{__GNUC_MINOR__} to 2, and @code{__GNUC_PATCHLEVEL__} to 1. These
1976 macros are also defined if you invoke the preprocessor directly.
1978 @code{__GNUC_PATCHLEVEL__} is new to GCC 3.0; it is also present in the
1979 widely-used development snapshots leading up to 3.0 (which identify
1980 themselves as GCC 2.96 or 2.97, depending on which snapshot you have).
1982 If all you need to know is whether or not your program is being compiled
1983 by GCC, or a non-GCC compiler that claims to accept the GNU C dialects,
1984 you can simply test @code{__GNUC__}. If you need to write code
1985 which depends on a specific version, you must be more careful. Each
1986 time the minor version is increased, the patch level is reset to zero;
1987 each time the major version is increased (which happens rarely), the
1988 minor version and patch level are reset. If you wish to use the
1989 predefined macros directly in the conditional, you will need to write it
1993 /* @r{Test for GCC > 3.2.0} */
1994 #if __GNUC__ > 3 || \
1995 (__GNUC__ == 3 && (__GNUC_MINOR__ > 2 || \
1996 (__GNUC_MINOR__ == 2 && \
1997 __GNUC_PATCHLEVEL__ > 0))
2001 Another approach is to use the predefined macros to
2002 calculate a single number, then compare that against a threshold:
2005 #define GCC_VERSION (__GNUC__ * 10000 \
2006 + __GNUC_MINOR__ * 100 \
2007 + __GNUC_PATCHLEVEL__)
2009 /* @r{Test for GCC > 3.2.0} */
2010 #if GCC_VERSION > 30200
2014 Many people find this form easier to understand.
2017 The GNU C++ compiler defines this. Testing it is equivalent to
2018 testing @code{@w{(__GNUC__ && __cplusplus)}}.
2020 @item __STRICT_ANSI__
2021 GCC defines this macro if and only if the @option{-ansi} switch, or a
2022 @option{-std} switch specifying strict conformance to some version of ISO C,
2023 was specified when GCC was invoked. It is defined to @samp{1}.
2024 This macro exists primarily to direct GNU libc's header files to
2025 restrict their definitions to the minimal set found in the 1989 C
2029 This macro expands to the name of the main input file, in the form
2030 of a C string constant. This is the source file that was specified
2031 on the command line of the preprocessor or C compiler.
2033 @item __INCLUDE_LEVEL__
2034 This macro expands to a decimal integer constant that represents the
2035 depth of nesting in include files. The value of this macro is
2036 incremented on every @samp{#include} directive and decremented at the
2037 end of every included file. It starts out at 0, its value within the
2038 base file specified on the command line.
2041 This macro is defined if the target uses the ELF object format.
2044 This macro expands to a string constant which describes the version of
2045 the compiler in use. You should not rely on its contents having any
2046 particular form, but it can be counted on to contain at least the
2050 @itemx __OPTIMIZE_SIZE__
2051 @itemx __NO_INLINE__
2052 These macros describe the compilation mode. @code{__OPTIMIZE__} is
2053 defined in all optimizing compilations. @code{__OPTIMIZE_SIZE__} is
2054 defined if the compiler is optimizing for size, not speed.
2055 @code{__NO_INLINE__} is defined if no functions will be inlined into
2056 their callers (when not optimizing, or when inlining has been
2057 specifically disabled by @option{-fno-inline}).
2059 These macros cause certain GNU header files to provide optimized
2060 definitions, using macros or inline functions, of system library
2061 functions. You should not use these macros in any way unless you make
2062 sure that programs will execute with the same effect whether or not they
2063 are defined. If they are defined, their value is 1.
2065 @item __GNUC_GNU_INLINE__
2066 GCC defines this macro if functions declared @code{inline} will be
2067 handled in GCC's traditional gnu89 mode. Object files will contain
2068 externally visible definitions of all functions declared @code{inline}
2069 without @code{extern} or @code{static}. They will not contain any
2070 definitions of any functions declared @code{extern inline}.
2072 @item __GNUC_STDC_INLINE__
2073 GCC defines this macro if functions declared @code{inline} will be
2074 handled according to the ISO C99 standard. Object files will contain
2075 externally visible definitions of all functions declared @code{extern
2076 inline}. They will not contain definitions of any functions declared
2077 @code{inline} without @code{extern}.
2079 If this macro is defined, GCC supports the @code{gnu_inline} function
2080 attribute as a way to always get the gnu89 behavior. Support for
2081 this and @code{__GNUC_GNU_INLINE__} was added in GCC 4.1.3. If
2082 neither macro is defined, an older version of GCC is being used:
2083 @code{inline} functions will be compiled in gnu89 mode, and the
2084 @code{gnu_inline} function attribute will not be recognized.
2086 @item __CHAR_UNSIGNED__
2087 GCC defines this macro if and only if the data type @code{char} is
2088 unsigned on the target machine. It exists to cause the standard header
2089 file @file{limits.h} to work correctly. You should not use this macro
2090 yourself; instead, refer to the standard macros defined in @file{limits.h}.
2092 @item __WCHAR_UNSIGNED__
2093 Like @code{__CHAR_UNSIGNED__}, this macro is defined if and only if the
2094 data type @code{wchar_t} is unsigned and the front-end is in C++ mode.
2096 @item __REGISTER_PREFIX__
2097 This macro expands to a single token (not a string constant) which is
2098 the prefix applied to CPU register names in assembly language for this
2099 target. You can use it to write assembly that is usable in multiple
2100 environments. For example, in the @code{m68k-aout} environment it
2101 expands to nothing, but in the @code{m68k-coff} environment it expands
2102 to a single @samp{%}.
2104 @item __USER_LABEL_PREFIX__
2105 This macro expands to a single token which is the prefix applied to
2106 user labels (symbols visible to C code) in assembly. For example, in
2107 the @code{m68k-aout} environment it expands to an @samp{_}, but in the
2108 @code{m68k-coff} environment it expands to nothing.
2110 This macro will have the correct definition even if
2111 @option{-f(no-)underscores} is in use, but it will not be correct if
2112 target-specific options that adjust this prefix are used (e.g.@: the
2113 OSF/rose @option{-mno-underscores} option).
2116 @itemx __PTRDIFF_TYPE__
2117 @itemx __WCHAR_TYPE__
2118 @itemx __WINT_TYPE__
2119 @itemx __INTMAX_TYPE__
2120 @itemx __UINTMAX_TYPE__
2121 @itemx __SIG_ATOMIC_TYPE__
2122 @itemx __INT8_TYPE__
2123 @itemx __INT16_TYPE__
2124 @itemx __INT32_TYPE__
2125 @itemx __INT64_TYPE__
2126 @itemx __UINT8_TYPE__
2127 @itemx __UINT16_TYPE__
2128 @itemx __UINT32_TYPE__
2129 @itemx __UINT64_TYPE__
2130 @itemx __INT_LEAST8_TYPE__
2131 @itemx __INT_LEAST16_TYPE__
2132 @itemx __INT_LEAST32_TYPE__
2133 @itemx __INT_LEAST64_TYPE__
2134 @itemx __UINT_LEAST8_TYPE__
2135 @itemx __UINT_LEAST16_TYPE__
2136 @itemx __UINT_LEAST32_TYPE__
2137 @itemx __UINT_LEAST64_TYPE__
2138 @itemx __INT_FAST8_TYPE__
2139 @itemx __INT_FAST16_TYPE__
2140 @itemx __INT_FAST32_TYPE__
2141 @itemx __INT_FAST64_TYPE__
2142 @itemx __UINT_FAST8_TYPE__
2143 @itemx __UINT_FAST16_TYPE__
2144 @itemx __UINT_FAST32_TYPE__
2145 @itemx __UINT_FAST64_TYPE__
2146 @itemx __INTPTR_TYPE__
2147 @itemx __UINTPTR_TYPE__
2148 These macros are defined to the correct underlying types for the
2149 @code{size_t}, @code{ptrdiff_t}, @code{wchar_t}, @code{wint_t},
2150 @code{intmax_t}, @code{uintmax_t}, @code{sig_atomic_t}, @code{int8_t},
2151 @code{int16_t}, @code{int32_t}, @code{int64_t}, @code{uint8_t},
2152 @code{uint16_t}, @code{uint32_t}, @code{uint64_t},
2153 @code{int_least8_t}, @code{int_least16_t}, @code{int_least32_t},
2154 @code{int_least64_t}, @code{uint_least8_t}, @code{uint_least16_t},
2155 @code{uint_least32_t}, @code{uint_least64_t}, @code{int_fast8_t},
2156 @code{int_fast16_t}, @code{int_fast32_t}, @code{int_fast64_t},
2157 @code{uint_fast8_t}, @code{uint_fast16_t}, @code{uint_fast32_t},
2158 @code{uint_fast64_t}, @code{intptr_t}, and @code{uintptr_t} typedefs,
2159 respectively. They exist to make the standard header files
2160 @file{stddef.h}, @file{stdint.h}, and @file{wchar.h} work correctly.
2161 You should not use these macros directly; instead, include the
2162 appropriate headers and use the typedefs. Some of these macros may
2163 not be defined on particular systems if GCC does not provide a
2164 @file{stdint.h} header on those systems.
2167 Defined to the number of bits used in the representation of the
2168 @code{char} data type. It exists to make the standard header given
2169 numerical limits work correctly. You should not use
2170 this macro directly; instead, include the appropriate headers.
