![[BACK]](/cvsweb/icons/back.gif){kind=icon}
Return to cpp.texinfo CVS log ![[TXT]](/cvsweb/icons/text.gif){kind=icon}
![[DIR]](/cvsweb/icons/dir.gif){kind=icon}
Up to [GCC 1.x] / gcc|
Return to cpp.texinfo CVS log | Up to [GCC 1.x] / gcc |
1.1 root 1: \input texinfo
1.1.1.2 root 2: @setfilename cpp.info
1.1 root 3: @settitle The C Preprocessor
4: @ifinfo
5: This file documents the GNU C Preprocessor.
6:
1.1.1.6 root 7: Copyright (C) 1987, 1989 Free Software Foundation, Inc.
1.1 root 8:
9: Permission is granted to make and distribute verbatim copies of
10: this manual provided the copyright notice and this permission notice
11: are preserved on all copies.
12:
13: @ignore
14: Permission is granted to process this file through Tex and print the
15: results, provided the printed document carries copying permission
16: notice identical to this one except for the removal of this paragraph
17: (this paragraph not being relevant to the printed manual).
18:
19: @end ignore
20: Permission is granted to copy and distribute modified versions of this
21: manual under the conditions for verbatim copying, provided also that
22: the entire resulting derived work is distributed under the terms of a
23: permission notice identical to this one.
24:
25: Permission is granted to copy and distribute translations of this manual
26: into another language, under the above conditions for modified versions.
27: @end ifinfo
28:
29: @titlepage
30: @sp 6
31: @center @titlefont{The C Preprocessor}
32: @sp 4
33: @center First Edition
34: @sp 1
1.1.1.6 root 35: @center April 1989
1.1 root 36: @sp 5
37: @center Richard M. Stallman
38: @page
39: @vskip 0pt plus 1filll
1.1.1.6 root 40: Copyright @copyright{} 1987, 1989 Free Software Foundation, Inc.
1.1 root 41:
42: Permission is granted to make and distribute verbatim copies of
43: this manual provided the copyright notice and this permission notice
44: are preserved on all copies.
45:
46: Permission is granted to copy and distribute modified versions of this
47: manual under the conditions for verbatim copying, provided also that
48: the entire resulting derived work is distributed under the terms of a
49: permission notice identical to this one.
50:
51: Permission is granted to copy and distribute translations of this manual
52: into another language, under the above conditions for modified versions.
53: @end titlepage
54: @page
55:
56: @node Top, Global Actions,, (DIR)
57: @chapter The C Preprocessor
58:
59: The C preprocessor is a @dfn{macro processor} that is used automatically by
60: the C compiler to transform your program before actual compilation. It is
61: called a macro processor because it allows you to define @dfn{macros},
62: which are brief abbreviations for longer constructs.
63:
64: The C preprocessor provides four separate facilities that you can use as
65: you see fit:
66:
67: @itemize @bullet
68: @item
69: Inclusion of header files. These are files of declarations that can be
70: substituted into your program.
71:
72: @item
73: Macro expansion. You can define @dfn{macros}, which are abbreviations
74: for arbitrary fragments of C code, and then the C preprocessor will
75: replace the macros with their definitions throughout the program.
76:
77: @item
78: Conditional compilation. Using special preprocessor commands, you
79: can include or exclude parts of the program according to various
80: conditions.
81:
82: @item
83: Line control. If you use a program to combine or rearrange source files into
84: an intermediate file which is then compiled, you can use line control
85: to inform the compiler of where each source line originally came from.
86: @end itemize
87:
88: C preprocessors vary in some details. This manual discusses the GNU C
89: preprocessor, the C Compatible Compiler Preprocessor. The GNU C
90: preprocessor provides a superset of the features of ANSI Standard C.
91:
92: ANSI Standard C requires the rejection of many harmless constructs commonly
93: used by today's C programs. Such incompatibility would be inconvenient for
94: users, so the GNU C preprocessor is configured to accept these constructs
95: by default. Strictly speaking, to get ANSI Standard C, you must use the
1.1.1.6 root 96: options @samp{-trigraphs}, @samp{-undef} and @samp{-pedantic}, but in
97: practice the consequences of having strict ANSI Standard C make it
98: undesirable to do this. @xref{Invocation}.
1.1 root 99:
100: @menu
101: * Global Actions:: Actions made uniformly on all input files.
102: * Commands:: General syntax of preprocessor commands.
103: * Header Files:: How and why to use header files.
104: * Macros:: How and why to use macros.
105: * Conditionals:: How and why to use conditionals.
106: * Combining Sources:: Use of line control when you combine source files.
107: * Other Commands:: Miscellaneous preprocessor commands.
108: * Output:: Format of output from the C preprocessor.
1.1.1.2 root 109: * Invocation:: How to invoke the preprocessor; command options.
1.1 root 110: * Concept Index:: Index of concepts and terms.
1.1.1.2 root 111: * Index:: Index of commands, predefined macros and options.
1.1 root 112: @end menu
113:
114: @node Global Actions, Commands, Top, Top
115: @section Transformations Made Globally
116:
117: Most C preprocessor features are inactive unless you give specific commands
118: to request their use. (Preprocessor commands are lines starting with
119: @samp{#}; @pxref{Commands}). But there are three transformations that the
120: preprocessor always makes on all the input it receives, even in the absence
121: of commands.
122:
123: @itemize @bullet
124: @item
125: All C comments are replaced with single spaces.
126:
127: @item
128: Backslash-Newline sequences are deleted, no matter where. This
129: feature allows you to break long lines for cosmetic purposes without
130: changing their meaning.
131:
132: @item
133: Predefined macro names are replaced with their expansions
134: (@pxref{Predefined}).
135: @end itemize
136:
137: The first two transformations are done @emph{before} nearly all other parsing
138: and before preprocessor commands are recognized. Thus, for example, you
1.1.1.2 root 139: can split a line cosmetically with Backslash-Newline anywhere (except
140: when trigraphs are in use; see below).
1.1 root 141:
142: @example
1.1.1.2 root 143: /*
144: */ # /*
145: */ defi\
1.1 root 146: ne FO\
147: O 10\
148: 20
149: @end example
150:
151: @noindent
1.1.1.2 root 152: is equivalent into @samp{#define FOO 1020}. You can split even an escape
153: sequence with Backslash-Newline. For example, you can split @code{"foo\bar"}
154: between the @samp{\} and the @samp{b} to get
1.1 root 155:
1.1.1.2 root 156: @example
157: "foo\\
158: bar"
159: @end example
160:
161: @noindent
162: This behavior is unclean: in all other contexts, a Backslash can be
163: inserted in a string constant as an ordinary character by writing a double
164: Backslash, and this creates an exception. But the ANSI C standard requires
165: it. (Strict ANSI C does not allow Newlines in string constants, so they
166: do not consider this a problem.)
167:
168: But there are a few exceptions to all three transformations.
1.1 root 169:
170: @itemize @bullet
171: @item
172: C comments and predefined macro names are not recognized inside a
173: @samp{#include} command in which the file name is delimited with
174: @samp{<} and @samp{>}.
175:
176: @item
177: C comments and predefined macro names are never recognized within a
178: character or string constant. (Strictly speaking, this is the rule,
179: not an exception, but it is worth noting here anyway.)
180:
181: @item
1.1.1.2 root 182: Backslash-Newline may not safely be used within an ANSI ``trigraph''.
183: Trigraphs are converted before Backslash-Newline is deleted. If you
184: write what looks like a trigraph with a Backslash-Newline inside, the
185: Backslash-Newline is deleted as usual, but it is then too late to
186: recognize the trigraph.
1.1 root 187:
1.1.1.6 root 188: This exception is relevant only if you use the @samp{-trigraphs}
189: option to enable trigraph processing. @xref{Invocation}.
1.1 root 190: @end itemize
191:
192: @node Commands, Header Files, Global Actions, Top
193: @section Preprocessor Commands
194:
195: @cindex preprocessor commands
196: @cindex commands
197: Most preprocessor features are active only if you use preprocessor commands
198: to request their use.
199:
200: Preprocessor commands are lines in your program that start with @samp{#}.
201: The @samp{#} is followed by an identifier that is the @dfn{command name}.
202: For example, @samp{#define} is the command that defines a macro.
203: Whitespace is also allowed before and after the @samp{#}.
204:
205: The set of valid command names is fixed. Programs cannot define new
206: preprocessor commands.
207:
208: Some command names require arguments; these make up the rest of the command
209: line and must be separated from the command name by whitespace. For example,
210: @samp{#define} must be followed by a macro name and the intended expansion
211: of the macro.
212:
213: A preprocessor command cannot be more than one line in normal circumstances.
214: It may be split cosmetically with Backslash-Newline, but that has no effect
215: on its meaning. Comments containing Newlines can also divide the command into
216: multiple lines, but the comments are changed to Spaces before the command
217: is interpreted. The only way a significant Newline can occur in a preprocessor
218: command is within a string constant or character constant. Note that
219: most C compilers that might be applied to the output from the preprocessor
220: do not accept string or character constants containing Newlines.
221:
222: The @samp{#} and the command name cannot come from a macro expansion. For
223: example, if @samp{foo} is defined as a macro expanding to @samp{define},
224: that does not make @samp{#foo} a valid preprocessor command.
225:
226: @node Header Files, Macros, Commands, Top
227: @section Header Files
228:
229: @cindex header file
230: A header file is a file containing C declarations and macro definitions
231: (@pxref{Macros}) to be shared between several source files. You request
232: the use of a header file in your program with the C preprocessor command
233: @samp{#include}.
234:
1.1.1.6 root 235: @menu
236: * Header Uses:: What header files are used for.
237: * Include Syntax:: How to write @samp{#include} commands.
238: * Include Operation:: What @samp{#include} does.
239: * Once-Only:: Preventing multiple inclusion of one header file.
240: @end menu
241:
1.1 root 242: @node Header Uses, Include Syntax, Header Files, Header Files
243: @subsection Uses of Header Files
244:
245: Header files serve two kinds of purposes.
246:
247: @itemize @bullet
248: @item
249: @findex system header files
250: System header files declare the interfaces to parts of the operating
251: system. You include them in your program to supply the definitions
252: you need to invoke system calls and libraries.
253:
254: @item
255: Your own header files contain declarations for interfaces between the
256: source files of your program. Each time you have a group of related
257: declarations and macro definitions all or most of which are needed in
258: several different source files, it is a good idea to create a header
259: file for them.
260: @end itemize
261:
262: Including a header file produces the same results in C compilation as
263: copying the header file into each source file that needs it. But such
264: copying would be time-consuming and error-prone. With a header file, the
265: related declarations appear in only one place. If they need to be changed,
266: they can be changed in one place, and programs that include the header file
267: will automatically use the new version when next recompiled. The header
268: file eliminates the labor of finding and changing all the copies as well as
269: the risk that a failure to find one copy will result in inconsistencies
270: within a program.
271:
272: The usual convention is to give header files names that end with @file{.h}.
