1 This is doc/cpp.info, produced by makeinfo version 4.8 from
2 /scratch/mitchell/gcc-releases/gcc-4.1.2/gcc-4.1.2/gcc/doc/cpp.texi.
4 Copyright (C) 1987, 1989, 1991, 1992, 1993, 1994, 1995, 1996, 1997,
5 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software
8 Permission is granted to copy, distribute and/or modify this document
9 under the terms of the GNU Free Documentation License, Version 1.1 or
10 any later version published by the Free Software Foundation. A copy of
11 the license is included in the section entitled "GNU Free Documentation
14 This manual contains no Invariant Sections. The Front-Cover Texts
15 are (a) (see below), and the Back-Cover Texts are (b) (see below).
17 (a) The FSF's Front-Cover Text is:
21 (b) The FSF's Back-Cover Text is:
23 You have freedom to copy and modify this GNU Manual, like GNU
24 software. Copies published by the Free Software Foundation raise
25 funds for GNU development.
27 INFO-DIR-SECTION Programming
29 * Cpp: (cpp). The GNU C preprocessor.
33 File: cpp.info, Node: Top, Next: Overview, Up: (dir)
38 The C preprocessor implements the macro language used to transform C,
39 C++, and Objective-C programs before they are compiled. It can also be
52 * Preprocessor Output::
54 * Implementation Details::
56 * Environment Variables::
57 * GNU Free Documentation License::
58 * Index of Directives::
62 --- The Detailed Node Listing ---
67 * Initial processing::
69 * The preprocessing language::
83 * Object-like Macros::
84 * Function-like Macros::
90 * Undefining and Redefining Macros::
91 * Directives Within Macro Arguments::
96 * Standard Predefined Macros::
97 * Common Predefined Macros::
98 * System-specific Predefined Macros::
99 * C++ Named Operators::
104 * Operator Precedence Problems::
105 * Swallowing the Semicolon::
106 * Duplication of Side Effects::
107 * Self-Referential Macros::
109 * Newlines in Arguments::
114 * Conditional Syntax::
125 Implementation Details
127 * Implementation-defined behavior::
128 * Implementation limits::
129 * Obsolete Features::
130 * Differences from previous versions::
135 * Obsolete once-only headers::
137 Copyright (C) 1987, 1989, 1991, 1992, 1993, 1994, 1995, 1996, 1997,
138 1998, 1999, 2000, 2001, 2002, 2003, 2004, 2005 Free Software
141 Permission is granted to copy, distribute and/or modify this document
142 under the terms of the GNU Free Documentation License, Version 1.1 or
143 any later version published by the Free Software Foundation. A copy of
144 the license is included in the section entitled "GNU Free Documentation
147 This manual contains no Invariant Sections. The Front-Cover Texts
148 are (a) (see below), and the Back-Cover Texts are (b) (see below).
150 (a) The FSF's Front-Cover Text is:
154 (b) The FSF's Back-Cover Text is:
156 You have freedom to copy and modify this GNU Manual, like GNU
157 software. Copies published by the Free Software Foundation raise
158 funds for GNU development.
161 File: cpp.info, Node: Overview, Next: Header Files, Prev: Top, Up: Top
166 The C preprocessor, often known as "cpp", is a "macro processor" that
167 is used automatically by the C compiler to transform your program
168 before compilation. It is called a macro processor because it allows
169 you to define "macros", which are brief abbreviations for longer
172 The C preprocessor is intended to be used only with C, C++, and
173 Objective-C source code. In the past, it has been abused as a general
174 text processor. It will choke on input which does not obey C's lexical
175 rules. For example, apostrophes will be interpreted as the beginning of
176 character constants, and cause errors. Also, you cannot rely on it
177 preserving characteristics of the input which are not significant to
178 C-family languages. If a Makefile is preprocessed, all the hard tabs
179 will be removed, and the Makefile will not work.
181 Having said that, you can often get away with using cpp on things
182 which are not C. Other Algol-ish programming languages are often safe
183 (Pascal, Ada, etc.) So is assembly, with caution. `-traditional-cpp'
184 mode preserves more white space, and is otherwise more permissive. Many
185 of the problems can be avoided by writing C or C++ style comments
186 instead of native language comments, and keeping macros simple.
188 Wherever possible, you should use a preprocessor geared to the
189 language you are writing in. Modern versions of the GNU assembler have
190 macro facilities. Most high level programming languages have their own
191 conditional compilation and inclusion mechanism. If all else fails,
192 try a true general text processor, such as GNU M4.
194 C preprocessors vary in some details. This manual discusses the GNU
195 C preprocessor, which provides a small superset of the features of ISO
196 Standard C. In its default mode, the GNU C preprocessor does not do a
197 few things required by the standard. These are features which are
198 rarely, if ever, used, and may cause surprising changes to the meaning
199 of a program which does not expect them. To get strict ISO Standard C,
200 you should use the `-std=c89' or `-std=c99' options, depending on which
201 version of the standard you want. To get all the mandatory
202 diagnostics, you must also use `-pedantic'. *Note Invocation::.
204 This manual describes the behavior of the ISO preprocessor. To
205 minimize gratuitous differences, where the ISO preprocessor's behavior
206 does not conflict with traditional semantics, the traditional
207 preprocessor should behave the same way. The various differences that
208 do exist are detailed in the section *Note Traditional Mode::.
210 For clarity, unless noted otherwise, references to `CPP' in this
211 manual refer to GNU CPP.
216 * Initial processing::
218 * The preprocessing language::
221 File: cpp.info, Node: Character sets, Next: Initial processing, Up: Overview
226 Source code character set processing in C and related languages is
227 rather complicated. The C standard discusses two character sets, but
228 there are really at least four.
230 The files input to CPP might be in any character set at all. CPP's
231 very first action, before it even looks for line boundaries, is to
232 convert the file into the character set it uses for internal
233 processing. That set is what the C standard calls the "source"
234 character set. It must be isomorphic with ISO 10646, also known as
235 Unicode. CPP uses the UTF-8 encoding of Unicode.
237 At present, GNU CPP does not implement conversion from arbitrary file
238 encodings to the source character set. Use of any encoding other than
239 plain ASCII or UTF-8, except in comments, will cause errors. Use of
240 encodings that are not strict supersets of ASCII, such as Shift JIS,
241 may cause errors even if non-ASCII characters appear only in comments.
242 We plan to fix this in the near future.
244 All preprocessing work (the subject of the rest of this manual) is
245 carried out in the source character set. If you request textual output
246 from the preprocessor with the `-E' option, it will be in UTF-8.
248 After preprocessing is complete, string and character constants are
249 converted again, into the "execution" character set. This character
250 set is under control of the user; the default is UTF-8, matching the
251 source character set. Wide string and character constants have their
252 own character set, which is not called out specifically in the
253 standard. Again, it is under control of the user. The default is
254 UTF-16 or UTF-32, whichever fits in the target's `wchar_t' type, in the
255 target machine's byte order.(1) Octal and hexadecimal escape sequences
256 do not undergo conversion; '\x12' has the value 0x12 regardless of the
257 currently selected execution character set. All other escapes are
258 replaced by the character in the source character set that they
259 represent, then converted to the execution character set, just like
260 unescaped characters.
262 Unless the experimental `-fextended-identifiers' option is used, GCC
263 does not permit the use of characters outside the ASCII range, nor `\u'
264 and `\U' escapes, in identifiers. Even with that option, characters
265 outside the ASCII range can only be specified with the `\u' and `\U'
266 escapes, not used directly in identifiers.
268 ---------- Footnotes ----------
270 (1) UTF-16 does not meet the requirements of the C standard for a
271 wide character set, but the choice of 16-bit `wchar_t' is enshrined in
272 some system ABIs so we cannot fix this.
275 File: cpp.info, Node: Initial processing, Next: Tokenization, Prev: Character sets, Up: Overview
277 1.2 Initial processing
278 ======================
280 The preprocessor performs a series of textual transformations on its
281 input. These happen before all other processing. Conceptually, they
282 happen in a rigid order, and the entire file is run through each
283 transformation before the next one begins. CPP actually does them all
284 at once, for performance reasons. These transformations correspond
285 roughly to the first three "phases of translation" described in the C
288 1. The input file is read into memory and broken into lines.
290 Different systems use different conventions to indicate the end of
291 a line. GCC accepts the ASCII control sequences `LF', `CR LF' and
292 `CR' as end-of-line markers. These are the canonical sequences
293 used by Unix, DOS and VMS, and the classic Mac OS (before OSX)
294 respectively. You may therefore safely copy source code written
295 on any of those systems to a different one and use it without
296 conversion. (GCC may lose track of the current line number if a
297 file doesn't consistently use one convention, as sometimes happens
298 when it is edited on computers with different conventions that
299 share a network file system.)
301 If the last line of any input file lacks an end-of-line marker,
302 the end of the file is considered to implicitly supply one. The C
303 standard says that this condition provokes undefined behavior, so
304 GCC will emit a warning message.
306 2. If trigraphs are enabled, they are replaced by their corresponding
307 single characters. By default GCC ignores trigraphs, but if you
308 request a strictly conforming mode with the `-std' option, or you
309 specify the `-trigraphs' option, then it converts them.
311 These are nine three-character sequences, all starting with `??',
312 that are defined by ISO C to stand for single characters. They
313 permit obsolete systems that lack some of C's punctuation to use
314 C. For example, `??/' stands for `\', so '??/n' is a character
315 constant for a newline.
317 Trigraphs are not popular and many compilers implement them
318 incorrectly. Portable code should not rely on trigraphs being
319 either converted or ignored. With `-Wtrigraphs' GCC will warn you
320 when a trigraph may change the meaning of your program if it were
321 converted. *Note Wtrigraphs::.
323 In a string constant, you can prevent a sequence of question marks
324 from being confused with a trigraph by inserting a backslash
325 between the question marks, or by separating the string literal at
326 the trigraph and making use of string literal concatenation.
327 "(??\?)" is the string `(???)', not `(?]'. Traditional C
328 compilers do not recognize these idioms.
330 The nine trigraphs and their replacements are
332 Trigraph: ??( ??) ??< ??> ??= ??/ ??' ??! ??-
333 Replacement: [ ] { } # \ ^ | ~
335 3. Continued lines are merged into one long line.
337 A continued line is a line which ends with a backslash, `\'. The
338 backslash is removed and the following line is joined with the
339 current one. No space is inserted, so you may split a line
340 anywhere, even in the middle of a word. (It is generally more
341 readable to split lines only at white space.)
343 The trailing backslash on a continued line is commonly referred to
344 as a "backslash-newline".
346 If there is white space between a backslash and the end of a line,
347 that is still a continued line. However, as this is usually the
348 result of an editing mistake, and many compilers will not accept
349 it as a continued line, GCC will warn you about it.
351 4. All comments are replaced with single spaces.
353 There are two kinds of comments. "Block comments" begin with `/*'
354 and continue until the next `*/'. Block comments do not nest:
356 /* this is /* one comment */ text outside comment
358 "Line comments" begin with `//' and continue to the end of the
359 current line. Line comments do not nest either, but it does not
360 matter, because they would end in the same place anyway.
362 // this is // one comment
365 It is safe to put line comments inside block comments, or vice versa.
368 // contains line comment
372 // line comment /* contains block comment */
374 But beware of commenting out one end of a block comment with a line
377 // l.c. /* block comment begins
378 oops! this isn't a comment anymore */
380 Comments are not recognized within string literals. "/* blah */" is
381 the string constant `/* blah */', not an empty string.
383 Line comments are not in the 1989 edition of the C standard, but they
384 are recognized by GCC as an extension. In C++ and in the 1999 edition
385 of the C standard, they are an official part of the language.
387 Since these transformations happen before all other processing, you
388 can split a line mechanically with backslash-newline anywhere. You can
389 comment out the end of a line. You can continue a line comment onto the
390 next line with backslash-newline. You can even split `/*', `*/', and
391 `//' onto multiple lines with backslash-newline. For example:
401 is equivalent to `#define FOO 1020'. All these tricks are extremely
402 confusing and should not be used in code intended to be readable.
404 There is no way to prevent a backslash at the end of a line from
405 being interpreted as a backslash-newline. This cannot affect any
406 correct program, however.
409 File: cpp.info, Node: Tokenization, Next: The preprocessing language, Prev: Initial processing, Up: Overview
414 After the textual transformations are finished, the input file is
415 converted into a sequence of "preprocessing tokens". These mostly
416 correspond to the syntactic tokens used by the C compiler, but there are
417 a few differences. White space separates tokens; it is not itself a
418 token of any kind. Tokens do not have to be separated by white space,
419 but it is often necessary to avoid ambiguities.
421 When faced with a sequence of characters that has more than one
422 possible tokenization, the preprocessor is greedy. It always makes
423 each token, starting from the left, as big as possible before moving on
424 to the next token. For instance, `a+++++b' is interpreted as
425 `a ++ ++ + b', not as `a ++ + ++ b', even though the latter
426 tokenization could be part of a valid C program and the former could
429 Once the input file is broken into tokens, the token boundaries never
430 change, except when the `##' preprocessing operator is used to paste
431 tokens together. *Note Concatenation::. For example,
439 The compiler does not re-tokenize the preprocessor's output. Each
440 preprocessing token becomes one compiler token.
442 Preprocessing tokens fall into five broad classes: identifiers,
443 preprocessing numbers, string literals, punctuators, and other. An
444 "identifier" is the same as an identifier in C: any sequence of
445 letters, digits, or underscores, which begins with a letter or
446 underscore. Keywords of C have no significance to the preprocessor;
447 they are ordinary identifiers. You can define a macro whose name is a
448 keyword, for instance. The only identifier which can be considered a
449 preprocessing keyword is `defined'. *Note Defined::.
451 This is mostly true of other languages which use the C preprocessor.
452 However, a few of the keywords of C++ are significant even in the
453 preprocessor. *Note C++ Named Operators::.
455 In the 1999 C standard, identifiers may contain letters which are not
456 part of the "basic source character set", at the implementation's
457 discretion (such as accented Latin letters, Greek letters, or Chinese
458 ideograms). This may be done with an extended character set, or the
459 `\u' and `\U' escape sequences. The implementation of this feature in
460 GCC is experimental; such characters are only accepted in the `\u' and
461 `\U' forms and only if `-fextended-identifiers' is used.
463 As an extension, GCC treats `$' as a letter. This is for
464 compatibility with some systems, such as VMS, where `$' is commonly
465 used in system-defined function and object names. `$' is not a letter
466 in strictly conforming mode, or if you specify the `-$' option. *Note
469 A "preprocessing number" has a rather bizarre definition. The
470 category includes all the normal integer and floating point constants
471 one expects of C, but also a number of other things one might not
472 initially recognize as a number. Formally, preprocessing numbers begin
473 with an optional period, a required decimal digit, and then continue
474 with any sequence of letters, digits, underscores, periods, and
475 exponents. Exponents are the two-character sequences `e+', `e-', `E+',
476 `E-', `p+', `p-', `P+', and `P-'. (The exponents that begin with `p'
477 or `P' are new to C99. They are used for hexadecimal floating-point
480 The purpose of this unusual definition is to isolate the preprocessor
481 from the full complexity of numeric constants. It does not have to
482 distinguish between lexically valid and invalid floating-point numbers,
483 which is complicated. The definition also permits you to split an
484 identifier at any position and get exactly two tokens, which can then be
485 pasted back together with the `##' operator.
487 It's possible for preprocessing numbers to cause programs to be
488 misinterpreted. For example, `0xE+12' is a preprocessing number which
489 does not translate to any valid numeric constant, therefore a syntax
490 error. It does not mean `0xE + 12', which is what you might have
493 "String literals" are string constants, character constants, and
494 header file names (the argument of `#include').(1) String constants
495 and character constants are straightforward: "..." or '...'. In either
496 case embedded quotes should be escaped with a backslash: '\'' is the
497 character constant for `''. There is no limit on the length of a
498 character constant, but the value of a character constant that contains
499 more than one character is implementation-defined. *Note
500 Implementation Details::.
502 Header file names either look like string constants, "...", or are
503 written with angle brackets instead, <...>. In either case, backslash
504 is an ordinary character. There is no way to escape the closing quote
505 or angle bracket. The preprocessor looks for the header file in
506 different places depending on which form you use. *Note Include
509 No string literal may extend past the end of a line. Older versions
510 of GCC accepted multi-line string constants. You may use continued
511 lines instead, or string constant concatenation. *Note Differences
512 from previous versions::.
514 "Punctuators" are all the usual bits of punctuation which are
515 meaningful to C and C++. All but three of the punctuation characters in
516 ASCII are C punctuators. The exceptions are `@', `$', and ``'. In
517 addition, all the two- and three-character operators are punctuators.
518 There are also six "digraphs", which the C++ standard calls
519 "alternative tokens", which are merely alternate ways to spell other
520 punctuators. This is a second attempt to work around missing
521 punctuation in obsolete systems. It has no negative side effects,
522 unlike trigraphs, but does not cover as much ground. The digraphs and
523 their corresponding normal punctuators are:
525 Digraph: <% %> <: :> %: %:%:
526 Punctuator: { } [ ] # ##
528 Any other single character is considered "other". It is passed on to
529 the preprocessor's output unmolested. The C compiler will almost
530 certainly reject source code containing "other" tokens. In ASCII, the
531 only other characters are `@', `$', ``', and control characters other
532 than NUL (all bits zero). (Note that `$' is normally considered a
533 letter.) All characters with the high bit set (numeric range
534 0x7F-0xFF) are also "other" in the present implementation. This will
535 change when proper support for international character sets is added to
538 NUL is a special case because of the high probability that its
539 appearance is accidental, and because it may be invisible to the user
540 (many terminals do not display NUL at all). Within comments, NULs are
541 silently ignored, just as any other character would be. In running
542 text, NUL is considered white space. For example, these two directives
543 have the same meaning.
548 (where `^@' is ASCII NUL). Within string or character constants, NULs
549 are preserved. In the latter two cases the preprocessor emits a
552 ---------- Footnotes ----------
554 (1) The C standard uses the term "string literal" to refer only to
555 what we are calling "string constants".
558 File: cpp.info, Node: The preprocessing language, Prev: Tokenization, Up: Overview
560 1.4 The preprocessing language
561 ==============================
563 After tokenization, the stream of tokens may simply be passed straight
564 to the compiler's parser. However, if it contains any operations in the
565 "preprocessing language", it will be transformed first. This stage
566 corresponds roughly to the standard's "translation phase 4" and is what
567 most people think of as the preprocessor's job.
569 The preprocessing language consists of "directives" to be executed
570 and "macros" to be expanded. Its primary capabilities are:
572 * Inclusion of header files. These are files of declarations that
573 can be substituted into your program.
575 * Macro expansion. You can define "macros", which are abbreviations
576 for arbitrary fragments of C code. The preprocessor will replace
577 the macros with their definitions throughout the program. Some
578 macros are automatically defined for you.
580 * Conditional compilation. You can include or exclude parts of the
581 program according to various conditions.
583 * Line control. If you use a program to combine or rearrange source
584 files into an intermediate file which is then compiled, you can
585 use line control to inform the compiler where each source line
586 originally came from.
588 * Diagnostics. You can detect problems at compile time and issue
591 There are a few more, less useful, features.
593 Except for expansion of predefined macros, all these operations are
594 triggered with "preprocessing directives". Preprocessing directives
595 are lines in your program that start with `#'. Whitespace is allowed
596 before and after the `#'. The `#' is followed by an identifier, the
597 "directive name". It specifies the operation to perform. Directives
598 are commonly referred to as `#NAME' where NAME is the directive name.
599 For example, `#define' is the directive that defines a macro.
601 The `#' which begins a directive cannot come from a macro expansion.
602 Also, the directive name is not macro expanded. Thus, if `foo' is
603 defined as a macro expanding to `define', that does not make `#foo' a
604 valid preprocessing directive.
606 The set of valid directive names is fixed. Programs cannot define
607 new preprocessing directives.
609 Some directives require arguments; these make up the rest of the
610 directive line and must be separated from the directive name by
611 whitespace. For example, `#define' must be followed by a macro name
612 and the intended expansion of the macro.
614 A preprocessing directive cannot cover more than one line. The line
615 may, however, be continued with backslash-newline, or by a block comment
616 which extends past the end of the line. In either case, when the
617 directive is processed, the continuations have already been merged with
618 the first line to make one long line.
621 File: cpp.info, Node: Header Files, Next: Macros, Prev: Overview, Up: Top
626 A header file is a file containing C declarations and macro definitions
627 (*note Macros::) to be shared between several source files. You request
628 the use of a header file in your program by "including" it, with the C
629 preprocessing directive `#include'.
631 Header files serve two purposes.
633 * System header files declare the interfaces to parts of the
634 operating system. You include them in your program to supply the
635 definitions and declarations you need to invoke system calls and
638 * Your own header files contain declarations for interfaces between
639 the source files of your program. Each time you have a group of
640 related declarations and macro definitions all or most of which
641 are needed in several different source files, it is a good idea to
642 create a header file for them.
644 Including a header file produces the same results as copying the
645 header file into each source file that needs it. Such copying would be
646 time-consuming and error-prone. With a header file, the related
647 declarations appear in only one place. If they need to be changed, they
648 can be changed in one place, and programs that include the header file
649 will automatically use the new version when next recompiled. The header
650 file eliminates the labor of finding and changing all the copies as well
651 as the risk that a failure to find one copy will result in
652 inconsistencies within a program.
654 In C, the usual convention is to give header files names that end
655 with `.h'. It is most portable to use only letters, digits, dashes, and
656 underscores in header file names, and at most one dot.
661 * Include Operation::
663 * Once-Only Headers::
664 * Computed Includes::
669 File: cpp.info, Node: Include Syntax, Next: Include Operation, Up: Header Files
674 Both user and system header files are included using the preprocessing
675 directive `#include'. It has two variants:
678 This variant is used for system header files. It searches for a
679 file named FILE in a standard list of system directories. You can
680 prepend directories to this list with the `-I' option (*note
684 This variant is used for header files of your own program. It
685 searches for a file named FILE first in the directory containing
686 the current file, then in the quote directories and then the same
687 directories used for `<FILE>'. You can prepend directories to the
688 list of quote directories with the `-iquote' option.
690 The argument of `#include', whether delimited with quote marks or
691 angle brackets, behaves like a string constant in that comments are not
692 recognized, and macro names are not expanded. Thus, `#include <x/*y>'
693 specifies inclusion of a system header file named `x/*y'.
695 However, if backslashes occur within FILE, they are considered
696 ordinary text characters, not escape characters. None of the character
697 escape sequences appropriate to string constants in C are processed.