2173 @itemx __WCHAR_MAX__
2177 @itemx __LONG_LONG_MAX__
2180 @itemx __PTRDIFF_MAX__
2181 @itemx __INTMAX_MAX__
2182 @itemx __UINTMAX_MAX__
2183 @itemx __SIG_ATOMIC_MAX__
2185 @itemx __INT16_MAX__
2186 @itemx __INT32_MAX__
2187 @itemx __INT64_MAX__
2188 @itemx __UINT8_MAX__
2189 @itemx __UINT16_MAX__
2190 @itemx __UINT32_MAX__
2191 @itemx __UINT64_MAX__
2192 @itemx __INT_LEAST8_MAX__
2193 @itemx __INT_LEAST16_MAX__
2194 @itemx __INT_LEAST32_MAX__
2195 @itemx __INT_LEAST64_MAX__
2196 @itemx __UINT_LEAST8_MAX__
2197 @itemx __UINT_LEAST16_MAX__
2198 @itemx __UINT_LEAST32_MAX__
2199 @itemx __UINT_LEAST64_MAX__
2200 @itemx __INT_FAST8_MAX__
2201 @itemx __INT_FAST16_MAX__
2202 @itemx __INT_FAST32_MAX__
2203 @itemx __INT_FAST64_MAX__
2204 @itemx __UINT_FAST8_MAX__
2205 @itemx __UINT_FAST16_MAX__
2206 @itemx __UINT_FAST32_MAX__
2207 @itemx __UINT_FAST64_MAX__
2208 @itemx __INTPTR_MAX__
2209 @itemx __UINTPTR_MAX__
2210 @itemx __WCHAR_MIN__
2212 @itemx __SIG_ATOMIC_MIN__
2213 Defined to the maximum value of the @code{signed char}, @code{wchar_t},
2214 @code{signed short},
2215 @code{signed int}, @code{signed long}, @code{signed long long},
2216 @code{wint_t}, @code{size_t}, @code{ptrdiff_t},
2217 @code{intmax_t}, @code{uintmax_t}, @code{sig_atomic_t}, @code{int8_t},
2218 @code{int16_t}, @code{int32_t}, @code{int64_t}, @code{uint8_t},
2219 @code{uint16_t}, @code{uint32_t}, @code{uint64_t},
2220 @code{int_least8_t}, @code{int_least16_t}, @code{int_least32_t},
2221 @code{int_least64_t}, @code{uint_least8_t}, @code{uint_least16_t},
2222 @code{uint_least32_t}, @code{uint_least64_t}, @code{int_fast8_t},
2223 @code{int_fast16_t}, @code{int_fast32_t}, @code{int_fast64_t},
2224 @code{uint_fast8_t}, @code{uint_fast16_t}, @code{uint_fast32_t},
2225 @code{uint_fast64_t}, @code{intptr_t}, and @code{uintptr_t} types and
2226 to the minimum value of the @code{wchar_t}, @code{wint_t}, and
2227 @code{sig_atomic_t} types respectively. They exist to make the
2228 standard header given numerical limits work correctly. You should not
2229 use these macros directly; instead, include the appropriate headers.
2230 Some of these macros may not be defined on particular systems if GCC
2231 does not provide a @file{stdint.h} header on those systems.
2243 Defined to implementations of the standard @file{stdint.h} macros with
2244 the same names without the leading @code{__}. They exist the make the
2245 implementation of that header work correctly. You should not use
2246 these macros directly; instead, include the appropriate headers. Some
2247 of these macros may not be defined on particular systems if GCC does
2248 not provide a @file{stdint.h} header on those systems.
2250 @item __SIZEOF_INT__
2251 @itemx __SIZEOF_LONG__
2252 @itemx __SIZEOF_LONG_LONG__
2253 @itemx __SIZEOF_SHORT__
2254 @itemx __SIZEOF_POINTER__
2255 @itemx __SIZEOF_FLOAT__
2256 @itemx __SIZEOF_DOUBLE__
2257 @itemx __SIZEOF_LONG_DOUBLE__
2258 @itemx __SIZEOF_SIZE_T__
2259 @itemx __SIZEOF_WCHAR_T__
2260 @itemx __SIZEOF_WINT_T__
2261 @itemx __SIZEOF_PTRDIFF_T__
2262 Defined to the number of bytes of the C standard data types: @code{int},
2263 @code{long}, @code{long long}, @code{short}, @code{void *}, @code{float},
2264 @code{double}, @code{long double}, @code{size_t}, @code{wchar_t}, @code{wint_t}
2265 and @code{ptrdiff_t}.
2268 This macro is defined, with value 1, when compiling a C++ source file
2269 with warnings about deprecated constructs enabled. These warnings are
2270 enabled by default, but can be disabled with @option{-Wno-deprecated}.
2273 This macro is defined, with value 1, when compiling a C++ source file
2274 with exceptions enabled. If @option{-fno-exceptions} is used when
2275 compiling the file, then this macro is not defined.
2278 This macro is defined, with value 1, when compiling a C++ source file
2279 with runtime type identification enabled. If @option{-fno-rtti} is
2280 used when compiling the file, then this macro is not defined.
2282 @item __USING_SJLJ_EXCEPTIONS__
2283 This macro is defined, with value 1, if the compiler uses the old
2284 mechanism based on @code{setjmp} and @code{longjmp} for exception
2287 @item __GXX_EXPERIMENTAL_CXX0X__
2288 This macro is defined when compiling a C++ source file with the option
2289 @option{-std=c++0x} or @option{-std=gnu++0x}. It indicates that some
2290 features likely to be included in C++0x are available. Note that these
2291 features are experimental, and may change or be removed in future
2295 This macro is defined when compiling a C++ source file. It has the
2296 value 1 if the compiler will use weak symbols, COMDAT sections, or
2297 other similar techniques to collapse symbols with ``vague linkage''
2298 that are defined in multiple translation units. If the compiler will
2299 not collapse such symbols, this macro is defined with value 0. In
2300 general, user code should not need to make use of this macro; the
2301 purpose of this macro is to ease implementation of the C++ runtime
2302 library provided with G++.
2304 @item __NEXT_RUNTIME__
2305 This macro is defined, with value 1, if (and only if) the NeXT runtime
2306 (as in @option{-fnext-runtime}) is in use for Objective-C@. If the GNU
2307 runtime is used, this macro is not defined, so that you can use this
2308 macro to determine which runtime (NeXT or GNU) is being used.
2312 These macros are defined, with value 1, if (and only if) the compilation
2313 is for a target where @code{long int} and pointer both use 64-bits and
2314 @code{int} uses 32-bit.
2317 This macro is defined, with value 1, when @option{-fstack-protector} is in
2321 This macro is defined, with value 2, when @option{-fstack-protector-all} is
2325 This macro expands to a string constant that describes the date and time
2326 of the last modification of the current source file. The string constant
2327 contains abbreviated day of the week, month, day of the month, time in
2328 hh:mm:ss form, year and looks like @code{@w{"Sun Sep 16 01:03:52 1973"}}.
2329 If the day of the month is less than 10, it is padded with a space on the left.
2331 If GCC cannot determine the current date, it will emit a warning message
2332 (once per compilation) and @code{__TIMESTAMP__} will expand to
2333 @code{@w{"??? ??? ?? ??:??:?? ????"}}.
2335 @item __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1
2336 @itemx __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2
2337 @itemx __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4
2338 @itemx __GCC_HAVE_SYNC_COMPARE_AND_SWAP_8
2339 @itemx __GCC_HAVE_SYNC_COMPARE_AND_SWAP_16
2340 These macros are defined when the target processor supports atomic compare
2341 and swap operations on operands 1, 2, 4, 8 or 16 bytes in length, respectively.
2343 @item __GCC_HAVE_DWARF2_CFI_ASM
2344 This macro is defined when the compiler is emitting Dwarf2 CFI directives
2345 to the assembler. When this is defined, it is possible to emit those same
2346 directives in inline assembly.
2349 @node System-specific Predefined Macros
2350 @subsection System-specific Predefined Macros
2352 @cindex system-specific predefined macros
2353 @cindex predefined macros, system-specific
2354 @cindex reserved namespace
2356 The C preprocessor normally predefines several macros that indicate what
2357 type of system and machine is in use. They are obviously different on
2358 each target supported by GCC@. This manual, being for all systems and
2359 machines, cannot tell you what their names are, but you can use
2360 @command{cpp -dM} to see them all. @xref{Invocation}. All system-specific
2361 predefined macros expand to the constant 1, so you can test them with
2362 either @samp{#ifdef} or @samp{#if}.
2364 The C standard requires that all system-specific macros be part of the
2365 @dfn{reserved namespace}. All names which begin with two underscores,
2366 or an underscore and a capital letter, are reserved for the compiler and
2367 library to use as they wish. However, historically system-specific
2368 macros have had names with no special prefix; for instance, it is common
2369 to find @code{unix} defined on Unix systems. For all such macros, GCC
2370 provides a parallel macro with two underscores added at the beginning
2371 and the end. If @code{unix} is defined, @code{__unix__} will be defined
2372 too. There will never be more than two underscores; the parallel of
2373 @code{_mips} is @code{__mips__}.
2375 When the @option{-ansi} option, or any @option{-std} option that
2376 requests strict conformance, is given to the compiler, all the
2377 system-specific predefined macros outside the reserved namespace are
2378 suppressed. The parallel macros, inside the reserved namespace, remain
2381 We are slowly phasing out all predefined macros which are outside the
2382 reserved namespace. You should never use them in new programs, and we
2383 encourage you to correct older code to use the parallel macros whenever
2384 you find it. We don't recommend you use the system-specific macros that
2385 are in the reserved namespace, either. It is better in the long run to
2386 check specifically for features you need, using a tool such as
2389 @node C++ Named Operators
2390 @subsection C++ Named Operators
2391 @cindex named operators
2392 @cindex C++ named operators
2395 In C++, there are eleven keywords which are simply alternate spellings
2396 of operators normally written with punctuation. These keywords are
2397 treated as such even in the preprocessor. They function as operators in
2398 @samp{#if}, and they cannot be defined as macros or poisoned. In C, you
2399 can request that those keywords take their C++ meaning by including
2400 @file{iso646.h}. That header defines each one as a normal object-like
2401 macro expanding to the appropriate punctuator.
2403 These are the named operators and their corresponding punctuators:
2405 @multitable {Named Operator} {Punctuator}
2406 @item Named Operator @tab Punctuator
2407 @item @code{and} @tab @code{&&}
2408 @item @code{and_eq} @tab @code{&=}
2409 @item @code{bitand} @tab @code{&}
2410 @item @code{bitor} @tab @code{|}
2411 @item @code{compl} @tab @code{~}
2412 @item @code{not} @tab @code{!}
2413 @item @code{not_eq} @tab @code{!=}
2414 @item @code{or} @tab @code{||}
2415 @item @code{or_eq} @tab @code{|=}
2416 @item @code{xor} @tab @code{^}
2417 @item @code{xor_eq} @tab @code{^=}
2420 @node Undefining and Redefining Macros
2421 @section Undefining and Redefining Macros
2422 @cindex undefining macros
2423 @cindex redefining macros
2426 If a macro ceases to be useful, it may be @dfn{undefined} with the
2427 @samp{#undef} directive. @samp{#undef} takes a single argument, the
2428 name of the macro to undefine. You use the bare macro name, even if the
2429 macro is function-like. It is an error if anything appears on the line
2430 after the macro name. @samp{#undef} has no effect if the name is not a
2435 x = FOO; @expansion{} x = 4;
2437 x = FOO; @expansion{} x = FOO;
2440 Once a macro has been undefined, that identifier may be @dfn{redefined}
2441 as a macro by a subsequent @samp{#define} directive. The new definition
2442 need not have any resemblance to the old definition.
2444 However, if an identifier which is currently a macro is redefined, then
2445 the new definition must be @dfn{effectively the same} as the old one.
2446 Two macro definitions are effectively the same if:
2448 @item Both are the same type of macro (object- or function-like).
2449 @item All the tokens of the replacement list are the same.
2450 @item If there are any parameters, they are the same.
2451 @item Whitespace appears in the same places in both. It need not be
2452 exactly the same amount of whitespace, though. Remember that comments
2453 count as whitespace.