273:
274: @node Include Syntax, Include Operation, Header Uses, Header Files
275: @subsection The @samp{#include} Command
276:
277: @findex #include
278: Both user and system header files are included using the preprocessor
279: command @samp{#include}. It has three variants:
280:
281: @table @code
282: @item #include <@var{file}>
283: This variant is used for system header files. It searches for a file
1.1.1.2 root 284: named @var{file} in a list of directories specified by you, then in a
285: standard list of system directories. You specify directories to
286: search for header files with the command option @samp{-I}
1.1.1.3 root 287: (@pxref{Invocation}). The option @samp{-nostdinc} inhibits searching
288: the standard system directories; in this case only the directories
289: you specify are searched.
1.1 root 290:
291: The parsing of this form of @samp{#include} is slightly special
292: because comments are not recognized within the @samp{<@dots{}>}.
293: Thus, in @samp{#include <x/*y>} the @samp{/*} does not start a comment
294: and the command specifies inclusion of a system header file named
295: @file{x/*y}. Of course, a header file with such a name is unlikely to
296: exist on Unix, where shell wildcard features would make it hard to
1.1.1.2 root 297: manipulate.@refill
1.1 root 298:
1.1.1.2 root 299: The argument @var{file} may not contain a @samp{>} character. It may,
300: however, contain a @samp{<} character.
1.1 root 301:
302: @item #include "@var{file}"
303: This variant is used for header files of your own program. It
1.1.1.7 root 304: searches for a file named @var{file} first in the current directory,
305: then in the same directories used for system header files. The
306: current directory is the directory of the current input file. It is
307: tried first because it is presumed to be the location of the files
308: that the current input file refers to. (If the @samp{-I-} option is
309: used, the special treatment of the current directory is inhibited.)
1.1 root 310:
1.1.1.2 root 311: The argument @var{file} may not contain @samp{"} characters. If
312: backslashes occur within @var{file}, they are considered ordinary text
313: characters, not escape characters. None of the character escape
314: sequences appropriate to string constants in C are processed. Thus,
315: @samp{#include "x\n\\y"} specifies a filename containing three
316: backslashes. It is not clear why this behavior is ever useful, but
317: the ANSI standard specifies it.
1.1 root 318:
319: @item #include @var{anything else}
320: This variant is called a @dfn{computed #include}. Any @samp{#include}
321: command whose argument does not fit the above two forms is a computed
1.1.1.2 root 322: include. The text @var{anything else} is checked for macro calls,
323: which are expanded (@pxref{Macros}). When this is done, the result
324: must fit one of the above two variants.
1.1 root 325:
326: This feature allows you to define a macro which controls the file name
327: to be used at a later point in the program. One application of this
328: is to allow a site-configuration file for your program to specify the
329: names of the system include files to be used. This can help in
330: porting the program to various operating systems in which the
331: necessary system header files are found in different places.
332: @end table
333:
1.1.1.6 root 334: @node Include Operation, Once-Only, Include Syntax, Header Files
1.1 root 335: @subsection How @samp{#include} Works
336:
337: The @samp{#include} command works by directing the C preprocessor to scan
338: the specified file as input before continuing with the rest of the current
339: file. The output from the preprocessor contains the output already
340: generated, followed by the output resulting from the included file,
341: followed by the output that comes from the text after the @samp{#include}
342: command. For example, given two files as follows:
343:
344: @example
345: /* File program.c */
346: int x;
347: #include "header.h"
348:
349: main ()
350: @{
351: printf (test ());
352: @}
353:
354:
355: /* File header.h */
356: char *test ();
357: @end example
358:
359: @noindent
360: the output generated by the C preprocessor for @file{program.c} as input
361: would be
362:
363: @example
364: int x;
365: char *test ();
366:
367: main ()
368: @{
369: printf (test ());
370: @}
371: @end example
372:
373: Included files are not limited to declarations and macro definitions; they
374: are merely the typical use. Any fragment of a C program can be included
375: from another file. The include file could even contain the beginning of a
376: statement that is concluded in the containing file, or the end of a
377: statement that was started in the including file. However, a comment or a
378: string or character constant may not start in the included file and finish
379: in the including file. An unterminated comment, string constant or
380: character constant in an included file is considered to end (with an error
381: message) at the end of the file.
382:
383: The line following the @samp{#include} command is always treated as a
384: separate line by the C preprocessor even if the included file lacks a final
385: newline.
386:
1.1.1.6 root 387: @node Once-Only,, Include Operation, Header Files
388: @subsection Once-Only Include Files
389: @cindex repeated inclusion
390:
391: Very often, one header file includes another. It can easily result that a
392: certain header file is included more than once. This may lead to errors,
393: if the header file defines structure types or typedefs, and is certainly
394: wasteful. Therefore, we often wish to prevent multiple inclusion of a
395: header file.
396:
397: The standard way to do this is to enclose the entire real contents of the
398: file in a conditional, like this:
399:
400: @example
401: #ifndef __FILE_FOO_SEEN__
402: #define __FILE_FOO_SEEN__
403:
404: @var{the entire file}
405:
406: #endif /* __FILE_FOO_SEEN__ */
407: @end example
408:
409: The macro @code{__FILE_FOO_SEEN__} indicates that the file has been
410: included once already; its name should begin with @samp{__}, and should
411: contain the name of the file to avoid accidental conflicts.
412:
1.1.1.8 ! root 413: One drawback of this method is that the preprocessor must scan the input
! 414: file completely in order to determine that all of it is to be ignored.
! 415: This makes compilation slower. You can avoid the delay by inserting the
! 416: following command near the beginning of file @emph{in addition to the
! 417: conditionals described above}:
1.1.1.6 root 418:
419: @example
420: #pragma once
421: @end example
422:
1.1.1.8 ! root 423: This command tells the GNU C preprocessor to ignore any future commands
! 424: to include the same file (whichever file the @samp{#pragma} appears in).
! 425:
! 426: You should not @emph{rely} on @samp{#pragma once} to prevent multiple
! 427: inclusion of the file. It is just a hint, and a nonstandard one at
! 428: that. Most C compilers will ignore it entirely. For this reason, you
! 429: still need the conditionals if you want to make certain that the file's
! 430: contents are not included twice.
1.1.1.6 root 431:
432: Note that @samp{#pragma once} works by file name; if a file has more
1.1.1.8 ! root 433: than one name, it can be included once under each name, even in GNU CC,
! 434: despite @samp{#pragma once}.
1.1.1.6 root 435:
1.1 root 436: @node Macros, Conditionals, Header Files, Top
437: @section Macros
438:
439: A macro is a sort of abbreviation which you can define once and then
440: use later. There are many complicated features associated with macros
441: in the C preprocessor.
442:
443: @menu
444: * Simple Macros:: Macros that always expand the same way.
445: * Argument Macros:: Macros that accept arguments that are substituted
446: into the macro expansion.
447: * Predefined:: Predefined macros that are always available.
448: * Stringification:: Macro arguments converted into string constants.
449: * Concatenation:: Building tokens from parts taken from macro arguments.
450: * Undefining:: Cancelling a macro's definition.
451: * Redefining:: Changing a macro's definition.
452: * Macro Pitfalls:: Macros can confuse the unwary. Here we explain
453: several common problems and strange features.
454: @end menu
455:
456: @node Simple Macros, Argument Macros, Macros, Macros
457: @subsection Simple Macros
458:
459: A @dfn{simple macro} is a kind of abbreviation. It is a name which stands
460: for a fragment of code.
461:
462: Before you can use a macro, you must @dfn{define} it explicitly with the
463: @samp{#define} command. @samp{#define} is followed by the name of the
464: macro and then the code it should be an abbreviation for. For example,
465:
466: @example
467: #define BUFFER_SIZE 1020
468: @end example
469:
470: @noindent
471: defines a macro named @samp{BUFFER_SIZE} as an abbreviation for the text
472: @samp{1020}. Therefore, if somewhere after this @samp{#define} command
473: there comes a C statement of the form
474:
475: @example
476: foo = (char *) xmalloc (BUFFER_SIZE);
477: @end example
478:
479: @noindent
480: then the C preprocessor will recognize and @dfn{expand} the macro
481: @samp{BUFFER_SIZE}, resulting in
482:
483: @example
484: foo = (char *) xmalloc (1020);
485: @end example
486:
487: @noindent
488: the definition must be a single line; however, it may not end in the
489: middle of a multi-line string constant or character constant.
490:
491: The use of all upper case for macro names is a standard convention.
492: Programs are easier to read when it is possible to tell at a glance which
493: names are macros.
494:
495: Normally, a macro definition must be a single line, like all C preprocessor
496: commands. (You can split a long macro definition cosmetically with
497: Backslash-Newline.) There is one exception: Newlines can be included in
498: the macro definition if within a string or character constant. By the same
499: token, it is not possible for a macro definition to contain an unbalanced
500: quote character; the definition automatically extends to include the
501: matching quote character that ends the string or character constant.
502: Comments within a macro definition may contain Newlines, which make no
503: difference since the comments are entirely replaced with Spaces regardless
504: of their contents.
505:
506: Aside from the above, there is no restriction on what can go in a macro
507: body. Parentheses need not balance. The body need not resemble valid C
508: code. (Of course, you might get error messages from the C compiler when
509: you use the macro.)
510:
511: The C preprocessor scans your program sequentially, so macro definitions
512: take effect at the place you write them. Therefore, the following input to
513: the C preprocessor
514:
515: @example
516: foo = X;
517: #define X 4
518: bar = X;
519: @end example
520:
521: @noindent
522: produces as output
523:
524: @example
525: foo = X;
526:
527: bar = 4;
528: @end example
529:
530: After the preprocessor expands a macro name, the macro's definition body is
531: appended to the front of the remaining input, and the check for macro calls
532: continues. Therefore, the macro body can contain calls to other macros.
533: For example, after
534:
535: @example
536: #define BUFSIZE 1020
537: #define TABLESIZE BUFSIZE
538: @end example
539:
540: @noindent
541: the name @samp{TABLESIZE} when used in the program would go through two
542: stages of expansion, resulting ultimately in @samp{1020}.
543:
544: This is not at all the same as defining @samp{TABLESIZE} to be @samp{1020}.
545: The @samp{#define} for @samp{TABLESIZE} uses exactly the body you
546: specify---in this case, @samp{BUFSIZE}---and does not check to see whether
547: it too is the name of a macro. It's only when you @emph{use} @samp{TABLESIZE}
548: that the result of its expansion is checked for more macro names.
549: @xref{Cascaded Macros}.
550:
551: @node Argument Macros, Predefined, Simple Macros, Macros
552: @subsection Macros with Arguments
553:
554: A simple macro always stands for exactly the same text, each time it is
555: used. Macros can be more flexible when they accept @dfn{arguments}.
556: Arguments are fragments of code that you supply each time the macro is
557: used. These fragments are included in the expansion of the macro according
558: to the directions in the macro definition.
559:
560: To define a macro that uses arguments, you write a @samp{#define} command
561: with a list of @dfn{argument names} in parentheses after the name of the
562: macro. The argument names may be any valid C identifiers, separated by
563: commas and optionally whitespace. The open-parenthesis must follow the
564: macro name immediately, with no space in between.
565:
566: For example, here is a macro that computes the minimum of two numeric
1.1.1.2 root 567: values, as it is defined in many C programs:
1.1 root 568:
569: @example
570: #define min(X, Y) ((X) < (Y) ? (X) : (Y))
571: @end example
572:
1.1.1.2 root 573: @noindent
574: (This is not the best way to define a ``minimum'' macro in GNU C.