698 Thus, `#include "x\n\\y"' specifies a filename containing three
699 backslashes. (Some systems interpret `\' as a pathname separator. All
700 of these also interpret `/' the same way. It is most portable to use
703 It is an error if there is anything (other than comments) on the line
707 File: cpp.info, Node: Include Operation, Next: Search Path, Prev: Include Syntax, Up: Header Files
709 2.2 Include Operation
710 =====================
712 The `#include' directive works by directing the C preprocessor to scan
713 the specified file as input before continuing with the rest of the
714 current file. The output from the preprocessor contains the output
715 already generated, followed by the output resulting from the included
716 file, followed by the output that comes from the text after the
717 `#include' directive. For example, if you have a header file
718 `header.h' as follows,
722 and a main program called `program.c' that uses the header file, like
734 the compiler will see the same token stream as it would if `program.c'
746 Included files are not limited to declarations and macro definitions;
747 those are merely the typical uses. Any fragment of a C program can be
748 included from another file. The include file could even contain the
749 beginning of a statement that is concluded in the containing file, or
750 the end of a statement that was started in the including file. However,
751 an included file must consist of complete tokens. Comments and string
752 literals which have not been closed by the end of an included file are
753 invalid. For error recovery, they are considered to end at the end of
756 To avoid confusion, it is best if header files contain only complete
757 syntactic units--function declarations or definitions, type
760 The line following the `#include' directive is always treated as a
761 separate line by the C preprocessor, even if the included file lacks a
765 File: cpp.info, Node: Search Path, Next: Once-Only Headers, Prev: Include Operation, Up: Header Files
770 GCC looks in several different places for headers. On a normal Unix
771 system, if you do not instruct it otherwise, it will look for headers
772 requested with `#include <FILE>' in:
775 LIBDIR/gcc/TARGET/VERSION/include
779 For C++ programs, it will also look in `/usr/include/g++-v3', first.
780 In the above, TARGET is the canonical name of the system GCC was
781 configured to compile code for; often but not always the same as the
782 canonical name of the system it runs on. VERSION is the version of GCC
785 You can add to this list with the `-IDIR' command line option. All
786 the directories named by `-I' are searched, in left-to-right order,
787 _before_ the default directories. The only exception is when `dir' is
788 already searched by default. In this case, the option is ignored and
789 the search order for system directories remains unchanged.
791 Duplicate directories are removed from the quote and bracket search
792 chains before the two chains are merged to make the final search chain.
793 Thus, it is possible for a directory to occur twice in the final search
794 chain if it was specified in both the quote and bracket chains.
796 You can prevent GCC from searching any of the default directories
797 with the `-nostdinc' option. This is useful when you are compiling an
798 operating system kernel or some other program that does not use the
799 standard C library facilities, or the standard C library itself. `-I'
800 options are not ignored as described above when `-nostdinc' is in
803 GCC looks for headers requested with `#include "FILE"' first in the
804 directory containing the current file, then in the directories as
805 specified by `-iquote' options, then in the same places it would have
806 looked for a header requested with angle brackets. For example, if
807 `/usr/include/sys/stat.h' contains `#include "types.h"', GCC looks for
808 `types.h' first in `/usr/include/sys', then in its usual search path.
810 `#line' (*note Line Control::) does not change GCC's idea of the
811 directory containing the current file.
813 You may put `-I-' at any point in your list of `-I' options. This
814 has two effects. First, directories appearing before the `-I-' in the
815 list are searched only for headers requested with quote marks.
816 Directories after `-I-' are searched for all headers. Second, the
817 directory containing the current file is not searched for anything,
818 unless it happens to be one of the directories named by an `-I' switch.
819 `-I-' is deprecated, `-iquote' should be used instead.
821 `-I. -I-' is not the same as no `-I' options at all, and does not
822 cause the same behavior for `<>' includes that `""' includes get with
823 no special options. `-I.' searches the compiler's current working
824 directory for header files. That may or may not be the same as the
825 directory containing the current file.
827 If you need to look for headers in a directory named `-', write
830 There are several more ways to adjust the header search path. They
831 are generally less useful. *Note Invocation::.
834 File: cpp.info, Node: Once-Only Headers, Next: Computed Includes, Prev: Search Path, Up: Header Files
836 2.4 Once-Only Headers
837 =====================
839 If a header file happens to be included twice, the compiler will process
840 its contents twice. This is very likely to cause an error, e.g. when
841 the compiler sees the same structure definition twice. Even if it does
842 not, it will certainly waste time.
844 The standard way to prevent this is to enclose the entire real
845 contents of the file in a conditional, like this:
848 #ifndef FILE_FOO_SEEN
849 #define FILE_FOO_SEEN
853 #endif /* !FILE_FOO_SEEN */
855 This construct is commonly known as a "wrapper #ifndef". When the
856 header is included again, the conditional will be false, because
857 `FILE_FOO_SEEN' is defined. The preprocessor will skip over the entire
858 contents of the file, and the compiler will not see it twice.
860 CPP optimizes even further. It remembers when a header file has a
861 wrapper `#ifndef'. If a subsequent `#include' specifies that header,
862 and the macro in the `#ifndef' is still defined, it does not bother to
863 rescan the file at all.
865 You can put comments outside the wrapper. They will not interfere
866 with this optimization.
868 The macro `FILE_FOO_SEEN' is called the "controlling macro" or
869 "guard macro". In a user header file, the macro name should not begin
870 with `_'. In a system header file, it should begin with `__' to avoid
871 conflicts with user programs. In any kind of header file, the macro
872 name should contain the name of the file and some additional text, to
873 avoid conflicts with other header files.
876 File: cpp.info, Node: Computed Includes, Next: Wrapper Headers, Prev: Once-Only Headers, Up: Header Files
878 2.5 Computed Includes
879 =====================
881 Sometimes it is necessary to select one of several different header
882 files to be included into your program. They might specify
883 configuration parameters to be used on different sorts of operating
884 systems, for instance. You could do this with a series of conditionals,
887 # include "system_1.h"
889 # include "system_2.h"
894 That rapidly becomes tedious. Instead, the preprocessor offers the
895 ability to use a macro for the header name. This is called a "computed
896 include". Instead of writing a header name as the direct argument of
897 `#include', you simply put a macro name there instead:
899 #define SYSTEM_H "system_1.h"
903 `SYSTEM_H' will be expanded, and the preprocessor will look for
904 `system_1.h' as if the `#include' had been written that way originally.
905 `SYSTEM_H' could be defined by your Makefile with a `-D' option.
907 You must be careful when you define the macro. `#define' saves
908 tokens, not text. The preprocessor has no way of knowing that the macro
909 will be used as the argument of `#include', so it generates ordinary
910 tokens, not a header name. This is unlikely to cause problems if you
911 use double-quote includes, which are close enough to string constants.
912 If you use angle brackets, however, you may have trouble.
914 The syntax of a computed include is actually a bit more general than
915 the above. If the first non-whitespace character after `#include' is
916 not `"' or `<', then the entire line is macro-expanded like running
919 If the line expands to a single string constant, the contents of that
920 string constant are the file to be included. CPP does not re-examine
921 the string for embedded quotes, but neither does it process backslash
922 escapes in the string. Therefore
924 #define HEADER "a\"b"
927 looks for a file named `a\"b'. CPP searches for the file according to
928 the rules for double-quoted includes.
930 If the line expands to a token stream beginning with a `<' token and
931 including a `>' token, then the tokens between the `<' and the first
932 `>' are combined to form the filename to be included. Any whitespace
933 between tokens is reduced to a single space; then any space after the
934 initial `<' is retained, but a trailing space before the closing `>' is
935 ignored. CPP searches for the file according to the rules for
936 angle-bracket includes.
938 In either case, if there are any tokens on the line after the file
939 name, an error occurs and the directive is not processed. It is also
940 an error if the result of expansion does not match either of the two
943 These rules are implementation-defined behavior according to the C
944 standard. To minimize the risk of different compilers interpreting your
945 computed includes differently, we recommend you use only a single
946 object-like macro which expands to a string constant. This will also
947 minimize confusion for people reading your program.
950 File: cpp.info, Node: Wrapper Headers, Next: System Headers, Prev: Computed Includes, Up: Header Files
955 Sometimes it is necessary to adjust the contents of a system-provided
956 header file without editing it directly. GCC's `fixincludes' operation
957 does this, for example. One way to do that would be to create a new
958 header file with the same name and insert it in the search path before
959 the original header. That works fine as long as you're willing to
960 replace the old header entirely. But what if you want to refer to the
961 old header from the new one?
963 You cannot simply include the old header with `#include'. That will
964 start from the beginning, and find your new header again. If your
965 header is not protected from multiple inclusion (*note Once-Only
966 Headers::), it will recurse infinitely and cause a fatal error.
968 You could include the old header with an absolute pathname:
969 #include "/usr/include/old-header.h"
970 This works, but is not clean; should the system headers ever move,
971 you would have to edit the new headers to match.
973 There is no way to solve this problem within the C standard, but you
974 can use the GNU extension `#include_next'. It means, "Include the
975 _next_ file with this name". This directive works like `#include'
976 except in searching for the specified file: it starts searching the
977 list of header file directories _after_ the directory in which the
978 current file was found.
980 Suppose you specify `-I /usr/local/include', and the list of
981 directories to search also includes `/usr/include'; and suppose both
982 directories contain `signal.h'. Ordinary `#include <signal.h>' finds
983 the file under `/usr/local/include'. If that file contains
984 `#include_next <signal.h>', it starts searching after that directory,
985 and finds the file in `/usr/include'.
987 `#include_next' does not distinguish between `<FILE>' and `"FILE"'
988 inclusion, nor does it check that the file you specify has the same
989 name as the current file. It simply looks for the file named, starting
990 with the directory in the search path after the one where the current
993 The use of `#include_next' can lead to great confusion. We
994 recommend it be used only when there is no other alternative. In
995 particular, it should not be used in the headers belonging to a specific
996 program; it should be used only to make global corrections along the
997 lines of `fixincludes'.
1000 File: cpp.info, Node: System Headers, Prev: Wrapper Headers, Up: Header Files
1005 The header files declaring interfaces to the operating system and
1006 runtime libraries often cannot be written in strictly conforming C.
1007 Therefore, GCC gives code found in "system headers" special treatment.
1008 All warnings, other than those generated by `#warning' (*note
1009 Diagnostics::), are suppressed while GCC is processing a system header.
1010 Macros defined in a system header are immune to a few warnings
1011 wherever they are expanded. This immunity is granted on an ad-hoc
1012 basis, when we find that a warning generates lots of false positives
1013 because of code in macros defined in system headers.
1015 Normally, only the headers found in specific directories are
1016 considered system headers. These directories are determined when GCC
1017 is compiled. There are, however, two ways to make normal headers into
1020 The `-isystem' command line option adds its argument to the list of
1021 directories to search for headers, just like `-I'. Any headers found
1022 in that directory will be considered system headers.
1024 All directories named by `-isystem' are searched _after_ all
1025 directories named by `-I', no matter what their order was on the
1026 command line. If the same directory is named by both `-I' and
1027 `-isystem', the `-I' option is ignored. GCC provides an informative
1028 message when this occurs if `-v' is used.
1030 There is also a directive, `#pragma GCC system_header', which tells
1031 GCC to consider the rest of the current include file a system header,
1032 no matter where it was found. Code that comes before the `#pragma' in
1033 the file will not be affected. `#pragma GCC system_header' has no
1034 effect in the primary source file.
1036 On very old systems, some of the pre-defined system header
1037 directories get even more special treatment. GNU C++ considers code in
1038 headers found in those directories to be surrounded by an `extern "C"'
1039 block. There is no way to request this behavior with a `#pragma', or
1040 from the command line.
1043 File: cpp.info, Node: Macros, Next: Conditionals, Prev: Header Files, Up: Top
1048 A "macro" is a fragment of code which has been given a name. Whenever
1049 the name is used, it is replaced by the contents of the macro. There
1050 are two kinds of macros. They differ mostly in what they look like
1051 when they are used. "Object-like" macros resemble data objects when
1052 used, "function-like" macros resemble function calls.
1054 You may define any valid identifier as a macro, even if it is a C
1055 keyword. The preprocessor does not know anything about keywords. This
1056 can be useful if you wish to hide a keyword such as `const' from an
1057 older compiler that does not understand it. However, the preprocessor
1058 operator `defined' (*note Defined::) can never be defined as a macro,
1059 and C++'s named operators (*note C++ Named Operators::) cannot be
1060 macros when you are compiling C++.
1064 * Object-like Macros::
1065 * Function-like Macros::
1070 * Predefined Macros::
1071 * Undefining and Redefining Macros::
1072 * Directives Within Macro Arguments::
1076 File: cpp.info, Node: Object-like Macros, Next: Function-like Macros, Up: Macros
1078 3.1 Object-like Macros
1079 ======================
1081 An "object-like macro" is a simple identifier which will be replaced by
1082 a code fragment. It is called object-like because it looks like a data
1083 object in code that uses it. They are most commonly used to give
1084 symbolic names to numeric constants.
1086 You create macros with the `#define' directive. `#define' is
1087 followed by the name of the macro and then the token sequence it should
1088 be an abbreviation for, which is variously referred to as the macro's
1089 "body", "expansion" or "replacement list". For example,
1091 #define BUFFER_SIZE 1024
1093 defines a macro named `BUFFER_SIZE' as an abbreviation for the token
1094 `1024'. If somewhere after this `#define' directive there comes a C
1095 statement of the form
1097 foo = (char *) malloc (BUFFER_SIZE);
1099 then the C preprocessor will recognize and "expand" the macro
1100 `BUFFER_SIZE'. The C compiler will see the same tokens as it would if
1103 foo = (char *) malloc (1024);
1105 By convention, macro names are written in uppercase. Programs are
1106 easier to read when it is possible to tell at a glance which names are
1109 The macro's body ends at the end of the `#define' line. You may
1110 continue the definition onto multiple lines, if necessary, using
1111 backslash-newline. When the macro is expanded, however, it will all
1112 come out on one line. For example,
1114 #define NUMBERS 1, \
1117 int x[] = { NUMBERS };
1118 ==> int x[] = { 1, 2, 3 };
1120 The most common visible consequence of this is surprising line numbers
1123 There is no restriction on what can go in a macro body provided it
1124 decomposes into valid preprocessing tokens. Parentheses need not
1125 balance, and the body need not resemble valid C code. (If it does not,
1126 you may get error messages from the C compiler when you use the macro.)
1128 The C preprocessor scans your program sequentially. Macro
1129 definitions take effect at the place you write them. Therefore, the
1130 following input to the C preprocessor
1141 When the preprocessor expands a macro name, the macro's expansion
1142 replaces the macro invocation, then the expansion is examined for more
1143 macros to expand. For example,
1145 #define TABLESIZE BUFSIZE
1146 #define BUFSIZE 1024
1151 `TABLESIZE' is expanded first to produce `BUFSIZE', then that macro is
1152 expanded to produce the final result, `1024'.
1154 Notice that `BUFSIZE' was not defined when `TABLESIZE' was defined.
1155 The `#define' for `TABLESIZE' uses exactly the expansion you
1156 specify--in this case, `BUFSIZE'--and does not check to see whether it
1157 too contains macro names. Only when you _use_ `TABLESIZE' is the
1158 result of its expansion scanned for more macro names.
1160 This makes a difference if you change the definition of `BUFSIZE' at
1161 some point in the source file. `TABLESIZE', defined as shown, will
1162 always expand using the definition of `BUFSIZE' that is currently in
1165 #define BUFSIZE 1020
1166 #define TABLESIZE BUFSIZE
1170 Now `TABLESIZE' expands (in two stages) to `37'.
1172 If the expansion of a macro contains its own name, either directly or
1173 via intermediate macros, it is not expanded again when the expansion is
1174 examined for more macros. This prevents infinite recursion. *Note
1175 Self-Referential Macros::, for the precise details.
1178 File: cpp.info, Node: Function-like Macros, Next: Macro Arguments, Prev: Object-like Macros, Up: Macros
1180 3.2 Function-like Macros
1181 ========================
1183 You can also define macros whose use looks like a function call. These
1184 are called "function-like macros". To define a function-like macro,
1185 you use the same `#define' directive, but you put a pair of parentheses
1186 immediately after the macro name. For example,
1188 #define lang_init() c_init()
1192 A function-like macro is only expanded if its name appears with a
1193 pair of parentheses after it. If you write just the name, it is left
1194 alone. This can be useful when you have a function and a macro of the
1195 same name, and you wish to use the function sometimes.
1197 extern void foo(void);
1198 #define foo() /* optimized inline version */
1203 Here the call to `foo()' will use the macro, but the function
1204 pointer will get the address of the real function. If the macro were to
1205 be expanded, it would cause a syntax error.
1207 If you put spaces between the macro name and the parentheses in the
1208 macro definition, that does not define a function-like macro, it defines
1209 an object-like macro whose expansion happens to begin with a pair of
1212 #define lang_init () c_init()
1216 The first two pairs of parentheses in this expansion come from the
1217 macro. The third is the pair that was originally after the macro
1218 invocation. Since `lang_init' is an object-like macro, it does not
1219 consume those parentheses.
1222 File: cpp.info, Node: Macro Arguments, Next: Stringification, Prev: Function-like Macros, Up: Macros
1227 Function-like macros can take "arguments", just like true functions.
1228 To define a macro that uses arguments, you insert "parameters" between
1229 the pair of parentheses in the macro definition that make the macro
1230 function-like. The parameters must be valid C identifiers, separated
1231 by commas and optionally whitespace.
1233 To invoke a macro that takes arguments, you write the name of the
1234 macro followed by a list of "actual arguments" in parentheses, separated
1235 by commas. The invocation of the macro need not be restricted to a
1236 single logical line--it can cross as many lines in the source file as
1237 you wish. The number of arguments you give must match the number of
1238 parameters in the macro definition. When the macro is expanded, each
1239 use of a parameter in its body is replaced by the tokens of the
1240 corresponding argument. (You need not use all of the parameters in the
1243 As an example, here is a macro that computes the minimum of two
1244 numeric values, as it is defined in many C programs, and some uses.
1246 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
1247 x = min(a, b); ==> x = ((a) < (b) ? (a) : (b));
1248 y = min(1, 2); ==> y = ((1) < (2) ? (1) : (2));
1249 z = min(a + 28, *p); ==> z = ((a + 28) < (*p) ? (a + 28) : (*p));
1251 (In this small example you can already see several of the dangers of
1252 macro arguments. *Note Macro Pitfalls::, for detailed explanations.)
1254 Leading and trailing whitespace in each argument is dropped, and all
1255 whitespace between the tokens of an argument is reduced to a single
1256 space. Parentheses within each argument must balance; a comma within
1257 such parentheses does not end the argument. However, there is no
1258 requirement for square brackets or braces to balance, and they do not
1259 prevent a comma from separating arguments. Thus,
1261 macro (array[x = y, x + 1])
1263 passes two arguments to `macro': `array[x = y' and `x + 1]'. If you
1264 want to supply `array[x = y, x + 1]' as an argument, you can write it
1265 as `array[(x = y, x + 1)]', which is equivalent C code.
1267 All arguments to a macro are completely macro-expanded before they
1268 are substituted into the macro body. After substitution, the complete
1269 text is scanned again for macros to expand, including the arguments.
1270 This rule may seem strange, but it is carefully designed so you need
1271 not worry about whether any function call is actually a macro
1272 invocation. You can run into trouble if you try to be too clever,
1273 though. *Note Argument Prescan::, for detailed discussion.
1275 For example, `min (min (a, b), c)' is first expanded to
1277 min (((a) < (b) ? (a) : (b)), (c))
1281 ((((a) < (b) ? (a) : (b))) < (c)
1282 ? (((a) < (b) ? (a) : (b)))
1285 (Line breaks shown here for clarity would not actually be generated.)
1287 You can leave macro arguments empty; this is not an error to the
1288 preprocessor (but many macros will then expand to invalid code). You
1289 cannot leave out arguments entirely; if a macro takes two arguments,
1290 there must be exactly one comma at the top level of its argument list.
1291 Here are some silly examples using `min':
1293 min(, b) ==> (( ) < (b) ? ( ) : (b))
1294 min(a, ) ==> ((a ) < ( ) ? (a ) : ( ))
1295 min(,) ==> (( ) < ( ) ? ( ) : ( ))
1296 min((,),) ==> (((,)) < ( ) ? ((,)) : ( ))
1298 min() error--> macro "min" requires 2 arguments, but only 1 given
1299 min(,,) error--> macro "min" passed 3 arguments, but takes just 2
1301 Whitespace is not a preprocessing token, so if a macro `foo' takes
1302 one argument, `foo ()' and `foo ( )' both supply it an empty argument.
1303 Previous GNU preprocessor implementations and documentation were
1304 incorrect on this point, insisting that a function-like macro that
1305 takes a single argument be passed a space if an empty argument was
1308 Macro parameters appearing inside string literals are not replaced by
1309 their corresponding actual arguments.
1311 #define foo(x) x, "x"
1312 foo(bar) ==> bar, "x"
1315 File: cpp.info, Node: Stringification, Next: Concatenation, Prev: Macro Arguments, Up: Macros
1320 Sometimes you may want to convert a macro argument into a string
1321 constant. Parameters are not replaced inside string constants, but you
1322 can use the `#' preprocessing operator instead. When a macro parameter
1323 is used with a leading `#', the preprocessor replaces it with the
1324 literal text of the actual argument, converted to a string constant.
1325 Unlike normal parameter replacement, the argument is not macro-expanded
1326 first. This is called "stringification".
1328 There is no way to combine an argument with surrounding text and
1329 stringify it all together. Instead, you can write a series of adjacent
1330 string constants and stringified arguments. The preprocessor will
1331 replace the stringified arguments with string constants. The C
1332 compiler will then combine all the adjacent string constants into one
1335 Here is an example of a macro definition that uses stringification:
1337 #define WARN_IF(EXP) \
1339 fprintf (stderr, "Warning: " #EXP "\n"); } \
1342 ==> do { if (x == 0)
1343 fprintf (stderr, "Warning: " "x == 0" "\n"); } while (0);
1345 The argument for `EXP' is substituted once, as-is, into the `if'
1346 statement, and once, stringified, into the argument to `fprintf'. If
1347 `x' were a macro, it would be expanded in the `if' statement, but not
1350 The `do' and `while (0)' are a kludge to make it possible to write
1351 `WARN_IF (ARG);', which the resemblance of `WARN_IF' to a function
1352 would make C programmers want to do; see *Note Swallowing the
1355 Stringification in C involves more than putting double-quote
1356 characters around the fragment. The preprocessor backslash-escapes the
1357 quotes surrounding embedded string constants, and all backslashes
1358 within string and character constants, in order to get a valid C string
1359 constant with the proper contents. Thus, stringifying `p = "foo\n";'
1360 results in "p = \"foo\\n\";". However, backslashes that are not inside
1361 string or character constants are not duplicated: `\n' by itself
1362 stringifies to "\n".
1364 All leading and trailing whitespace in text being stringified is
1365 ignored. Any sequence of whitespace in the middle of the text is
1366 converted to a single space in the stringified result. Comments are
1367 replaced by whitespace long before stringification happens, so they
1368 never appear in stringified text.
1370 There is no way to convert a macro argument into a character
1373 If you want to stringify the result of expansion of a macro argument,
1374 you have to use two levels of macros.
1376 #define xstr(s) str(s)
1386 `s' is stringified when it is used in `str', so it is not
1387 macro-expanded first. But `s' is an ordinary argument to `xstr', so it
1388 is completely macro-expanded before `xstr' itself is expanded (*note
1389 Argument Prescan::). Therefore, by the time `str' gets to its
1390 argument, it has already been macro-expanded.