2457 These definitions are effectively the same:
2459 #define FOUR (2 + 2)
2460 #define FOUR (2 + 2)
2461 #define FOUR (2 /* @r{two} */ + 2)
2466 #define FOUR (2 + 2)
2467 #define FOUR ( 2+2 )
2468 #define FOUR (2 * 2)
2469 #define FOUR(score,and,seven,years,ago) (2 + 2)
2472 If a macro is redefined with a definition that is not effectively the
2473 same as the old one, the preprocessor issues a warning and changes the
2474 macro to use the new definition. If the new definition is effectively
2475 the same, the redefinition is silently ignored. This allows, for
2476 instance, two different headers to define a common macro. The
2477 preprocessor will only complain if the definitions do not match.
2479 @node Directives Within Macro Arguments
2480 @section Directives Within Macro Arguments
2481 @cindex macro arguments and directives
2483 Occasionally it is convenient to use preprocessor directives within
2484 the arguments of a macro. The C and C++ standards declare that
2485 behavior in these cases is undefined.
2487 Versions of CPP prior to 3.2 would reject such constructs with an
2488 error message. This was the only syntactic difference between normal
2489 functions and function-like macros, so it seemed attractive to remove
2490 this limitation, and people would often be surprised that they could
2491 not use macros in this way. Moreover, sometimes people would use
2492 conditional compilation in the argument list to a normal library
2493 function like @samp{printf}, only to find that after a library upgrade
2494 @samp{printf} had changed to be a function-like macro, and their code
2495 would no longer compile. So from version 3.2 we changed CPP to
2496 successfully process arbitrary directives within macro arguments in
2497 exactly the same way as it would have processed the directive were the
2498 function-like macro invocation not present.
2500 If, within a macro invocation, that macro is redefined, then the new
2501 definition takes effect in time for argument pre-expansion, but the
2502 original definition is still used for argument replacement. Here is a
2503 pathological example:
2521 with the semantics described above.
2523 @node Macro Pitfalls
2524 @section Macro Pitfalls
2525 @cindex problems with macros
2526 @cindex pitfalls of macros
2528 In this section we describe some special rules that apply to macros and
2529 macro expansion, and point out certain cases in which the rules have
2530 counter-intuitive consequences that you must watch out for.
2534 * Operator Precedence Problems::
2535 * Swallowing the Semicolon::
2536 * Duplication of Side Effects::
2537 * Self-Referential Macros::
2538 * Argument Prescan::
2539 * Newlines in Arguments::
2543 @subsection Misnesting
2545 When a macro is called with arguments, the arguments are substituted
2546 into the macro body and the result is checked, together with the rest of
2547 the input file, for more macro calls. It is possible to piece together
2548 a macro call coming partially from the macro body and partially from the
2549 arguments. For example,
2552 #define twice(x) (2*(x))
2553 #define call_with_1(x) x(1)
2555 @expansion{} twice(1)
2556 @expansion{} (2*(1))
2559 Macro definitions do not have to have balanced parentheses. By writing
2560 an unbalanced open parenthesis in a macro body, it is possible to create
2561 a macro call that begins inside the macro body but ends outside of it.
2565 #define strange(file) fprintf (file, "%s %d",
2567 strange(stderr) p, 35)
2568 @expansion{} fprintf (stderr, "%s %d", p, 35)
2571 The ability to piece together a macro call can be useful, but the use of
2572 unbalanced open parentheses in a macro body is just confusing, and
2575 @node Operator Precedence Problems
2576 @subsection Operator Precedence Problems
2577 @cindex parentheses in macro bodies
2579 You may have noticed that in most of the macro definition examples shown
2580 above, each occurrence of a macro argument name had parentheses around
2581 it. In addition, another pair of parentheses usually surround the
2582 entire macro definition. Here is why it is best to write macros that
2585 Suppose you define a macro as follows,
2588 #define ceil_div(x, y) (x + y - 1) / y
2592 whose purpose is to divide, rounding up. (One use for this operation is
2593 to compute how many @code{int} objects are needed to hold a certain
2594 number of @code{char} objects.) Then suppose it is used as follows:
2597 a = ceil_div (b & c, sizeof (int));
2598 @expansion{} a = (b & c + sizeof (int) - 1) / sizeof (int);
2602 This does not do what is intended. The operator-precedence rules of
2603 C make it equivalent to this:
2606 a = (b & (c + sizeof (int) - 1)) / sizeof (int);
2610 What we want is this:
2613 a = ((b & c) + sizeof (int) - 1)) / sizeof (int);
2617 Defining the macro as
2620 #define ceil_div(x, y) ((x) + (y) - 1) / (y)
2624 provides the desired result.
2626 Unintended grouping can result in another way. Consider @code{sizeof
2627 ceil_div(1, 2)}. That has the appearance of a C expression that would
2628 compute the size of the type of @code{ceil_div (1, 2)}, but in fact it
2629 means something very different. Here is what it expands to:
2632 sizeof ((1) + (2) - 1) / (2)
2636 This would take the size of an integer and divide it by two. The
2637 precedence rules have put the division outside the @code{sizeof} when it
2638 was intended to be inside.
2640 Parentheses around the entire macro definition prevent such problems.
2641 Here, then, is the recommended way to define @code{ceil_div}:
2644 #define ceil_div(x, y) (((x) + (y) - 1) / (y))
2647 @node Swallowing the Semicolon
2648 @subsection Swallowing the Semicolon
2649 @cindex semicolons (after macro calls)
2651 Often it is desirable to define a macro that expands into a compound
2652 statement. Consider, for example, the following macro, that advances a
2653 pointer (the argument @code{p} says where to find it) across whitespace
2657 #define SKIP_SPACES(p, limit) \
2658 @{ char *lim = (limit); \
2659 while (p < lim) @{ \
2660 if (*p++ != ' ') @{ \
2665 Here backslash-newline is used to split the macro definition, which must
2666 be a single logical line, so that it resembles the way such code would
2667 be laid out if not part of a macro definition.
2669 A call to this macro might be @code{SKIP_SPACES (p, lim)}. Strictly
2670 speaking, the call expands to a compound statement, which is a complete
2671 statement with no need for a semicolon to end it. However, since it
2672 looks like a function call, it minimizes confusion if you can use it
2673 like a function call, writing a semicolon afterward, as in
2674 @code{SKIP_SPACES (p, lim);}
2676 This can cause trouble before @code{else} statements, because the
2677 semicolon is actually a null statement. Suppose you write
2681 SKIP_SPACES (p, lim);
2686 The presence of two statements---the compound statement and a null
2687 statement---in between the @code{if} condition and the @code{else}
2688 makes invalid C code.
2690 The definition of the macro @code{SKIP_SPACES} can be altered to solve
2691 this problem, using a @code{do @dots{} while} statement. Here is how:
2694 #define SKIP_SPACES(p, limit) \
2695 do @{ char *lim = (limit); \
2696 while (p < lim) @{ \
2697 if (*p++ != ' ') @{ \
2698 p--; break; @}@}@} \
2702 Now @code{SKIP_SPACES (p, lim);} expands into
2705 do @{@dots{}@} while (0);
2709 which is one statement. The loop executes exactly once; most compilers
2710 generate no extra code for it.
2712 @node Duplication of Side Effects
2713 @subsection Duplication of Side Effects
2715 @cindex side effects (in macro arguments)
2716 @cindex unsafe macros
2717 Many C programs define a macro @code{min}, for ``minimum'', like this:
2720 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
2723 When you use this macro with an argument containing a side effect,
2727 next = min (x + y, foo (z));
2731 it expands as follows:
2734 next = ((x + y) < (foo (z)) ? (x + y) : (foo (z)));
2738 where @code{x + y} has been substituted for @code{X} and @code{foo (z)}
2741 The function @code{foo} is used only once in the statement as it appears
2742 in the program, but the expression @code{foo (z)} has been substituted
2743 twice into the macro expansion. As a result, @code{foo} might be called
2744 two times when the statement is executed. If it has side effects or if
2745 it takes a long time to compute, the results might not be what you
2746 intended. We say that @code{min} is an @dfn{unsafe} macro.
2748 The best solution to this problem is to define @code{min} in a way that
2749 computes the value of @code{foo (z)} only once. The C language offers
2750 no standard way to do this, but it can be done with GNU extensions as
2755 (@{ typeof (X) x_ = (X); \
2756 typeof (Y) y_ = (Y); \
2757 (x_ < y_) ? x_ : y_; @})
2760 The @samp{(@{ @dots{} @})} notation produces a compound statement that
2761 acts as an expression. Its value is the value of its last statement.
2762 This permits us to define local variables and assign each argument to
2763 one. The local variables have underscores after their names to reduce
2764 the risk of conflict with an identifier of wider scope (it is impossible
2765 to avoid this entirely). Now each argument is evaluated exactly once.
2767 If you do not wish to use GNU C extensions, the only solution is to be
2768 careful when @emph{using} the macro @code{min}. For example, you can
2769 calculate the value of @code{foo (z)}, save it in a variable, and use
2770 that variable in @code{min}:
2774 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
2778 next = min (x + y, tem);
2784 (where we assume that @code{foo} returns type @code{int}).
2786 @node Self-Referential Macros
2787 @subsection Self-Referential Macros
2788 @cindex self-reference
2790 A @dfn{self-referential} macro is one whose name appears in its
2791 definition. Recall that all macro definitions are rescanned for more
2792 macros to replace. If the self-reference were considered a use of the
2793 macro, it would produce an infinitely large expansion. To prevent this,
2794 the self-reference is not considered a macro call. It is passed into
2795 the preprocessor output unchanged. Consider an example:
2798 #define foo (4 + foo)
2802 where @code{foo} is also a variable in your program.
2804 Following the ordinary rules, each reference to @code{foo} will expand
2805 into @code{(4 + foo)}; then this will be rescanned and will expand into
2806 @code{(4 + (4 + foo))}; and so on until the computer runs out of memory.
2808 The self-reference rule cuts this process short after one step, at
2809 @code{(4 + foo)}. Therefore, this macro definition has the possibly
2810 useful effect of causing the program to add 4 to the value of @code{foo}
2811 wherever @code{foo} is referred to.
2813 In most cases, it is a bad idea to take advantage of this feature. A
2814 person reading the program who sees that @code{foo} is a variable will
2815 not expect that it is a macro as well. The reader will come across the
2816 identifier @code{foo} in the program and think its value should be that
2817 of the variable @code{foo}, whereas in fact the value is four greater.
2819 One common, useful use of self-reference is to create a macro which
2820 expands to itself. If you write
2827 then the macro @code{EPERM} expands to @code{EPERM}. Effectively, it is
2828 left alone by the preprocessor whenever it's used in running text. You
2829 can tell that it's a macro with @samp{#ifdef}. You might do this if you
2830 want to define numeric constants with an @code{enum}, but have
2831 @samp{#ifdef} be true for each constant.