575: @xref{Side Effects}, for more information.)
576:
1.1 root 577: To use a macro that expects arguments, you write the name of the macro
578: followed by a list of @dfn{actual arguments} in parentheses. separated by
579: commas. The number of actual arguments you give must match the number of
580: arguments the macro expects. Examples of use of the macro @samp{min}
581: include @samp{min (1, 2)} and @samp{min (x + 28, *p)}.
582:
583: The expansion text of the macro depends on the arguments you use.
584: Each of the argument names of the macro is replaced, throughout the
585: macro definition, with the corresponding actual argument. Using the
586: same macro @samp{min} defined above, @samp{min (1, 2)} expands into
587:
588: @example
589: ((1) < (2) ? (1) : (2))
590: @end example
591:
592: @noindent
593: where @samp{1} has been substituted for @samp{X} and @samp{2} for @samp{Y}.
594:
595: Likewise, @samp{min (x + 28, *p)} expands into
596:
597: @example
598: ((x + 28) < (*p) ? (x + 28) : (*p))
599: @end example
600:
601: Parentheses in the actual arguments must balance; a comma within
602: parentheses does not end an argument. However, there is no requirement for
603: brackets or braces to balance; thus, if you want to supply @samp{array[x =
604: y, x + 1]} as an argument, you must write it as @samp{array[(x = y, x +
605: 1)]}, which is equivalent C code.
606:
607: After the actual arguments are substituted into the macro body, the entire
608: result is appended to the front of the remaining input, and the check for
609: macro calls continues. Therefore, the actual arguments can contain calls
610: to other macros, either with or without arguments, or even to the same
611: macro. The macro body can also contain calls to other macros. For
612: example, @samp{min (min (a, b), c)} expands into
613:
614: @example
615: ((((a) < (b) ? (a) : (b))) < (c)
616: ? (((a) < (b) ? (a) : (b)))
617: : (c))
618: @end example
619:
620: @noindent
621: (Line breaks shown here for clarity would not actually be generated.)
622:
623: If you use the macro name followed by something other than an
1.1.1.2 root 624: open-parenthesis (after ignoring any spaces, tabs and comments that
1.1.1.8 ! root 625: follow), it is not a call to the macro, and the preprocessor leaves the
! 626: name unaltered. Therefore, it is possible for the same name to be a
! 627: variable or function in your program as well as a macro, and you can
! 628: choose in each instance whether to refer to the macro (if an actual
1.1.1.2 root 629: argument list follows) or the variable or function (if an argument list
630: does not follow).
1.1 root 631:
632: Such dual use of one name could be confusing and should be avoided
633: except when the two meanings are effectively synonymous: that is, when the
634: name is both a macro and a function and the two have similar effects. You
1.1.1.2 root 635: can think of the name simply as a function; use of the name for purposes
1.1 root 636: other than calling it (such as, to take the address) will refer to the
637: function, while calls will expand the macro and generate better but
638: equivalent code. For example, you can use a function named @samp{min} in
639: the same source file that defines the macro. If you write @samp{&min} with
640: no argument list, you refer to the function. If you write @samp{min (x,
641: bb)}, with an argument list, the macro is expanded. If you write
642: @samp{(min) (a, bb)}, where the name @samp{min} is not followed by an
643: open-parenthesis, the macro is not expanded, so you wind up with a call to
644: the function @samp{min}.
645:
646: It is not allowed to define the same name as both a simple macro and
647: a macro with arguments.
648:
649: In the definition of a macro with arguments, the list of argument names
650: must follow the macro name immediately with no space in between. If there
651: is a space after the macro name, the macro is defined as taking no
1.1.1.2 root 652: arguments, and all the rest of the name is taken to be the expansion. The
653: reason for this is that it is often useful to define a macro that takes no
654: arguments and whose definition begins with an identifier in parentheses.
655: This rule about spaces makes it possible for you to do either this:
1.1 root 656:
657: @example
658: #define FOO(x) - 1 / (x)
659: @end example
660:
661: @noindent
1.1.1.2 root 662: (which defines @samp{FOO} to take an argument and expand into minus the
663: reciprocal of that argument) or this:
1.1 root 664:
665: @example
666: #define BAR (x) - 1 / (x)
667: @end example
668:
669: @noindent
1.1.1.2 root 670: (which defines @samp{BAR} to take no argument and always expand into
671: @samp{(x) - 1 / (x)}).
672:
673: Note that the @emph{uses} of a macro with arguments can have spaces before
674: the left parenthesis; it's the @emph{definition} where it matters whether
675: there is a space.
1.1 root 676:
677: @node Predefined, Stringification, Argument Macros, Macros
678: @subsection Predefined Macros
679:
680: @cindex predefined macros
681: Several simple macros are predefined. You can use them without giving
682: definitions for them. They fall into two classes: standard macros and
683: system-specific macros.
684:
685: @menu
686: * Standard Predefined:: Standard predefined macros.
687: * Nonstandard Predefined:: Nonstandard predefined macros.
688: @end menu
689:
690: @node Standard Predefined, Nonstandard Predefined, Predefined, Predefined
691: @subsubsection Standard Predefined Macros
692:
1.1.1.2 root 693: The standard predefined macros are available with the same meanings
694: regardless of the machine or operating system on which you are using GNU C.
695: Their names all start and end with double underscores. Those preceding
696: @code{__GNUC__} in this table are standardized by ANSI C; the rest are
697: GNU C extensions.
1.1 root 698:
699: @table @code
700: @item __FILE__
701: @findex __FILE__
702: This macro expands to the name of the current input file, in the form
703: of a C string constant.
704:
1.1.1.6 root 705: @item __BASE_FILE__
706: @findex __BASE_FILE__
707: This macro expands to the name of the main input file, in the form
708: of a C string constant. This is the source file that was specified
709: as an argument when the C compiler was invoked.
710:
1.1 root 711: @item __LINE__
712: @findex __LINE__
713: This macro expands to the current input line number, in the form of a
714: decimal integer constant. While we call it a predefined macro, it's
715: a pretty strange macro, since its ``definition'' changes with each
716: new line of source code.
717:
718: This and @samp{__FILE__} are useful in generating an error message to
719: report an inconsistency detected by the program; the message can state
720: the source line at which the inconsistency was detected. For example,
721:
722: @example
723: fprintf (stderr, "Internal error: negative string length "
724: "%d at %s, line %d.",
725: length, __FILE__, __LINE__);
726: @end example
727:
728: A @samp{#include} command changes the expansions of @samp{__FILE__}
729: and @samp{__LINE__} to correspond to the included file. At the end of
730: that file, when processing resumes on the input file that contained
731: the @samp{#include} command, the expansions of @samp{__FILE__} and
732: @samp{__LINE__} revert to the values they had before the
733: @samp{#include} (but @samp{__LINE__} is then incremented by one as
734: processing moves to the line after the @samp{#include}).
735:
736: The expansions of both @samp{__FILE__} and @samp{__LINE__} are altered
737: if a @samp{#line} command is used. @xref{Combining Sources}.
738:
739: @item __DATE__
740: @findex __DATE__
741: This macro expands to a string constant that describes the date on
742: which the preprocessor is being run. The string constant contains
1.1.1.3 root 743: eleven characters and looks like @samp{"Jan 29 1987"} or @w{@samp{"Apr
744: 1 1905"}}.
1.1 root 745:
746: @item __TIME__
747: @findex __TIME__
748: This macro expands to a string constant that describes the time at
749: which the preprocessor is being run. The string constant contains
750: eight characters and looks like @samp{"23:59:01"}.
751:
752: @item __STDC__
753: @findex __STDC__
754: This macro expands to the constant 1, to signify that this is ANSI
755: Standard C. (Whether that is actually true depends on what C compiler
756: will operate on the output from the preprocessor.)
1.1.1.2 root 757:
758: @item __GNUC__
759: This macro is defined if and only if this is GNU C. This macro is
760: defined only when the entire GNU C compiler is in use; if you invoke
761: the preprocessor directly, @samp{__GNUC__} is undefined.
762:
763: @item __STRICT_ANSI__
764: This macro is defined if and only if the @samp{-ansi} switch was
765: specified when GNU C was invoked. Its definition is the null string.
766: This macro exists primarily to direct certain GNU header files not to
767: define certain traditional Unix constructs which are incompatible with
768: ANSI C.
769:
770: @item __VERSION__
771: This macro expands to a string which describes the version number of
772: GNU C. The string is normally a sequence of decimal numbers separated
773: by periods, such as @samp{"1.18"}. The only reasonable use of this
774: macro is to incorporate it into a string constant.
775:
776: @item __OPTIMIZE__
777: This macro is defined in optimizing compilations. It causes certain
778: GNU header files to define alternative macro definitions for some
779: system library functions. It is unwise to refer to or test the
780: definition of this macro unless you make very sure that programs will
781: execute with the same effect regardless.
782:
783: @item __CHAR_UNSIGNED__
784: This macro is defined if and only if the data type @code{char} is
785: unsigned on the target machine. It exists to cause the standard
786: header file @file{limit.h} to work correctly. It is bad practice
787: to refer to this macro yourself; instead, refer to the standard
788: macros defined in @file{limit.h}.
1.1 root 789: @end table
790:
791: @node Nonstandard Predefined,, Standard Predefined, Predefined
792: @subsubsection Nonstandard Predefined Macros
793:
794: The C preprocessor normally has several predefined macros that vary between
795: machines because their purpose is to indicate what type of system and
796: machine is in use. This manual, being for all systems and machines, cannot
797: tell you exactly what their names are; instead, we offer a list of some
798: typical ones.
799:
800: Some nonstandard predefined macros describe the operating system in use,
801: with more or less specificity. For example,
802:
803: @table @code
804: @item unix
805: @findex unix
806: @samp{unix} is normally predefined on all Unix systems.
807:
808: @item BSD
809: @findex BSD
810: @samp{BSD} is predefined on recent versions of Berkeley Unix
811: (perhaps only in version 4.3).
812: @end table
813:
814: Other nonstandard predefined macros describe the kind of CPU, with more or
815: less specificity. For example,
816:
817: @table @code
818: @item vax
819: @findex vax
820: @samp{vax} is predefined on Vax computers.
821:
822: @item mc68000
823: @findex mc68000
1.1.1.2 root 824: @samp{mc68000} is predefined on most computers whose CPU is a Motorola
825: 68000, 68010 or 68020.
1.1 root 826:
827: @item m68k
828: @findex m68k
1.1.1.2 root 829: @samp{m68k} is also predefined on most computers whose CPU is a 68000,
1.1 root 830: 68010 or 68020; however, some makers use @samp{mc68000} and some use
1.1.1.2 root 831: @samp{m68k}. Some predefine both names. What happens in GNU C
832: depends on the system you are using it on.
1.1 root 833:
834: @item M68020
835: @findex M68020
836: @samp{M68020} has been observed to be predefined on some systems that
837: use 68020 CPUs---in addition to @samp{mc68000} and @samp{m68k} that
838: are less specific.
1.1.1.2 root 839:
840: @item ns32000
841: @findex ns32000
842: @samp{ns32000} is predefined on computers which use the National
843: Semiconductor 32000 series CPU.