1393 File: cpp.info, Node: Concatenation, Next: Variadic Macros, Prev: Stringification, Up: Macros
1398 It is often useful to merge two tokens into one while expanding macros.
1399 This is called "token pasting" or "token concatenation". The `##'
1400 preprocessing operator performs token pasting. When a macro is
1401 expanded, the two tokens on either side of each `##' operator are
1402 combined into a single token, which then replaces the `##' and the two
1403 original tokens in the macro expansion. Usually both will be
1404 identifiers, or one will be an identifier and the other a preprocessing
1405 number. When pasted, they make a longer identifier. This isn't the
1406 only valid case. It is also possible to concatenate two numbers (or a
1407 number and a name, such as `1.5' and `e3') into a number. Also,
1408 multi-character operators such as `+=' can be formed by token pasting.
1410 However, two tokens that don't together form a valid token cannot be
1411 pasted together. For example, you cannot concatenate `x' with `+' in
1412 either order. If you try, the preprocessor issues a warning and emits
1413 the two tokens. Whether it puts white space between the tokens is
1414 undefined. It is common to find unnecessary uses of `##' in complex
1415 macros. If you get this warning, it is likely that you can simply
1418 Both the tokens combined by `##' could come from the macro body, but
1419 you could just as well write them as one token in the first place.
1420 Token pasting is most useful when one or both of the tokens comes from a
1421 macro argument. If either of the tokens next to an `##' is a parameter
1422 name, it is replaced by its actual argument before `##' executes. As
1423 with stringification, the actual argument is not macro-expanded first.
1424 If the argument is empty, that `##' has no effect.
1426 Keep in mind that the C preprocessor converts comments to whitespace
1427 before macros are even considered. Therefore, you cannot create a
1428 comment by concatenating `/' and `*'. You can put as much whitespace
1429 between `##' and its operands as you like, including comments, and you
1430 can put comments in arguments that will be concatenated. However, it
1431 is an error if `##' appears at either end of a macro body.
1433 Consider a C program that interprets named commands. There probably
1434 needs to be a table of commands, perhaps an array of structures declared
1440 void (*function) (void);
1443 struct command commands[] =
1445 { "quit", quit_command },
1446 { "help", help_command },
1450 It would be cleaner not to have to give each command name twice,
1451 once in the string constant and once in the function name. A macro
1452 which takes the name of a command as an argument can make this
1453 unnecessary. The string constant can be created with stringification,
1454 and the function name by concatenating the argument with `_command'.
1455 Here is how it is done:
1457 #define COMMAND(NAME) { #NAME, NAME ## _command }
1459 struct command commands[] =
1467 File: cpp.info, Node: Variadic Macros, Next: Predefined Macros, Prev: Concatenation, Up: Macros
1472 A macro can be declared to accept a variable number of arguments much as
1473 a function can. The syntax for defining the macro is similar to that of
1474 a function. Here is an example:
1476 #define eprintf(...) fprintf (stderr, __VA_ARGS__)
1478 This kind of macro is called "variadic". When the macro is invoked,
1479 all the tokens in its argument list after the last named argument (this
1480 macro has none), including any commas, become the "variable argument".
1481 This sequence of tokens replaces the identifier `__VA_ARGS__' in the
1482 macro body wherever it appears. Thus, we have this expansion:
1484 eprintf ("%s:%d: ", input_file, lineno)
1485 ==> fprintf (stderr, "%s:%d: ", input_file, lineno)
1487 The variable argument is completely macro-expanded before it is
1488 inserted into the macro expansion, just like an ordinary argument. You
1489 may use the `#' and `##' operators to stringify the variable argument
1490 or to paste its leading or trailing token with another token. (But see
1491 below for an important special case for `##'.)
1493 If your macro is complicated, you may want a more descriptive name
1494 for the variable argument than `__VA_ARGS__'. CPP permits this, as an
1495 extension. You may write an argument name immediately before the
1496 `...'; that name is used for the variable argument. The `eprintf'
1497 macro above could be written
1499 #define eprintf(args...) fprintf (stderr, args)
1501 using this extension. You cannot use `__VA_ARGS__' and this extension
1504 You can have named arguments as well as variable arguments in a
1505 variadic macro. We could define `eprintf' like this, instead:
1507 #define eprintf(format, ...) fprintf (stderr, format, __VA_ARGS__)
1509 This formulation looks more descriptive, but unfortunately it is less
1510 flexible: you must now supply at least one argument after the format
1511 string. In standard C, you cannot omit the comma separating the named
1512 argument from the variable arguments. Furthermore, if you leave the
1513 variable argument empty, you will get a syntax error, because there
1514 will be an extra comma after the format string.
1516 eprintf("success!\n", );
1517 ==> fprintf(stderr, "success!\n", );
1519 GNU CPP has a pair of extensions which deal with this problem.
1520 First, you are allowed to leave the variable argument out entirely:
1522 eprintf ("success!\n")
1523 ==> fprintf(stderr, "success!\n", );
1525 Second, the `##' token paste operator has a special meaning when placed
1526 between a comma and a variable argument. If you write
1528 #define eprintf(format, ...) fprintf (stderr, format, ##__VA_ARGS__)
1530 and the variable argument is left out when the `eprintf' macro is used,
1531 then the comma before the `##' will be deleted. This does _not_ happen
1532 if you pass an empty argument, nor does it happen if the token
1533 preceding `##' is anything other than a comma.
1535 eprintf ("success!\n")
1536 ==> fprintf(stderr, "success!\n");
1538 The above explanation is ambiguous about the case where the only macro
1539 parameter is a variable arguments parameter, as it is meaningless to
1540 try to distinguish whether no argument at all is an empty argument or a
1541 missing argument. In this case the C99 standard is clear that the
1542 comma must remain, however the existing GCC extension used to swallow
1543 the comma. So CPP retains the comma when conforming to a specific C
1544 standard, and drops it otherwise.
1546 C99 mandates that the only place the identifier `__VA_ARGS__' can
1547 appear is in the replacement list of a variadic macro. It may not be
1548 used as a macro name, macro argument name, or within a different type
1549 of macro. It may also be forbidden in open text; the standard is
1550 ambiguous. We recommend you avoid using it except for its defined
1553 Variadic macros are a new feature in C99. GNU CPP has supported them
1554 for a long time, but only with a named variable argument (`args...',
1555 not `...' and `__VA_ARGS__'). If you are concerned with portability to
1556 previous versions of GCC, you should use only named variable arguments.
1557 On the other hand, if you are concerned with portability to other
1558 conforming implementations of C99, you should use only `__VA_ARGS__'.
1560 Previous versions of CPP implemented the comma-deletion extension
1561 much more generally. We have restricted it in this release to minimize
1562 the differences from C99. To get the same effect with both this and
1563 previous versions of GCC, the token preceding the special `##' must be
1564 a comma, and there must be white space between that comma and whatever
1565 comes immediately before it:
1567 #define eprintf(format, args...) fprintf (stderr, format , ##args)
1569 *Note Differences from previous versions::, for the gory details.
1572 File: cpp.info, Node: Predefined Macros, Next: Undefining and Redefining Macros, Prev: Variadic Macros, Up: Macros
1574 3.7 Predefined Macros
1575 =====================
1577 Several object-like macros are predefined; you use them without
1578 supplying their definitions. They fall into three classes: standard,
1579 common, and system-specific.
1581 In C++, there is a fourth category, the named operators. They act
1582 like predefined macros, but you cannot undefine them.
1586 * Standard Predefined Macros::
1587 * Common Predefined Macros::
1588 * System-specific Predefined Macros::
1589 * C++ Named Operators::
1592 File: cpp.info, Node: Standard Predefined Macros, Next: Common Predefined Macros, Up: Predefined Macros
1594 3.7.1 Standard Predefined Macros
1595 --------------------------------
1597 The standard predefined macros are specified by the relevant language
1598 standards, so they are available with all compilers that implement
1599 those standards. Older compilers may not provide all of them. Their
1600 names all start with double underscores.
1603 This macro expands to the name of the current input file, in the
1604 form of a C string constant. This is the path by which the
1605 preprocessor opened the file, not the short name specified in
1606 `#include' or as the input file name argument. For example,
1607 `"/usr/local/include/myheader.h"' is a possible expansion of this
1611 This macro expands to the current input line number, in the form
1612 of a decimal integer constant. While we call it a predefined
1613 macro, it's a pretty strange macro, since its "definition" changes
1614 with each new line of source code.
1616 `__FILE__' and `__LINE__' are useful in generating an error message
1617 to report an inconsistency detected by the program; the message can
1618 state the source line at which the inconsistency was detected. For
1621 fprintf (stderr, "Internal error: "
1622 "negative string length "
1623 "%d at %s, line %d.",
1624 length, __FILE__, __LINE__);
1626 An `#include' directive changes the expansions of `__FILE__' and
1627 `__LINE__' to correspond to the included file. At the end of that
1628 file, when processing resumes on the input file that contained the
1629 `#include' directive, the expansions of `__FILE__' and `__LINE__'
1630 revert to the values they had before the `#include' (but `__LINE__' is
1631 then incremented by one as processing moves to the line after the
1634 A `#line' directive changes `__LINE__', and may change `__FILE__' as
1635 well. *Note Line Control::.
1637 C99 introduces `__func__', and GCC has provided `__FUNCTION__' for a
1638 long time. Both of these are strings containing the name of the
1639 current function (there are slight semantic differences; see the GCC
1640 manual). Neither of them is a macro; the preprocessor does not know the
1641 name of the current function. They tend to be useful in conjunction
1642 with `__FILE__' and `__LINE__', though.
1645 This macro expands to a string constant that describes the date on
1646 which the preprocessor is being run. The string constant contains
1647 eleven characters and looks like `"Feb 12 1996"'. If the day of
1648 the month is less than 10, it is padded with a space on the left.
1650 If GCC cannot determine the current date, it will emit a warning
1651 message (once per compilation) and `__DATE__' will expand to
1655 This macro expands to a string constant that describes the time at
1656 which the preprocessor is being run. The string constant contains
1657 eight characters and looks like `"23:59:01"'.
1659 If GCC cannot determine the current time, it will emit a warning
1660 message (once per compilation) and `__TIME__' will expand to
1664 In normal operation, this macro expands to the constant 1, to
1665 signify that this compiler conforms to ISO Standard C. If GNU CPP
1666 is used with a compiler other than GCC, this is not necessarily
1667 true; however, the preprocessor always conforms to the standard
1668 unless the `-traditional-cpp' option is used.
1670 This macro is not defined if the `-traditional-cpp' option is used.
1672 On some hosts, the system compiler uses a different convention,
1673 where `__STDC__' is normally 0, but is 1 if the user specifies
1674 strict conformance to the C Standard. CPP follows the host
1675 convention when processing system header files, but when
1676 processing user files `__STDC__' is always 1. This has been
1677 reported to cause problems; for instance, some versions of Solaris
1678 provide X Windows headers that expect `__STDC__' to be either
1679 undefined or 1. *Note Invocation::.
1682 This macro expands to the C Standard's version number, a long
1683 integer constant of the form `YYYYMML' where YYYY and MM are the
1684 year and month of the Standard version. This signifies which
1685 version of the C Standard the compiler conforms to. Like
1686 `__STDC__', this is not necessarily accurate for the entire
1687 implementation, unless GNU CPP is being used with GCC.
1689 The value `199409L' signifies the 1989 C standard as amended in
1690 1994, which is the current default; the value `199901L' signifies
1691 the 1999 revision of the C standard. Support for the 1999
1692 revision is not yet complete.
1694 This macro is not defined if the `-traditional-cpp' option is
1695 used, nor when compiling C++ or Objective-C.
1698 This macro is defined, with value 1, if the compiler's target is a
1699 "hosted environment". A hosted environment has the complete
1700 facilities of the standard C library available.
1703 This macro is defined when the C++ compiler is in use. You can use
1704 `__cplusplus' to test whether a header is compiled by a C compiler
1705 or a C++ compiler. This macro is similar to `__STDC_VERSION__', in
1706 that it expands to a version number. A fully conforming
1707 implementation of the 1998 C++ standard will define this macro to
1708 `199711L'. The GNU C++ compiler is not yet fully conforming, so
1709 it uses `1' instead. It is hoped to complete the implementation
1710 of standard C++ in the near future.
1713 This macro is defined, with value 1, when the Objective-C compiler
1714 is in use. You can use `__OBJC__' to test whether a header is
1715 compiled by a C compiler or a Objective-C compiler.
1718 This macro is defined with value 1 when preprocessing assembly
1723 File: cpp.info, Node: Common Predefined Macros, Next: System-specific Predefined Macros, Prev: Standard Predefined Macros, Up: Predefined Macros
1725 3.7.2 Common Predefined Macros
1726 ------------------------------
1728 The common predefined macros are GNU C extensions. They are available
1729 with the same meanings regardless of the machine or operating system on
1730 which you are using GNU C. Their names all start with double
1735 `__GNUC_PATCHLEVEL__'
1736 These macros are defined by all GNU compilers that use the C
1737 preprocessor: C, C++, and Objective-C. Their values are the major
1738 version, minor version, and patch level of the compiler, as integer
1739 constants. For example, GCC 3.2.1 will define `__GNUC__' to 3,
1740 `__GNUC_MINOR__' to 2, and `__GNUC_PATCHLEVEL__' to 1. These
1741 macros are also defined if you invoke the preprocessor directly.
1743 `__GNUC_PATCHLEVEL__' is new to GCC 3.0; it is also present in the
1744 widely-used development snapshots leading up to 3.0 (which identify
1745 themselves as GCC 2.96 or 2.97, depending on which snapshot you
1748 If all you need to know is whether or not your program is being
1749 compiled by GCC, or a non-GCC compiler that claims to accept the
1750 GNU C dialects, you can simply test `__GNUC__'. If you need to
1751 write code which depends on a specific version, you must be more
1752 careful. Each time the minor version is increased, the patch
1753 level is reset to zero; each time the major version is increased
1754 (which happens rarely), the minor version and patch level are
1755 reset. If you wish to use the predefined macros directly in the
1756 conditional, you will need to write it like this:
1758 /* Test for GCC > 3.2.0 */
1759 #if __GNUC__ > 3 || \
1760 (__GNUC__ == 3 && (__GNUC_MINOR__ > 2 || \
1761 (__GNUC_MINOR__ == 2 && \
1762 __GNUC_PATCHLEVEL__ > 0))
1764 Another approach is to use the predefined macros to calculate a
1765 single number, then compare that against a threshold:
1767 #define GCC_VERSION (__GNUC__ * 10000 \
1768 + __GNUC_MINOR__ * 100 \
1769 + __GNUC_PATCHLEVEL__)
1771 /* Test for GCC > 3.2.0 */
1772 #if GCC_VERSION > 30200
1774 Many people find this form easier to understand.
1777 The GNU C++ compiler defines this. Testing it is equivalent to
1778 testing `(__GNUC__ && __cplusplus)'.
1781 GCC defines this macro if and only if the `-ansi' switch, or a
1782 `-std' switch specifying strict conformance to some version of ISO
1783 C, was specified when GCC was invoked. It is defined to `1'.
1784 This macro exists primarily to direct GNU libc's header files to
1785 restrict their definitions to the minimal set found in the 1989 C
1789 This macro expands to the name of the main input file, in the form
1790 of a C string constant. This is the source file that was specified
1791 on the command line of the preprocessor or C compiler.
1794 This macro expands to a decimal integer constant that represents
1795 the depth of nesting in include files. The value of this macro is
1796 incremented on every `#include' directive and decremented at the
1797 end of every included file. It starts out at 0, it's value within
1798 the base file specified on the command line.
1801 This macro is defined if the target uses the ELF object format.
1804 This macro expands to a string constant which describes the
1805 version of the compiler in use. You should not rely on its
1806 contents having any particular form, but it can be counted on to
1807 contain at least the release number.
1812 These macros describe the compilation mode. `__OPTIMIZE__' is
1813 defined in all optimizing compilations. `__OPTIMIZE_SIZE__' is
1814 defined if the compiler is optimizing for size, not speed.
1815 `__NO_INLINE__' is defined if no functions will be inlined into
1816 their callers (when not optimizing, or when inlining has been
1817 specifically disabled by `-fno-inline').
1819 These macros cause certain GNU header files to provide optimized
1820 definitions, using macros or inline functions, of system library
1821 functions. You should not use these macros in any way unless you
1822 make sure that programs will execute with the same effect whether
1823 or not they are defined. If they are defined, their value is 1.
1826 GCC defines this macro if and only if the data type `char' is
1827 unsigned on the target machine. It exists to cause the standard
1828 header file `limits.h' to work correctly. You should not use this
1829 macro yourself; instead, refer to the standard macros defined in
1832 `__WCHAR_UNSIGNED__'
1833 Like `__CHAR_UNSIGNED__', this macro is defined if and only if the
1834 data type `wchar_t' is unsigned and the front-end is in C++ mode.
1836 `__REGISTER_PREFIX__'
1837 This macro expands to a single token (not a string constant) which
1838 is the prefix applied to CPU register names in assembly language
1839 for this target. You can use it to write assembly that is usable
1840 in multiple environments. For example, in the `m68k-aout'
1841 environment it expands to nothing, but in the `m68k-coff'
1842 environment it expands to a single `%'.
1844 `__USER_LABEL_PREFIX__'
1845 This macro expands to a single token which is the prefix applied to
1846 user labels (symbols visible to C code) in assembly. For example,
1847 in the `m68k-aout' environment it expands to an `_', but in the
1848 `m68k-coff' environment it expands to nothing.
1850 This macro will have the correct definition even if
1851 `-f(no-)underscores' is in use, but it will not be correct if
1852 target-specific options that adjust this prefix are used (e.g. the
1853 OSF/rose `-mno-underscores' option).
1861 These macros are defined to the correct underlying types for the
1862 `size_t', `ptrdiff_t', `wchar_t', `wint_t', `intmax_t', and
1863 `uintmax_t' typedefs, respectively. They exist to make the
1864 standard header files `stddef.h' and `wchar.h' work correctly.
1865 You should not use these macros directly; instead, include the
1866 appropriate headers and use the typedefs.
1869 Defined to the number of bits used in the representation of the
1870 `char' data type. It exists to make the standard header given
1871 numerical limits work correctly. You should not use this macro
1872 directly; instead, include the appropriate headers.
1881 Defined to the maximum value of the `signed char', `wchar_t',
1882 `signed short', `signed int', `signed long', `signed long long',
1883 and `intmax_t' types respectively. They exist to make the
1884 standard header given numerical limits work correctly. You should
1885 not use these macros directly; instead, include the appropriate
1889 This macro is defined, with value 1, when compiling a C++ source
1890 file with warnings about deprecated constructs enabled. These
1891 warnings are enabled by default, but can be disabled with
1895 This macro is defined, with value 1, when compiling a C++ source
1896 file with exceptions enabled. If `-fno-exceptions' was used when
1897 compiling the file, then this macro will not be defined.
1899 `__USING_SJLJ_EXCEPTIONS__'
1900 This macro is defined, with value 1, if the compiler uses the old
1901 mechanism based on `setjmp' and `longjmp' for exception handling.
1904 This macro is defined when compiling a C++ source file. It has the
1905 value 1 if the compiler will use weak symbols, COMDAT sections, or
1906 other similar techniques to collapse symbols with "vague linkage"
1907 that are defined in multiple translation units. If the compiler
1908 will not collapse such symbols, this macro is defined with value
1909 0. In general, user code should not need to make use of this
1910 macro; the purpose of this macro is to ease implementation of the
1911 C++ runtime library provided with G++.
1914 This macro is defined, with value 1, if (and only if) the NeXT
1915 runtime (as in `-fnext-runtime') is in use for Objective-C. If
1916 the GNU runtime is used, this macro is not defined, so that you
1917 can use this macro to determine which runtime (NeXT or GNU) is
1922 These macros are defined, with value 1, if (and only if) the
1923 compilation is for a target where `long int' and pointer both use
1924 64-bits and `int' uses 32-bit.
1927 This macro is defined, with value 1, when `-fstack-protector' is in
1931 This macro is defined, with value 2, when `-fstack-protector-all'
1936 File: cpp.info, Node: System-specific Predefined Macros, Next: C++ Named Operators, Prev: Common Predefined Macros, Up: Predefined Macros
1938 3.7.3 System-specific Predefined Macros
1939 ---------------------------------------
1941 The C preprocessor normally predefines several macros that indicate what
1942 type of system and machine is in use. They are obviously different on
1943 each target supported by GCC. This manual, being for all systems and
1944 machines, cannot tell you what their names are, but you can use `cpp
1945 -dM' to see them all. *Note Invocation::. All system-specific
1946 predefined macros expand to the constant 1, so you can test them with
1947 either `#ifdef' or `#if'.
1949 The C standard requires that all system-specific macros be part of
1950 the "reserved namespace". All names which begin with two underscores,
1951 or an underscore and a capital letter, are reserved for the compiler and
1952 library to use as they wish. However, historically system-specific
1953 macros have had names with no special prefix; for instance, it is common
1954 to find `unix' defined on Unix systems. For all such macros, GCC
1955 provides a parallel macro with two underscores added at the beginning
1956 and the end. If `unix' is defined, `__unix__' will be defined too.
1957 There will never be more than two underscores; the parallel of `_mips'
1960 When the `-ansi' option, or any `-std' option that requests strict
1961 conformance, is given to the compiler, all the system-specific
1962 predefined macros outside the reserved namespace are suppressed. The
1963 parallel macros, inside the reserved namespace, remain defined.
1965 We are slowly phasing out all predefined macros which are outside the
1966 reserved namespace. You should never use them in new programs, and we
1967 encourage you to correct older code to use the parallel macros whenever
1968 you find it. We don't recommend you use the system-specific macros that
1969 are in the reserved namespace, either. It is better in the long run to
1970 check specifically for features you need, using a tool such as
1974 File: cpp.info, Node: C++ Named Operators, Prev: System-specific Predefined Macros, Up: Predefined Macros
1976 3.7.4 C++ Named Operators
1977 -------------------------
1979 In C++, there are eleven keywords which are simply alternate spellings
1980 of operators normally written with punctuation. These keywords are
1981 treated as such even in the preprocessor. They function as operators in
1982 `#if', and they cannot be defined as macros or poisoned. In C, you can
1983 request that those keywords take their C++ meaning by including
1984 `iso646.h'. That header defines each one as a normal object-like macro
1985 expanding to the appropriate punctuator.
1987 These are the named operators and their corresponding punctuators:
1989 Named Operator Punctuator
2003 File: cpp.info, Node: Undefining and Redefining Macros, Next: Directives Within Macro Arguments, Prev: Predefined Macros, Up: Macros
2005 3.8 Undefining and Redefining Macros
2006 ====================================
2008 If a macro ceases to be useful, it may be "undefined" with the `#undef'
2009 directive. `#undef' takes a single argument, the name of the macro to
2010 undefine. You use the bare macro name, even if the macro is
2011 function-like. It is an error if anything appears on the line after
2012 the macro name. `#undef' has no effect if the name is not a macro.
2017 x = FOO; ==> x = FOO;
2019 Once a macro has been undefined, that identifier may be "redefined"
2020 as a macro by a subsequent `#define' directive. The new definition
2021 need not have any resemblance to the old definition.