2833 If a macro @code{x} expands to use a macro @code{y}, and the expansion of
2834 @code{y} refers to the macro @code{x}, that is an @dfn{indirect
2835 self-reference} of @code{x}. @code{x} is not expanded in this case
2836 either. Thus, if we have
2844 then @code{x} and @code{y} expand as follows:
2848 x @expansion{} (4 + y)
2849 @expansion{} (4 + (2 * x))
2851 y @expansion{} (2 * x)
2852 @expansion{} (2 * (4 + y))
2857 Each macro is expanded when it appears in the definition of the other
2858 macro, but not when it indirectly appears in its own definition.
2860 @node Argument Prescan
2861 @subsection Argument Prescan
2862 @cindex expansion of arguments
2863 @cindex macro argument expansion
2864 @cindex prescan of macro arguments
2866 Macro arguments are completely macro-expanded before they are
2867 substituted into a macro body, unless they are stringified or pasted
2868 with other tokens. After substitution, the entire macro body, including
2869 the substituted arguments, is scanned again for macros to be expanded.
2870 The result is that the arguments are scanned @emph{twice} to expand
2871 macro calls in them.
2873 Most of the time, this has no effect. If the argument contained any
2874 macro calls, they are expanded during the first scan. The result
2875 therefore contains no macro calls, so the second scan does not change
2876 it. If the argument were substituted as given, with no prescan, the
2877 single remaining scan would find the same macro calls and produce the
2880 You might expect the double scan to change the results when a
2881 self-referential macro is used in an argument of another macro
2882 (@pxref{Self-Referential Macros}): the self-referential macro would be
2883 expanded once in the first scan, and a second time in the second scan.
2884 However, this is not what happens. The self-references that do not
2885 expand in the first scan are marked so that they will not expand in the
2888 You might wonder, ``Why mention the prescan, if it makes no difference?
2889 And why not skip it and make the preprocessor faster?'' The answer is
2890 that the prescan does make a difference in three special cases:
2894 Nested calls to a macro.
2896 We say that @dfn{nested} calls to a macro occur when a macro's argument
2897 contains a call to that very macro. For example, if @code{f} is a macro
2898 that expects one argument, @code{f (f (1))} is a nested pair of calls to
2899 @code{f}. The desired expansion is made by expanding @code{f (1)} and
2900 substituting that into the definition of @code{f}. The prescan causes
2901 the expected result to happen. Without the prescan, @code{f (1)} itself
2902 would be substituted as an argument, and the inner use of @code{f} would
2903 appear during the main scan as an indirect self-reference and would not
2907 Macros that call other macros that stringify or concatenate.
2909 If an argument is stringified or concatenated, the prescan does not
2910 occur. If you @emph{want} to expand a macro, then stringify or
2911 concatenate its expansion, you can do that by causing one macro to call
2912 another macro that does the stringification or concatenation. For
2913 instance, if you have
2916 #define AFTERX(x) X_ ## x
2917 #define XAFTERX(x) AFTERX(x)
2918 #define TABLESIZE 1024
2919 #define BUFSIZE TABLESIZE
2922 then @code{AFTERX(BUFSIZE)} expands to @code{X_BUFSIZE}, and
2923 @code{XAFTERX(BUFSIZE)} expands to @code{X_1024}. (Not to
2924 @code{X_TABLESIZE}. Prescan always does a complete expansion.)
2927 Macros used in arguments, whose expansions contain unshielded commas.
2929 This can cause a macro expanded on the second scan to be called with the
2930 wrong number of arguments. Here is an example:
2934 #define bar(x) lose(x)
2935 #define lose(x) (1 + (x))
2938 We would like @code{bar(foo)} to turn into @code{(1 + (foo))}, which
2939 would then turn into @code{(1 + (a,b))}. Instead, @code{bar(foo)}
2940 expands into @code{lose(a,b)}, and you get an error because @code{lose}
2941 requires a single argument. In this case, the problem is easily solved
2942 by the same parentheses that ought to be used to prevent misnesting of
2943 arithmetic operations:
2948 #define bar(x) lose((x))
2951 The extra pair of parentheses prevents the comma in @code{foo}'s
2952 definition from being interpreted as an argument separator.
2956 @node Newlines in Arguments
2957 @subsection Newlines in Arguments
2958 @cindex newlines in macro arguments
2960 The invocation of a function-like macro can extend over many logical
2961 lines. However, in the present implementation, the entire expansion
2962 comes out on one line. Thus line numbers emitted by the compiler or
2963 debugger refer to the line the invocation started on, which might be
2964 different to the line containing the argument causing the problem.
2966 Here is an example illustrating this:
2969 #define ignore_second_arg(a,b,c) a; c
2971 ignore_second_arg (foo (),
2977 The syntax error triggered by the tokens @code{syntax error} results in
2978 an error message citing line three---the line of ignore_second_arg---
2979 even though the problematic code comes from line five.
2981 We consider this a bug, and intend to fix it in the near future.
2984 @chapter Conditionals
2985 @cindex conditionals
2987 A @dfn{conditional} is a directive that instructs the preprocessor to
2988 select whether or not to include a chunk of code in the final token
2989 stream passed to the compiler. Preprocessor conditionals can test
2990 arithmetic expressions, or whether a name is defined as a macro, or both
2991 simultaneously using the special @code{defined} operator.
2993 A conditional in the C preprocessor resembles in some ways an @code{if}
2994 statement in C, but it is important to understand the difference between
2995 them. The condition in an @code{if} statement is tested during the
2996 execution of your program. Its purpose is to allow your program to
2997 behave differently from run to run, depending on the data it is
2998 operating on. The condition in a preprocessing conditional directive is
2999 tested when your program is compiled. Its purpose is to allow different
3000 code to be included in the program depending on the situation at the
3001 time of compilation.
3003 However, the distinction is becoming less clear. Modern compilers often
3004 do test @code{if} statements when a program is compiled, if their
3005 conditions are known not to vary at run time, and eliminate code which
3006 can never be executed. If you can count on your compiler to do this,
3007 you may find that your program is more readable if you use @code{if}
3008 statements with constant conditions (perhaps determined by macros). Of
3009 course, you can only use this to exclude code, not type definitions or
3010 other preprocessing directives, and you can only do it if the code
3011 remains syntactically valid when it is not to be used.
3013 GCC version 3 eliminates this kind of never-executed code even when
3014 not optimizing. Older versions did it only when optimizing.
3017 * Conditional Uses::
3018 * Conditional Syntax::
3022 @node Conditional Uses
3023 @section Conditional Uses
3025 There are three general reasons to use a conditional.
3029 A program may need to use different code depending on the machine or
3030 operating system it is to run on. In some cases the code for one
3031 operating system may be erroneous on another operating system; for
3032 example, it might refer to data types or constants that do not exist on
3033 the other system. When this happens, it is not enough to avoid
3034 executing the invalid code. Its mere presence will cause the compiler
3035 to reject the program. With a preprocessing conditional, the offending
3036 code can be effectively excised from the program when it is not valid.
3039 You may want to be able to compile the same source file into two
3040 different programs. One version might make frequent time-consuming
3041 consistency checks on its intermediate data, or print the values of
3042 those data for debugging, and the other not.
3045 A conditional whose condition is always false is one way to exclude code
3046 from the program but keep it as a sort of comment for future reference.
3049 Simple programs that do not need system-specific logic or complex
3050 debugging hooks generally will not need to use preprocessing
3053 @node Conditional Syntax
3054 @section Conditional Syntax
3057 A conditional in the C preprocessor begins with a @dfn{conditional
3058 directive}: @samp{#if}, @samp{#ifdef} or @samp{#ifndef}.
3073 The simplest sort of conditional is
3079 @var{controlled text}
3081 #endif /* @var{MACRO} */
3085 @cindex conditional group
3086 This block is called a @dfn{conditional group}. @var{controlled text}
3087 will be included in the output of the preprocessor if and only if
3088 @var{MACRO} is defined. We say that the conditional @dfn{succeeds} if
3089 @var{MACRO} is defined, @dfn{fails} if it is not.
3091 The @var{controlled text} inside of a conditional can include
3092 preprocessing directives. They are executed only if the conditional
3093 succeeds. You can nest conditional groups inside other conditional
3094 groups, but they must be completely nested. In other words,
3095 @samp{#endif} always matches the nearest @samp{#ifdef} (or
3096 @samp{#ifndef}, or @samp{#if}). Also, you cannot start a conditional
3097 group in one file and end it in another.
3099 Even if a conditional fails, the @var{controlled text} inside it is
3100 still run through initial transformations and tokenization. Therefore,
3101 it must all be lexically valid C@. Normally the only way this matters is
3102 that all comments and string literals inside a failing conditional group
3103 must still be properly ended.
3105 The comment following the @samp{#endif} is not required, but it is a
3106 good practice if there is a lot of @var{controlled text}, because it
3107 helps people match the @samp{#endif} to the corresponding @samp{#ifdef}.
3108 Older programs sometimes put @var{MACRO} directly after the
3109 @samp{#endif} without enclosing it in a comment. This is invalid code
3110 according to the C standard. CPP accepts it with a warning. It
3111 never affects which @samp{#ifndef} the @samp{#endif} matches.
3114 Sometimes you wish to use some code if a macro is @emph{not} defined.
3115 You can do this by writing @samp{#ifndef} instead of @samp{#ifdef}.
3116 One common use of @samp{#ifndef} is to include code only the first
3117 time a header file is included. @xref{Once-Only Headers}.
3119 Macro definitions can vary between compilations for several reasons.
3120 Here are some samples.
3124 Some macros are predefined on each kind of machine
3125 (@pxref{System-specific Predefined Macros}). This allows you to provide
3126 code specially tuned for a particular machine.
3129 System header files define more macros, associated with the features
3130 they implement. You can test these macros with conditionals to avoid
3131 using a system feature on a machine where it is not implemented.
3134 Macros can be defined or undefined with the @option{-D} and @option{-U}
3135 command line options when you compile the program. You can arrange to
3136 compile the same source file into two different programs by choosing a
3137 macro name to specify which program you want, writing conditionals to
3138 test whether or how this macro is defined, and then controlling the
3139 state of the macro with command line options, perhaps set in the
3140 Makefile. @xref{Invocation}.
3143 Your program might have a special header file (often called
3144 @file{config.h}) that is adjusted when the program is compiled. It can
3145 define or not define macros depending on the features of the system and
3146 the desired capabilities of the program. The adjustment can be
3147 automated by a tool such as @command{autoconf}, or done by hand.
3153 The @samp{#if} directive allows you to test the value of an arithmetic
3154 expression, rather than the mere existence of one macro. Its syntax is
3158 #if @var{expression}
3160 @var{controlled text}
3162 #endif /* @var{expression} */
3166 @var{expression} is a C expression of integer type, subject to stringent
3167 restrictions. It may contain
3174 Character constants, which are interpreted as they would be in normal
3178 Arithmetic operators for addition, subtraction, multiplication,
3179 division, bitwise operations, shifts, comparisons, and logical
3180 operations (@code{&&} and @code{||}). The latter two obey the usual
3181 short-circuiting rules of standard C@.