1.1 root 844: @end table
845:
846: Yet other nonstandard predefined macros describe the manufacturer of
847: the system. For example,
848:
849: @table @code
850: @item sun
851: @findex sun
852: @samp{sun} is predefined on all models of Sun computers.
853:
854: @item pyr
855: @findex pyr
1.1.1.2 root 856: @samp{pyr} is predefined on all models of Pyramid computers.
857:
858: @item sequent
859: @findex sequent
860: @samp{sequent} is predefined on all models of Sequent computers.
1.1 root 861: @end table
862:
863: These predefined symbols are not only nonstandard, they are contrary to the
1.1.1.6 root 864: ANSI standard because their names do not start with underscores.
865: Therefore, the option @samp{-ansi} inhibits the definition of these
866: symbols.
867:
868: This tends to make @samp{-ansi} useless, since many programs depend on the
869: customary nonstandard predefined symbols. Even system header files check
870: them and will generate incorrect declarations if they do not find the names
871: that are expected. You might think that the header files supplied for the
872: Uglix computer would not need to test what machine they are running on,
873: because they can simply assume it is the Uglix; but often they do, and they
1.1.1.8 ! root 874: do so using the customary names. As a result, very few C programs will
1.1.1.6 root 875: compile with @samp{-ansi}. We intend to avoid such problems on the GNU
876: system.
877:
878: What, then, should you do in an ANSI C program to test the type of machine
1.1.1.8 ! root 879: it is to run on?
1.1.1.6 root 880:
881: GNU C offers a parallel series of symbols for this purpose, whose names
882: are made from the customary ones by adding @samp{__} at the beginning
883: and end. Thus, the symbol @code{__vax__} would be available on a vax,
884: and so on.
1.1 root 885:
1.1.1.2 root 886: The set of nonstandard predefined names in the GNU C preprocessor is
887: controlled by the macro @samp{CPP_PREDEFINES}, which should be a string
1.1.1.6 root 888: containing @samp{-D} options, separated by spaces. For example, on the
889: Sun 3, we use the following definition:
1.1.1.2 root 890:
891: @example
892: #define CPP_PREDEFINES "-Dmc68000 -Dsun -Dunix -Dm68k"
893: @end example
894:
1.1 root 895: @node Stringification, Concatenation, Predefined, Macros
896: @subsection Stringification
897:
898: @cindex stringification
1.1.1.2 root 899: @dfn{Stringification} means turning a code fragment into a string constant
1.1 root 900: whose contents are the text for the code fragment. For example,
901: stringifying @samp{foo (z)} results in @samp{"foo (z)"}.
902:
903: In the C preprocessor, stringification is an option available when macro
904: arguments are substituted into the macro definition. In the body of the
905: definition, when an argument name appears, the character @samp{#} before
906: the name specifies stringification of the corresponding actual argument
907: when it is substituted at that point in the definition. The same argument
908: may be substituted in other places in the definition without
909: stringification if the argument name appears in those places with no
910: @samp{#}.
911:
912: Here is an example of a macro definition that uses stringification:
913:
914: @example
915: #define WARN_IF(EXP) \
916: do @{ if (EXP) fprintf (stderr, "Warning: " #EXP "\n"); @} while (0)
917: @end example
918:
919: @noindent
920: Here the actual argument for @samp{EXP} is substituted once as given,
921: into the @samp{if} statement, and once as stringified, into the
922: argument to @samp{fprintf}. The @samp{do} and @samp{while (0)} are
923: a kludge to make it possible to write @samp{WARN_IF (@var{arg});},
924: which the resemblance of @samp{WARN_IF} to a function would make
925: C programmers want to do; @pxref{Swallow Semicolon}).
926:
927: The stringification feature is limited to transforming one macro argument
928: into one string constant: there is no way to combine the argument with
929: other text and then stringify it all together. But the example above shows
930: how an equivalent result can be obtained in ANSI Standard C using the
931: feature that adjacent string constants are concatenated as one string
932: constant. The preprocessor stringifies @samp{EXP}'s actual argument
933: into a separate string constant, resulting in text like
934:
935: @example
936: do @{ if (x == 0) fprintf (stderr, "Warning: " "x == 0" "\n"); @} while (0)
937: @end example
938:
939: @noindent
940: but the C compiler then sees three consecutive string constants and
941: concatenates them into one, producing effectively
942:
943: @example
944: do @{ if (x == 0) fprintf (stderr, "Warning: x == 0\n"); @} while (0)
945: @end example
946:
947: Stringification in C involves more than putting doublequote characters
948: around the fragment; it is necessary to put backslashes in front of all
949: doublequote characters, and all backslashes in string and character
950: constants, in order to get a valid C string constant with the proper
951: contents. Thus, stringifying @samp{p = "foo\n";} results in @samp{"p =
952: \"foo\\n\";"}. However, backslashes that are not inside of string or
953: character constants are not duplicated: @samp{\n} by itself stringifies to
954: @samp{"\n"}.
955:
956: Whitespace (including comments) in the text being stringified is handled
957: according to precise rules. All leading and trailing whitespace is ignored.
958: Any sequence of whitespace in the middle of the text is converted to
959: a single space in the stringified result.
960:
961: @node Concatenation, Undefining, Stringification, Macros
962: @subsection Concatenation
963:
964: @cindex concatenation
965: @dfn{Concatenation} means joining two strings into one. In the context
966: of macro expansion, concatenation refers to joining two lexical units
967: into one longer one. Specifically, an actual argument to the macro can be
968: concatenated with another actual argument or with fixed text to produce
969: a longer name. The longer name might be the name of a function,
970: variable or type, or a C keyword; it might even be the name of another
971: macro, in which case it will be expanded.
972:
973: When you define a macro, you request concatenation with the special
974: operator @samp{##} in the macro body. When the macro is called,
975: after actual arguments are substituted, all @samp{##} operators are
976: deleted, and so is any whitespace next to them (including whitespace
977: that was part of an actual argument). The result is to concatenate
978: the syntactic tokens on either side of the @samp{##}.
979:
980: Consider a C program that interprets named commands. There probably needs
981: to be a table of commands, perhaps an array of structures declared as
982: follows:
983:
984: @example
985: struct command
986: @{
987: char *name;
988: void (*function) ();
989: @};
990:
991: struct command commands[] =
992: @{
993: @{ "quit", quit_command@},
994: @{ "help", help_command@},
995: @dots{}
996: @};
997: @end example
998:
999: It would be cleaner not to have to give each command name twice, once in
1000: the string constant and once in the function name. A macro which takes the
1001: name of a command as an argument can make this unnecessary. The string
1002: constant can be created with stringification, and the function name by
1003: concatenating the argument with @samp{_command}. Here is how it is done:
1004:
1005: @example
1006: #define COMMAND(NAME) @{ #NAME, NAME ## _command @}
1007:
1008: struct command commands[] =
1009: @{
1010: COMMAND (quit),
1011: COMMAND (help),
1012: @dots{}
1013: @};
1014: @end example
1015:
1016: The usual case of concatenation is concatenating two names (or a name and a
1017: number) into a longer name. But this isn't the only valid case. It is
1018: also possible to concatenate two numbers (or a number and a name, such as
1019: @samp{1.5} and @samp{e3}) into a number. Also, multi-character operators
1020: such as @samp{+=} can be formed by concatenation. In some cases it is even
1021: possible to piece together a string constant. However, two pieces of text
1022: that don't together form a valid lexical unit cannot be concatenated. For
1023: example, concatenation with @samp{x} on one side and @samp{+} on the other
1024: is not meaningful because those two characters can't fit together in any
1025: lexical unit of C. The ANSI standard says that such attempts at
1026: concatenation are undefined, but in the GNU C preprocessor it is well
1027: defined: it puts the @samp{x} and @samp{+} side by side with no particular
1028: special results.
1029:
1030: Keep in mind that the C preprocessor converts comments to whitespace before
1031: macros are even considered. Therefore, you cannot create a comment by
1032: concatenating @samp{/} and @samp{*}: the @samp{/*} sequence that starts a
1033: comment is not a lexical unit, but rather the beginning of a ``long'' space
1034: character. Also, you can freely use comments next to a @samp{##} in a
1035: macro definition, or in actual arguments that will be concatenated, because
1036: the comments will be converted to spaces at first sight, and concatenation
1037: will later discard the spaces.
1038:
1039: @node Undefining, Redefining, Concatenation, Macros
1040: @subsection Undefining Macros
1041:
1042: @cindex undefining macros
1043: To @dfn{undefine} a macro means to cancel its definition. This is done
1044: with the @samp{#undef} command. @samp{#undef} is followed by the macro
1045: name to be undefined.
1046:
1047: Like definition, undefinition occurs at a specific point in the source
1048: file, and it applies starting from that point. The name ceases to be a
1049: macro name, and from that point on it is treated by the preprocessor as if
1050: it had never been a macro name.
1051:
1052: For example,
1053:
1054: @example
1055: #define FOO 4
1056: x = FOO;
1057: #undef FOO
1058: x = FOO;
1059: @end example
1060:
1061: @noindent
1062: expands into
1063:
1064: @example
1065: x = 4;
1066:
1067: x = FOO;
1068: @end example
1069:
1070: @noindent
1071: In this example, @samp{FOO} had better be a variable or function as well
1072: as (temporarily) a macro, in order for the result of the expansion to be
1073: valid C code.
1074:
1075: The same form of @samp{#undef} command will cancel definitions with
1076: arguments or definitions that don't expect arguments. The @samp{#undef}
1077: command has no effect when used on a name not currently defined as a macro.
1078:
1079: @node Redefining, Macro Pitfalls, Undefining, Macros
1080: @subsection Redefining Macros
1081:
1082: @cindex redefining macros
1083: @dfn{Redefining} a macro means defining (with @samp{#define}) a name that
1084: is already defined as a macro.
1085:
1086: A redefinition is trivial if the new definition is transparently identical
1087: to the old one. You probably wouldn't deliberately write a trivial
1088: redefinition, but they can happen automatically when a header file is
1089: included more than once (@pxref{Header Files}), so they are accepted
1090: silently and without effect.
1091:
1092: Nontrivial redefinition is considered likely to be an error, so
1093: it provokes a warning message from the preprocessor. However, sometimes it
1094: is useful to change the definition of a macro in mid-compilation. You can
1095: inhibit the warning by undefining the macro with @samp{#undef} before the
1096: second definition.
1097:
1.1.1.8 ! root 1098: In order for a redefinition to be trivial, the new definition must
! 1099: exactly match the one already in effect, with two possible exceptions:
1.1 root 1100:
1101: @itemize @bullet
1102: @item
1103: Whitespace may be added or deleted at the beginning or the end.
1104:
1105: @item
1106: Whitespace may be changed in the middle (but not inside strings).
1107: However, it may not be eliminated entirely, and it may not be added
1108: where there was no whitespace at all.
1109: @end itemize
1110:
1111: Recall that a comment counts as whitespace.
1112:
1113: @node Macro Pitfalls,, Redefining, Macros
1114: @subsection Pitfalls and Subtleties of Macros
1115:
1116: In this section we describe some special rules that apply to macros and
1.1.1.8 ! root 1117: macro expansion, and point out certain cases in which the rules have
1.1 root 1118: counterintuitive consequences that you must watch out for.
1119:
1120: @menu
1121: * Misnesting:: Macros can contain unmatched parentheses.