2023 However, if an identifier which is currently a macro is redefined,
2024 then the new definition must be "effectively the same" as the old one.
2025 Two macro definitions are effectively the same if:
2026 * Both are the same type of macro (object- or function-like).
2028 * All the tokens of the replacement list are the same.
2030 * If there are any parameters, they are the same.
2032 * Whitespace appears in the same places in both. It need not be
2033 exactly the same amount of whitespace, though. Remember that
2034 comments count as whitespace.
2036 These definitions are effectively the same:
2037 #define FOUR (2 + 2)
2038 #define FOUR (2 + 2)
2039 #define FOUR (2 /* two */ + 2)
2041 #define FOUR (2 + 2)
2042 #define FOUR ( 2+2 )
2043 #define FOUR (2 * 2)
2044 #define FOUR(score,and,seven,years,ago) (2 + 2)
2046 If a macro is redefined with a definition that is not effectively the
2047 same as the old one, the preprocessor issues a warning and changes the
2048 macro to use the new definition. If the new definition is effectively
2049 the same, the redefinition is silently ignored. This allows, for
2050 instance, two different headers to define a common macro. The
2051 preprocessor will only complain if the definitions do not match.
2054 File: cpp.info, Node: Directives Within Macro Arguments, Next: Macro Pitfalls, Prev: Undefining and Redefining Macros, Up: Macros
2056 3.9 Directives Within Macro Arguments
2057 =====================================
2059 Occasionally it is convenient to use preprocessor directives within the
2060 arguments of a macro. The C and C++ standards declare that behavior in
2061 these cases is undefined.
2063 Versions of CPP prior to 3.2 would reject such constructs with an
2064 error message. This was the only syntactic difference between normal
2065 functions and function-like macros, so it seemed attractive to remove
2066 this limitation, and people would often be surprised that they could
2067 not use macros in this way. Moreover, sometimes people would use
2068 conditional compilation in the argument list to a normal library
2069 function like `printf', only to find that after a library upgrade
2070 `printf' had changed to be a function-like macro, and their code would
2071 no longer compile. So from version 3.2 we changed CPP to successfully
2072 process arbitrary directives within macro arguments in exactly the same
2073 way as it would have processed the directive were the function-like
2074 macro invocation not present.
2076 If, within a macro invocation, that macro is redefined, then the new
2077 definition takes effect in time for argument pre-expansion, but the
2078 original definition is still used for argument replacement. Here is a
2079 pathological example:
2091 with the semantics described above.
2094 File: cpp.info, Node: Macro Pitfalls, Prev: Directives Within Macro Arguments, Up: Macros
2099 In this section we describe some special rules that apply to macros and
2100 macro expansion, and point out certain cases in which the rules have
2101 counter-intuitive consequences that you must watch out for.
2106 * Operator Precedence Problems::
2107 * Swallowing the Semicolon::
2108 * Duplication of Side Effects::
2109 * Self-Referential Macros::
2110 * Argument Prescan::
2111 * Newlines in Arguments::
2114 File: cpp.info, Node: Misnesting, Next: Operator Precedence Problems, Up: Macro Pitfalls
2119 When a macro is called with arguments, the arguments are substituted
2120 into the macro body and the result is checked, together with the rest of
2121 the input file, for more macro calls. It is possible to piece together
2122 a macro call coming partially from the macro body and partially from the
2123 arguments. For example,
2125 #define twice(x) (2*(x))
2126 #define call_with_1(x) x(1)
2131 Macro definitions do not have to have balanced parentheses. By
2132 writing an unbalanced open parenthesis in a macro body, it is possible
2133 to create a macro call that begins inside the macro body but ends
2134 outside of it. For example,
2136 #define strange(file) fprintf (file, "%s %d",
2138 strange(stderr) p, 35)
2139 ==> fprintf (stderr, "%s %d", p, 35)
2141 The ability to piece together a macro call can be useful, but the
2142 use of unbalanced open parentheses in a macro body is just confusing,
2143 and should be avoided.
2146 File: cpp.info, Node: Operator Precedence Problems, Next: Swallowing the Semicolon, Prev: Misnesting, Up: Macro Pitfalls
2148 3.10.2 Operator Precedence Problems
2149 -----------------------------------
2151 You may have noticed that in most of the macro definition examples shown
2152 above, each occurrence of a macro argument name had parentheses around
2153 it. In addition, another pair of parentheses usually surround the
2154 entire macro definition. Here is why it is best to write macros that
2157 Suppose you define a macro as follows,
2159 #define ceil_div(x, y) (x + y - 1) / y
2161 whose purpose is to divide, rounding up. (One use for this operation is
2162 to compute how many `int' objects are needed to hold a certain number
2163 of `char' objects.) Then suppose it is used as follows:
2165 a = ceil_div (b & c, sizeof (int));
2166 ==> a = (b & c + sizeof (int) - 1) / sizeof (int);
2168 This does not do what is intended. The operator-precedence rules of C
2169 make it equivalent to this:
2171 a = (b & (c + sizeof (int) - 1)) / sizeof (int);
2173 What we want is this:
2175 a = ((b & c) + sizeof (int) - 1)) / sizeof (int);
2177 Defining the macro as
2179 #define ceil_div(x, y) ((x) + (y) - 1) / (y)
2181 provides the desired result.
2183 Unintended grouping can result in another way. Consider `sizeof
2184 ceil_div(1, 2)'. That has the appearance of a C expression that would
2185 compute the size of the type of `ceil_div (1, 2)', but in fact it means
2186 something very different. Here is what it expands to:
2188 sizeof ((1) + (2) - 1) / (2)
2190 This would take the size of an integer and divide it by two. The
2191 precedence rules have put the division outside the `sizeof' when it was
2192 intended to be inside.
2194 Parentheses around the entire macro definition prevent such problems.
2195 Here, then, is the recommended way to define `ceil_div':
2197 #define ceil_div(x, y) (((x) + (y) - 1) / (y))
2200 File: cpp.info, Node: Swallowing the Semicolon, Next: Duplication of Side Effects, Prev: Operator Precedence Problems, Up: Macro Pitfalls
2202 3.10.3 Swallowing the Semicolon
2203 -------------------------------
2205 Often it is desirable to define a macro that expands into a compound
2206 statement. Consider, for example, the following macro, that advances a
2207 pointer (the argument `p' says where to find it) across whitespace
2210 #define SKIP_SPACES(p, limit) \
2211 { char *lim = (limit); \
2213 if (*p++ != ' ') { \
2216 Here backslash-newline is used to split the macro definition, which must
2217 be a single logical line, so that it resembles the way such code would
2218 be laid out if not part of a macro definition.
2220 A call to this macro might be `SKIP_SPACES (p, lim)'. Strictly
2221 speaking, the call expands to a compound statement, which is a complete
2222 statement with no need for a semicolon to end it. However, since it
2223 looks like a function call, it minimizes confusion if you can use it
2224 like a function call, writing a semicolon afterward, as in `SKIP_SPACES
2227 This can cause trouble before `else' statements, because the
2228 semicolon is actually a null statement. Suppose you write
2231 SKIP_SPACES (p, lim);
2234 The presence of two statements--the compound statement and a null
2235 statement--in between the `if' condition and the `else' makes invalid C
2238 The definition of the macro `SKIP_SPACES' can be altered to solve
2239 this problem, using a `do ... while' statement. Here is how:
2241 #define SKIP_SPACES(p, limit) \
2242 do { char *lim = (limit); \
2244 if (*p++ != ' ') { \
2248 Now `SKIP_SPACES (p, lim);' expands into
2252 which is one statement. The loop executes exactly once; most compilers
2253 generate no extra code for it.
2256 File: cpp.info, Node: Duplication of Side Effects, Next: Self-Referential Macros, Prev: Swallowing the Semicolon, Up: Macro Pitfalls
2258 3.10.4 Duplication of Side Effects
2259 ----------------------------------
2261 Many C programs define a macro `min', for "minimum", like this:
2263 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
2265 When you use this macro with an argument containing a side effect,
2268 next = min (x + y, foo (z));
2270 it expands as follows:
2272 next = ((x + y) < (foo (z)) ? (x + y) : (foo (z)));
2274 where `x + y' has been substituted for `X' and `foo (z)' for `Y'.
2276 The function `foo' is used only once in the statement as it appears
2277 in the program, but the expression `foo (z)' has been substituted twice
2278 into the macro expansion. As a result, `foo' might be called two times
2279 when the statement is executed. If it has side effects or if it takes
2280 a long time to compute, the results might not be what you intended. We
2281 say that `min' is an "unsafe" macro.
2283 The best solution to this problem is to define `min' in a way that
2284 computes the value of `foo (z)' only once. The C language offers no
2285 standard way to do this, but it can be done with GNU extensions as
2289 ({ typeof (X) x_ = (X); \
2290 typeof (Y) y_ = (Y); \
2291 (x_ < y_) ? x_ : y_; })
2293 The `({ ... })' notation produces a compound statement that acts as
2294 an expression. Its value is the value of its last statement. This
2295 permits us to define local variables and assign each argument to one.
2296 The local variables have underscores after their names to reduce the
2297 risk of conflict with an identifier of wider scope (it is impossible to
2298 avoid this entirely). Now each argument is evaluated exactly once.
2300 If you do not wish to use GNU C extensions, the only solution is to
2301 be careful when _using_ the macro `min'. For example, you can
2302 calculate the value of `foo (z)', save it in a variable, and use that
2305 #define min(X, Y) ((X) < (Y) ? (X) : (Y))
2309 next = min (x + y, tem);
2312 (where we assume that `foo' returns type `int').
2315 File: cpp.info, Node: Self-Referential Macros, Next: Argument Prescan, Prev: Duplication of Side Effects, Up: Macro Pitfalls
2317 3.10.5 Self-Referential Macros
2318 ------------------------------
2320 A "self-referential" macro is one whose name appears in its definition.
2321 Recall that all macro definitions are rescanned for more macros to
2322 replace. If the self-reference were considered a use of the macro, it
2323 would produce an infinitely large expansion. To prevent this, the
2324 self-reference is not considered a macro call. It is passed into the
2325 preprocessor output unchanged. Consider an example:
2327 #define foo (4 + foo)
2329 where `foo' is also a variable in your program.
2331 Following the ordinary rules, each reference to `foo' will expand
2332 into `(4 + foo)'; then this will be rescanned and will expand into `(4
2333 + (4 + foo))'; and so on until the computer runs out of memory.
2335 The self-reference rule cuts this process short after one step, at
2336 `(4 + foo)'. Therefore, this macro definition has the possibly useful
2337 effect of causing the program to add 4 to the value of `foo' wherever
2338 `foo' is referred to.
2340 In most cases, it is a bad idea to take advantage of this feature. A
2341 person reading the program who sees that `foo' is a variable will not
2342 expect that it is a macro as well. The reader will come across the
2343 identifier `foo' in the program and think its value should be that of
2344 the variable `foo', whereas in fact the value is four greater.
2346 One common, useful use of self-reference is to create a macro which
2347 expands to itself. If you write
2351 then the macro `EPERM' expands to `EPERM'. Effectively, it is left
2352 alone by the preprocessor whenever it's used in running text. You can
2353 tell that it's a macro with `#ifdef'. You might do this if you want to
2354 define numeric constants with an `enum', but have `#ifdef' be true for
2357 If a macro `x' expands to use a macro `y', and the expansion of `y'
2358 refers to the macro `x', that is an "indirect self-reference" of `x'.
2359 `x' is not expanded in this case either. Thus, if we have
2364 then `x' and `y' expand as follows:
2372 Each macro is expanded when it appears in the definition of the other
2373 macro, but not when it indirectly appears in its own definition.
2376 File: cpp.info, Node: Argument Prescan, Next: Newlines in Arguments, Prev: Self-Referential Macros, Up: Macro Pitfalls
2378 3.10.6 Argument Prescan
2379 -----------------------
2381 Macro arguments are completely macro-expanded before they are
2382 substituted into a macro body, unless they are stringified or pasted
2383 with other tokens. After substitution, the entire macro body, including
2384 the substituted arguments, is scanned again for macros to be expanded.
2385 The result is that the arguments are scanned _twice_ to expand macro
2388 Most of the time, this has no effect. If the argument contained any
2389 macro calls, they are expanded during the first scan. The result
2390 therefore contains no macro calls, so the second scan does not change
2391 it. If the argument were substituted as given, with no prescan, the
2392 single remaining scan would find the same macro calls and produce the
2395 You might expect the double scan to change the results when a
2396 self-referential macro is used in an argument of another macro (*note
2397 Self-Referential Macros::): the self-referential macro would be
2398 expanded once in the first scan, and a second time in the second scan.
2399 However, this is not what happens. The self-references that do not
2400 expand in the first scan are marked so that they will not expand in the
2403 You might wonder, "Why mention the prescan, if it makes no
2404 difference? And why not skip it and make the preprocessor faster?"
2405 The answer is that the prescan does make a difference in three special
2408 * Nested calls to a macro.
2410 We say that "nested" calls to a macro occur when a macro's argument
2411 contains a call to that very macro. For example, if `f' is a macro
2412 that expects one argument, `f (f (1))' is a nested pair of calls to
2413 `f'. The desired expansion is made by expanding `f (1)' and
2414 substituting that into the definition of `f'. The prescan causes
2415 the expected result to happen. Without the prescan, `f (1)' itself
2416 would be substituted as an argument, and the inner use of `f' would
2417 appear during the main scan as an indirect self-reference and
2418 would not be expanded.
2420 * Macros that call other macros that stringify or concatenate.
2422 If an argument is stringified or concatenated, the prescan does not
2423 occur. If you _want_ to expand a macro, then stringify or
2424 concatenate its expansion, you can do that by causing one macro to
2425 call another macro that does the stringification or concatenation.
2426 For instance, if you have
2428 #define AFTERX(x) X_ ## x
2429 #define XAFTERX(x) AFTERX(x)
2430 #define TABLESIZE 1024
2431 #define BUFSIZE TABLESIZE
2433 then `AFTERX(BUFSIZE)' expands to `X_BUFSIZE', and
2434 `XAFTERX(BUFSIZE)' expands to `X_1024'. (Not to `X_TABLESIZE'.
2435 Prescan always does a complete expansion.)
2437 * Macros used in arguments, whose expansions contain unshielded
2440 This can cause a macro expanded on the second scan to be called
2441 with the wrong number of arguments. Here is an example:
2444 #define bar(x) lose(x)
2445 #define lose(x) (1 + (x))
2447 We would like `bar(foo)' to turn into `(1 + (foo))', which would
2448 then turn into `(1 + (a,b))'. Instead, `bar(foo)' expands into
2449 `lose(a,b)', and you get an error because `lose' requires a single
2450 argument. In this case, the problem is easily solved by the same
2451 parentheses that ought to be used to prevent misnesting of
2452 arithmetic operations:
2456 #define bar(x) lose((x))
2458 The extra pair of parentheses prevents the comma in `foo''s
2459 definition from being interpreted as an argument separator.
2463 File: cpp.info, Node: Newlines in Arguments, Prev: Argument Prescan, Up: Macro Pitfalls
2465 3.10.7 Newlines in Arguments
2466 ----------------------------
2468 The invocation of a function-like macro can extend over many logical
2469 lines. However, in the present implementation, the entire expansion
2470 comes out on one line. Thus line numbers emitted by the compiler or
2471 debugger refer to the line the invocation started on, which might be
2472 different to the line containing the argument causing the problem.
2474 Here is an example illustrating this:
2476 #define ignore_second_arg(a,b,c) a; c
2478 ignore_second_arg (foo (),
2482 The syntax error triggered by the tokens `syntax error' results in an
2483 error message citing line three--the line of ignore_second_arg-- even
2484 though the problematic code comes from line five.
2486 We consider this a bug, and intend to fix it in the near future.
2489 File: cpp.info, Node: Conditionals, Next: Diagnostics, Prev: Macros, Up: Top
2494 A "conditional" is a directive that instructs the preprocessor to
2495 select whether or not to include a chunk of code in the final token
2496 stream passed to the compiler. Preprocessor conditionals can test
2497 arithmetic expressions, or whether a name is defined as a macro, or both
2498 simultaneously using the special `defined' operator.
2500 A conditional in the C preprocessor resembles in some ways an `if'
2501 statement in C, but it is important to understand the difference between
2502 them. The condition in an `if' statement is tested during the
2503 execution of your program. Its purpose is to allow your program to
2504 behave differently from run to run, depending on the data it is
2505 operating on. The condition in a preprocessing conditional directive is
2506 tested when your program is compiled. Its purpose is to allow different
2507 code to be included in the program depending on the situation at the
2508 time of compilation.
2510 However, the distinction is becoming less clear. Modern compilers
2511 often do test `if' statements when a program is compiled, if their
2512 conditions are known not to vary at run time, and eliminate code which
2513 can never be executed. If you can count on your compiler to do this,
2514 you may find that your program is more readable if you use `if'
2515 statements with constant conditions (perhaps determined by macros). Of
2516 course, you can only use this to exclude code, not type definitions or
2517 other preprocessing directives, and you can only do it if the code
2518 remains syntactically valid when it is not to be used.
2520 GCC version 3 eliminates this kind of never-executed code even when
2521 not optimizing. Older versions did it only when optimizing.
2525 * Conditional Uses::
2526 * Conditional Syntax::
2530 File: cpp.info, Node: Conditional Uses, Next: Conditional Syntax, Up: Conditionals
2532 4.1 Conditional Uses
2533 ====================
2535 There are three general reasons to use a conditional.
2537 * A program may need to use different code depending on the machine
2538 or operating system it is to run on. In some cases the code for
2539 one operating system may be erroneous on another operating system;
2540 for example, it might refer to data types or constants that do not
2541 exist on the other system. When this happens, it is not enough to
2542 avoid executing the invalid code. Its mere presence will cause
2543 the compiler to reject the program. With a preprocessing
2544 conditional, the offending code can be effectively excised from
2545 the program when it is not valid.
2547 * You may want to be able to compile the same source file into two
2548 different programs. One version might make frequent time-consuming
2549 consistency checks on its intermediate data, or print the values of
2550 those data for debugging, and the other not.
2552 * A conditional whose condition is always false is one way to
2553 exclude code from the program but keep it as a sort of comment for
2556 Simple programs that do not need system-specific logic or complex
2557 debugging hooks generally will not need to use preprocessing
2561 File: cpp.info, Node: Conditional Syntax, Next: Deleted Code, Prev: Conditional Uses, Up: Conditionals
2563 4.2 Conditional Syntax
2564 ======================
2566 A conditional in the C preprocessor begins with a "conditional
2567 directive": `#if', `#ifdef' or `#ifndef'.
2578 File: cpp.info, Node: Ifdef, Next: If, Up: Conditional Syntax
2583 The simplest sort of conditional is
2591 This block is called a "conditional group". CONTROLLED TEXT will be
2592 included in the output of the preprocessor if and only if MACRO is
2593 defined. We say that the conditional "succeeds" if MACRO is defined,
2594 "fails" if it is not.
2596 The CONTROLLED TEXT inside of a conditional can include
2597 preprocessing directives. They are executed only if the conditional
2598 succeeds. You can nest conditional groups inside other conditional
2599 groups, but they must be completely nested. In other words, `#endif'
2600 always matches the nearest `#ifdef' (or `#ifndef', or `#if'). Also,
2601 you cannot start a conditional group in one file and end it in another.
2603 Even if a conditional fails, the CONTROLLED TEXT inside it is still
2604 run through initial transformations and tokenization. Therefore, it
2605 must all be lexically valid C. Normally the only way this matters is
2606 that all comments and string literals inside a failing conditional group
2607 must still be properly ended.
2609 The comment following the `#endif' is not required, but it is a good
2610 practice if there is a lot of CONTROLLED TEXT, because it helps people
2611 match the `#endif' to the corresponding `#ifdef'. Older programs
2612 sometimes put MACRO directly after the `#endif' without enclosing it in
2613 a comment. This is invalid code according to the C standard. CPP
2614 accepts it with a warning. It never affects which `#ifndef' the
2617 Sometimes you wish to use some code if a macro is _not_ defined.
2618 You can do this by writing `#ifndef' instead of `#ifdef'. One common
2619 use of `#ifndef' is to include code only the first time a header file
2620 is included. *Note Once-Only Headers::.
2622 Macro definitions can vary between compilations for several reasons.
2623 Here are some samples.
2625 * Some macros are predefined on each kind of machine (*note
2626 System-specific Predefined Macros::). This allows you to provide
2627 code specially tuned for a particular machine.
2629 * System header files define more macros, associated with the
2630 features they implement. You can test these macros with
2631 conditionals to avoid using a system feature on a machine where it
2634 * Macros can be defined or undefined with the `-D' and `-U' command
2635 line options when you compile the program. You can arrange to
2636 compile the same source file into two different programs by
2637 choosing a macro name to specify which program you want, writing
2638 conditionals to test whether or how this macro is defined, and
2639 then controlling the state of the macro with command line options,
2640 perhaps set in the Makefile. *Note Invocation::.
2642 * Your program might have a special header file (often called
2643 `config.h') that is adjusted when the program is compiled. It can
2644 define or not define macros depending on the features of the
2645 system and the desired capabilities of the program. The
2646 adjustment can be automated by a tool such as `autoconf', or done
2650 File: cpp.info, Node: If, Next: Defined, Prev: Ifdef, Up: Conditional Syntax
2655 The `#if' directive allows you to test the value of an arithmetic
2656 expression, rather than the mere existence of one macro. Its syntax is
2662 #endif /* EXPRESSION */
2664 EXPRESSION is a C expression of integer type, subject to stringent
2665 restrictions. It may contain
2667 * Integer constants.
2669 * Character constants, which are interpreted as they would be in
2672 * Arithmetic operators for addition, subtraction, multiplication,
2673 division, bitwise operations, shifts, comparisons, and logical
2674 operations (`&&' and `||'). The latter two obey the usual
2675 short-circuiting rules of standard C.
2677 * Macros. All macros in the expression are expanded before actual
2678 computation of the expression's value begins.
2680 * Uses of the `defined' operator, which lets you check whether macros
2681 are defined in the middle of an `#if'.
2683 * Identifiers that are not macros, which are all considered to be the
2684 number zero. This allows you to write `#if MACRO' instead of
2685 `#ifdef MACRO', if you know that MACRO, when defined, will always
2686 have a nonzero value. Function-like macros used without their
2687 function call parentheses are also treated as zero.
2689 In some contexts this shortcut is undesirable. The `-Wundef'
2690 option causes GCC to warn whenever it encounters an identifier
2691 which is not a macro in an `#if'.
2693 The preprocessor does not know anything about types in the language.
2694 Therefore, `sizeof' operators are not recognized in `#if', and neither
2695 are `enum' constants. They will be taken as identifiers which are not
2696 macros, and replaced by zero. In the case of `sizeof', this is likely
2697 to cause the expression to be invalid.
2699 The preprocessor calculates the value of EXPRESSION. It carries out
2700 all calculations in the widest integer type known to the compiler; on
2701 most machines supported by GCC this is 64 bits. This is not the same
2702 rule as the compiler uses to calculate the value of a constant
2703 expression, and may give different results in some cases. If the value
2704 comes out to be nonzero, the `#if' succeeds and the CONTROLLED TEXT is
2705 included; otherwise it is skipped.