3184 Macros. All macros in the expression are expanded before actual
3185 computation of the expression's value begins.
3188 Uses of the @code{defined} operator, which lets you check whether macros
3189 are defined in the middle of an @samp{#if}.
3192 Identifiers that are not macros, which are all considered to be the
3193 number zero. This allows you to write @code{@w{#if MACRO}} instead of
3194 @code{@w{#ifdef MACRO}}, if you know that MACRO, when defined, will
3195 always have a nonzero value. Function-like macros used without their
3196 function call parentheses are also treated as zero.
3198 In some contexts this shortcut is undesirable. The @option{-Wundef}
3199 option causes GCC to warn whenever it encounters an identifier which is
3200 not a macro in an @samp{#if}.
3203 The preprocessor does not know anything about types in the language.
3204 Therefore, @code{sizeof} operators are not recognized in @samp{#if}, and
3205 neither are @code{enum} constants. They will be taken as identifiers
3206 which are not macros, and replaced by zero. In the case of
3207 @code{sizeof}, this is likely to cause the expression to be invalid.
3209 The preprocessor calculates the value of @var{expression}. It carries
3210 out all calculations in the widest integer type known to the compiler;
3211 on most machines supported by GCC this is 64 bits. This is not the same
3212 rule as the compiler uses to calculate the value of a constant
3213 expression, and may give different results in some cases. If the value
3214 comes out to be nonzero, the @samp{#if} succeeds and the @var{controlled
3215 text} is included; otherwise it is skipped.
3220 @cindex @code{defined}
3221 The special operator @code{defined} is used in @samp{#if} and
3222 @samp{#elif} expressions to test whether a certain name is defined as a
3223 macro. @code{defined @var{name}} and @code{defined (@var{name})} are
3224 both expressions whose value is 1 if @var{name} is defined as a macro at
3225 the current point in the program, and 0 otherwise. Thus, @code{@w{#if
3226 defined MACRO}} is precisely equivalent to @code{@w{#ifdef MACRO}}.
3228 @code{defined} is useful when you wish to test more than one macro for
3229 existence at once. For example,
3232 #if defined (__vax__) || defined (__ns16000__)
3236 would succeed if either of the names @code{__vax__} or
3237 @code{__ns16000__} is defined as a macro.
3239 Conditionals written like this:
3242 #if defined BUFSIZE && BUFSIZE >= 1024
3246 can generally be simplified to just @code{@w{#if BUFSIZE >= 1024}},
3247 since if @code{BUFSIZE} is not defined, it will be interpreted as having
3250 If the @code{defined} operator appears as a result of a macro expansion,
3251 the C standard says the behavior is undefined. GNU cpp treats it as a
3252 genuine @code{defined} operator and evaluates it normally. It will warn
3253 wherever your code uses this feature if you use the command-line option
3254 @option{-pedantic}, since other compilers may handle it differently.
3260 The @samp{#else} directive can be added to a conditional to provide
3261 alternative text to be used if the condition fails. This is what it
3266 #if @var{expression}
3268 #else /* Not @var{expression} */
3270 #endif /* Not @var{expression} */
3275 If @var{expression} is nonzero, the @var{text-if-true} is included and
3276 the @var{text-if-false} is skipped. If @var{expression} is zero, the
3279 You can use @samp{#else} with @samp{#ifdef} and @samp{#ifndef}, too.
3285 One common case of nested conditionals is used to check for more than two
3286 possible alternatives. For example, you might have
3300 Another conditional directive, @samp{#elif}, allows this to be
3301 abbreviated as follows:
3308 #else /* X != 2 and X != 1*/
3310 #endif /* X != 2 and X != 1*/
3313 @samp{#elif} stands for ``else if''. Like @samp{#else}, it goes in the
3314 middle of a conditional group and subdivides it; it does not require a
3315 matching @samp{#endif} of its own. Like @samp{#if}, the @samp{#elif}
3316 directive includes an expression to be tested. The text following the
3317 @samp{#elif} is processed only if the original @samp{#if}-condition
3318 failed and the @samp{#elif} condition succeeds.
3320 More than one @samp{#elif} can go in the same conditional group. Then
3321 the text after each @samp{#elif} is processed only if the @samp{#elif}
3322 condition succeeds after the original @samp{#if} and all previous
3323 @samp{#elif} directives within it have failed.
3325 @samp{#else} is allowed after any number of @samp{#elif} directives, but
3326 @samp{#elif} may not follow @samp{#else}.
3329 @section Deleted Code
3330 @cindex commenting out code
3332 If you replace or delete a part of the program but want to keep the old
3333 code around for future reference, you often cannot simply comment it
3334 out. Block comments do not nest, so the first comment inside the old
3335 code will end the commenting-out. The probable result is a flood of
3338 One way to avoid this problem is to use an always-false conditional
3339 instead. For instance, put @code{#if 0} before the deleted code and
3340 @code{#endif} after it. This works even if the code being turned
3341 off contains conditionals, but they must be entire conditionals
3342 (balanced @samp{#if} and @samp{#endif}).
3344 Some people use @code{#ifdef notdef} instead. This is risky, because
3345 @code{notdef} might be accidentally defined as a macro, and then the
3346 conditional would succeed. @code{#if 0} can be counted on to fail.
3348 Do not use @code{#if 0} for comments which are not C code. Use a real
3349 comment, instead. The interior of @code{#if 0} must consist of complete
3350 tokens; in particular, single-quote characters must balance. Comments
3351 often contain unbalanced single-quote characters (known in English as
3352 apostrophes). These confuse @code{#if 0}. They don't confuse
3356 @chapter Diagnostics
3358 @cindex reporting errors
3359 @cindex reporting warnings
3362 The directive @samp{#error} causes the preprocessor to report a fatal
3363 error. The tokens forming the rest of the line following @samp{#error}
3364 are used as the error message.
3366 You would use @samp{#error} inside of a conditional that detects a
3367 combination of parameters which you know the program does not properly
3368 support. For example, if you know that the program will not run
3369 properly on a VAX, you might write
3374 #error "Won't work on VAXen. See comments at get_last_object."
3379 If you have several configuration parameters that must be set up by
3380 the installation in a consistent way, you can use conditionals to detect
3381 an inconsistency and report it with @samp{#error}. For example,
3384 #if !defined(UNALIGNED_INT_ASM_OP) && defined(DWARF2_DEBUGGING_INFO)
3385 #error "DWARF2_DEBUGGING_INFO requires UNALIGNED_INT_ASM_OP."
3390 The directive @samp{#warning} is like @samp{#error}, but causes the
3391 preprocessor to issue a warning and continue preprocessing. The tokens
3392 following @samp{#warning} are used as the warning message.
3394 You might use @samp{#warning} in obsolete header files, with a message
3395 directing the user to the header file which should be used instead.
3397 Neither @samp{#error} nor @samp{#warning} macro-expands its argument.
3398 Internal whitespace sequences are each replaced with a single space.
3399 The line must consist of complete tokens. It is wisest to make the
3400 argument of these directives be a single string constant; this avoids
3401 problems with apostrophes and the like.
3404 @chapter Line Control
3405 @cindex line control
3407 The C preprocessor informs the C compiler of the location in your source
3408 code where each token came from. Presently, this is just the file name
3409 and line number. All the tokens resulting from macro expansion are
3410 reported as having appeared on the line of the source file where the
3411 outermost macro was used. We intend to be more accurate in the future.
3413 If you write a program which generates source code, such as the
3414 @command{bison} parser generator, you may want to adjust the preprocessor's
3415 notion of the current file name and line number by hand. Parts of the
3416 output from @command{bison} are generated from scratch, other parts come
3417 from a standard parser file. The rest are copied verbatim from
3418 @command{bison}'s input. You would like compiler error messages and
3419 symbolic debuggers to be able to refer to @code{bison}'s input file.
3422 @command{bison} or any such program can arrange this by writing
3423 @samp{#line} directives into the output file. @samp{#line} is a
3424 directive that specifies the original line number and source file name
3425 for subsequent input in the current preprocessor input file.
3426 @samp{#line} has three variants:
3429 @item #line @var{linenum}
3430 @var{linenum} is a non-negative decimal integer constant. It specifies
3431 the line number which should be reported for the following line of
3432 input. Subsequent lines are counted from @var{linenum}.
3434 @item #line @var{linenum} @var{filename}
3435 @var{linenum} is the same as for the first form, and has the same
3436 effect. In addition, @var{filename} is a string constant. The
3437 following line and all subsequent lines are reported to come from the
3438 file it specifies, until something else happens to change that.
3439 @var{filename} is interpreted according to the normal rules for a string
3440 constant: backslash escapes are interpreted. This is different from
3443 Previous versions of CPP did not interpret escapes in @samp{#line};
3444 we have changed it because the standard requires they be interpreted,
3445 and most other compilers do.
3447 @item #line @var{anything else}
3448 @var{anything else} is checked for macro calls, which are expanded.
3449 The result should match one of the above two forms.
3452 @samp{#line} directives alter the results of the @code{__FILE__} and
3453 @code{__LINE__} predefined macros from that point on. @xref{Standard
3454 Predefined Macros}. They do not have any effect on @samp{#include}'s
3455 idea of the directory containing the current file. This is a change
3456 from GCC 2.95. Previously, a file reading
3459 #line 1 "../src/gram.y"
3463 would search for @file{gram.h} in @file{../src}, then the @option{-I}
3464 chain; the directory containing the physical source file would not be
3465 searched. In GCC 3.0 and later, the @samp{#include} is not affected by
3466 the presence of a @samp{#line} referring to a different directory.
3468 We made this change because the old behavior caused problems when
3469 generated source files were transported between machines. For instance,
3470 it is common practice to ship generated parsers with a source release,
3471 so that people building the distribution do not need to have yacc or
3472 Bison installed. These files frequently have @samp{#line} directives
3473 referring to the directory tree of the system where the distribution was
3474 created. If GCC tries to search for headers in those directories, the
3475 build is likely to fail.
3477 The new behavior can cause failures too, if the generated file is not
3478 in the same directory as its source and it attempts to include a header
3479 which would be visible searching from the directory containing the
3480 source file. However, this problem is easily solved with an additional
3481 @option{-I} switch on the command line. The failures caused by the old
3482 semantics could sometimes be corrected only by editing the generated
3483 files, which is difficult and error-prone.
3488 The @samp{#pragma} directive is the method specified by the C standard
3489 for providing additional information to the compiler, beyond what is
3490 conveyed in the language itself. Three forms of this directive
3491 (commonly known as @dfn{pragmas}) are specified by the 1999 C standard.
3492 A C compiler is free to attach any meaning it likes to other pragmas.
3494 GCC has historically preferred to use extensions to the syntax of the
3495 language, such as @code{__attribute__}, for this purpose. However, GCC
3496 does define a few pragmas of its own. These mostly have effects on the
3497 entire translation unit or source file.