1122: * Macro Parentheses:: Why apparently superfluous parentheses
1123: may be necessary to avoid incorrect grouping.
1124: * Swallow Semicolon:: Macros that look like functions
1125: but expand into compound statements.
1126: * Side Effects:: Unsafe macros that cause trouble when
1127: arguments contain side effects.
1128: * Self-Reference:: Macros whose definitions use the macros' own names.
1129: * Argument Prescan:: Actual arguments are checked for macro calls
1130: before they are substituted.
1131: * Cascaded Macros:: Macros whose definitions use other macros.
1132: @end menu
1133:
1134: @node Misnesting, Macro Parentheses, Macro Pitfalls, Macro Pitfalls
1135: @subsubsection Improperly Nested Constructs
1136:
1137: Recall that when a macro is called with arguments, the arguments are
1138: substituted into the macro body and the result is checked, together with
1139: the rest of the input file, for more macro calls.
1140:
1141: It is possible to piece together a macro call coming partially from the
1142: macro body and partially from the actual arguments. For example,
1143:
1144: @example
1145: #define double(x) (2*(x))
1146: #define call_with_1(x) x(1)
1147: @end example
1148:
1149: @noindent
1150: would expand @samp{call_with_1 (double)} into @samp{(2*(1))}.
1151:
1152: Macro definitions do not have to have balanced parentheses. By writing an
1153: unbalanced open parenthesis in a macro body, it is possible to create a
1154: macro call that begins inside the macro body but ends outside of it. For
1155: example,
1156:
1157: @example
1158: #define strange(file) fprintf (file, "%s %d",
1159: @dots{}
1160: strange(stderr) p, 35)
1161: @end example
1162:
1163: @noindent
1164: This bizarre example expands to @samp{fprintf (stderr, "%s %d", p, 35)}!
1165:
1166: @node Macro Parentheses, Swallow Semicolon, Misnesting, Macro Pitfalls
1167: @subsubsection Unintended Grouping of Arithmetic
1168:
1169: You may have noticed that in most of the macro definition examples shown
1170: above, each occurrence of a macro argument name had parentheses around it.
1171: In addition, another pair of parentheses usually surround the entire macro
1172: definition. Here is why it is best to write macros that way.
1173:
1.1.1.8 ! root 1174: Suppose you define a macro as follows,
1.1 root 1175:
1176: @example
1177: #define ceil_div(x, y) (x + y - 1) / y
1178: @end example
1179:
1180: @noindent
1181: whose purpose is to divide, rounding up. (One use for this
1182: operation is to compute how many @samp{int}'s are needed to hold
1183: a certain number of @samp{char}'s.) Then suppose it is used as follows:
1184:
1185: @example
1186: a = ceil_div (b & c, sizeof (int));
1187: @end example
1188:
1189: @noindent
1190: This expands into
1191:
1192: @example
1193: a = (b & c + sizeof (int) - 1) / sizeof (int);
1194: @end example
1195:
1196: @noindent
1.1.1.2 root 1197: which does not do what is intended. The operator-precedence rules of
1.1.1.4 root 1198: C make it equivalent to this:
1.1 root 1199:
1200: @example
1201: a = (b & (c + sizeof (int) - 1)) / sizeof (int);
1202: @end example
1203:
1.1.1.2 root 1204: @noindent
1205: But what we want is this:
1.1 root 1206:
1207: @example
1208: a = ((b & c) + sizeof (int) - 1)) / sizeof (int);
1209: @end example
1210:
1211: @noindent
1212: Defining the macro as
1213:
1214: @example
1215: #define ceil_div(x, y) ((x) + (y) - 1) / (y)
1216: @end example
1217:
1218: @noindent
1219: provides the desired result.
1220:
1221: However, unintended grouping can result in another way. Consider
1222: @samp{sizeof ceil_div(1, 2)}. That has the appearance of a C expression
1223: that would compute the size of the type of @samp{ceil_div (1, 2)}, but in
1224: fact it means something very different. Here is what it expands to:
1225:
1226: @example
1227: sizeof ((1) + (2) - 1) / (2)
1228: @end example
1229:
1230: @noindent
1231: This would take the size of an integer and divide it by two. The precedence
1232: rules have put the division outside the @samp{sizeof} when it was intended
1233: to be inside.
1234:
1235: Parentheses around the entire macro definition can prevent such problems.
1236: Here, then, is the recommended way to define @samp{ceil_div}:
1237:
1238: @example
1239: #define ceil_div(x, y) (((x) + (y) - 1) / (y))
1240: @end example
1241:
1242: @node Swallow Semicolon, Side Effects, Macro Parentheses, Macro Pitfalls
1243: @subsubsection Swallowing the Semicolon
1244:
1245: @cindex semicolons (after macro calls)
1246: Often it is desirable to define a macro that expands into a compound
1247: statement. Consider, for example, the following macro, that advances a
1248: pointer (the argument @samp{p} says where to find it) across whitespace
1249: characters:
1250:
1251: @example
1252: #define SKIP_SPACES (p, limit) \
1253: @{ register char *lim = (limit); \
1254: while (p != lim) @{ \
1255: if (*p++ != ' ') @{ \
1256: p--; break; @}@}@}
1257: @end example
1258:
1259: @noindent
1260: Here Backslash-Newline is used to split the macro definition, which must
1261: be a single line, so that it resembles the way such C code would be
1.1.1.8 ! root 1262: laid out if not part of a macro definition.
1.1 root 1263:
1264: A call to this macro might be @samp{SKIP_SPACES (p, lim)}. Strictly
1265: speaking, the call expands to a compound statement, which is a complete
1266: statement with no need for a semicolon to end it. But it looks like a
1267: function call. So it minimizes confusion if you can use it like a function
1268: call, writing a semicolon afterward, as in @samp{SKIP_SPACES (p, lim);}
1269:
1270: But this can cause trouble before @samp{else} statements, because the
1271: semicolon is actually a null statement. Suppose you write
1272:
1273: @example
1274: if (*p != 0)
1275: SKIP_SPACES (p, lim);
1276: else @dots{}
1277: @end example
1278:
1.1.1.2 root 1279: @noindent
1.1 root 1280: The presence of two statements---the compound statement and a null
1281: statement---in between the @samp{if} condition and the @samp{else}
1282: makes invalid C code.
1283:
1284: The definition of the macro @samp{SKIP_SPACES} can be altered to solve
1285: this problem, using a @samp{do @dots{} while} statement. Here is how:
1.1.1.8 ! root 1286:
1.1 root 1287: @example
1288: #define SKIP_SPACES (p, limit) \
1289: do @{ register char *lim = (limit); \
1290: while (p != lim) @{ \
1291: if (*p++ != ' ') @{ \
1292: p--; break; @}@}@} \
1293: while (0)
1294: @end example
1295:
1296: Now @samp{SKIP_SPACES (p, lim);} expands into
1297:
1298: @example
1299: do @{@dots{}@} while (0);
1300: @end example
1301:
1.1.1.2 root 1302: @noindent
1.1 root 1303: which is one statement.
1304:
1305: @node Side Effects, Self-Reference, Swallow Semicolon, Macro Pitfalls
1306: @subsubsection Duplication of Side Effects
1307:
1308: @cindex side effects (in macro arguments)
1309: @cindex unsafe macros
1.1.1.2 root 1310: Many C programs define a macro @samp{min}, for ``minimum'', like this:
1.1 root 1311:
1312: @example
1313: #define min(X, Y) ((X) < (Y) ? (X) : (Y))
1.1.1.2 root 1314: @end example
1315:
1316: When you use this macro with an argument containing a side effect,
1317: as shown here,
1318:
1319: @example
1.1 root 1320: next = min (x + y, foo (z));
1321: @end example
1322:
1.1.1.2 root 1323: @noindent
1324: it expands as follows:
1.1 root 1325:
1326: @example
1327: next = ((x + y) < (foo (z)) ? (x + y) : (foo (z)));
1328: @end example
1329:
1.1.1.2 root 1330: @noindent
1.1 root 1331: where @samp{x + y} has been substituted for @samp{X} and @samp{foo (z)}
1332: for @samp{Y}.
1333:
1334: The function @samp{foo} is used only once in the statement as it appears
1335: in the program, but the expression @samp{foo (z)} has been substituted
1336: twice into the macro expansion. As a result, @samp{foo} might be called
1337: two times when the statement is executed. If it has side effects or
1338: if it takes a long time to compute, the results might not be what you
1339: intended. We say that @samp{min} is an @dfn{unsafe} macro.
1340:
1.1.1.2 root 1341: The best solution to this problem is to define @samp{min} in a way that
1342: computes the value of @samp{foo (z)} only once. The C language offers no
1.1.1.8 ! root 1343: standard way to do this, but it can be done with GNU C extensions as
1.1.1.2 root 1344: follows:
1345:
1346: @example
1347: #define min(X, Y) \
1348: (@{ typeof (X) __x = (X), __y = (Y); \
1349: (__x < __y) ? __x : __y; @})
1350: @end example
1351:
1352: If you do not wish to use GNU C extensions, the only solution is to be
1353: careful when @emph{using} the macro @samp{min}. For example, you can
1354: calculate the value of @samp{foo (z)}, save it in a variable, and use that
1355: variable in @samp{min}:
1.1 root 1356:
1357: @example
1358: #define min(X, Y) ((X) < (Y) ? (X) : (Y))
1359: @dots{}
1360: @{
1361: int tem = foo (z);
1362: next = min (x + y, tem);
1363: @}
1364: @end example
1365:
1366: @noindent
1367: (where I assume that @samp{foo} returns type @samp{int}).
1368:
1369: @node Self-Reference, Argument Prescan, Side Effects, Macro Pitfalls
1370: @subsubsection Self-Referential Macros
1371:
1372: @cindex self-reference
1373: A @dfn{self-referential} macro is one whose name appears in its definition.
1374: A special feature of ANSI Standard C is that the self-reference is not
1375: considered a macro call. It is passed into the preprocessor output
1376: unchanged.
1377:
1378: Let's consider an example:
1379:
1380: @example
1381: #define foo (4 + foo)
1382: @end example
1383:
1384: @noindent
1385: where @samp{foo} is also a variable in your program.
1386:
1387: Following the ordinary rules, each reference to @samp{foo} will expand into
1388: @samp{(4 + foo)}; then this will be rescanned and will expand into @samp{(4
1389: + (4 + foo))}; and so on until it causes a fatal error (memory full) in the
1390: preprocessor.
1391:
1392: However, the special rule about self-reference cuts this process short
1393: after one step, at @samp{(4 + foo)}. Therefore, this macro definition
1394: has the possibly useful effect of causing the program to add 4 to
1395: the value of @samp{foo} wherever @samp{foo} is referred to.
1396:
1.1.1.8 ! root 1397: In most cases, it is a bad idea to take advantage of this feature. A
! 1398: person reading the program who sees that @samp{foo} is a variable will
! 1399: not expect that it is a macro as well. The reader will come across the
! 1400: identifier @samp{foo} in the program and think its value should be that
! 1401: of the variable @samp{foo}, whereas in fact the value is four greater.