2708 File: cpp.info, Node: Defined, Next: Else, Prev: If, Up: Conditional Syntax
2713 The special operator `defined' is used in `#if' and `#elif' expressions
2714 to test whether a certain name is defined as a macro. `defined NAME'
2715 and `defined (NAME)' are both expressions whose value is 1 if NAME is
2716 defined as a macro at the current point in the program, and 0
2717 otherwise. Thus, `#if defined MACRO' is precisely equivalent to
2720 `defined' is useful when you wish to test more than one macro for
2721 existence at once. For example,
2723 #if defined (__vax__) || defined (__ns16000__)
2725 would succeed if either of the names `__vax__' or `__ns16000__' is
2728 Conditionals written like this:
2730 #if defined BUFSIZE && BUFSIZE >= 1024
2732 can generally be simplified to just `#if BUFSIZE >= 1024', since if
2733 `BUFSIZE' is not defined, it will be interpreted as having the value
2736 If the `defined' operator appears as a result of a macro expansion,
2737 the C standard says the behavior is undefined. GNU cpp treats it as a
2738 genuine `defined' operator and evaluates it normally. It will warn
2739 wherever your code uses this feature if you use the command-line option
2740 `-pedantic', since other compilers may handle it differently.
2743 File: cpp.info, Node: Else, Next: Elif, Prev: Defined, Up: Conditional Syntax
2748 The `#else' directive can be added to a conditional to provide
2749 alternative text to be used if the condition fails. This is what it
2754 #else /* Not EXPRESSION */
2756 #endif /* Not EXPRESSION */
2758 If EXPRESSION is nonzero, the TEXT-IF-TRUE is included and the
2759 TEXT-IF-FALSE is skipped. If EXPRESSION is zero, the opposite happens.
2761 You can use `#else' with `#ifdef' and `#ifndef', too.
2764 File: cpp.info, Node: Elif, Prev: Else, Up: Conditional Syntax
2769 One common case of nested conditionals is used to check for more than
2770 two possible alternatives. For example, you might have
2782 Another conditional directive, `#elif', allows this to be
2783 abbreviated as follows:
2789 #else /* X != 2 and X != 1*/
2791 #endif /* X != 2 and X != 1*/
2793 `#elif' stands for "else if". Like `#else', it goes in the middle
2794 of a conditional group and subdivides it; it does not require a
2795 matching `#endif' of its own. Like `#if', the `#elif' directive
2796 includes an expression to be tested. The text following the `#elif' is
2797 processed only if the original `#if'-condition failed and the `#elif'
2800 More than one `#elif' can go in the same conditional group. Then
2801 the text after each `#elif' is processed only if the `#elif' condition
2802 succeeds after the original `#if' and all previous `#elif' directives
2803 within it have failed.
2805 `#else' is allowed after any number of `#elif' directives, but
2806 `#elif' may not follow `#else'.
2809 File: cpp.info, Node: Deleted Code, Prev: Conditional Syntax, Up: Conditionals
2814 If you replace or delete a part of the program but want to keep the old
2815 code around for future reference, you often cannot simply comment it
2816 out. Block comments do not nest, so the first comment inside the old
2817 code will end the commenting-out. The probable result is a flood of
2820 One way to avoid this problem is to use an always-false conditional
2821 instead. For instance, put `#if 0' before the deleted code and
2822 `#endif' after it. This works even if the code being turned off
2823 contains conditionals, but they must be entire conditionals (balanced
2824 `#if' and `#endif').
2826 Some people use `#ifdef notdef' instead. This is risky, because
2827 `notdef' might be accidentally defined as a macro, and then the
2828 conditional would succeed. `#if 0' can be counted on to fail.
2830 Do not use `#if 0' for comments which are not C code. Use a real
2831 comment, instead. The interior of `#if 0' must consist of complete
2832 tokens; in particular, single-quote characters must balance. Comments
2833 often contain unbalanced single-quote characters (known in English as
2834 apostrophes). These confuse `#if 0'. They don't confuse `/*'.
2837 File: cpp.info, Node: Diagnostics, Next: Line Control, Prev: Conditionals, Up: Top
2842 The directive `#error' causes the preprocessor to report a fatal error.
2843 The tokens forming the rest of the line following `#error' are used as
2846 You would use `#error' inside of a conditional that detects a
2847 combination of parameters which you know the program does not properly
2848 support. For example, if you know that the program will not run
2849 properly on a VAX, you might write
2852 #error "Won't work on VAXen. See comments at get_last_object."
2855 If you have several configuration parameters that must be set up by
2856 the installation in a consistent way, you can use conditionals to detect
2857 an inconsistency and report it with `#error'. For example,
2859 #if !defined(UNALIGNED_INT_ASM_OP) && defined(DWARF2_DEBUGGING_INFO)
2860 #error "DWARF2_DEBUGGING_INFO requires UNALIGNED_INT_ASM_OP."
2863 The directive `#warning' is like `#error', but causes the
2864 preprocessor to issue a warning and continue preprocessing. The tokens
2865 following `#warning' are used as the warning message.
2867 You might use `#warning' in obsolete header files, with a message
2868 directing the user to the header file which should be used instead.
2870 Neither `#error' nor `#warning' macro-expands its argument.
2871 Internal whitespace sequences are each replaced with a single space.
2872 The line must consist of complete tokens. It is wisest to make the
2873 argument of these directives be a single string constant; this avoids
2874 problems with apostrophes and the like.
2877 File: cpp.info, Node: Line Control, Next: Pragmas, Prev: Diagnostics, Up: Top
2882 The C preprocessor informs the C compiler of the location in your source
2883 code where each token came from. Presently, this is just the file name
2884 and line number. All the tokens resulting from macro expansion are
2885 reported as having appeared on the line of the source file where the
2886 outermost macro was used. We intend to be more accurate in the future.
2888 If you write a program which generates source code, such as the
2889 `bison' parser generator, you may want to adjust the preprocessor's
2890 notion of the current file name and line number by hand. Parts of the
2891 output from `bison' are generated from scratch, other parts come from a
2892 standard parser file. The rest are copied verbatim from `bison''s
2893 input. You would like compiler error messages and symbolic debuggers
2894 to be able to refer to `bison''s input file.
2896 `bison' or any such program can arrange this by writing `#line'
2897 directives into the output file. `#line' is a directive that specifies
2898 the original line number and source file name for subsequent input in
2899 the current preprocessor input file. `#line' has three variants:
2902 LINENUM is a non-negative decimal integer constant. It specifies
2903 the line number which should be reported for the following line of
2904 input. Subsequent lines are counted from LINENUM.
2906 `#line LINENUM FILENAME'
2907 LINENUM is the same as for the first form, and has the same
2908 effect. In addition, FILENAME is a string constant. The
2909 following line and all subsequent lines are reported to come from
2910 the file it specifies, until something else happens to change that.
2911 FILENAME is interpreted according to the normal rules for a string
2912 constant: backslash escapes are interpreted. This is different
2915 Previous versions of CPP did not interpret escapes in `#line'; we
2916 have changed it because the standard requires they be interpreted,
2917 and most other compilers do.
2919 `#line ANYTHING ELSE'
2920 ANYTHING ELSE is checked for macro calls, which are expanded. The
2921 result should match one of the above two forms.
2923 `#line' directives alter the results of the `__FILE__' and
2924 `__LINE__' predefined macros from that point on. *Note Standard
2925 Predefined Macros::. They do not have any effect on `#include''s idea
2926 of the directory containing the current file. This is a change from
2927 GCC 2.95. Previously, a file reading
2929 #line 1 "../src/gram.y"
2932 would search for `gram.h' in `../src', then the `-I' chain; the
2933 directory containing the physical source file would not be searched.
2934 In GCC 3.0 and later, the `#include' is not affected by the presence of
2935 a `#line' referring to a different directory.
2937 We made this change because the old behavior caused problems when
2938 generated source files were transported between machines. For instance,
2939 it is common practice to ship generated parsers with a source release,
2940 so that people building the distribution do not need to have yacc or
2941 Bison installed. These files frequently have `#line' directives
2942 referring to the directory tree of the system where the distribution was
2943 created. If GCC tries to search for headers in those directories, the
2944 build is likely to fail.
2946 The new behavior can cause failures too, if the generated file is not
2947 in the same directory as its source and it attempts to include a header
2948 which would be visible searching from the directory containing the
2949 source file. However, this problem is easily solved with an additional
2950 `-I' switch on the command line. The failures caused by the old
2951 semantics could sometimes be corrected only by editing the generated
2952 files, which is difficult and error-prone.
2955 File: cpp.info, Node: Pragmas, Next: Other Directives, Prev: Line Control, Up: Top
2960 The `#pragma' directive is the method specified by the C standard for
2961 providing additional information to the compiler, beyond what is
2962 conveyed in the language itself. Three forms of this directive
2963 (commonly known as "pragmas") are specified by the 1999 C standard. A
2964 C compiler is free to attach any meaning it likes to other pragmas.
2966 GCC has historically preferred to use extensions to the syntax of the
2967 language, such as `__attribute__', for this purpose. However, GCC does
2968 define a few pragmas of its own. These mostly have effects on the
2969 entire translation unit or source file.
2971 In GCC version 3, all GNU-defined, supported pragmas have been given
2972 a `GCC' prefix. This is in line with the `STDC' prefix on all pragmas
2973 defined by C99. For backward compatibility, pragmas which were
2974 recognized by previous versions are still recognized without the `GCC'
2975 prefix, but that usage is deprecated. Some older pragmas are
2976 deprecated in their entirety. They are not recognized with the `GCC'
2977 prefix. *Note Obsolete Features::.
2979 C99 introduces the `_Pragma' operator. This feature addresses a
2980 major problem with `#pragma': being a directive, it cannot be produced
2981 as the result of macro expansion. `_Pragma' is an operator, much like
2982 `sizeof' or `defined', and can be embedded in a macro.
2984 Its syntax is `_Pragma (STRING-LITERAL)', where STRING-LITERAL can
2985 be either a normal or wide-character string literal. It is
2986 destringized, by replacing all `\\' with a single `\' and all `\"' with
2987 a `"'. The result is then processed as if it had appeared as the right
2988 hand side of a `#pragma' directive. For example,
2990 _Pragma ("GCC dependency \"parse.y\"")
2992 has the same effect as `#pragma GCC dependency "parse.y"'. The same
2993 effect could be achieved using macros, for example
2995 #define DO_PRAGMA(x) _Pragma (#x)
2996 DO_PRAGMA (GCC dependency "parse.y")
2998 The standard is unclear on where a `_Pragma' operator can appear.
2999 The preprocessor does not accept it within a preprocessing conditional
3000 directive like `#if'. To be safe, you are probably best keeping it out
3001 of directives other than `#define', and putting it on a line of its own.
3003 This manual documents the pragmas which are meaningful to the
3004 preprocessor itself. Other pragmas are meaningful to the C or C++
3005 compilers. They are documented in the GCC manual.
3007 `#pragma GCC dependency'
3008 `#pragma GCC dependency' allows you to check the relative dates of
3009 the current file and another file. If the other file is more
3010 recent than the current file, a warning is issued. This is useful
3011 if the current file is derived from the other file, and should be
3012 regenerated. The other file is searched for using the normal
3013 include search path. Optional trailing text can be used to give
3014 more information in the warning message.
3016 #pragma GCC dependency "parse.y"
3017 #pragma GCC dependency "/usr/include/time.h" rerun fixincludes
3019 `#pragma GCC poison'
3020 Sometimes, there is an identifier that you want to remove
3021 completely from your program, and make sure that it never creeps
3022 back in. To enforce this, you can "poison" the identifier with
3023 this pragma. `#pragma GCC poison' is followed by a list of
3024 identifiers to poison. If any of those identifiers appears
3025 anywhere in the source after the directive, it is a hard error.
3028 #pragma GCC poison printf sprintf fprintf
3029 sprintf(some_string, "hello");
3031 will produce an error.
3033 If a poisoned identifier appears as part of the expansion of a
3034 macro which was defined before the identifier was poisoned, it
3035 will _not_ cause an error. This lets you poison an identifier
3036 without worrying about system headers defining macros that use it.
3040 #define strrchr rindex
3041 #pragma GCC poison rindex
3042 strrchr(some_string, 'h');
3044 will not produce an error.
3046 `#pragma GCC system_header'
3047 This pragma takes no arguments. It causes the rest of the code in
3048 the current file to be treated as if it came from a system header.
3049 *Note System Headers::.
3053 File: cpp.info, Node: Other Directives, Next: Preprocessor Output, Prev: Pragmas, Up: Top
3058 The `#ident' directive takes one argument, a string constant. On some
3059 systems, that string constant is copied into a special segment of the
3060 object file. On other systems, the directive is ignored. The `#sccs'
3061 directive is a synonym for `#ident'.
3063 These directives are not part of the C standard, but they are not
3064 official GNU extensions either. What historical information we have
3065 been able to find, suggests they originated with System V.
3067 The "null directive" consists of a `#' followed by a newline, with
3068 only whitespace (including comments) in between. A null directive is
3069 understood as a preprocessing directive but has no effect on the
3070 preprocessor output. The primary significance of the existence of the
3071 null directive is that an input line consisting of just a `#' will
3072 produce no output, rather than a line of output containing just a `#'.
3073 Supposedly some old C programs contain such lines.
3076 File: cpp.info, Node: Preprocessor Output, Next: Traditional Mode, Prev: Other Directives, Up: Top
3078 9 Preprocessor Output
3079 *********************
3081 When the C preprocessor is used with the C, C++, or Objective-C
3082 compilers, it is integrated into the compiler and communicates a stream
3083 of binary tokens directly to the compiler's parser. However, it can
3084 also be used in the more conventional standalone mode, where it produces
3087 The output from the C preprocessor looks much like the input, except
3088 that all preprocessing directive lines have been replaced with blank
3089 lines and all comments with spaces. Long runs of blank lines are
3092 The ISO standard specifies that it is implementation defined whether
3093 a preprocessor preserves whitespace between tokens, or replaces it with
3094 e.g. a single space. In GNU CPP, whitespace between tokens is collapsed
3095 to become a single space, with the exception that the first token on a
3096 non-directive line is preceded with sufficient spaces that it appears in
3097 the same column in the preprocessed output that it appeared in the
3098 original source file. This is so the output is easy to read. *Note
3099 Differences from previous versions::. CPP does not insert any
3100 whitespace where there was none in the original source, except where
3101 necessary to prevent an accidental token paste.
3103 Source file name and line number information is conveyed by lines of
3106 # LINENUM FILENAME FLAGS
3108 These are called "linemarkers". They are inserted as needed into the
3109 output (but never within a string or character constant). They mean
3110 that the following line originated in file FILENAME at line LINENUM.
3111 FILENAME will never contain any non-printing characters; they are
3112 replaced with octal escape sequences.
3114 After the file name comes zero or more flags, which are `1', `2',
3115 `3', or `4'. If there are multiple flags, spaces separate them. Here
3116 is what the flags mean:
3119 This indicates the start of a new file.
3122 This indicates returning to a file (after having included another
3126 This indicates that the following text comes from a system header
3127 file, so certain warnings should be suppressed.
3130 This indicates that the following text should be treated as being
3131 wrapped in an implicit `extern "C"' block.
3133 As an extension, the preprocessor accepts linemarkers in
3134 non-assembler input files. They are treated like the corresponding
3135 `#line' directive, (*note Line Control::), except that trailing flags
3136 are permitted, and are interpreted with the meanings described above.
3137 If multiple flags are given, they must be in ascending order.
3139 Some directives may be duplicated in the output of the preprocessor.
3140 These are `#ident' (always), `#pragma' (only if the preprocessor does
3141 not handle the pragma itself), and `#define' and `#undef' (with certain
3142 debugging options). If this happens, the `#' of the directive will
3143 always be in the first column, and there will be no space between the
3144 `#' and the directive name. If macro expansion happens to generate
3145 tokens which might be mistaken for a duplicated directive, a space will
3146 be inserted between the `#' and the directive name.
3149 File: cpp.info, Node: Traditional Mode, Next: Implementation Details, Prev: Preprocessor Output, Up: Top
3154 Traditional (pre-standard) C preprocessing is rather different from the
3155 preprocessing specified by the standard. When GCC is given the
3156 `-traditional-cpp' option, it attempts to emulate a traditional
3159 GCC versions 3.2 and later only support traditional mode semantics in
3160 the preprocessor, and not in the compiler front ends. This chapter
3161 outlines the traditional preprocessor semantics we implemented.
3163 The implementation does not correspond precisely to the behavior of
3164 earlier versions of GCC, nor to any true traditional preprocessor.
3165 After all, inconsistencies among traditional implementations were a
3166 major motivation for C standardization. However, we intend that it
3167 should be compatible with true traditional preprocessors in all ways
3168 that actually matter.
3172 * Traditional lexical analysis::
3173 * Traditional macros::
3174 * Traditional miscellany::
3175 * Traditional warnings::
3178 File: cpp.info, Node: Traditional lexical analysis, Next: Traditional macros, Up: Traditional Mode
3180 10.1 Traditional lexical analysis
3181 =================================
3183 The traditional preprocessor does not decompose its input into tokens
3184 the same way a standards-conforming preprocessor does. The input is
3185 simply treated as a stream of text with minimal internal form.
3187 This implementation does not treat trigraphs (*note trigraphs::)
3188 specially since they were an invention of the standards committee. It
3189 handles arbitrarily-positioned escaped newlines properly and splices
3190 the lines as you would expect; many traditional preprocessors did not
3193 The form of horizontal whitespace in the input file is preserved in
3194 the output. In particular, hard tabs remain hard tabs. This can be
3195 useful if, for example, you are preprocessing a Makefile.
3197 Traditional CPP only recognizes C-style block comments, and treats
3198 the `/*' sequence as introducing a comment only if it lies outside
3199 quoted text. Quoted text is introduced by the usual single and double
3200 quotes, and also by an initial `<' in a `#include' directive.
3202 Traditionally, comments are completely removed and are not replaced
3203 with a space. Since a traditional compiler does its own tokenization
3204 of the output of the preprocessor, this means that comments can
3205 effectively be used as token paste operators. However, comments behave
3206 like separators for text handled by the preprocessor itself, since it
3207 doesn't re-lex its input. For example, in
3211 `foo' and `bar' are distinct identifiers and expanded separately if
3212 they happen to be macros. In other words, this directive is equivalent
3221 Generally speaking, in traditional mode an opening quote need not
3222 have a matching closing quote. In particular, a macro may be defined
3223 with replacement text that contains an unmatched quote. Of course, if
3224 you attempt to compile preprocessed output containing an unmatched quote
3225 you will get a syntax error.
3227 However, all preprocessing directives other than `#define' require
3228 matching quotes. For example:
3230 #define m This macro's fine and has an unmatched quote
3231 "/* This is not a comment. */
3232 /* This is a comment. The following #include directive
3236 Just as for the ISO preprocessor, what would be a closing quote can
3237 be escaped with a backslash to prevent the quoted text from closing.
3240 File: cpp.info, Node: Traditional macros, Next: Traditional miscellany, Prev: Traditional lexical analysis, Up: Traditional Mode
3242 10.2 Traditional macros
3243 =======================
3245 The major difference between traditional and ISO macros is that the
3246 former expand to text rather than to a token sequence. CPP removes all
3247 leading and trailing horizontal whitespace from a macro's replacement
3248 text before storing it, but preserves the form of internal whitespace.
3250 One consequence is that it is legitimate for the replacement text to
3251 contain an unmatched quote (*note Traditional lexical analysis::). An
3252 unclosed string or character constant continues into the text following
3253 the macro call. Similarly, the text at the end of a macro's expansion
3254 can run together with the text after the macro invocation to produce a
3257 Normally comments are removed from the replacement text after the
3258 macro is expanded, but if the `-CC' option is passed on the command
3259 line comments are preserved. (In fact, the current implementation
3260 removes comments even before saving the macro replacement text, but it
3261 careful to do it in such a way that the observed effect is identical
3262 even in the function-like macro case.)
3264 The ISO stringification operator `#' and token paste operator `##'
3265 have no special meaning. As explained later, an effect similar to
3266 these operators can be obtained in a different way. Macro names that
3267 are embedded in quotes, either from the main file or after macro
3268 replacement, do not expand.
3270 CPP replaces an unquoted object-like macro name with its replacement
3271 text, and then rescans it for further macros to replace. Unlike
3272 standard macro expansion, traditional macro expansion has no provision
3273 to prevent recursion. If an object-like macro appears unquoted in its
3274 replacement text, it will be replaced again during the rescan pass, and
3275 so on _ad infinitum_. GCC detects when it is expanding recursive
3276 macros, emits an error message, and continues after the offending macro
3280 #define INC(x) PLUS+x
3284 Function-like macros are similar in form but quite different in
3285 behavior to their ISO counterparts. Their arguments are contained
3286 within parentheses, are comma-separated, and can cross physical lines.
3287 Commas within nested parentheses are not treated as argument
3288 separators. Similarly, a quote in an argument cannot be left unclosed;
3289 a following comma or parenthesis that comes before the closing quote is
3290 treated like any other character. There is no facility for handling
3293 This implementation removes all comments from macro arguments, unless
3294 the `-C' option is given. The form of all other horizontal whitespace
3295 in arguments is preserved, including leading and trailing whitespace.
3300 is treated as an invocation of the macro `f' with a single argument
3301 consisting of a single space. If you want to invoke a function-like
3302 macro that takes no arguments, you must not leave any whitespace
3303 between the parentheses.
3305 If a macro argument crosses a new line, the new line is replaced with
3306 a space when forming the argument. If the previous line contained an
3307 unterminated quote, the following line inherits the quoted state.
3309 Traditional preprocessors replace parameters in the replacement text
3310 with their arguments regardless of whether the parameters are within
3311 quotes or not. This provides a way to stringize arguments. For example
3314 str(/* A comment */some text )
3317 Note that the comment is removed, but that the trailing space is
3318 preserved. Here is an example of using a comment to effect token
3321 #define suffix(x) foo_/**/x
3326 File: cpp.info, Node: Traditional miscellany, Next: Traditional warnings, Prev: Traditional macros, Up: Traditional Mode
3328 10.3 Traditional miscellany
3329 ===========================
3331 Here are some things to be aware of when using the traditional
3334 * Preprocessing directives are recognized only when their leading
3335 `#' appears in the first column. There can be no whitespace
3336 between the beginning of the line and the `#', but whitespace can
3339 * A true traditional C preprocessor does not recognize `#error' or
3340 `#pragma', and may not recognize `#elif'. CPP supports all the
3341 directives in traditional mode that it supports in ISO mode,
3342 including extensions, with the exception that the effects of
3343 `#pragma GCC poison' are undefined.
3345 * __STDC__ is not defined.
3347 * If you use digraphs the behavior is undefined.
3349 * If a line that looks like a directive appears within macro
3350 arguments, the behavior is undefined.
3354 File: cpp.info, Node: Traditional warnings, Prev: Traditional miscellany, Up: Traditional Mode
3356 10.4 Traditional warnings
3357 =========================
3359 You can request warnings about features that did not exist, or worked
3360 differently, in traditional C with the `-Wtraditional' option. GCC
3361 does not warn about features of ISO C which you must use when you are
3362 using a conforming compiler, such as the `#' and `##' operators.