3499 In GCC version 3, all GNU-defined, supported pragmas have been given a
3500 @code{GCC} prefix. This is in line with the @code{STDC} prefix on all
3501 pragmas defined by C99. For backward compatibility, pragmas which were
3502 recognized by previous versions are still recognized without the
3503 @code{GCC} prefix, but that usage is deprecated. Some older pragmas are
3504 deprecated in their entirety. They are not recognized with the
3505 @code{GCC} prefix. @xref{Obsolete Features}.
3507 @cindex @code{_Pragma}
3508 C99 introduces the @code{@w{_Pragma}} operator. This feature addresses a
3509 major problem with @samp{#pragma}: being a directive, it cannot be
3510 produced as the result of macro expansion. @code{@w{_Pragma}} is an
3511 operator, much like @code{sizeof} or @code{defined}, and can be embedded
3514 Its syntax is @code{@w{_Pragma (@var{string-literal})}}, where
3515 @var{string-literal} can be either a normal or wide-character string
3516 literal. It is destringized, by replacing all @samp{\\} with a single
3517 @samp{\} and all @samp{\"} with a @samp{"}. The result is then
3518 processed as if it had appeared as the right hand side of a
3519 @samp{#pragma} directive. For example,
3522 _Pragma ("GCC dependency \"parse.y\"")
3526 has the same effect as @code{#pragma GCC dependency "parse.y"}. The
3527 same effect could be achieved using macros, for example
3530 #define DO_PRAGMA(x) _Pragma (#x)
3531 DO_PRAGMA (GCC dependency "parse.y")
3534 The standard is unclear on where a @code{_Pragma} operator can appear.
3535 The preprocessor does not accept it within a preprocessing conditional
3536 directive like @samp{#if}. To be safe, you are probably best keeping it
3537 out of directives other than @samp{#define}, and putting it on a line of
3540 This manual documents the pragmas which are meaningful to the
3541 preprocessor itself. Other pragmas are meaningful to the C or C++
3542 compilers. They are documented in the GCC manual.
3545 @item #pragma GCC dependency
3546 @code{#pragma GCC dependency} allows you to check the relative dates of
3547 the current file and another file. If the other file is more recent than
3548 the current file, a warning is issued. This is useful if the current
3549 file is derived from the other file, and should be regenerated. The
3550 other file is searched for using the normal include search path.
3551 Optional trailing text can be used to give more information in the
3555 #pragma GCC dependency "parse.y"
3556 #pragma GCC dependency "/usr/include/time.h" rerun fixincludes
3559 @item #pragma GCC poison
3560 Sometimes, there is an identifier that you want to remove completely
3561 from your program, and make sure that it never creeps back in. To
3562 enforce this, you can @dfn{poison} the identifier with this pragma.
3563 @code{#pragma GCC poison} is followed by a list of identifiers to
3564 poison. If any of those identifiers appears anywhere in the source
3565 after the directive, it is a hard error. For example,
3568 #pragma GCC poison printf sprintf fprintf
3569 sprintf(some_string, "hello");
3573 will produce an error.
3575 If a poisoned identifier appears as part of the expansion of a macro
3576 which was defined before the identifier was poisoned, it will @emph{not}
3577 cause an error. This lets you poison an identifier without worrying
3578 about system headers defining macros that use it.
3583 #define strrchr rindex
3584 #pragma GCC poison rindex
3585 strrchr(some_string, 'h');
3589 will not produce an error.
3591 @item #pragma GCC system_header
3592 This pragma takes no arguments. It causes the rest of the code in the
3593 current file to be treated as if it came from a system header.
3594 @xref{System Headers}.
3598 @node Other Directives
3599 @chapter Other Directives
3603 The @samp{#ident} directive takes one argument, a string constant. On
3604 some systems, that string constant is copied into a special segment of
3605 the object file. On other systems, the directive is ignored. The
3606 @samp{#sccs} directive is a synonym for @samp{#ident}.
3608 These directives are not part of the C standard, but they are not
3609 official GNU extensions either. What historical information we have
3610 been able to find, suggests they originated with System V@.
3612 Both @samp{#ident} and @samp{#sccs} are deprecated extensions.
3614 @cindex null directive
3615 The @dfn{null directive} consists of a @samp{#} followed by a newline,
3616 with only whitespace (including comments) in between. A null directive
3617 is understood as a preprocessing directive but has no effect on the
3618 preprocessor output. The primary significance of the existence of the
3619 null directive is that an input line consisting of just a @samp{#} will
3620 produce no output, rather than a line of output containing just a
3621 @samp{#}. Supposedly some old C programs contain such lines.
3623 @node Preprocessor Output
3624 @chapter Preprocessor Output
3626 When the C preprocessor is used with the C, C++, or Objective-C
3627 compilers, it is integrated into the compiler and communicates a stream
3628 of binary tokens directly to the compiler's parser. However, it can
3629 also be used in the more conventional standalone mode, where it produces
3631 @c FIXME: Document the library interface.
3633 @cindex output format
3634 The output from the C preprocessor looks much like the input, except
3635 that all preprocessing directive lines have been replaced with blank
3636 lines and all comments with spaces. Long runs of blank lines are
3639 The ISO standard specifies that it is implementation defined whether a
3640 preprocessor preserves whitespace between tokens, or replaces it with
3641 e.g.@: a single space. In GNU CPP, whitespace between tokens is collapsed
3642 to become a single space, with the exception that the first token on a
3643 non-directive line is preceded with sufficient spaces that it appears in
3644 the same column in the preprocessed output that it appeared in the
3645 original source file. This is so the output is easy to read.
3646 @xref{Differences from previous versions}. CPP does not insert any
3647 whitespace where there was none in the original source, except where
3648 necessary to prevent an accidental token paste.
3651 Source file name and line number information is conveyed by lines
3655 # @var{linenum} @var{filename} @var{flags}
3659 These are called @dfn{linemarkers}. They are inserted as needed into
3660 the output (but never within a string or character constant). They mean
3661 that the following line originated in file @var{filename} at line
3662 @var{linenum}. @var{filename} will never contain any non-printing
3663 characters; they are replaced with octal escape sequences.
3665 After the file name comes zero or more flags, which are @samp{1},
3666 @samp{2}, @samp{3}, or @samp{4}. If there are multiple flags, spaces
3667 separate them. Here is what the flags mean:
3671 This indicates the start of a new file.
3673 This indicates returning to a file (after having included another file).
3675 This indicates that the following text comes from a system header file,
3676 so certain warnings should be suppressed.
3678 This indicates that the following text should be treated as being
3679 wrapped in an implicit @code{extern "C"} block.
3680 @c maybe cross reference NO_IMPLICIT_EXTERN_C
3683 As an extension, the preprocessor accepts linemarkers in non-assembler
3684 input files. They are treated like the corresponding @samp{#line}
3685 directive, (@pxref{Line Control}), except that trailing flags are
3686 permitted, and are interpreted with the meanings described above. If
3687 multiple flags are given, they must be in ascending order.
3689 Some directives may be duplicated in the output of the preprocessor.
3690 These are @samp{#ident} (always), @samp{#pragma} (only if the
3691 preprocessor does not handle the pragma itself), and @samp{#define} and
3692 @samp{#undef} (with certain debugging options). If this happens, the
3693 @samp{#} of the directive will always be in the first column, and there
3694 will be no space between the @samp{#} and the directive name. If macro
3695 expansion happens to generate tokens which might be mistaken for a
3696 duplicated directive, a space will be inserted between the @samp{#} and
3699 @node Traditional Mode
3700 @chapter Traditional Mode
3702 Traditional (pre-standard) C preprocessing is rather different from
3703 the preprocessing specified by the standard. When GCC is given the
3704 @option{-traditional-cpp} option, it attempts to emulate a traditional
3707 GCC versions 3.2 and later only support traditional mode semantics in
3708 the preprocessor, and not in the compiler front ends. This chapter
3709 outlines the traditional preprocessor semantics we implemented.
3711 The implementation does not correspond precisely to the behavior of
3712 earlier versions of GCC, nor to any true traditional preprocessor.
3713 After all, inconsistencies among traditional implementations were a
3714 major motivation for C standardization. However, we intend that it
3715 should be compatible with true traditional preprocessors in all ways
3716 that actually matter.
3719 * Traditional lexical analysis::
3720 * Traditional macros::
3721 * Traditional miscellany::
3722 * Traditional warnings::
3725 @node Traditional lexical analysis
3726 @section Traditional lexical analysis
3728 The traditional preprocessor does not decompose its input into tokens
3729 the same way a standards-conforming preprocessor does. The input is
3730 simply treated as a stream of text with minimal internal form.
3732 This implementation does not treat trigraphs (@pxref{trigraphs})
3733 specially since they were an invention of the standards committee. It
3734 handles arbitrarily-positioned escaped newlines properly and splices
3735 the lines as you would expect; many traditional preprocessors did not
3738 The form of horizontal whitespace in the input file is preserved in
3739 the output. In particular, hard tabs remain hard tabs. This can be
3740 useful if, for example, you are preprocessing a Makefile.
3742 Traditional CPP only recognizes C-style block comments, and treats the
3743 @samp{/*} sequence as introducing a comment only if it lies outside
3744 quoted text. Quoted text is introduced by the usual single and double
3745 quotes, and also by an initial @samp{<} in a @code{#include}
3748 Traditionally, comments are completely removed and are not replaced
3749 with a space. Since a traditional compiler does its own tokenization
3750 of the output of the preprocessor, this means that comments can
3751 effectively be used as token paste operators. However, comments
3752 behave like separators for text handled by the preprocessor itself,
3753 since it doesn't re-lex its input. For example, in
3760 @samp{foo} and @samp{bar} are distinct identifiers and expanded
3761 separately if they happen to be macros. In other words, this
3762 directive is equivalent to
3775 Generally speaking, in traditional mode an opening quote need not have
3776 a matching closing quote. In particular, a macro may be defined with
3777 replacement text that contains an unmatched quote. Of course, if you
3778 attempt to compile preprocessed output containing an unmatched quote
3779 you will get a syntax error.
3781 However, all preprocessing directives other than @code{#define}
3782 require matching quotes. For example:
3785 #define m This macro's fine and has an unmatched quote
3786 "/* This is not a comment. */
3787 /* @r{This is a comment. The following #include directive
3792 Just as for the ISO preprocessor, what would be a closing quote can be
3793 escaped with a backslash to prevent the quoted text from closing.
3795 @node Traditional macros
3796 @section Traditional macros
3798 The major difference between traditional and ISO macros is that the
3799 former expand to text rather than to a token sequence. CPP removes
3800 all leading and trailing horizontal whitespace from a macro's
3801 replacement text before storing it, but preserves the form of internal
3804 One consequence is that it is legitimate for the replacement text to
3805 contain an unmatched quote (@pxref{Traditional lexical analysis}). An
3806 unclosed string or character constant continues into the text
3807 following the macro call. Similarly, the text at the end of a macro's
3808 expansion can run together with the text after the macro invocation to
3809 produce a single token.