1.1 root 1402:
1403: The special rule for self-reference applies also to @dfn{indirect}
1404: self-reference. This is the case where a macro @var{x} expands to use a
1405: macro @samp{y}, and @samp{y}'s expansion refers to the macro @samp{x}. The
1406: resulting reference to @samp{x} comes indirectly from the expansion of
1407: @samp{x}, so it is a self-reference and is not further expanded. Thus,
1408: after
1409:
1410: @example
1411: #define x (4 + y)
1412: #define y (2 * x)
1413: @end example
1414:
1415: @noindent
1416: @samp{x} would expand into @samp{(4 + (2 * x))}. Clear?
1417:
1418: But suppose @samp{y} is used elsewhere, not from the definition of @samp{x}.
1419: Then the use of @samp{x} in the expansion of @samp{y} is not a self-reference
1420: because @samp{x} is not ``in progress''. So it does expand. However,
1421: the expansion of @samp{x} contains a reference to @samp{y}, and that
1422: is an indirect self-reference now because @samp{y} is ``in progress''.
1423: The result is that @samp{y} expands to @samp{(2 * (4 + y))}.
1424:
1425: It is not clear that this behavior would ever be useful, but it is specified
1426: by the ANSI C standard, so you need to understand it.
1427:
1428: @node Argument Prescan, Cascaded Macros, Self-Reference, Macro Pitfalls
1429: @subsubsection Separate Expansion of Macro Arguments
1430:
1431: We have explained that the expansion of a macro, including the substituted
1432: actual arguments, is scanned over again for macro calls to be expanded.
1433:
1434: What really happens is more subtle: first each actual argument text is scanned
1435: separately for macro calls. Then the results of this are substituted into
1436: the macro body to produce the macro expansion, and the macro expansion
1437: is scanned again for macros to expand.
1438:
1439: The result is that the actual arguments are scanned @emph{twice} to expand
1440: macro calls in them.
1441:
1442: Most of the time, this has no effect. If the actual argument contained
1443: any macro calls, they are expanded during the first scan. The result
1444: therefore contains no macro calls, so the second scan does not change it.
1445: If the actual argument were substituted as given, with no prescan,
1446: the single remaining scan would find the same macro calls and produce
1447: the same results.
1448:
1449: You might expect the double scan to change the results when a
1450: self-referential macro is used in an actual argument of another macro
1451: (@pxref{Self-Reference}): the self-referential macro would be expanded once
1452: in the first scan, and a second time in the second scan. But this is not
1453: what happens. The self-references that do not expand in the first scan are
1454: marked so that they will not expand in the second scan either.
1455:
1456: The prescan is not done when an argument is stringified or concatenated.
1.1.1.5 root 1457: Thus,
1.1 root 1458:
1459: @example
1460: #define str(s) #s
1461: #define foo 4
1462: str (foo)
1463: @end example
1464:
1465: @noindent
1466: expands to @samp{"foo"}. Once more, prescan has been prevented from
1467: having any noticeable effect.
1468:
1.1.1.5 root 1469: More precisely, stringification and concatenation use the argument as
1470: written, in un-prescanned form. The same actual argument would be used in
1471: prescanned form if it is substituted elsewhere without stringification or
1472: concatenation.
1473:
1474: @example
1475: #define str(s) #s lose(s)
1476: #define foo 4
1477: str (foo)
1478: @end example
1479:
1480: expands to @samp{"foo" lose(4)}.
1481:
1.1 root 1482: You might now ask, ``Why mention the prescan, if it makes no difference?
1.1.1.5 root 1483: And why not skip it and make the preprocessor faster?'' The answer is
1484: that the prescan does make a difference in three special cases:
1.1 root 1485:
1486: @itemize @bullet
1487: @item
1488: Nested calls to a macro.
1489:
1490: @item
1491: Macros that call other macros that stringify or concatenate.
1.1.1.5 root 1492:
1493: @item
1494: Macros whose expansions contain unshielded commas.
1.1 root 1495: @end itemize
1496:
1497: We say that @dfn{nested} calls to a macro occur when a macro's actual
1498: argument contains a call to that very macro. For example, if @samp{f}
1499: is a macro that expects one argument, @samp{f (f (1))} is a nested
1500: pair of calls to @samp{f}. The desired expansion is made by
1501: expanding @samp{f (1)} and substituting that into the definition of
1502: @samp{f}. The prescan causes the expected result to happen.
1503: Without the prescan, @samp{f (1)} itself would be substituted as
1504: an actual argument, and the inner use of @samp{f} would appear
1505: during the main scan as an indirect self-reference and would not
1506: be expanded. Here, the prescan cancels an undesirable side effect
1507: (in the medical, not computational, sense of the term) of the special
1508: rule for self-referential macros.
1509:
1.1.1.5 root 1510: But prescan causes trouble in certain other cases of nested macro calls.
1511: Here is an example:
1512:
1513: @example
1514: #define foo a,b
1515: #define bar(x) lose(x)
1516: #define lose(x) (1 + (x))
1517:
1518: bar(foo)
1519: @end example
1520:
1521: @noindent
1522: We would like @samp{bar(foo)} to turn into @samp{(1 + (foo))}, which
1523: would then turn into @samp{(1 + (a,b))}. But instead, @samp{bar(foo)}
1524: expands into @samp{lose(a,b)}, and you get an error because @code{lose}
1525: requires a single argument. In this case, the problem is easily solved
1526: by the same parentheses that ought to be used to prevent misnesting of
1527: arithmetic operations:
1528:
1529: @example
1530: #define foo (a,b)
1531: #define bar(x) lose((x))
1532: @end example
1533:
1534: The problem is more serious when the operands of the macro are not
1535: expressions; for example, when they are statements. Then parentheses
1536: are unacceptable because they would make for invalid C code:
1537:
1538: @example
1539: #define foo @{ int a, b; @dots{} @}
1540: @end example
1541:
1542: @noindent
1543: In GNU C you can shield the commas using the @samp{(@{@dots{}@})}
1544: construct which turns a compound statement into an expression:
1545:
1546: @example
1547: #define foo (@{ int a, b; @dots{} @})
1548: @end example
1549:
1550: Or you can rewrite the macro definition to avoid such commas:
1551:
1552: @example
1553: #define foo @{ int a; int b; @dots{} @}
1554: @end example
1555:
1.1 root 1556: There is also one case where prescan is useful. It is possible
1557: to use prescan to expand an argument and then stringify it---if you use
1558: two levels of macros. Let's add a new macro @samp{xstr} to the
1559: example shown above:
1560:
1561: @example
1562: #define xstr(s) str(s)
1563: #define str(s) #s
1564: #define foo 4
1565: xstr (foo)
1566: @end example
1567:
1568: This expands into @samp{"4"}, not @samp{"foo"}. The reason for the
1569: difference is that the argument of @samp{xstr} is expanded at prescan
1570: (because @samp{xstr} does not specify stringification or concatenation of
1571: the argument). The result of prescan then forms the actual argument for
1572: @samp{str}. @samp{str} uses its argument without prescan because it
1.1.1.8 ! root 1573: performs stringification; but it cannot prevent or undo the prescanning
1.1 root 1574: already done by @samp{xstr}.
1575:
1576: @node Cascaded Macros,, Argument Prescan, Macro Pitfalls
1577: @subsubsection Cascaded Use of Macros
1578:
1579: @cindex cascaded macros
1580: @cindex macro body uses macro
1581: A @dfn{cascade} of macros is when one macro's body contains a reference
1582: to another macro. This is very common practice. For example,
1583:
1584: @example
1585: #define BUFSIZE 1020
1586: #define TABLESIZE BUFSIZE
1587: @end example
1588:
1589: This is not at all the same as defining @samp{TABLESIZE} to be @samp{1020}.
1590: The @samp{#define} for @samp{TABLESIZE} uses exactly the body you
1591: specify---in this case, @samp{BUFSIZE}---and does not check to see whether
1592: it too is the name of a macro.
1593:
1594: It's only when you @emph{use} @samp{TABLESIZE} that the result of its expansion
1595: is checked for more macro names.
1596:
1597: This makes a difference if you change the definition of @samp{BUFSIZE}
1598: at some point in the source file. @samp{TABLESIZE}, defined as shown,
1599: will always expand using the definition of @samp{BUFSIZE} that is
1600: currently in effect:
1601:
1602: @example
1603: #define BUFSIZE 1020
1604: #define TABLESIZE BUFSIZE
1605: #undef BUFSIZE
1606: #define BUFSIZE 37
1607: @end example
1608:
1.1.1.2 root 1609: @noindent
1.1 root 1610: Now @samp{TABLESIZE} expands (in two stages) to @samp{37}.
1611:
1612: @node Conditionals, Combining Sources, Macros, Top
1613: @section Conditionals
1614:
1615: @cindex conditionals
1616: In a macro processor, a @dfn{conditional} is a command that allows a part
1617: of the program to be ignored during compilation, on some conditions.
1618: In the C preprocessor, a conditional can test either an arithmetic expression
1619: or whether a name is defined as a macro.
1620:
1621: A conditional in the C preprocessor resembles in some ways an @samp{if}
1622: statement in C, but it is important to understand the difference between
1623: them. The condition in an @samp{if} statement is tested during the execution
1624: of your program. Its purpose is to allow your program to behave differently
1625: from run to run, depending on the data it is operating on. The condition
1626: in a preprocessor conditional command is tested when your program is compiled.
1627: Its purpose is to allow different code to be included in the program depending
1628: on the situation at the time of compilation.
1629:
1630: @menu
1631: * Uses: Conditional Uses. What conditionals are for.
1632: * Syntax: Conditional Syntax. How conditionals are written.
1633: * Deletion: Deleted Code. Making code into a comment.
1634: * Macros: Conditionals-Macros. Why conditionals are used with macros.
1635: * Errors: #error Command. Detecting inconsistent compilation parameters.
1636: @end menu
1637:
1638: @node Conditional Uses, Conditional Syntax, Conditionals, Conditionals
1.1.1.6 root 1639: @subsection Why Conditionals are Used
1640:
1.1 root 1641: Generally there are three kinds of reason to use a conditional.
1642:
1643: @itemize @bullet
1644: @item
1645: A program may need to use different code depending on the machine or
1646: operating system it is to run on. In some cases the code for one
1647: operating system may be erroneous on another operating system; for
1648: example, it might refer to library routines that do not exist on the
1649: other system. When this happens, it is not enough to avoid executing
1650: the invalid code: merely having it in the program makes it impossible
1651: to link the program and run it. With a preprocessor conditional, the
1652: offending code can be effectively excised from the program when it is
1653: not valid.
1654:
1655: @item
1656: You may want to be able to compile the same source file into two
1657: different programs. Sometimes the difference between the programs is
1658: that one makes frequent time-consuming consistency checks on its
1659: intermediate data while the other does not.
1660:
1661: @item
1662: A conditional whose condition is always false is a good way to exclude
1663: code from the program but keep it as a sort of comment for future
1664: reference.
1665: @end itemize
1666:
1667: Most simple programs that are intended to run on only one machine will
1668: not need to use preprocessor conditionals.
1669:
1.1.1.8 ! root 1670: @node Conditional Syntax, Deleted Code, Conditional Uses, Conditionals
1.1 root 1671: @subsection Syntax of Conditionals
1672:
1673: @findex #if
1674: A conditional in the C preprocessor begins with a @dfn{conditional
1.1.1.8 ! root 1675: command}: @samp{#if}, @samp{#ifdef} or @samp{#ifndef}.