3364 Presently `-Wtraditional' warns about:
3366 * Macro parameters that appear within string literals in the macro
3367 body. In traditional C macro replacement takes place within
3368 string literals, but does not in ISO C.
3370 * In traditional C, some preprocessor directives did not exist.
3371 Traditional preprocessors would only consider a line to be a
3372 directive if the `#' appeared in column 1 on the line. Therefore
3373 `-Wtraditional' warns about directives that traditional C
3374 understands but would ignore because the `#' does not appear as the
3375 first character on the line. It also suggests you hide directives
3376 like `#pragma' not understood by traditional C by indenting them.
3377 Some traditional implementations would not recognize `#elif', so it
3378 suggests avoiding it altogether.
3380 * A function-like macro that appears without an argument list. In
3381 some traditional preprocessors this was an error. In ISO C it
3382 merely means that the macro is not expanded.
3384 * The unary plus operator. This did not exist in traditional C.
3386 * The `U' and `LL' integer constant suffixes, which were not
3387 available in traditional C. (Traditional C does support the `L'
3388 suffix for simple long integer constants.) You are not warned
3389 about uses of these suffixes in macros defined in system headers.
3390 For instance, `UINT_MAX' may well be defined as `4294967295U', but
3391 you will not be warned if you use `UINT_MAX'.
3393 You can usually avoid the warning, and the related warning about
3394 constants which are so large that they are unsigned, by writing the
3395 integer constant in question in hexadecimal, with no U suffix.
3396 Take care, though, because this gives the wrong result in exotic
3400 File: cpp.info, Node: Implementation Details, Next: Invocation, Prev: Traditional Mode, Up: Top
3402 11 Implementation Details
3403 *************************
3405 Here we document details of how the preprocessor's implementation
3406 affects its user-visible behavior. You should try to avoid undue
3407 reliance on behavior described here, as it is possible that it will
3408 change subtly in future implementations.
3410 Also documented here are obsolete features and changes from previous
3415 * Implementation-defined behavior::
3416 * Implementation limits::
3417 * Obsolete Features::
3418 * Differences from previous versions::
3421 File: cpp.info, Node: Implementation-defined behavior, Next: Implementation limits, Up: Implementation Details
3423 11.1 Implementation-defined behavior
3424 ====================================
3426 This is how CPP behaves in all the cases which the C standard describes
3427 as "implementation-defined". This term means that the implementation
3428 is free to do what it likes, but must document its choice and stick to
3431 * The mapping of physical source file multi-byte characters to the
3432 execution character set.
3434 Currently, CPP requires its input to be ASCII or UTF-8. The
3435 execution character set may be controlled by the user, with the
3436 `-fexec-charset' and `-fwide-exec-charset' options.
3438 * Identifier characters.
3440 The C and C++ standards allow identifiers to be composed of `_'
3441 and the alphanumeric characters. C++ and C99 also allow universal
3442 character names, and C99 further permits implementation-defined
3443 characters. GCC currently only permits universal character names
3444 if `-fextended-identifiers' is used, because the implementation of
3445 universal character names in identifiers is experimental.
3447 GCC allows the `$' character in identifiers as an extension for
3448 most targets. This is true regardless of the `std=' switch, since
3449 this extension cannot conflict with standards-conforming programs.
3450 When preprocessing assembler, however, dollars are not identifier
3451 characters by default.
3453 Currently the targets that by default do not permit `$' are AVR,
3454 IP2K, MMIX, MIPS Irix 3, ARM aout, and PowerPC targets for the AIX
3455 and BeOS operating systems.
3457 You can override the default with `-fdollars-in-identifiers' or
3458 `fno-dollars-in-identifiers'. *Note fdollars-in-identifiers::.
3460 * Non-empty sequences of whitespace characters.
3462 In textual output, each whitespace sequence is collapsed to a
3463 single space. For aesthetic reasons, the first token on each
3464 non-directive line of output is preceded with sufficient spaces
3465 that it appears in the same column as it did in the original
3468 * The numeric value of character constants in preprocessor
3471 The preprocessor and compiler interpret character constants in the
3472 same way; i.e. escape sequences such as `\a' are given the values
3473 they would have on the target machine.
3475 The compiler values a multi-character character constant a
3476 character at a time, shifting the previous value left by the
3477 number of bits per target character, and then or-ing in the
3478 bit-pattern of the new character truncated to the width of a
3479 target character. The final bit-pattern is given type `int', and
3480 is therefore signed, regardless of whether single characters are
3481 signed or not (a slight change from versions 3.1 and earlier of
3482 GCC). If there are more characters in the constant than would fit
3483 in the target `int' the compiler issues a warning, and the excess
3484 leading characters are ignored.
3486 For example, `'ab'' for a target with an 8-bit `char' would be
3488 `(int) ((unsigned char) 'a' * 256 + (unsigned char) 'b')', and
3490 `(int) ((unsigned char) '\234' * 256 + (unsigned char) 'a')'.
3492 * Source file inclusion.
3494 For a discussion on how the preprocessor locates header files,
3495 *Note Include Operation::.
3497 * Interpretation of the filename resulting from a macro-expanded
3498 `#include' directive.
3500 *Note Computed Includes::.
3502 * Treatment of a `#pragma' directive that after macro-expansion
3503 results in a standard pragma.
3505 No macro expansion occurs on any `#pragma' directive line, so the
3506 question does not arise.
3508 Note that GCC does not yet implement any of the standard pragmas.
3512 File: cpp.info, Node: Implementation limits, Next: Obsolete Features, Prev: Implementation-defined behavior, Up: Implementation Details
3514 11.2 Implementation limits
3515 ==========================
3517 CPP has a small number of internal limits. This section lists the
3518 limits which the C standard requires to be no lower than some minimum,
3519 and all the others known. It is intended that there should be as few
3520 limits as possible. If you encounter an undocumented or inconvenient
3521 limit, please report that as a bug. *Note Reporting Bugs: (gcc)Bugs.
3523 Where we say something is limited "only by available memory", that
3524 means that internal data structures impose no intrinsic limit, and space
3525 is allocated with `malloc' or equivalent. The actual limit will
3526 therefore depend on many things, such as the size of other things
3527 allocated by the compiler at the same time, the amount of memory
3528 consumed by other processes on the same computer, etc.
3530 * Nesting levels of `#include' files.
3532 We impose an arbitrary limit of 200 levels, to avoid runaway
3533 recursion. The standard requires at least 15 levels.
3535 * Nesting levels of conditional inclusion.
3537 The C standard mandates this be at least 63. CPP is limited only
3538 by available memory.
3540 * Levels of parenthesized expressions within a full expression.
3542 The C standard requires this to be at least 63. In preprocessor
3543 conditional expressions, it is limited only by available memory.
3545 * Significant initial characters in an identifier or macro name.
3547 The preprocessor treats all characters as significant. The C
3548 standard requires only that the first 63 be significant.
3550 * Number of macros simultaneously defined in a single translation
3553 The standard requires at least 4095 be possible. CPP is limited
3554 only by available memory.
3556 * Number of parameters in a macro definition and arguments in a
3559 We allow `USHRT_MAX', which is no smaller than 65,535. The minimum
3560 required by the standard is 127.
3562 * Number of characters on a logical source line.
3564 The C standard requires a minimum of 4096 be permitted. CPP places
3565 no limits on this, but you may get incorrect column numbers
3566 reported in diagnostics for lines longer than 65,535 characters.
3568 * Maximum size of a source file.
3570 The standard does not specify any lower limit on the maximum size
3571 of a source file. GNU cpp maps files into memory, so it is
3572 limited by the available address space. This is generally at
3573 least two gigabytes. Depending on the operating system, the size
3574 of physical memory may or may not be a limitation.
3578 File: cpp.info, Node: Obsolete Features, Next: Differences from previous versions, Prev: Implementation limits, Up: Implementation Details
3580 11.3 Obsolete Features
3581 ======================
3583 CPP has a number of features which are present mainly for compatibility
3584 with older programs. We discourage their use in new code. In some
3585 cases, we plan to remove the feature in a future version of GCC.
3590 * Obsolete once-only headers::
3593 File: cpp.info, Node: Assertions, Next: Obsolete once-only headers, Up: Obsolete Features
3598 "Assertions" are a deprecated alternative to macros in writing
3599 conditionals to test what sort of computer or system the compiled
3600 program will run on. Assertions are usually predefined, but you can
3601 define them with preprocessing directives or command-line options.
3603 Assertions were intended to provide a more systematic way to describe
3604 the compiler's target system. However, in practice they are just as
3605 unpredictable as the system-specific predefined macros. In addition,
3606 they are not part of any standard, and only a few compilers support
3607 them. Therefore, the use of assertions is *less* portable than the use
3608 of system-specific predefined macros. We recommend you do not use them
3611 An assertion looks like this:
3615 PREDICATE must be a single identifier. ANSWER can be any sequence of
3616 tokens; all characters are significant except for leading and trailing
3617 whitespace, and differences in internal whitespace sequences are
3618 ignored. (This is similar to the rules governing macro redefinition.)
3619 Thus, `(x + y)' is different from `(x+y)' but equivalent to
3620 `( x + y )'. Parentheses do not nest inside an answer.
3622 To test an assertion, you write it in an `#if'. For example, this
3623 conditional succeeds if either `vax' or `ns16000' has been asserted as
3624 an answer for `machine'.
3626 #if #machine (vax) || #machine (ns16000)
3628 You can test whether _any_ answer is asserted for a predicate by
3629 omitting the answer in the conditional:
3633 Assertions are made with the `#assert' directive. Its sole argument
3634 is the assertion to make, without the leading `#' that identifies
3635 assertions in conditionals.
3637 #assert PREDICATE (ANSWER)
3639 You may make several assertions with the same predicate and different
3640 answers. Subsequent assertions do not override previous ones for the
3641 same predicate. All the answers for any given predicate are
3642 simultaneously true.
3644 Assertions can be canceled with the `#unassert' directive. It has
3645 the same syntax as `#assert'. In that form it cancels only the answer
3646 which was specified on the `#unassert' line; other answers for that
3647 predicate remain true. You can cancel an entire predicate by leaving
3652 In either form, if no such assertion has been made, `#unassert' has no
3655 You can also make or cancel assertions using command line options.
3659 File: cpp.info, Node: Obsolete once-only headers, Prev: Assertions, Up: Obsolete Features
3661 11.3.2 Obsolete once-only headers
3662 ---------------------------------
3664 CPP supports two more ways of indicating that a header file should be
3665 read only once. Neither one is as portable as a wrapper `#ifndef', and
3666 we recommend you do not use them in new programs.
3668 In the Objective-C language, there is a variant of `#include' called
3669 `#import' which includes a file, but does so at most once. If you use
3670 `#import' instead of `#include', then you don't need the conditionals
3671 inside the header file to prevent multiple inclusion of the contents.
3672 GCC permits the use of `#import' in C and C++ as well as Objective-C.
3673 However, it is not in standard C or C++ and should therefore not be
3674 used by portable programs.
3676 `#import' is not a well designed feature. It requires the users of
3677 a header file to know that it should only be included once. It is much
3678 better for the header file's implementor to write the file so that users
3679 don't need to know this. Using a wrapper `#ifndef' accomplishes this
3682 In the present implementation, a single use of `#import' will
3683 prevent the file from ever being read again, by either `#import' or
3684 `#include'. You should not rely on this; do not use both `#import' and
3685 `#include' to refer to the same header file.
3687 Another way to prevent a header file from being included more than
3688 once is with the `#pragma once' directive. If `#pragma once' is seen
3689 when scanning a header file, that file will never be read again, no
3692 `#pragma once' does not have the problems that `#import' does, but
3693 it is not recognized by all preprocessors, so you cannot rely on it in
3697 File: cpp.info, Node: Differences from previous versions, Prev: Obsolete Features, Up: Implementation Details
3699 11.4 Differences from previous versions
3700 =======================================
3702 This section details behavior which has changed from previous versions
3703 of CPP. We do not plan to change it again in the near future, but we
3704 do not promise not to, either.
3706 The "previous versions" discussed here are 2.95 and before. The
3707 behavior of GCC 3.0 is mostly the same as the behavior of the widely
3708 used 2.96 and 2.97 development snapshots. Where there are differences,
3709 they generally represent bugs in the snapshots.
3713 This option has been deprecated in 4.0. `-iquote' is meant to
3714 replace the need for this option.
3716 * Order of evaluation of `#' and `##' operators
3718 The standard does not specify the order of evaluation of a chain of
3719 `##' operators, nor whether `#' is evaluated before, after, or at
3720 the same time as `##'. You should therefore not write any code
3721 which depends on any specific ordering. It is possible to
3722 guarantee an ordering, if you need one, by suitable use of nested
3725 An example of where this might matter is pasting the arguments `1',
3726 `e' and `-2'. This would be fine for left-to-right pasting, but
3727 right-to-left pasting would produce an invalid token `e-2'.
3729 GCC 3.0 evaluates `#' and `##' at the same time and strictly left
3730 to right. Older versions evaluated all `#' operators first, then
3731 all `##' operators, in an unreliable order.
3733 * The form of whitespace between tokens in preprocessor output
3735 *Note Preprocessor Output::, for the current textual format. This
3736 is also the format used by stringification. Normally, the
3737 preprocessor communicates tokens directly to the compiler's
3738 parser, and whitespace does not come up at all.
3740 Older versions of GCC preserved all whitespace provided by the
3741 user and inserted lots more whitespace of their own, because they
3742 could not accurately predict when extra spaces were needed to
3743 prevent accidental token pasting.
3745 * Optional argument when invoking rest argument macros
3747 As an extension, GCC permits you to omit the variable arguments
3748 entirely when you use a variable argument macro. This is
3749 forbidden by the 1999 C standard, and will provoke a pedantic
3750 warning with GCC 3.0. Previous versions accepted it silently.
3752 * `##' swallowing preceding text in rest argument macros
3754 Formerly, in a macro expansion, if `##' appeared before a variable
3755 arguments parameter, and the set of tokens specified for that
3756 argument in the macro invocation was empty, previous versions of
3757 CPP would back up and remove the preceding sequence of
3758 non-whitespace characters (*not* the preceding token). This
3759 extension is in direct conflict with the 1999 C standard and has
3760 been drastically pared back.
3762 In the current version of the preprocessor, if `##' appears between
3763 a comma and a variable arguments parameter, and the variable
3764 argument is omitted entirely, the comma will be removed from the
3765 expansion. If the variable argument is empty, or the token before
3766 `##' is not a comma, then `##' behaves as a normal token paste.
3768 * `#line' and `#include'
3770 The `#line' directive used to change GCC's notion of the
3771 "directory containing the current file", used by `#include' with a
3772 double-quoted header file name. In 3.0 and later, it does not.
3773 *Note Line Control::, for further explanation.
3777 In GCC 2.95 and previous, the string constant argument to `#line'
3778 was treated the same way as the argument to `#include': backslash
3779 escapes were not honored, and the string ended at the second `"'.
3780 This is not compliant with the C standard. In GCC 3.0, an attempt
3781 was made to correct the behavior, so that the string was treated
3782 as a real string constant, but it turned out to be buggy. In 3.1,
3783 the bugs have been fixed. (We are not fixing the bugs in 3.0
3784 because they affect relatively few people and the fix is quite
3789 File: cpp.info, Node: Invocation, Next: Environment Variables, Prev: Implementation Details, Up: Top
3794 Most often when you use the C preprocessor you will not have to invoke
3795 it explicitly: the C compiler will do so automatically. However, the
3796 preprocessor is sometimes useful on its own. All the options listed
3797 here are also acceptable to the C compiler and have the same meaning,
3798 except that the C compiler has different rules for specifying the output
3801 _Note:_ Whether you use the preprocessor by way of `gcc' or `cpp',
3802 the "compiler driver" is run first. This program's purpose is to
3803 translate your command into invocations of the programs that do the
3804 actual work. Their command line interfaces are similar but not
3805 identical to the documented interface, and may change without notice.
3807 The C preprocessor expects two file names as arguments, INFILE and
3808 OUTFILE. The preprocessor reads INFILE together with any other files
3809 it specifies with `#include'. All the output generated by the combined
3810 input files is written in OUTFILE.
3812 Either INFILE or OUTFILE may be `-', which as INFILE means to read
3813 from standard input and as OUTFILE means to write to standard output.
3814 Also, if either file is omitted, it means the same as if `-' had been
3815 specified for that file.
3817 Unless otherwise noted, or the option ends in `=', all options which
3818 take an argument may have that argument appear either immediately after
3819 the option, or with a space between option and argument: `-Ifoo' and
3820 `-I foo' have the same effect.
3822 Many options have multi-letter names; therefore multiple
3823 single-letter options may _not_ be grouped: `-dM' is very different from
3827 Predefine NAME as a macro, with definition `1'.
3829 `-D NAME=DEFINITION'
3830 The contents of DEFINITION are tokenized and processed as if they
3831 appeared during translation phase three in a `#define' directive.
3832 In particular, the definition will be truncated by embedded
3835 If you are invoking the preprocessor from a shell or shell-like
3836 program you may need to use the shell's quoting syntax to protect
3837 characters such as spaces that have a meaning in the shell syntax.
3839 If you wish to define a function-like macro on the command line,
3840 write its argument list with surrounding parentheses before the
3841 equals sign (if any). Parentheses are meaningful to most shells,
3842 so you will need to quote the option. With `sh' and `csh',
3843 `-D'NAME(ARGS...)=DEFINITION'' works.
3845 `-D' and `-U' options are processed in the order they are given on
3846 the command line. All `-imacros FILE' and `-include FILE' options
3847 are processed after all `-D' and `-U' options.
3850 Cancel any previous definition of NAME, either built in or
3851 provided with a `-D' option.
3854 Do not predefine any system-specific or GCC-specific macros. The
3855 standard predefined macros remain defined. *Note Standard
3856 Predefined Macros::.
3859 Add the directory DIR to the list of directories to be searched
3860 for header files. *Note Search Path::. Directories named by `-I'
3861 are searched before the standard system include directories. If
3862 the directory DIR is a standard system include directory, the
3863 option is ignored to ensure that the default search order for
3864 system directories and the special treatment of system headers are
3865 not defeated (*note System Headers::) .
3868 Write output to FILE. This is the same as specifying FILE as the
3869 second non-option argument to `cpp'. `gcc' has a different
3870 interpretation of a second non-option argument, so you must use
3871 `-o' to specify the output file.
3874 Turns on all optional warnings which are desirable for normal code.
3875 At present this is `-Wcomment', `-Wtrigraphs', `-Wmultichar' and a
3876 warning about integer promotion causing a change of sign in `#if'
3877 expressions. Note that many of the preprocessor's warnings are on
3878 by default and have no options to control them.
3882 Warn whenever a comment-start sequence `/*' appears in a `/*'
3883 comment, or whenever a backslash-newline appears in a `//' comment.
3884 (Both forms have the same effect.)
3887 Most trigraphs in comments cannot affect the meaning of the
3888 program. However, a trigraph that would form an escaped newline
3889 (`??/' at the end of a line) can, by changing where the comment
3890 begins or ends. Therefore, only trigraphs that would form escaped
3891 newlines produce warnings inside a comment.
3893 This option is implied by `-Wall'. If `-Wall' is not given, this
3894 option is still enabled unless trigraphs are enabled. To get
3895 trigraph conversion without warnings, but get the other `-Wall'
3896 warnings, use `-trigraphs -Wall -Wno-trigraphs'.
3899 Warn about certain constructs that behave differently in
3900 traditional and ISO C. Also warn about ISO C constructs that have
3901 no traditional C equivalent, and problematic constructs which
3902 should be avoided. *Note Traditional Mode::.
3905 Warn the first time `#import' is used.
3908 Warn whenever an identifier which is not a macro is encountered in
3909 an `#if' directive, outside of `defined'. Such identifiers are
3913 Warn about macros defined in the main file that are unused. A
3914 macro is "used" if it is expanded or tested for existence at least
3915 once. The preprocessor will also warn if the macro has not been
3916 used at the time it is redefined or undefined.
3918 Built-in macros, macros defined on the command line, and macros
3919 defined in include files are not warned about.
3921 _Note:_ If a macro is actually used, but only used in skipped
3922 conditional blocks, then CPP will report it as unused. To avoid
3923 the warning in such a case, you might improve the scope of the
3924 macro's definition by, for example, moving it into the first
3925 skipped block. Alternatively, you could provide a dummy use with
3928 #if defined the_macro_causing_the_warning
3932 Warn whenever an `#else' or an `#endif' are followed by text.
3933 This usually happens in code of the form
3941 The second and third `FOO' should be in comments, but often are not
3942 in older programs. This warning is on by default.
3945 Make all warnings into hard errors. Source code which triggers
3946 warnings will be rejected.
3949 Issue warnings for code in system headers. These are normally
3950 unhelpful in finding bugs in your own code, therefore suppressed.
3951 If you are responsible for the system library, you may want to see
3955 Suppress all warnings, including those which GNU CPP issues by
3959 Issue all the mandatory diagnostics listed in the C standard.
3960 Some of them are left out by default, since they trigger
3961 frequently on harmless code.
3964 Issue all the mandatory diagnostics, and make all mandatory
3965 diagnostics into errors. This includes mandatory diagnostics that
3966 GCC issues without `-pedantic' but treats as warnings.
3969 Instead of outputting the result of preprocessing, output a rule
3970 suitable for `make' describing the dependencies of the main source
3971 file. The preprocessor outputs one `make' rule containing the
3972 object file name for that source file, a colon, and the names of
3973 all the included files, including those coming from `-include' or
3974 `-imacros' command line options.
3976 Unless specified explicitly (with `-MT' or `-MQ'), the object file
3977 name consists of the basename of the source file with any suffix
3978 replaced with object file suffix. If there are many included
3979 files then the rule is split into several lines using `\'-newline.
3980 The rule has no commands.
3982 This option does not suppress the preprocessor's debug output,
3983 such as `-dM'. To avoid mixing such debug output with the
3984 dependency rules you should explicitly specify the dependency
3985 output file with `-MF', or use an environment variable like
3986 `DEPENDENCIES_OUTPUT' (*note Environment Variables::). Debug
3987 output will still be sent to the regular output stream as normal.
3989 Passing `-M' to the driver implies `-E', and suppresses warnings
3990 with an implicit `-w'.
3993 Like `-M' but do not mention header files that are found in system
3994 header directories, nor header files that are included, directly
3995 or indirectly, from such a header.
3997 This implies that the choice of angle brackets or double quotes in
3998 an `#include' directive does not in itself determine whether that
3999 header will appear in `-MM' dependency output. This is a slight
4000 change in semantics from GCC versions 3.0 and earlier.
4003 When used with `-M' or `-MM', specifies a file to write the
4004 dependencies to. If no `-MF' switch is given the preprocessor
4005 sends the rules to the same place it would have sent preprocessed
4008 When used with the driver options `-MD' or `-MMD', `-MF' overrides
4009 the default dependency output file.
4012 In conjunction with an option such as `-M' requesting dependency
4013 generation, `-MG' assumes missing header files are generated files
4014 and adds them to the dependency list without raising an error.