3811 Normally comments are removed from the replacement text after the
3812 macro is expanded, but if the @option{-CC} option is passed on the
3813 command line comments are preserved. (In fact, the current
3814 implementation removes comments even before saving the macro
3815 replacement text, but it careful to do it in such a way that the
3816 observed effect is identical even in the function-like macro case.)
3818 The ISO stringification operator @samp{#} and token paste operator
3819 @samp{##} have no special meaning. As explained later, an effect
3820 similar to these operators can be obtained in a different way. Macro
3821 names that are embedded in quotes, either from the main file or after
3822 macro replacement, do not expand.
3824 CPP replaces an unquoted object-like macro name with its replacement
3825 text, and then rescans it for further macros to replace. Unlike
3826 standard macro expansion, traditional macro expansion has no provision
3827 to prevent recursion. If an object-like macro appears unquoted in its
3828 replacement text, it will be replaced again during the rescan pass,
3829 and so on @emph{ad infinitum}. GCC detects when it is expanding
3830 recursive macros, emits an error message, and continues after the
3831 offending macro invocation.
3835 #define INC(x) PLUS+x
3840 Function-like macros are similar in form but quite different in
3841 behavior to their ISO counterparts. Their arguments are contained
3842 within parentheses, are comma-separated, and can cross physical lines.
3843 Commas within nested parentheses are not treated as argument
3844 separators. Similarly, a quote in an argument cannot be left
3845 unclosed; a following comma or parenthesis that comes before the
3846 closing quote is treated like any other character. There is no
3847 facility for handling variadic macros.
3849 This implementation removes all comments from macro arguments, unless
3850 the @option{-C} option is given. The form of all other horizontal
3851 whitespace in arguments is preserved, including leading and trailing
3852 whitespace. In particular
3859 is treated as an invocation of the macro @samp{f} with a single
3860 argument consisting of a single space. If you want to invoke a
3861 function-like macro that takes no arguments, you must not leave any
3862 whitespace between the parentheses.
3864 If a macro argument crosses a new line, the new line is replaced with
3865 a space when forming the argument. If the previous line contained an
3866 unterminated quote, the following line inherits the quoted state.
3868 Traditional preprocessors replace parameters in the replacement text
3869 with their arguments regardless of whether the parameters are within
3870 quotes or not. This provides a way to stringize arguments. For
3875 str(/* @r{A comment} */some text )
3876 @expansion{} "some text "
3880 Note that the comment is removed, but that the trailing space is
3881 preserved. Here is an example of using a comment to effect token
3885 #define suffix(x) foo_/**/x
3887 @expansion{} foo_bar
3890 @node Traditional miscellany
3891 @section Traditional miscellany
3893 Here are some things to be aware of when using the traditional
3898 Preprocessing directives are recognized only when their leading
3899 @samp{#} appears in the first column. There can be no whitespace
3900 between the beginning of the line and the @samp{#}, but whitespace can
3901 follow the @samp{#}.
3904 A true traditional C preprocessor does not recognize @samp{#error} or
3905 @samp{#pragma}, and may not recognize @samp{#elif}. CPP supports all
3906 the directives in traditional mode that it supports in ISO mode,
3907 including extensions, with the exception that the effects of
3908 @samp{#pragma GCC poison} are undefined.
3911 __STDC__ is not defined.
3914 If you use digraphs the behavior is undefined.
3917 If a line that looks like a directive appears within macro arguments,
3918 the behavior is undefined.
3922 @node Traditional warnings
3923 @section Traditional warnings
3924 You can request warnings about features that did not exist, or worked
3925 differently, in traditional C with the @option{-Wtraditional} option.
3926 GCC does not warn about features of ISO C which you must use when you
3927 are using a conforming compiler, such as the @samp{#} and @samp{##}
3930 Presently @option{-Wtraditional} warns about:
3934 Macro parameters that appear within string literals in the macro body.
3935 In traditional C macro replacement takes place within string literals,
3936 but does not in ISO C@.
3939 In traditional C, some preprocessor directives did not exist.
3940 Traditional preprocessors would only consider a line to be a directive
3941 if the @samp{#} appeared in column 1 on the line. Therefore
3942 @option{-Wtraditional} warns about directives that traditional C
3943 understands but would ignore because the @samp{#} does not appear as the
3944 first character on the line. It also suggests you hide directives like
3945 @samp{#pragma} not understood by traditional C by indenting them. Some
3946 traditional implementations would not recognize @samp{#elif}, so it
3947 suggests avoiding it altogether.
3950 A function-like macro that appears without an argument list. In some
3951 traditional preprocessors this was an error. In ISO C it merely means
3952 that the macro is not expanded.
3955 The unary plus operator. This did not exist in traditional C@.
3958 The @samp{U} and @samp{LL} integer constant suffixes, which were not
3959 available in traditional C@. (Traditional C does support the @samp{L}
3960 suffix for simple long integer constants.) You are not warned about
3961 uses of these suffixes in macros defined in system headers. For
3962 instance, @code{UINT_MAX} may well be defined as @code{4294967295U}, but
3963 you will not be warned if you use @code{UINT_MAX}.
3965 You can usually avoid the warning, and the related warning about
3966 constants which are so large that they are unsigned, by writing the
3967 integer constant in question in hexadecimal, with no U suffix. Take
3968 care, though, because this gives the wrong result in exotic cases.
3971 @node Implementation Details
3972 @chapter Implementation Details
3974 Here we document details of how the preprocessor's implementation
3975 affects its user-visible behavior. You should try to avoid undue
3976 reliance on behavior described here, as it is possible that it will
3977 change subtly in future implementations.
3979 Also documented here are obsolete features and changes from previous
3983 * Implementation-defined behavior::
3984 * Implementation limits::
3985 * Obsolete Features::
3986 * Differences from previous versions::
3989 @node Implementation-defined behavior
3990 @section Implementation-defined behavior
3991 @cindex implementation-defined behavior
3993 This is how CPP behaves in all the cases which the C standard
3994 describes as @dfn{implementation-defined}. This term means that the
3995 implementation is free to do what it likes, but must document its choice
3997 @c FIXME: Check the C++ standard for more implementation-defined stuff.
4001 @item The mapping of physical source file multi-byte characters to the
4002 execution character set.
4004 The input character set can be specified using the
4005 @option{-finput-charset} option, while the execution character set may
4006 be controlled using the @option{-fexec-charset} and
4007 @option{-fwide-exec-charset} options.
4009 @item Identifier characters.
4010 @anchor{Identifier characters}
4012 The C and C++ standards allow identifiers to be composed of @samp{_}
4013 and the alphanumeric characters. C++ and C99 also allow universal
4014 character names, and C99 further permits implementation-defined
4015 characters. GCC currently only permits universal character names if
4016 @option{-fextended-identifiers} is used, because the implementation of
4017 universal character names in identifiers is experimental.
4019 GCC allows the @samp{$} character in identifiers as an extension for
4020 most targets. This is true regardless of the @option{std=} switch,
4021 since this extension cannot conflict with standards-conforming
4022 programs. When preprocessing assembler, however, dollars are not
4023 identifier characters by default.
4025 Currently the targets that by default do not permit @samp{$} are AVR,
4026 IP2K, MMIX, MIPS Irix 3, ARM aout, and PowerPC targets for the AIX
4029 You can override the default with @option{-fdollars-in-identifiers} or
4030 @option{fno-dollars-in-identifiers}. @xref{fdollars-in-identifiers}.
4032 @item Non-empty sequences of whitespace characters.
4034 In textual output, each whitespace sequence is collapsed to a single
4035 space. For aesthetic reasons, the first token on each non-directive
4036 line of output is preceded with sufficient spaces that it appears in the
4037 same column as it did in the original source file.
4039 @item The numeric value of character constants in preprocessor expressions.
4041 The preprocessor and compiler interpret character constants in the
4042 same way; i.e.@: escape sequences such as @samp{\a} are given the
4043 values they would have on the target machine.
4045 The compiler evaluates a multi-character character constant a character
4046 at a time, shifting the previous value left by the number of bits per
4047 target character, and then or-ing in the bit-pattern of the new
4048 character truncated to the width of a target character. The final
4049 bit-pattern is given type @code{int}, and is therefore signed,
4050 regardless of whether single characters are signed or not (a slight
4051 change from versions 3.1 and earlier of GCC)@. If there are more
4052 characters in the constant than would fit in the target @code{int} the
4053 compiler issues a warning, and the excess leading characters are
4056 For example, @code{'ab'} for a target with an 8-bit @code{char} would be
4057 interpreted as @w{@samp{(int) ((unsigned char) 'a' * 256 + (unsigned char)
4058 'b')}}, and @code{'\234a'} as @w{@samp{(int) ((unsigned char) '\234' *
4059 256 + (unsigned char) 'a')}}.
4061 @item Source file inclusion.
4063 For a discussion on how the preprocessor locates header files,
4064 @ref{Include Operation}.
4066 @item Interpretation of the filename resulting from a macro-expanded
4067 @samp{#include} directive.
4069 @xref{Computed Includes}.
4071 @item Treatment of a @samp{#pragma} directive that after macro-expansion
4072 results in a standard pragma.
4074 No macro expansion occurs on any @samp{#pragma} directive line, so the
4075 question does not arise.
4077 Note that GCC does not yet implement any of the standard
4082 @node Implementation limits
4083 @section Implementation limits
4084 @cindex implementation limits
4086 CPP has a small number of internal limits. This section lists the
4087 limits which the C standard requires to be no lower than some minimum,
4088 and all the others known. It is intended that there should be as few limits
4089 as possible. If you encounter an undocumented or inconvenient limit,
4090 please report that as a bug. @xref{Bugs, , Reporting Bugs, gcc, Using
4091 the GNU Compiler Collection (GCC)}.
4093 Where we say something is limited @dfn{only by available memory}, that
4094 means that internal data structures impose no intrinsic limit, and space
4095 is allocated with @code{malloc} or equivalent. The actual limit will
4096 therefore depend on many things, such as the size of other things
4097 allocated by the compiler at the same time, the amount of memory
4098 consumed by other processes on the same computer, etc.
4102 @item Nesting levels of @samp{#include} files.
4104 We impose an arbitrary limit of 200 levels, to avoid runaway recursion.
4105 The standard requires at least 15 levels.
4107 @item Nesting levels of conditional inclusion.
4109 The C standard mandates this be at least 63. CPP is limited only by
4112 @item Levels of parenthesized expressions within a full expression.
4114 The C standard requires this to be at least 63. In preprocessor
4115 conditional expressions, it is limited only by available memory.
4117 @item Significant initial characters in an identifier or macro name.