! 1676: @xref{Conditionals-Macros}, for info on @samp{#ifdef} and
! 1677: @samp{#ifndef}; only @samp{#if} is explained here.
1.1 root 1678:
1679: @menu
1680: * If: #if Command. Basic conditionals using @samp{#if} and @samp{#endif}.
1681: * Else: #else Command. Including some text if the condition fails.
1682: * Elif: #elif Command. Testing several alternative possibilities.
1683: @end menu
1684:
1685: @node #if Command, #else Command, Conditional Syntax, Conditional Syntax
1686: @subsubsection The @samp{#if} Command
1687:
1688: The @samp{#if} command in its simplest form consists of
1689:
1690: @example
1691: #if @var{expression}
1692: @var{controlled text}
1693: #endif /* @var{expression} */
1694: @end example
1695:
1696: The comment following the @samp{#endif} is not required, but it is a good
1697: practice because it helps people match the @samp{#endif} to the
1698: corresponding @samp{#if}. Such comments should always be used, except in
1699: short conditionals that are not nested. In fact, you can put anything at
1700: all after the @samp{#endif} and it will be ignored by the GNU C preprocessor,
1701: but only comments are acceptable in ANSI Standard C.
1702:
1703: @var{expression} is a C expression of integer type, subject to stringent
1704: restrictions. It may contain
1705:
1706: @itemize @bullet
1707: @item
1708: Integer constants, which are all regarded as @code{long} or
1709: @code{unsigned long}.
1710:
1711: @item
1712: Character constants, which are interpreted according to the character
1713: set and conventions of the machine and operating system on which the
1714: preprocessor is running. The GNU C preprocessor uses the C data type
1715: @samp{char} for these character constants; therefore, whether some
1716: character codes are negative is determined by the C compiler used to
1717: compile the preprocessor. If it treats @samp{char} as signed, then
1718: character codes large enough to set the sign bit will be considered
1719: negative; otherwise, no character code is considered negative.
1720:
1721: @item
1722: Arithmetic operators for addition, subtraction, multiplication,
1723: division, bitwise operations, shifts, comparisons, and @samp{&&} and
1724: @samp{||}.
1725:
1726: @item
1727: Identifiers that are not macros, which are all treated as zero(!).
1728:
1729: @item
1730: Macro calls. All macro calls in the expression are expanded before
1731: actual computation of the expression's value begins.
1732: @end itemize
1733:
1734: Note that @samp{sizeof} operators and @code{enum}-type values are not allowed.
1735: @code{enum}-type values, like all other identifiers that are not taken
1736: as macro calls and expanded, are treated as zero.
1737:
1738: The text inside of a conditional can include preprocessor commands. Then
1739: the commands inside the conditional are obeyed only if that branch of the
1740: conditional succeeds. The text can also contain other conditional groups.
1741: However, the @samp{#if}'s and @samp{#endif}'s must balance.
1742:
1743: @node #else Command, #elif Command, #if Command, Conditional Syntax
1744: @subsubsection The @samp{#else} Command
1745:
1746: @findex #else
1.1.1.8 ! root 1747: The @samp{#else} command can be added to a conditional to provide alternative
1.1 root 1748: text to be used if the condition is false. This looks like
1749:
1750: @example
1751: #if @var{expression}
1752: @var{text-if-true}
1753: #else /* Not @var{expression} */
1754: @var{text-if-false}
1755: #endif /* Not @var{expression} */
1756: @end example
1757:
1758: If @var{expression} is nonzero, and the @var{text-if-true} is considered
1759: included, then @samp{#else} acts like a failing conditional and the
1760: @var{text-if-false} is ignored. Contrariwise, if the @samp{#if}
1761: conditional fails, the @var{text-if-false} is considered included.
1762:
1763: @node #elif Command,, #else Command, Conditional Syntax
1764: @subsubsection The @samp{#elif} Command
1765:
1766: @findex #elif
1767: One common case of nested conditionals is used to check for more than two
1768: possible alternatives. For example, you might have
1769:
1770: @example
1771: #if X == 1
1772: @dots{}
1773: #else /* X != 1 */
1774: #if X == 2
1775: @dots{}
1776: #else /* X != 2 */
1777: @dots{}
1778: #endif /* X != 2 */
1779: #endif /* X != 1 */
1780: @end example
1781:
1782: Another conditional command, @samp{#elif}, allows this to be abbreviated
1783: as follows:
1784:
1785: @example
1786: #if X == 1
1787: @dots{}
1788: #elif X == 2
1789: @dots{}
1790: #else /* X != 2 and X != 1*/
1791: @dots{}
1792: #endif /* X != 2 and X != 1*/
1793: @end example
1794:
1795: @samp{#elif} stands for ``else if''. Like @samp{#else}, it goes in the
1796: middle of a @samp{#if}-@samp{#endif} pair and subdivides it; it does not
1797: require a matching @samp{#endif} of its own. Like @samp{#if}, the
1798: @samp{#elif} command includes an expression to be tested.
1799:
1800: The text following the @samp{#elif} is processed only if the original
1801: @samp{#if}-condition failed and the @samp{#elif} condition succeeeds. More
1802: than one @samp{#elif} can go in the same @samp{#if}-@samp{#endif} group.
1803: Then the text after each @samp{#elif} is processed only if the @samp{#elif}
1804: condition succeeds after the original @samp{#if} and any previous
1805: @samp{#elif}'s within it have failed. @samp{#else} is equivalent to
1806: @samp{#elif 1}, and @samp{#else} is allowed after any number of
1807: @samp{#elif}'s, but @samp{#elif} may not follow a @samp{#else}.
1808:
1809: @node Deleted Code, Conditionals-Macros, Conditional Syntax, Conditionals
1810: @subsection Keeping Deleted Code for Future Reference
1811:
1812: If you replace or delete a part of the program but want to keep the old
1813: code around as a comment for future reference, the easy way to do this is
1814: to put @samp{#if 0} before it and @samp{#endif} after it.
1815:
1816: This works even if the code being turned off contains conditionals, but
1817: they must be entire conditionals (balanced @samp{#if} and @samp{#endif}).
1818:
1819: @node Conditionals-Macros, #error Command, Deleted Code, Conditionals
1820: @subsection Conditionals and Macros
1821:
1822: Conditionals are rarely useful except in connection with macros. A
1823: @samp{#if} command whose expression uses no macros is equivalent to
1824: @samp{#if 1} or @samp{#if 0}; you might as well determine which one, by
1825: computing the value of the expression yourself, and then simplify the
1826: program. But when the expression uses macros, its value can vary from
1827: compilation to compilation.
1828:
1829: For example, here is a conditional that tests the expression
1830: @samp{BUFSIZE == 1020}, where @samp{BUFSIZE} must be a macro.
1831:
1832: @example
1833: #if BUFSIZE == 1020
1834: printf ("Large buffers!\n");
1835: #endif /* BUFSIZE is large */
1836: @end example
1837:
1838: @findex defined
1839: The special operator @samp{defined} may be used in @samp{#if} expressions
1.1.1.2 root 1840: to test whether a certain name is defined as a macro. Either @samp{defined
1841: @var{name}} or @samp{defined (@var{name})} is an expression whose value is
1842: 1 if @var{name} is defined as macro at the current point in the program,
1843: and 0 otherwise. For the @samp{defined} operator it makes no difference
1844: what the definition of the macro is; all that matters is whether there is a
1845: definition. Thus, for example,@refill
1.1 root 1846:
1847: @example
1848: #if defined (vax) || defined (ns16000)
1849: @end example
1850:
1851: @noindent
1852: would include the following code if either of the names @samp{vax} and
1853: @samp{ns16000} is defined as a macro.
1854:
1855: If a macro is defined and later undefined with @samp{#undef},
1856: subsequent use of the @samp{defined} operator will return 0, because
1857: the name is no longer defined. If the macro is defined again with
1858: another @samp{#define}, @samp{defined} will recommence returning 1.
1859:
1860: @findex #ifdef
1861: @findex #ifndef
1862: Conditionals that test just the definedness of one name are very common, so
1863: there are two special short conditional commands for this case. They are
1864:
1865: @table @code
1866: @item #ifdef @var{name}
1867: is equivalent to @samp{#if defined (@var{name})}.
1868:
1869: @item #ifndef @var{name}
1870: is equivalent to @samp{#if ! defined (@var{name})}.
1871: @end table
1872:
1873: Macro definitions can vary between compilations for several reasons.
1874:
1875: @itemize @bullet
1876: @item
1877: Some macros are predefined on each kind of machine. For example, on a
1878: Vax, the name @samp{vax} is a predefined macro. On other machines, it
1879: would not be defined.
1880:
1881: @item
1882: Many more macros are defined by system header files. Different
1883: systems and machines define different macros, or give them different
1884: values. It is useful to test these macros with conditionals to avoid
1885: using a system feature on a machine where it is not implemented.
1886:
1887: @item
1888: Macros are a common way of allowing users to customize a program for
1889: different machines or applications. For example, the macro
1890: @samp{BUFSIZE} might be defined in a configuration file for your
1891: program that is included as a header file in each source file. You
1892: would use @samp{BUFSIZE} in a preprocessor conditional in order to
1893: generate different code depending on the chosen configuration.
1894:
1895: @item
1896: Macros can be defined or undefined with @samp{-D} and @samp{-U}
1.1.1.2 root 1897: command options when you compile the program. You can arrange to
1.1 root 1898: compile the same source file into two different programs by choosing
1899: a macro name to specify which program you want, writing conditionals
1900: to test whether or how this macro is defined, and then controlling
1.1.1.2 root 1901: the state of the macro with compiler command options.
1.1 root 1902: @xref{Invocation}.
1903: @end itemize
1904:
1905: @node #error Command,, Conditionals-Macros, Conditionals
1906: @subsection The @samp{#error} Command
1907:
1908: @findex #error
1909: The command @samp{#error} causes the preprocessor to report a fatal
1910: error. The rest of the line that follows @samp{#error} is used as the
1911: error message.
1912:
1913: You would use @samp{#error} inside of a conditional that detects a
1914: combination of parameters which you know the program does not properly
1915: support. For example, if you know that the program will not run
1916: properly on a Vax, you might write
1917:
1918: @example
1919: #ifdef vax
1920: #error Won't work on Vaxen. See comments at get_last_object.
1921: #endif
1922: @end example
1923:
1924: @noindent
1925: @xref{Nonstandard Predefined}, for why this works.
1926:
1927: If you have several configuration parameters that must be set up by
1928: the installation in a consistent way, you can use conditionals to detect
1929: an inconsistency and report it with @samp{#error}. For example,
1930:
1931: @example
1932: #if HASH_TABLE_SIZE % 2 == 0 || HASH_TABLE_SIZE % 3 == 0 \
1933: || HASH_TABLE_SIZE % 5 == 0
1934: #error HASH_TABLE_SIZE should not be divisible by a small prime
1935: #endif
1936: @end example
1937:
1938: @node Combining Sources, Other Commands, Conditionals, Top
1939: @section Combining Source Files
1940:
1941: @cindex line control
1942: @findex #line
1943: One of the jobs of the C preprocessor is to inform the C compiler of where
1944: each line of C code came from: which source file and which line number.