4015 The dependency filename is taken directly from the `#include'
4016 directive without prepending any path. `-MG' also suppresses
4017 preprocessed output, as a missing header file renders this useless.
4019 This feature is used in automatic updating of makefiles.
4022 This option instructs CPP to add a phony target for each dependency
4023 other than the main file, causing each to depend on nothing. These
4024 dummy rules work around errors `make' gives if you remove header
4025 files without updating the `Makefile' to match.
4027 This is typical output:
4029 test.o: test.c test.h
4034 Change the target of the rule emitted by dependency generation. By
4035 default CPP takes the name of the main input file, including any
4036 path, deletes any file suffix such as `.c', and appends the
4037 platform's usual object suffix. The result is the target.
4039 An `-MT' option will set the target to be exactly the string you
4040 specify. If you want multiple targets, you can specify them as a
4041 single argument to `-MT', or use multiple `-MT' options.
4043 For example, `-MT '$(objpfx)foo.o'' might give
4045 $(objpfx)foo.o: foo.c
4048 Same as `-MT', but it quotes any characters which are special to
4049 Make. `-MQ '$(objpfx)foo.o'' gives
4051 $$(objpfx)foo.o: foo.c
4053 The default target is automatically quoted, as if it were given
4057 `-MD' is equivalent to `-M -MF FILE', except that `-E' is not
4058 implied. The driver determines FILE based on whether an `-o'
4059 option is given. If it is, the driver uses its argument but with
4060 a suffix of `.d', otherwise it take the basename of the input file
4061 and applies a `.d' suffix.
4063 If `-MD' is used in conjunction with `-E', any `-o' switch is
4064 understood to specify the dependency output file (but *note -MF:
4065 dashMF.), but if used without `-E', each `-o' is understood to
4066 specify a target object file.
4068 Since `-E' is not implied, `-MD' can be used to generate a
4069 dependency output file as a side-effect of the compilation process.
4072 Like `-MD' except mention only user header files, not system
4078 `-x assembler-with-cpp'
4079 Specify the source language: C, C++, Objective-C, or assembly.
4080 This has nothing to do with standards conformance or extensions;
4081 it merely selects which base syntax to expect. If you give none
4082 of these options, cpp will deduce the language from the extension
4083 of the source file: `.c', `.cc', `.m', or `.S'. Some other common
4084 extensions for C++ and assembly are also recognized. If cpp does
4085 not recognize the extension, it will treat the file as C; this is
4086 the most generic mode.
4088 _Note:_ Previous versions of cpp accepted a `-lang' option which
4089 selected both the language and the standards conformance level.
4090 This option has been removed, because it conflicts with the `-l'
4095 Specify the standard to which the code should conform. Currently
4096 CPP knows about C and C++ standards; others may be added in the
4099 STANDARD may be one of:
4102 The ISO C standard from 1990. `c89' is the customary
4103 shorthand for this version of the standard.
4105 The `-ansi' option is equivalent to `-std=c89'.
4108 The 1990 C standard, as amended in 1994.
4114 The revised ISO C standard, published in December 1999.
4115 Before publication, this was known as C9X.
4118 The 1990 C standard plus GNU extensions. This is the default.
4122 The 1999 C standard plus GNU extensions.
4125 The 1998 ISO C++ standard plus amendments.
4128 The same as `-std=c++98' plus GNU extensions. This is the
4129 default for C++ code.
4132 Split the include path. Any directories specified with `-I'
4133 options before `-I-' are searched only for headers requested with
4134 `#include "FILE"'; they are not searched for `#include <FILE>'.
4135 If additional directories are specified with `-I' options after
4136 the `-I-', those directories are searched for all `#include'
4139 In addition, `-I-' inhibits the use of the directory of the current
4140 file directory as the first search directory for `#include "FILE"'.
4141 *Note Search Path::. This option has been deprecated.
4144 Do not search the standard system directories for header files.
4145 Only the directories you have specified with `-I' options (and the
4146 directory of the current file, if appropriate) are searched.
4149 Do not search for header files in the C++-specific standard
4150 directories, but do still search the other standard directories.
4151 (This option is used when building the C++ library.)
4154 Process FILE as if `#include "file"' appeared as the first line of
4155 the primary source file. However, the first directory searched
4156 for FILE is the preprocessor's working directory _instead of_ the
4157 directory containing the main source file. If not found there, it
4158 is searched for in the remainder of the `#include "..."' search
4161 If multiple `-include' options are given, the files are included
4162 in the order they appear on the command line.
4165 Exactly like `-include', except that any output produced by
4166 scanning FILE is thrown away. Macros it defines remain defined.
4167 This allows you to acquire all the macros from a header without
4168 also processing its declarations.
4170 All files specified by `-imacros' are processed before all files
4171 specified by `-include'.
4174 Search DIR for header files, but do it _after_ all directories
4175 specified with `-I' and the standard system directories have been
4176 exhausted. DIR is treated as a system include directory.
4179 Specify PREFIX as the prefix for subsequent `-iwithprefix'
4180 options. If the prefix represents a directory, you should include
4184 `-iwithprefixbefore DIR'
4185 Append DIR to the prefix specified previously with `-iprefix', and
4186 add the resulting directory to the include search path.
4187 `-iwithprefixbefore' puts it in the same place `-I' would;
4188 `-iwithprefix' puts it where `-idirafter' would.
4191 This option is like the `--sysroot' option, but applies only to
4192 header files. See the `--sysroot' option for more information.
4195 Search DIR for header files, after all directories specified by
4196 `-I' but before the standard system directories. Mark it as a
4197 system directory, so that it gets the same special treatment as is
4198 applied to the standard system directories. *Note System
4202 Search DIR only for header files requested with `#include "FILE"';
4203 they are not searched for `#include <FILE>', before all
4204 directories specified by `-I' and before the standard system
4205 directories. *Note Search Path::.
4207 `-fdollars-in-identifiers'
4208 Accept `$' in identifiers. *Note Identifier characters::.
4210 `-fextended-identifiers'
4211 Accept universal character names in identifiers. This option is
4212 experimental; in a future version of GCC, it will be enabled by
4213 default for C99 and C++.
4216 Indicate to the preprocessor that the input file has already been
4217 preprocessed. This suppresses things like macro expansion,
4218 trigraph conversion, escaped newline splicing, and processing of
4219 most directives. The preprocessor still recognizes and removes
4220 comments, so that you can pass a file preprocessed with `-C' to
4221 the compiler without problems. In this mode the integrated
4222 preprocessor is little more than a tokenizer for the front ends.
4224 `-fpreprocessed' is implicit if the input file has one of the
4225 extensions `.i', `.ii' or `.mi'. These are the extensions that
4226 GCC uses for preprocessed files created by `-save-temps'.
4229 Set the distance between tab stops. This helps the preprocessor
4230 report correct column numbers in warnings or errors, even if tabs
4231 appear on the line. If the value is less than 1 or greater than
4232 100, the option is ignored. The default is 8.
4234 `-fexec-charset=CHARSET'
4235 Set the execution character set, used for string and character
4236 constants. The default is UTF-8. CHARSET can be any encoding
4237 supported by the system's `iconv' library routine.
4239 `-fwide-exec-charset=CHARSET'
4240 Set the wide execution character set, used for wide string and
4241 character constants. The default is UTF-32 or UTF-16, whichever
4242 corresponds to the width of `wchar_t'. As with `-fexec-charset',
4243 CHARSET can be any encoding supported by the system's `iconv'
4244 library routine; however, you will have problems with encodings
4245 that do not fit exactly in `wchar_t'.
4247 `-finput-charset=CHARSET'
4248 Set the input character set, used for translation from the
4249 character set of the input file to the source character set used
4250 by GCC. If the locale does not specify, or GCC cannot get this
4251 information from the locale, the default is UTF-8. This can be
4252 overridden by either the locale or this command line option.
4253 Currently the command line option takes precedence if there's a
4254 conflict. CHARSET can be any encoding supported by the system's
4255 `iconv' library routine.
4257 `-fworking-directory'
4258 Enable generation of linemarkers in the preprocessor output that
4259 will let the compiler know the current working directory at the
4260 time of preprocessing. When this option is enabled, the
4261 preprocessor will emit, after the initial linemarker, a second
4262 linemarker with the current working directory followed by two
4263 slashes. GCC will use this directory, when it's present in the
4264 preprocessed input, as the directory emitted as the current
4265 working directory in some debugging information formats. This
4266 option is implicitly enabled if debugging information is enabled,
4267 but this can be inhibited with the negated form
4268 `-fno-working-directory'. If the `-P' flag is present in the
4269 command line, this option has no effect, since no `#line'
4270 directives are emitted whatsoever.
4273 Do not print column numbers in diagnostics. This may be necessary
4274 if diagnostics are being scanned by a program that does not
4275 understand the column numbers, such as `dejagnu'.
4277 `-A PREDICATE=ANSWER'
4278 Make an assertion with the predicate PREDICATE and answer ANSWER.
4279 This form is preferred to the older form `-A PREDICATE(ANSWER)',
4280 which is still supported, because it does not use shell special
4281 characters. *Note Assertions::.
4283 `-A -PREDICATE=ANSWER'
4284 Cancel an assertion with the predicate PREDICATE and answer ANSWER.
4287 CHARS is a sequence of one or more of the following characters,
4288 and must not be preceded by a space. Other characters are
4289 interpreted by the compiler proper, or reserved for future
4290 versions of GCC, and so are silently ignored. If you specify
4291 characters whose behavior conflicts, the result is undefined.
4294 Instead of the normal output, generate a list of `#define'
4295 directives for all the macros defined during the execution of
4296 the preprocessor, including predefined macros. This gives
4297 you a way of finding out what is predefined in your version
4298 of the preprocessor. Assuming you have no file `foo.h', the
4301 touch foo.h; cpp -dM foo.h
4303 will show all the predefined macros.
4306 Like `M' except in two respects: it does _not_ include the
4307 predefined macros, and it outputs _both_ the `#define'
4308 directives and the result of preprocessing. Both kinds of
4309 output go to the standard output file.
4312 Like `D', but emit only the macro names, not their expansions.
4315 Output `#include' directives in addition to the result of
4319 Inhibit generation of linemarkers in the output from the
4320 preprocessor. This might be useful when running the preprocessor
4321 on something that is not C code, and will be sent to a program
4322 which might be confused by the linemarkers. *Note Preprocessor
4326 Do not discard comments. All comments are passed through to the
4327 output file, except for comments in processed directives, which
4328 are deleted along with the directive.
4330 You should be prepared for side effects when using `-C'; it causes
4331 the preprocessor to treat comments as tokens in their own right.
4332 For example, comments appearing at the start of what would be a
4333 directive line have the effect of turning that line into an
4334 ordinary source line, since the first token on the line is no
4338 Do not discard comments, including during macro expansion. This is
4339 like `-C', except that comments contained within macros are also
4340 passed through to the output file where the macro is expanded.
4342 In addition to the side-effects of the `-C' option, the `-CC'
4343 option causes all C++-style comments inside a macro to be
4344 converted to C-style comments. This is to prevent later use of
4345 that macro from inadvertently commenting out the remainder of the
4348 The `-CC' option is generally used to support lint comments.
4351 Try to imitate the behavior of old-fashioned C preprocessors, as
4352 opposed to ISO C preprocessors. *Note Traditional Mode::.
4355 Process trigraph sequences. *Note Initial processing::.
4358 Enable special code to work around file systems which only permit
4359 very short file names, such as MS-DOS.
4363 Print text describing all the command line options instead of
4364 preprocessing anything.
4367 Verbose mode. Print out GNU CPP's version number at the beginning
4368 of execution, and report the final form of the include path.
4371 Print the name of each header file used, in addition to other
4372 normal activities. Each name is indented to show how deep in the
4373 `#include' stack it is. Precompiled header files are also
4374 printed, even if they are found to be invalid; an invalid
4375 precompiled header file is printed with `...x' and a valid one
4380 Print out GNU CPP's version number. With one dash, proceed to
4381 preprocess as normal. With two dashes, exit immediately.
4384 File: cpp.info, Node: Environment Variables, Next: GNU Free Documentation License, Prev: Invocation, Up: Top
4386 13 Environment Variables
4387 ************************
4389 This section describes the environment variables that affect how CPP
4390 operates. You can use them to specify directories or prefixes to use
4391 when searching for include files, or to control dependency output.
4393 Note that you can also specify places to search using options such as
4394 `-I', and control dependency output with options like `-M' (*note
4395 Invocation::). These take precedence over environment variables, which
4396 in turn take precedence over the configuration of GCC.
4400 `CPLUS_INCLUDE_PATH'
4402 Each variable's value is a list of directories separated by a
4403 special character, much like `PATH', in which to look for header
4404 files. The special character, `PATH_SEPARATOR', is
4405 target-dependent and determined at GCC build time. For Microsoft
4406 Windows-based targets it is a semicolon, and for almost all other
4407 targets it is a colon.
4409 `CPATH' specifies a list of directories to be searched as if
4410 specified with `-I', but after any paths given with `-I' options
4411 on the command line. This environment variable is used regardless
4412 of which language is being preprocessed.
4414 The remaining environment variables apply only when preprocessing
4415 the particular language indicated. Each specifies a list of
4416 directories to be searched as if specified with `-isystem', but
4417 after any paths given with `-isystem' options on the command line.
4419 In all these variables, an empty element instructs the compiler to
4420 search its current working directory. Empty elements can appear
4421 at the beginning or end of a path. For instance, if the value of
4422 `CPATH' is `:/special/include', that has the same effect as
4423 `-I. -I/special/include'.
4425 See also *Note Search Path::.
4427 `DEPENDENCIES_OUTPUT'
4428 If this variable is set, its value specifies how to output
4429 dependencies for Make based on the non-system header files
4430 processed by the compiler. System header files are ignored in the
4433 The value of `DEPENDENCIES_OUTPUT' can be just a file name, in
4434 which case the Make rules are written to that file, guessing the
4435 target name from the source file name. Or the value can have the
4436 form `FILE TARGET', in which case the rules are written to file
4437 FILE using TARGET as the target name.
4439 In other words, this environment variable is equivalent to
4440 combining the options `-MM' and `-MF' (*note Invocation::), with
4441 an optional `-MT' switch too.
4443 `SUNPRO_DEPENDENCIES'
4444 This variable is the same as `DEPENDENCIES_OUTPUT' (see above),
4445 except that system header files are not ignored, so it implies
4446 `-M' rather than `-MM'. However, the dependence on the main input
4447 file is omitted. *Note Invocation::.
4450 File: cpp.info, Node: GNU Free Documentation License, Next: Index of Directives, Prev: Environment Variables, Up: Top
4452 GNU Free Documentation License
4453 ******************************
4455 Version 1.2, November 2002
4457 Copyright (C) 2000,2001,2002 Free Software Foundation, Inc.
4458 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA
4460 Everyone is permitted to copy and distribute verbatim copies
4461 of this license document, but changing it is not allowed.
4465 The purpose of this License is to make a manual, textbook, or other
4466 functional and useful document "free" in the sense of freedom: to
4467 assure everyone the effective freedom to copy and redistribute it,
4468 with or without modifying it, either commercially or
4469 noncommercially. Secondarily, this License preserves for the
4470 author and publisher a way to get credit for their work, while not
4471 being considered responsible for modifications made by others.
4473 This License is a kind of "copyleft", which means that derivative
4474 works of the document must themselves be free in the same sense.
4475 It complements the GNU General Public License, which is a copyleft
4476 license designed for free software.
4478 We have designed this License in order to use it for manuals for
4479 free software, because free software needs free documentation: a
4480 free program should come with manuals providing the same freedoms
4481 that the software does. But this License is not limited to
4482 software manuals; it can be used for any textual work, regardless
4483 of subject matter or whether it is published as a printed book.
4484 We recommend this License principally for works whose purpose is
4485 instruction or reference.
4487 1. APPLICABILITY AND DEFINITIONS
4489 This License applies to any manual or other work, in any medium,
4490 that contains a notice placed by the copyright holder saying it
4491 can be distributed under the terms of this License. Such a notice
4492 grants a world-wide, royalty-free license, unlimited in duration,
4493 to use that work under the conditions stated herein. The
4494 "Document", below, refers to any such manual or work. Any member
4495 of the public is a licensee, and is addressed as "you". You
4496 accept the license if you copy, modify or distribute the work in a
4497 way requiring permission under copyright law.
4499 A "Modified Version" of the Document means any work containing the
4500 Document or a portion of it, either copied verbatim, or with
4501 modifications and/or translated into another language.
4503 A "Secondary Section" is a named appendix or a front-matter section
4504 of the Document that deals exclusively with the relationship of the
4505 publishers or authors of the Document to the Document's overall
4506 subject (or to related matters) and contains nothing that could
4507 fall directly within that overall subject. (Thus, if the Document
4508 is in part a textbook of mathematics, a Secondary Section may not
4509 explain any mathematics.) The relationship could be a matter of
4510 historical connection with the subject or with related matters, or
4511 of legal, commercial, philosophical, ethical or political position
4514 The "Invariant Sections" are certain Secondary Sections whose
4515 titles are designated, as being those of Invariant Sections, in
4516 the notice that says that the Document is released under this
4517 License. If a section does not fit the above definition of
4518 Secondary then it is not allowed to be designated as Invariant.
4519 The Document may contain zero Invariant Sections. If the Document
4520 does not identify any Invariant Sections then there are none.
4522 The "Cover Texts" are certain short passages of text that are
4523 listed, as Front-Cover Texts or Back-Cover Texts, in the notice
4524 that says that the Document is released under this License. A
4525 Front-Cover Text may be at most 5 words, and a Back-Cover Text may
4526 be at most 25 words.
4528 A "Transparent" copy of the Document means a machine-readable copy,
4529 represented in a format whose specification is available to the
4530 general public, that is suitable for revising the document
4531 straightforwardly with generic text editors or (for images
4532 composed of pixels) generic paint programs or (for drawings) some
4533 widely available drawing editor, and that is suitable for input to
4534 text formatters or for automatic translation to a variety of
4535 formats suitable for input to text formatters. A copy made in an
4536 otherwise Transparent file format whose markup, or absence of
4537 markup, has been arranged to thwart or discourage subsequent
4538 modification by readers is not Transparent. An image format is
4539 not Transparent if used for any substantial amount of text. A
4540 copy that is not "Transparent" is called "Opaque".
4542 Examples of suitable formats for Transparent copies include plain
4543 ASCII without markup, Texinfo input format, LaTeX input format,
4544 SGML or XML using a publicly available DTD, and
4545 standard-conforming simple HTML, PostScript or PDF designed for
4546 human modification. Examples of transparent image formats include
4547 PNG, XCF and JPG. Opaque formats include proprietary formats that
4548 can be read and edited only by proprietary word processors, SGML or
4549 XML for which the DTD and/or processing tools are not generally
4550 available, and the machine-generated HTML, PostScript or PDF
4551 produced by some word processors for output purposes only.
4553 The "Title Page" means, for a printed book, the title page itself,
4554 plus such following pages as are needed to hold, legibly, the
4555 material this License requires to appear in the title page. For
4556 works in formats which do not have any title page as such, "Title
4557 Page" means the text near the most prominent appearance of the
4558 work's title, preceding the beginning of the body of the text.
4560 A section "Entitled XYZ" means a named subunit of the Document
4561 whose title either is precisely XYZ or contains XYZ in parentheses
4562 following text that translates XYZ in another language. (Here XYZ
4563 stands for a specific section name mentioned below, such as
4564 "Acknowledgements", "Dedications", "Endorsements", or "History".)
4565 To "Preserve the Title" of such a section when you modify the
4566 Document means that it remains a section "Entitled XYZ" according
4569 The Document may include Warranty Disclaimers next to the notice
4570 which states that this License applies to the Document. These
4571 Warranty Disclaimers are considered to be included by reference in
4572 this License, but only as regards disclaiming warranties: any other
4573 implication that these Warranty Disclaimers may have is void and
4574 has no effect on the meaning of this License.
4578 You may copy and distribute the Document in any medium, either
4579 commercially or noncommercially, provided that this License, the
4580 copyright notices, and the license notice saying this License
4581 applies to the Document are reproduced in all copies, and that you
4582 add no other conditions whatsoever to those of this License. You
4583 may not use technical measures to obstruct or control the reading
4584 or further copying of the copies you make or distribute. However,
4585 you may accept compensation in exchange for copies. If you
4586 distribute a large enough number of copies you must also follow
4587 the conditions in section 3.
4589 You may also lend copies, under the same conditions stated above,
4590 and you may publicly display copies.
4592 3. COPYING IN QUANTITY
4594 If you publish printed copies (or copies in media that commonly
4595 have printed covers) of the Document, numbering more than 100, and
4596 the Document's license notice requires Cover Texts, you must
4597 enclose the copies in covers that carry, clearly and legibly, all
4598 these Cover Texts: Front-Cover Texts on the front cover, and
4599 Back-Cover Texts on the back cover. Both covers must also clearly
4600 and legibly identify you as the publisher of these copies. The
4601 front cover must present the full title with all words of the
4602 title equally prominent and visible. You may add other material
4603 on the covers in addition. Copying with changes limited to the
4604 covers, as long as they preserve the title of the Document and
4605 satisfy these conditions, can be treated as verbatim copying in
4608 If the required texts for either cover are too voluminous to fit
4609 legibly, you should put the first ones listed (as many as fit
4610 reasonably) on the actual cover, and continue the rest onto
4613 If you publish or distribute Opaque copies of the Document
4614 numbering more than 100, you must either include a
4615 machine-readable Transparent copy along with each Opaque copy, or
4616 state in or with each Opaque copy a computer-network location from
4617 which the general network-using public has access to download
4618 using public-standard network protocols a complete Transparent
4619 copy of the Document, free of added material. If you use the
4620 latter option, you must take reasonably prudent steps, when you
4621 begin distribution of Opaque copies in quantity, to ensure that
4622 this Transparent copy will remain thus accessible at the stated
4623 location until at least one year after the last time you
4624 distribute an Opaque copy (directly or through your agents or
4625 retailers) of that edition to the public.
4627 It is requested, but not required, that you contact the authors of
4628 the Document well before redistributing any large number of
4629 copies, to give them a chance to provide you with an updated
4630 version of the Document.
4634 You may copy and distribute a Modified Version of the Document
4635 under the conditions of sections 2 and 3 above, provided that you
4636 release the Modified Version under precisely this License, with
4637 the Modified Version filling the role of the Document, thus
4638 licensing distribution and modification of the Modified Version to
4639 whoever possesses a copy of it. In addition, you must do these
4640 things in the Modified Version:
4642 A. Use in the Title Page (and on the covers, if any) a title
4643 distinct from that of the Document, and from those of
4644 previous versions (which should, if there were any, be listed
4645 in the History section of the Document). You may use the
4646 same title as a previous version if the original publisher of
4647 that version gives permission.
4649 B. List on the Title Page, as authors, one or more persons or
4650 entities responsible for authorship of the modifications in
4651 the Modified Version, together with at least five of the
4652 principal authors of the Document (all of its principal
4653 authors, if it has fewer than five), unless they release you
4654 from this requirement.