4119 The preprocessor treats all characters as significant. The C standard
4120 requires only that the first 63 be significant.
4122 @item Number of macros simultaneously defined in a single translation unit.
4124 The standard requires at least 4095 be possible. CPP is limited only
4125 by available memory.
4127 @item Number of parameters in a macro definition and arguments in a macro call.
4129 We allow @code{USHRT_MAX}, which is no smaller than 65,535. The minimum
4130 required by the standard is 127.
4132 @item Number of characters on a logical source line.
4134 The C standard requires a minimum of 4096 be permitted. CPP places
4135 no limits on this, but you may get incorrect column numbers reported in
4136 diagnostics for lines longer than 65,535 characters.
4138 @item Maximum size of a source file.
4140 The standard does not specify any lower limit on the maximum size of a
4141 source file. GNU cpp maps files into memory, so it is limited by the
4142 available address space. This is generally at least two gigabytes.
4143 Depending on the operating system, the size of physical memory may or
4144 may not be a limitation.
4148 @node Obsolete Features
4149 @section Obsolete Features
4151 CPP has some features which are present mainly for compatibility with
4152 older programs. We discourage their use in new code. In some cases,
4153 we plan to remove the feature in a future version of GCC@.
4155 @subsection Assertions
4158 @dfn{Assertions} are a deprecated alternative to macros in writing
4159 conditionals to test what sort of computer or system the compiled
4160 program will run on. Assertions are usually predefined, but you can
4161 define them with preprocessing directives or command-line options.
4163 Assertions were intended to provide a more systematic way to describe
4164 the compiler's target system. However, in practice they are just as
4165 unpredictable as the system-specific predefined macros. In addition, they
4166 are not part of any standard, and only a few compilers support them.
4167 Therefore, the use of assertions is @strong{less} portable than the use
4168 of system-specific predefined macros. We recommend you do not use them at
4172 An assertion looks like this:
4175 #@var{predicate} (@var{answer})
4179 @var{predicate} must be a single identifier. @var{answer} can be any
4180 sequence of tokens; all characters are significant except for leading
4181 and trailing whitespace, and differences in internal whitespace
4182 sequences are ignored. (This is similar to the rules governing macro
4183 redefinition.) Thus, @code{(x + y)} is different from @code{(x+y)} but
4184 equivalent to @code{@w{( x + y )}}. Parentheses do not nest inside an
4187 @cindex testing predicates
4188 To test an assertion, you write it in an @samp{#if}. For example, this
4189 conditional succeeds if either @code{vax} or @code{ns16000} has been
4190 asserted as an answer for @code{machine}.
4193 #if #machine (vax) || #machine (ns16000)
4197 You can test whether @emph{any} answer is asserted for a predicate by
4198 omitting the answer in the conditional:
4205 Assertions are made with the @samp{#assert} directive. Its sole
4206 argument is the assertion to make, without the leading @samp{#} that
4207 identifies assertions in conditionals.
4210 #assert @var{predicate} (@var{answer})
4214 You may make several assertions with the same predicate and different
4215 answers. Subsequent assertions do not override previous ones for the
4216 same predicate. All the answers for any given predicate are
4217 simultaneously true.
4219 @cindex assertions, canceling
4221 Assertions can be canceled with the @samp{#unassert} directive. It
4222 has the same syntax as @samp{#assert}. In that form it cancels only the
4223 answer which was specified on the @samp{#unassert} line; other answers
4224 for that predicate remain true. You can cancel an entire predicate by
4225 leaving out the answer:
4228 #unassert @var{predicate}
4232 In either form, if no such assertion has been made, @samp{#unassert} has
4235 You can also make or cancel assertions using command line options.
4238 @node Differences from previous versions
4239 @section Differences from previous versions
4240 @cindex differences from previous versions
4242 This section details behavior which has changed from previous versions
4243 of CPP@. We do not plan to change it again in the near future, but
4244 we do not promise not to, either.
4246 The ``previous versions'' discussed here are 2.95 and before. The
4247 behavior of GCC 3.0 is mostly the same as the behavior of the widely
4248 used 2.96 and 2.97 development snapshots. Where there are differences,
4249 they generally represent bugs in the snapshots.
4253 @item -I- deprecated
4255 This option has been deprecated in 4.0. @option{-iquote} is meant to
4256 replace the need for this option.
4258 @item Order of evaluation of @samp{#} and @samp{##} operators
4260 The standard does not specify the order of evaluation of a chain of
4261 @samp{##} operators, nor whether @samp{#} is evaluated before, after, or
4262 at the same time as @samp{##}. You should therefore not write any code
4263 which depends on any specific ordering. It is possible to guarantee an
4264 ordering, if you need one, by suitable use of nested macros.
4266 An example of where this might matter is pasting the arguments @samp{1},
4267 @samp{e} and @samp{-2}. This would be fine for left-to-right pasting,
4268 but right-to-left pasting would produce an invalid token @samp{e-2}.
4270 GCC 3.0 evaluates @samp{#} and @samp{##} at the same time and strictly
4271 left to right. Older versions evaluated all @samp{#} operators first,
4272 then all @samp{##} operators, in an unreliable order.
4274 @item The form of whitespace between tokens in preprocessor output
4276 @xref{Preprocessor Output}, for the current textual format. This is
4277 also the format used by stringification. Normally, the preprocessor
4278 communicates tokens directly to the compiler's parser, and whitespace
4279 does not come up at all.
4281 Older versions of GCC preserved all whitespace provided by the user and
4282 inserted lots more whitespace of their own, because they could not
4283 accurately predict when extra spaces were needed to prevent accidental
4286 @item Optional argument when invoking rest argument macros
4288 As an extension, GCC permits you to omit the variable arguments entirely
4289 when you use a variable argument macro. This is forbidden by the 1999 C
4290 standard, and will provoke a pedantic warning with GCC 3.0. Previous
4291 versions accepted it silently.
4293 @item @samp{##} swallowing preceding text in rest argument macros
4295 Formerly, in a macro expansion, if @samp{##} appeared before a variable
4296 arguments parameter, and the set of tokens specified for that argument
4297 in the macro invocation was empty, previous versions of CPP would
4298 back up and remove the preceding sequence of non-whitespace characters
4299 (@strong{not} the preceding token). This extension is in direct
4300 conflict with the 1999 C standard and has been drastically pared back.
4302 In the current version of the preprocessor, if @samp{##} appears between
4303 a comma and a variable arguments parameter, and the variable argument is
4304 omitted entirely, the comma will be removed from the expansion. If the
4305 variable argument is empty, or the token before @samp{##} is not a
4306 comma, then @samp{##} behaves as a normal token paste.
4308 @item @samp{#line} and @samp{#include}
4310 The @samp{#line} directive used to change GCC's notion of the
4311 ``directory containing the current file'', used by @samp{#include} with
4312 a double-quoted header file name. In 3.0 and later, it does not.
4313 @xref{Line Control}, for further explanation.
4315 @item Syntax of @samp{#line}
4317 In GCC 2.95 and previous, the string constant argument to @samp{#line}
4318 was treated the same way as the argument to @samp{#include}: backslash
4319 escapes were not honored, and the string ended at the second @samp{"}.
4320 This is not compliant with the C standard. In GCC 3.0, an attempt was
4321 made to correct the behavior, so that the string was treated as a real
4322 string constant, but it turned out to be buggy. In 3.1, the bugs have
4323 been fixed. (We are not fixing the bugs in 3.0 because they affect
4324 relatively few people and the fix is quite invasive.)
4331 @cindex command line
4333 Most often when you use the C preprocessor you will not have to invoke it
4334 explicitly: the C compiler will do so automatically. However, the
4335 preprocessor is sometimes useful on its own. All the options listed
4336 here are also acceptable to the C compiler and have the same meaning,
4337 except that the C compiler has different rules for specifying the output
4340 @emph{Note:} Whether you use the preprocessor by way of @command{gcc}
4341 or @command{cpp}, the @dfn{compiler driver} is run first. This
4342 program's purpose is to translate your command into invocations of the
4343 programs that do the actual work. Their command line interfaces are
4344 similar but not identical to the documented interface, and may change
4348 @c man begin SYNOPSIS
4349 cpp [@option{-D}@var{macro}[=@var{defn}]@dots{}] [@option{-U}@var{macro}]
4350 [@option{-I}@var{dir}@dots{}] [@option{-iquote}@var{dir}@dots{}]
4351 [@option{-W}@var{warn}@dots{}]
4352 [@option{-M}|@option{-MM}] [@option{-MG}] [@option{-MF} @var{filename}]
4353 [@option{-MP}] [@option{-MQ} @var{target}@dots{}]
4354 [@option{-MT} @var{target}@dots{}]
4355 [@option{-P}] [@option{-fno-working-directory}]
4356 [@option{-x} @var{language}] [@option{-std=}@var{standard}]
4357 @var{infile} @var{outfile}
4359 Only the most useful options are listed here; see below for the remainder.
4361 @c man begin SEEALSO
4362 gpl(7), gfdl(7), fsf-funding(7),
4363 gcc(1), as(1), ld(1), and the Info entries for @file{cpp}, @file{gcc}, and
4368 @c man begin OPTIONS
4369 The C preprocessor expects two file names as arguments, @var{infile} and
4370 @var{outfile}. The preprocessor reads @var{infile} together with any
4371 other files it specifies with @samp{#include}. All the output generated
4372 by the combined input files is written in @var{outfile}.
4374 Either @var{infile} or @var{outfile} may be @option{-}, which as
4375 @var{infile} means to read from standard input and as @var{outfile}
4376 means to write to standard output. Also, if either file is omitted, it
4377 means the same as if @option{-} had been specified for that file.
4379 Unless otherwise noted, or the option ends in @samp{=}, all options
4380 which take an argument may have that argument appear either immediately
4381 after the option, or with a space between option and argument:
4382 @option{-Ifoo} and @option{-I foo} have the same effect.
4384 @cindex grouping options
4385 @cindex options, grouping
4386 Many options have multi-letter names; therefore multiple single-letter
4387 options may @emph{not} be grouped: @option{-dM} is very different from
4391 @include cppopts.texi
4394 @node Environment Variables
4395 @chapter Environment Variables
4396 @cindex environment variables
4397 @c man begin ENVIRONMENT
4399 This section describes the environment variables that affect how CPP
4400 operates. You can use them to specify directories or prefixes to use
4401 when searching for include files, or to control dependency output.
4403 Note that you can also specify places to search using options such as
4404 @option{-I}, and control dependency output with options like
4405 @option{-M} (@pxref{Invocation}). These take precedence over
4406 environment variables, which in turn take precedence over the
4407 configuration of GCC@.
4409 @include cppenv.texi
4416 @node Index of Directives
4417 @unnumbered Index of Directives
4421 @unnumbered Option Index
4423 CPP's command line options and environment variables are indexed here
4424 without any initial @samp{-} or @samp{--}.
4429 @unnumbered Concept Index