1945:
1946: C code can come from multiple source files if you use @samp{#include};
1947: both @samp{#include} and the use of conditionals and macros can cause
1948: the line number of a line in the preprocessor output to be different
1949: from the line's number in the original source file. You will appreciate
1950: the value of making both the C compiler (in error messages) and symbolic
1951: debuggers such as GDB use the line numbers in your source file.
1952:
1953: The C preprocessor builds on this feature by offering a command by which
1954: you can control the feature explicitly. This is useful when a file for
1955: input to the C preprocessor is the output from another program such as the
1956: @code{bison} parser generator, which operates on another file that is the
1957: true source file. Parts of the output from @code{bison} are generated from
1958: scratch, other parts come from a standard parser file. The rest are copied
1959: nearly verbatim from the source file, but their line numbers in the
1960: @code{bison} output are not the same as their original line numbers.
1961: Naturally you would like compiler error messages and symbolic debuggers to
1962: know the original source file and line number of each line in the
1963: @code{bison} output.
1964:
1965: @code{bison} arranges this by writing @samp{#line} commands into the output
1966: file. @samp{#line} is a command that specifies the original line number
1967: and source file name for subsequent input in the current preprocessor input
1968: file. @samp{#line} has three variants:
1969:
1970: @table @code
1971: @item #line @var{linenum}
1972: Here @var{linenum} is a decimal integer constant. This specifies that
1973: the line number of the following line of input, in its original source file,
1974: was @var{linenum}.
1975:
1976: @item #line @var{linenum} @var{filename}
1977: Here @var{linenum} is a decimal integer constant and @var{filename}
1978: is a string constant. This specifies that the following line of input
1979: came originally from source file @var{filename} and its line number there
1980: was @var{linenum}. Keep in mind that @var{filename} is not just a
1.1.1.2 root 1981: file name; it is surrounded by doublequote characters so that it looks
1982: like a string constant.
1.1 root 1983:
1984: @item #line @var{anything else}
1985: @var{anything else} is checked for macro calls, which are expanded.
1986: The result should be a decimal integer constant followed optionally
1987: by a string constant, as described above.
1988: @end table
1989:
1990: @samp{#line} commands alter the results of the @samp{__FILE__} and
1991: @samp{__LINE__} predefined macros from that point on. @xref{Standard
1992: Predefined}.
1993:
1994: @node Other Commands, Output, Combining Sources, Top
1995: @section Miscellaneous Preprocessor Commands
1996:
1997: @findex #pragma
1.1.1.6 root 1998: @findex #ident
1.1 root 1999: @cindex null command
1.1.1.8 ! root 2000: This section describes three additional preprocessor commands. They are
1.1 root 2001: not very useful, but are mentioned for completeness.
2002:
2003: The @dfn{null command} consists of a @samp{#} followed by a Newline, with
2004: only whitespace (including comments) in between. A null command is
2005: understood as a preprocessor command but has no effect on the preprocessor
2006: output. The primary significance of the existence of the null command is
2007: that an input line consisting of just a @samp{#} will produce no output,
2008: rather than a line of output containing just a @samp{#}. Supposedly
2009: some old C programs contain such lines.
2010:
2011: The @samp{#pragma} command is specified in the ANSI standard to have an
2012: arbitrary implementation-defined effect. In the GNU C preprocessor,
1.1.1.6 root 2013: @samp{#pragma} commands are ignored, except for @samp{#pragma once}
2014: (@pxref{Once-Only}).
2015:
1.1.1.8 ! root 2016: The @samp{#ident} command is supported for compatibility with certain
! 2017: other systems. It is followed by a line of text. On certain systems,
! 2018: the text is copied into a special place in the object file; on most
! 2019: systems, the text is ignored and this directive has no effect.
1.1 root 2020:
2021: @node Output, Invocation, Other Commands, Top
2022: @section C Preprocessor Output
2023:
2024: @cindex output format
2025: The output from the C preprocessor looks much like the input, except
2026: that all preprocessor command lines have been replaced with blank lines
2027: and all comments with spaces. Whitespace within a line is not altered;
2028: however, a space is inserted after the expansions of most macro calls.
2029:
2030: Source file name and line number information is conveyed by lines of
2031: the form
2032:
2033: @example
1.1.1.7 root 2034: # @var{linenum} @var{filename} @var{flag}
1.1 root 2035: @end example
2036:
2037: @noindent
1.1.1.7 root 2038: which are inserted as needed into the middle of the input (but never
2039: within a string or character constant). Such a line means that the
2040: following line originated in file @var{filename} at line @var{linenum}.
2041:
2042: The third field, @var{flag}, may be a number, or may be absent. It is
2043: @samp{1} for the beginning of a new source file, and @samp{2} for return
2044: to an old source file at the end of an included file. It is absent
2045: otherwise.
1.1 root 2046:
1.1.1.8 ! root 2047: @node Invocation, Concept Index, Output, Top
1.1 root 2048: @section Invoking the C Preprocessor
2049:
2050: Most often when you use the C preprocessor you will not have to invoke it
2051: explicitly: the C compiler will do so automatically. However, the
2052: preprocessor is sometimes useful individually.
2053:
2054: The C preprocessor expects two file names as arguments, @var{infile} and
2055: @var{outfile}. The preprocessor reads @var{infile} together with any other
2056: files it specifies with @samp{#include}. All the output generated by the
2057: combined input files is written in @var{outfile}.
2058:
2059: Either @var{infile} or @var{outfile} may be @samp{-}, which as @var{infile}
2060: means to read from standard input and as @var{outfile} means to write to
2061: standard output. Also, if @var{outfile} or both file names are omitted,
2062: the standard output and standard input are used for the omitted file names.
2063:
1.1.1.2 root 2064: @cindex options
2065: Here is a table of command options accepted by the C preprocessor. Most
1.1 root 2066: of them can also be given when compiling a C program; they are passed along
2067: automatically to the preprocessor when it is invoked by the compiler.
2068:
2069: @table @samp
2070: @item -P
2071: @findex -P
2072: Inhibit generation of @samp{#}-lines with line-number information in
2073: the output from the preprocessor (@pxref{Output}). This might be
2074: useful when running the preprocessor on something that is not C code
2075: and will be sent to a program which might be confused by the
2076: @samp{#}-lines
2077:
2078: @item -C
2079: @findex -C
2080: Do not discard comments: pass them through to the output file.
2081: Comments appearing in arguments of a macro call will be copied to the
2082: output before the expansion of the macro call.
2083:
1.1.1.6 root 2084: @item -trigraphs
2085: @findex -trigraphs
1.1 root 2086: Process ANSI standard trigraph sequences. These are three-character
2087: sequences, all starting with @samp{??}, that are defined by ANSI C to
2088: stand for single characters. For example, @samp{??/} stands for
1.1.1.6 root 2089: @samp{\}, so @samp{'??/n'} is a character constant for a newline.
1.1 root 2090: Strictly speaking, the GNU C preprocessor does not support all
1.1.1.6 root 2091: programs in ANSI Standard C unless @samp{-trigraphs} is used, but if
2092: you ever notice the difference it will be with relief.
1.1 root 2093:
2094: You don't want to know any more about trigraphs.
2095:
2096: @item -pedantic
2097: @findex -pedantic
2098: Issue warnings required by the ANSI C standard in certain cases such
2099: as when text other than a comment follows @samp{#else} or @samp{#endif}.
2100:
2101: @item -I @var{directory}
2102: @findex -I
2103: Add the directory @var{directory} to the end of the list of
1.1.1.2 root 2104: directories to be searched for header files (@pxref{Include Syntax}).
2105: This can be used to override a system header file, substituting your
2106: own version, since these directories are searched before the system
2107: header file directories. If you use more than one @samp{-I} option,
2108: the directories are scanned in left-to-right order; the standard
2109: system directories come after.
1.1 root 2110:
1.1.1.3 root 2111: @item -I-
2112: Any directories specified with @samp{-I} options before the @samp{-I-}
2113: option are searched only for the case of @samp{#include "@var{file}"};
2114: they are not searched for @samp{#include <@var{file}>}.
2115:
2116: If additional directories are specified with @samp{-I} options after
2117: the @samp{-I-}, these directories are searched for all @samp{#include}
2118: directives.
2119:
2120: In addition, the @samp{-I-} option inhibits the use of the current
2121: directory as the first search directory for @samp{#include "@var{file}"}.
2122: Therefore, the current directory is searched only if it is requested
2123: explicitly with @samp{-I.}. Specifying both @samp{-I-} and @samp{-I.}
2124: allows you to control precisely which directories are searched before
2125: the current one and which are searched after.
2126:
2127: @item -nostdinc
2128: Do not search the standard system directories for header files.
2129: Only the directories you have specified with @samp{-I} options
2130: (and the current directory, if appropriate) are searched.
2131:
1.1 root 2132: @item -D @var{name}
2133: @findex -D
2134: Predefine @var{name} as a macro, with definition @samp{1}.
2135:
2136: @item -D @var{name}=@var{definition}
2137: Predefine @var{name} as a macro, with definition @var{definition}.
2138: There are no restrictions on the contents of @var{definition}, but if
2139: you are invoking the preprocessor from a shell or shell-like program
2140: you may need to use the shell's quoting syntax to protect characters
2141: such as spaces that have a meaning in the shell syntax.
2142:
2143: @item -U @var{name}
2144: @findex -U
2145: Do not predefine @var{name}. If both @samp{-U} and @samp{-D} are
2146: specified for one name, the @samp{-U} beats the @samp{-D} and the name
2147: is not predefined.
2148:
2149: @item -undef
2150: @findex -undef
2151: Do not predefine any nonstandard macros.
2152:
2153: @item -d
2154: @findex -d
2155: Instead of outputting the result of preprocessing, output a list of
2156: @samp{#define} commands for all the macros defined during the
2157: execution of the preprocessor.
2158:
1.1.1.2 root 2159: @item -M
2160: @findex -M
2161: Instead of outputting the result of preprocessing, output a rule
2162: suitable for @code{make} describing the dependencies of the main
2163: source file. The preprocessor outputs one @code{make} rule containing
2164: the object file name for that source file, a colon, and the names of
2165: all the included files. If there are many included files then the
2166: rule is split into several lines using @samp{\}-newline.
2167:
2168: This feature is used in automatic updating of makefiles.
2169:
2170: @item -MM
2171: @findex -MM
2172: Like @samp{-M} but mention only the files included with @samp{#include
2173: "@var{file}"}. System header files included with @samp{#include
2174: <@var{file}>} are omitted.
2175:
1.1 root 2176: @item -i @var{file}
2177: @findex -i
2178: Process @var{file} as input, discarding the resulting output, before
2179: processing the regular input file. Because the output generated from
2180: @var{file} is discarded, the only effect of @samp{-i @var{file}} is to
2181: make the macros defined in @var{file} available for use in the main
2182: input.
2183: @end table
2184:
2185: @node Concept Index, Index, Invocation, Top
2186: @unnumbered Concept Index
2187: @printindex cp
2188:
2189: @node Index,, Concept Index, Top
1.1.1.2 root 2190: @unnumbered Index of Commands, Macros and Options
1.1 root 2191: @printindex fn
2192:
2193: @contents
2194: @bye
This archive runs on limited infrastructure. Preserving old code on modern bandwidth. Automated agents are requested to crawl responsibly.