4656 C. State on the Title page the name of the publisher of the
4657 Modified Version, as the publisher.
4659 D. Preserve all the copyright notices of the Document.
4661 E. Add an appropriate copyright notice for your modifications
4662 adjacent to the other copyright notices.
4664 F. Include, immediately after the copyright notices, a license
4665 notice giving the public permission to use the Modified
4666 Version under the terms of this License, in the form shown in
4669 G. Preserve in that license notice the full lists of Invariant
4670 Sections and required Cover Texts given in the Document's
4673 H. Include an unaltered copy of this License.
4675 I. Preserve the section Entitled "History", Preserve its Title,
4676 and add to it an item stating at least the title, year, new
4677 authors, and publisher of the Modified Version as given on
4678 the Title Page. If there is no section Entitled "History" in
4679 the Document, create one stating the title, year, authors,
4680 and publisher of the Document as given on its Title Page,
4681 then add an item describing the Modified Version as stated in
4682 the previous sentence.
4684 J. Preserve the network location, if any, given in the Document
4685 for public access to a Transparent copy of the Document, and
4686 likewise the network locations given in the Document for
4687 previous versions it was based on. These may be placed in
4688 the "History" section. You may omit a network location for a
4689 work that was published at least four years before the
4690 Document itself, or if the original publisher of the version
4691 it refers to gives permission.
4693 K. For any section Entitled "Acknowledgements" or "Dedications",
4694 Preserve the Title of the section, and preserve in the
4695 section all the substance and tone of each of the contributor
4696 acknowledgements and/or dedications given therein.
4698 L. Preserve all the Invariant Sections of the Document,
4699 unaltered in their text and in their titles. Section numbers
4700 or the equivalent are not considered part of the section
4703 M. Delete any section Entitled "Endorsements". Such a section
4704 may not be included in the Modified Version.
4706 N. Do not retitle any existing section to be Entitled
4707 "Endorsements" or to conflict in title with any Invariant
4710 O. Preserve any Warranty Disclaimers.
4712 If the Modified Version includes new front-matter sections or
4713 appendices that qualify as Secondary Sections and contain no
4714 material copied from the Document, you may at your option
4715 designate some or all of these sections as invariant. To do this,
4716 add their titles to the list of Invariant Sections in the Modified
4717 Version's license notice. These titles must be distinct from any
4718 other section titles.
4720 You may add a section Entitled "Endorsements", provided it contains
4721 nothing but endorsements of your Modified Version by various
4722 parties--for example, statements of peer review or that the text
4723 has been approved by an organization as the authoritative
4724 definition of a standard.
4726 You may add a passage of up to five words as a Front-Cover Text,
4727 and a passage of up to 25 words as a Back-Cover Text, to the end
4728 of the list of Cover Texts in the Modified Version. Only one
4729 passage of Front-Cover Text and one of Back-Cover Text may be
4730 added by (or through arrangements made by) any one entity. If the
4731 Document already includes a cover text for the same cover,
4732 previously added by you or by arrangement made by the same entity
4733 you are acting on behalf of, you may not add another; but you may
4734 replace the old one, on explicit permission from the previous
4735 publisher that added the old one.
4737 The author(s) and publisher(s) of the Document do not by this
4738 License give permission to use their names for publicity for or to
4739 assert or imply endorsement of any Modified Version.
4741 5. COMBINING DOCUMENTS
4743 You may combine the Document with other documents released under
4744 this License, under the terms defined in section 4 above for
4745 modified versions, provided that you include in the combination
4746 all of the Invariant Sections of all of the original documents,
4747 unmodified, and list them all as Invariant Sections of your
4748 combined work in its license notice, and that you preserve all
4749 their Warranty Disclaimers.
4751 The combined work need only contain one copy of this License, and
4752 multiple identical Invariant Sections may be replaced with a single
4753 copy. If there are multiple Invariant Sections with the same name
4754 but different contents, make the title of each such section unique
4755 by adding at the end of it, in parentheses, the name of the
4756 original author or publisher of that section if known, or else a
4757 unique number. Make the same adjustment to the section titles in
4758 the list of Invariant Sections in the license notice of the
4761 In the combination, you must combine any sections Entitled
4762 "History" in the various original documents, forming one section
4763 Entitled "History"; likewise combine any sections Entitled
4764 "Acknowledgements", and any sections Entitled "Dedications". You
4765 must delete all sections Entitled "Endorsements."
4767 6. COLLECTIONS OF DOCUMENTS
4769 You may make a collection consisting of the Document and other
4770 documents released under this License, and replace the individual
4771 copies of this License in the various documents with a single copy
4772 that is included in the collection, provided that you follow the
4773 rules of this License for verbatim copying of each of the
4774 documents in all other respects.
4776 You may extract a single document from such a collection, and
4777 distribute it individually under this License, provided you insert
4778 a copy of this License into the extracted document, and follow
4779 this License in all other respects regarding verbatim copying of
4782 7. AGGREGATION WITH INDEPENDENT WORKS
4784 A compilation of the Document or its derivatives with other
4785 separate and independent documents or works, in or on a volume of
4786 a storage or distribution medium, is called an "aggregate" if the
4787 copyright resulting from the compilation is not used to limit the
4788 legal rights of the compilation's users beyond what the individual
4789 works permit. When the Document is included in an aggregate, this
4790 License does not apply to the other works in the aggregate which
4791 are not themselves derivative works of the Document.
4793 If the Cover Text requirement of section 3 is applicable to these
4794 copies of the Document, then if the Document is less than one half
4795 of the entire aggregate, the Document's Cover Texts may be placed
4796 on covers that bracket the Document within the aggregate, or the
4797 electronic equivalent of covers if the Document is in electronic
4798 form. Otherwise they must appear on printed covers that bracket
4799 the whole aggregate.
4803 Translation is considered a kind of modification, so you may
4804 distribute translations of the Document under the terms of section
4805 4. Replacing Invariant Sections with translations requires special
4806 permission from their copyright holders, but you may include
4807 translations of some or all Invariant Sections in addition to the
4808 original versions of these Invariant Sections. You may include a
4809 translation of this License, and all the license notices in the
4810 Document, and any Warranty Disclaimers, provided that you also
4811 include the original English version of this License and the
4812 original versions of those notices and disclaimers. In case of a
4813 disagreement between the translation and the original version of
4814 this License or a notice or disclaimer, the original version will
4817 If a section in the Document is Entitled "Acknowledgements",
4818 "Dedications", or "History", the requirement (section 4) to
4819 Preserve its Title (section 1) will typically require changing the
4824 You may not copy, modify, sublicense, or distribute the Document
4825 except as expressly provided for under this License. Any other
4826 attempt to copy, modify, sublicense or distribute the Document is
4827 void, and will automatically terminate your rights under this
4828 License. However, parties who have received copies, or rights,
4829 from you under this License will not have their licenses
4830 terminated so long as such parties remain in full compliance.
4832 10. FUTURE REVISIONS OF THIS LICENSE
4834 The Free Software Foundation may publish new, revised versions of
4835 the GNU Free Documentation License from time to time. Such new
4836 versions will be similar in spirit to the present version, but may
4837 differ in detail to address new problems or concerns. See
4838 `http://www.gnu.org/copyleft/'.
4840 Each version of the License is given a distinguishing version
4841 number. If the Document specifies that a particular numbered
4842 version of this License "or any later version" applies to it, you
4843 have the option of following the terms and conditions either of
4844 that specified version or of any later version that has been
4845 published (not as a draft) by the Free Software Foundation. If
4846 the Document does not specify a version number of this License,
4847 you may choose any version ever published (not as a draft) by the
4848 Free Software Foundation.
4850 ADDENDUM: How to use this License for your documents
4851 ====================================================
4853 To use this License in a document you have written, include a copy of
4854 the License in the document and put the following copyright and license
4855 notices just after the title page:
4857 Copyright (C) YEAR YOUR NAME.
4858 Permission is granted to copy, distribute and/or modify this document
4859 under the terms of the GNU Free Documentation License, Version 1.2
4860 or any later version published by the Free Software Foundation;
4861 with no Invariant Sections, no Front-Cover Texts, and no Back-Cover
4862 Texts. A copy of the license is included in the section entitled ``GNU
4863 Free Documentation License''.
4865 If you have Invariant Sections, Front-Cover Texts and Back-Cover
4866 Texts, replace the "with...Texts." line with this:
4868 with the Invariant Sections being LIST THEIR TITLES, with
4869 the Front-Cover Texts being LIST, and with the Back-Cover Texts
4872 If you have Invariant Sections without Cover Texts, or some other
4873 combination of the three, merge those two alternatives to suit the
4876 If your document contains nontrivial examples of program code, we
4877 recommend releasing these examples in parallel under your choice of
4878 free software license, such as the GNU General Public License, to
4879 permit their use in free software.
4882 File: cpp.info, Node: Index of Directives, Next: Option Index, Prev: GNU Free Documentation License, Up: Top
4890 * #assert: Assertions. (line 41)
4891 * #define: Object-like Macros. (line 11)
4892 * #elif: Elif. (line 6)
4893 * #else: Else. (line 6)
4894 * #endif: Ifdef. (line 6)
4895 * #error: Diagnostics. (line 6)
4896 * #ident: Other Directives. (line 6)
4897 * #if: Conditional Syntax. (line 6)
4898 * #ifdef: Ifdef. (line 6)
4899 * #ifndef: Ifdef. (line 40)
4900 * #import: Obsolete once-only headers.
4902 * #include: Include Syntax. (line 6)
4903 * #include_next: Wrapper Headers. (line 6)
4904 * #line: Line Control. (line 20)
4905 * #pragma GCC dependency: Pragmas. (line 53)
4906 * #pragma GCC poison: Pragmas. (line 65)
4907 * #pragma GCC system_header <1>: Pragmas. (line 92)
4908 * #pragma GCC system_header: System Headers. (line 31)
4909 * #sccs: Other Directives. (line 6)
4910 * #unassert: Assertions. (line 52)
4911 * #undef: Undefining and Redefining Macros.
4913 * #warning: Diagnostics. (line 27)
4916 File: cpp.info, Node: Option Index, Next: Concept Index, Prev: Index of Directives, Up: Top
4921 CPP's command line options and environment variables are indexed here
4922 without any initial `-' or `--'.
4927 * A: Invocation. (line 490)
4928 * ansi: Invocation. (line 307)
4929 * C: Invocation. (line 538)
4930 * C_INCLUDE_PATH: Environment Variables.
4932 * CPATH: Environment Variables.
4934 * CPLUS_INCLUDE_PATH: Environment Variables.
4936 * D: Invocation. (line 39)
4937 * dD: Invocation. (line 518)
4938 * DEPENDENCIES_OUTPUT: Environment Variables.
4940 * dI: Invocation. (line 527)
4941 * dM: Invocation. (line 506)
4942 * dN: Invocation. (line 524)
4943 * fdollars-in-identifiers: Invocation. (line 420)
4944 * fexec-charset: Invocation. (line 447)
4945 * fextended-identifiers: Invocation. (line 423)
4946 * finput-charset: Invocation. (line 460)
4947 * fno-show-column: Invocation. (line 485)
4948 * fno-working-directory: Invocation. (line 470)
4949 * fpreprocessed: Invocation. (line 428)
4950 * ftabstop: Invocation. (line 441)
4951 * fwide-exec-charset: Invocation. (line 452)
4952 * fworking-directory: Invocation. (line 470)
4953 * H: Invocation. (line 583)
4954 * help: Invocation. (line 575)
4955 * I: Invocation. (line 71)
4956 * I-: Invocation. (line 344)
4957 * idirafter: Invocation. (line 386)
4958 * imacros: Invocation. (line 377)
4959 * include: Invocation. (line 366)
4960 * iprefix: Invocation. (line 391)
4961 * iquote: Invocation. (line 414)
4962 * isysroot: Invocation. (line 403)
4963 * isystem: Invocation. (line 407)
4964 * iwithprefix: Invocation. (line 397)
4965 * iwithprefixbefore: Invocation. (line 397)
4966 * M: Invocation. (line 181)
4967 * MD: Invocation. (line 269)
4968 * MF: Invocation. (line 215)
4969 * MG: Invocation. (line 224)
4970 * MM: Invocation. (line 205)
4971 * MMD: Invocation. (line 284)
4972 * MP: Invocation. (line 234)
4973 * MQ: Invocation. (line 260)
4974 * MT: Invocation. (line 246)
4975 * nostdinc: Invocation. (line 356)
4976 * nostdinc++: Invocation. (line 361)
4977 * o: Invocation. (line 80)
4978 * OBJC_INCLUDE_PATH: Environment Variables.
4980 * P: Invocation. (line 531)
4981 * pedantic: Invocation. (line 171)
4982 * pedantic-errors: Invocation. (line 176)
4983 * remap: Invocation. (line 570)
4984 * std=: Invocation. (line 307)
4985 * SUNPRO_DEPENDENCIES: Environment Variables.
4987 * target-help: Invocation. (line 575)
4988 * traditional-cpp: Invocation. (line 563)
4989 * trigraphs: Invocation. (line 567)
4990 * U: Invocation. (line 62)
4991 * undef: Invocation. (line 66)
4992 * v: Invocation. (line 579)
4993 * version: Invocation. (line 592)
4994 * w: Invocation. (line 167)
4995 * Wall: Invocation. (line 86)
4996 * Wcomment: Invocation. (line 94)
4997 * Wcomments: Invocation. (line 94)
4998 * Wendif-labels: Invocation. (line 144)
4999 * Werror: Invocation. (line 157)
5000 * Wimport: Invocation. (line 117)
5001 * Wsystem-headers: Invocation. (line 161)
5002 * Wtraditional: Invocation. (line 111)
5003 * Wtrigraphs: Invocation. (line 99)
5004 * Wundef: Invocation. (line 120)
5005 * Wunused-macros: Invocation. (line 125)
5006 * x: Invocation. (line 291)
5009 File: cpp.info, Node: Concept Index, Prev: Option Index, Up: Top
5017 * # operator: Stringification. (line 6)
5018 * ## operator: Concatenation. (line 6)
5019 * _Pragma: Pragmas. (line 25)
5020 * alternative tokens: Tokenization. (line 106)
5021 * arguments: Macro Arguments. (line 6)
5022 * arguments in macro definitions: Macro Arguments. (line 6)
5023 * assertions: Assertions. (line 6)
5024 * assertions, canceling: Assertions. (line 52)
5025 * backslash-newline: Initial processing. (line 61)
5026 * block comments: Initial processing. (line 77)
5027 * C++ named operators: C++ Named Operators. (line 6)
5028 * character constants: Tokenization. (line 85)
5029 * character set, execution: Invocation. (line 447)
5030 * character set, input: Invocation. (line 460)
5031 * character set, wide execution: Invocation. (line 452)
5032 * command line: Invocation. (line 6)
5033 * commenting out code: Deleted Code. (line 6)
5034 * comments: Initial processing. (line 77)
5035 * common predefined macros: Common Predefined Macros.
5037 * computed includes: Computed Includes. (line 6)
5038 * concatenation: Concatenation. (line 6)
5039 * conditional group: Ifdef. (line 14)
5040 * conditionals: Conditionals. (line 6)
5041 * continued lines: Initial processing. (line 61)
5042 * controlling macro: Once-Only Headers. (line 35)
5043 * defined: Defined. (line 6)
5044 * dependencies for make as output: Environment Variables.
5046 * dependencies, make: Invocation. (line 181)
5047 * diagnostic: Diagnostics. (line 6)
5048 * differences from previous versions: Differences from previous versions.
5050 * digraphs: Tokenization. (line 106)
5051 * directive line: The preprocessing language.
5053 * directive name: The preprocessing language.
5055 * directives: The preprocessing language.
5057 * empty macro arguments: Macro Arguments. (line 66)
5058 * environment variables: Environment Variables.
5060 * expansion of arguments: Argument Prescan. (line 6)
5061 * FDL, GNU Free Documentation License: GNU Free Documentation License.
5063 * function-like macros: Function-like Macros.
5065 * grouping options: Invocation. (line 34)
5066 * guard macro: Once-Only Headers. (line 35)
5067 * header file: Header Files. (line 6)
5068 * header file names: Tokenization. (line 85)
5069 * identifiers: Tokenization. (line 34)
5070 * implementation limits: Implementation limits.
5072 * implementation-defined behavior: Implementation-defined behavior.
5074 * including just once: Once-Only Headers. (line 6)
5075 * invocation: Invocation. (line 6)
5076 * iso646.h: C++ Named Operators. (line 6)
5077 * line comments: Initial processing. (line 77)
5078 * line control: Line Control. (line 6)
5079 * line endings: Initial processing. (line 14)
5080 * linemarkers: Preprocessor Output. (line 28)
5081 * macro argument expansion: Argument Prescan. (line 6)
5082 * macro arguments and directives: Directives Within Macro Arguments.
5084 * macros in include: Computed Includes. (line 6)
5085 * macros with arguments: Macro Arguments. (line 6)
5086 * macros with variable arguments: Variadic Macros. (line 6)
5087 * make: Invocation. (line 181)
5088 * manifest constants: Object-like Macros. (line 6)
5089 * named operators: C++ Named Operators. (line 6)
5090 * newlines in macro arguments: Newlines in Arguments.
5092 * null directive: Other Directives. (line 15)
5093 * numbers: Tokenization. (line 61)
5094 * object-like macro: Object-like Macros. (line 6)
5095 * options: Invocation. (line 38)
5096 * options, grouping: Invocation. (line 34)
5097 * other tokens: Tokenization. (line 120)
5098 * output format: Preprocessor Output. (line 12)
5099 * overriding a header file: Wrapper Headers. (line 6)
5100 * parentheses in macro bodies: Operator Precedence Problems.
5102 * pitfalls of macros: Macro Pitfalls. (line 6)
5103 * predefined macros: Predefined Macros. (line 6)
5104 * predefined macros, system-specific: System-specific Predefined Macros.
5106 * predicates: Assertions. (line 19)
5107 * preprocessing directives: The preprocessing language.
5109 * preprocessing numbers: Tokenization. (line 61)
5110 * preprocessing tokens: Tokenization. (line 6)
5111 * prescan of macro arguments: Argument Prescan. (line 6)
5112 * problems with macros: Macro Pitfalls. (line 6)
5113 * punctuators: Tokenization. (line 106)
5114 * redefining macros: Undefining and Redefining Macros.
5116 * repeated inclusion: Once-Only Headers. (line 6)
5117 * reporting errors: Diagnostics. (line 6)
5118 * reporting warnings: Diagnostics. (line 6)
5119 * reserved namespace: System-specific Predefined Macros.
5121 * self-reference: Self-Referential Macros.
5123 * semicolons (after macro calls): Swallowing the Semicolon.
5125 * side effects (in macro arguments): Duplication of Side Effects.
5127 * standard predefined macros.: Standard Predefined Macros.
5129 * string constants: Tokenization. (line 85)
5130 * string literals: Tokenization. (line 85)
5131 * stringification: Stringification. (line 6)
5132 * symbolic constants: Object-like Macros. (line 6)
5133 * system header files <1>: System Headers. (line 6)
5134 * system header files: Header Files. (line 13)
5135 * system-specific predefined macros: System-specific Predefined Macros.
5137 * testing predicates: Assertions. (line 30)
5138 * token concatenation: Concatenation. (line 6)
5139 * token pasting: Concatenation. (line 6)
5140 * tokens: Tokenization. (line 6)
5141 * trigraphs: Initial processing. (line 32)
5142 * undefining macros: Undefining and Redefining Macros.
5144 * unsafe macros: Duplication of Side Effects.
5146 * variable number of arguments: Variadic Macros. (line 6)
5147 * variadic macros: Variadic Macros. (line 6)
5148 * wrapper #ifndef: Once-Only Headers. (line 6)
5149 * wrapper headers: Wrapper Headers. (line 6)
5155 Node: Overview
\7f3769
5156 Node: Character sets
\7f6590
5157 Ref: Character sets-Footnote-1
\7f9074
5158 Node: Initial processing
\7f9255
5159 Ref: trigraphs
\7f10814
5160 Node: Tokenization
\7f15016
5161 Ref: Tokenization-Footnote-1
\7f22152
5162 Node: The preprocessing language
\7f22263
5163 Node: Header Files
\7f25141
5164 Node: Include Syntax
\7f27021
5165 Node: Include Operation
\7f28658
5166 Node: Search Path
\7f30506
5167 Node: Once-Only Headers
\7f33696
5168 Node: Computed Includes
\7f35341
5169 Node: Wrapper Headers
\7f38485
5170 Node: System Headers
\7f40911
5171 Node: Macros
\7f42961
5172 Node: Object-like Macros
\7f44102
5173 Node: Function-like Macros
\7f47692
5174 Node: Macro Arguments
\7f49308
5175 Node: Stringification
\7f53453
5176 Node: Concatenation
\7f56659
5177 Node: Variadic Macros
\7f59767
5178 Node: Predefined Macros
\7f64554
5179 Node: Standard Predefined Macros
\7f65142
5180 Node: Common Predefined Macros
\7f71078
5181 Node: System-specific Predefined Macros
\7f80242
5182 Node: C++ Named Operators
\7f82263
5183 Node: Undefining and Redefining Macros
\7f83227
5184 Node: Directives Within Macro Arguments
\7f85331
5185 Node: Macro Pitfalls
\7f86879
5186 Node: Misnesting
\7f87412
5187 Node: Operator Precedence Problems
\7f88524
5188 Node: Swallowing the Semicolon
\7f90390
5189 Node: Duplication of Side Effects
\7f92413
5190 Node: Self-Referential Macros
\7f94596
5191 Node: Argument Prescan
\7f97005
5192 Node: Newlines in Arguments
\7f100759
5193 Node: Conditionals
\7f101710
5194 Node: Conditional Uses
\7f103540
5195 Node: Conditional Syntax
\7f104898
5196 Node: Ifdef
\7f105218
5198 Node: Defined
\7f110683
5201 Node: Deleted Code
\7f113825
5202 Node: Diagnostics
\7f115072
5203 Node: Line Control
\7f116689
5204 Node: Pragmas
\7f120493
5205 Node: Other Directives
\7f124763
5206 Node: Preprocessor Output
\7f125813
5207 Node: Traditional Mode
\7f129014
5208 Node: Traditional lexical analysis
\7f130072
5209 Node: Traditional macros
\7f132575
5210 Node: Traditional miscellany
\7f136377
5211 Node: Traditional warnings
\7f137374
5212 Node: Implementation Details
\7f139571
5213 Node: Implementation-defined behavior
\7f140192
5214 Ref: Identifier characters
\7f140925
5215 Node: Implementation limits
\7f144010
5216 Node: Obsolete Features
\7f146684
5217 Node: Assertions
\7f147139
5218 Node: Obsolete once-only headers
\7f149680
5219 Node: Differences from previous versions
\7f151415
5220 Node: Invocation
\7f155623
5221 Ref: Wtrigraphs
\7f159954
5222 Ref: dashMF
\7f164738
5223 Ref: fdollars-in-identifiers
\7f172653
5224 Node: Environment Variables
\7f180306
5225 Node: GNU Free Documentation License
\7f183272
5226 Node: Index of Directives
\7f205705
5227 Node: Option Index
\7f207634
5228 Node: Concept Index
\7f213672