1 /* Copyright (C) 1989, 2000 Aladdin Enterprises. All rights reserved. */
3 /*$Id: ansi2knr.c,v 1.14 2003/09/06 05:36:56 eggert Exp $*/
4 /* Convert ANSI C function definitions to K&R ("traditional C") syntax */
7 ansi2knr is distributed in the hope that it will be useful, but WITHOUT ANY
8 WARRANTY. No author or distributor accepts responsibility to anyone for the
9 consequences of using it or for whether it serves any particular purpose or
10 works at all, unless he says so in writing. Refer to the GNU General Public
11 License (the "GPL") for full details.
13 Everyone is granted permission to copy, modify and redistribute ansi2knr,
14 but only under the conditions described in the GPL. A copy of this license
15 is supposed to have been given to you along with ansi2knr so you can know
16 your rights and responsibilities. It should be in a file named COPYLEFT,
17 or, if there is no file named COPYLEFT, a file named COPYING. Among other
18 things, the copyright notice and this notice must be preserved on all
21 We explicitly state here what we believe is already implied by the GPL: if
22 the ansi2knr program is distributed as a separate set of sources and a
23 separate executable file which are aggregated on a storage medium together
24 with another program, this in itself does not bring the other program under
25 the GPL, nor does the mere fact that such a program or the procedures for
26 constructing it invoke the ansi2knr executable bring any other part of the
27 program under the GPL.
32 ansi2knr [--filename FILENAME] [INPUT_FILE [OUTPUT_FILE]]
33 * --filename provides the file name for the #line directive in the output,
34 * overriding input_file (if present).
35 * If no input_file is supplied, input is read from stdin.
36 * If no output_file is supplied, output goes to stdout.
37 * There are no error messages.
39 * ansi2knr recognizes function definitions by seeing a non-keyword
40 * identifier at the left margin, followed by a left parenthesis, with a
41 * right parenthesis as the last character on the line, and with a left
42 * brace as the first token on the following line (ignoring possible
43 * intervening comments and/or preprocessor directives), except that a line
45 * identifier1(identifier2)
46 * will not be considered a function definition unless identifier2 is
47 * the word "void", and a line consisting of
48 * identifier1(identifier2, <<arbitrary>>)
49 * will not be considered a function definition.
50 * ansi2knr will recognize a multi-line header provided that no intervening
51 * line ends with a left or right brace or a semicolon. These algorithms
52 * ignore whitespace, comments, and preprocessor directives, except that
53 * the function name must be the first thing on the line. The following
54 * constructs will confuse it:
55 * - Any other construct that starts at the left margin and
56 * follows the above syntax (such as a macro or function call).
57 * - Some macros that tinker with the syntax of function headers.
61 * The original and principal author of ansi2knr is L. Peter Deutsch
62 * <ghost@aladdin.com>. Other authors are noted in the change history
63 * that follows (in reverse chronological order):
65 lpd 2000-04-12 backs out Eggert's changes because of bugs:
66 - concatlits didn't declare the type of its bufend argument;
67 - concatlits didn't recognize when it was inside a comment;
68 - scanstring could scan backward past the beginning of the string; when
69 - the check for \ + newline in scanstring was unnecessary.
71 2000-03-05 Paul Eggert <eggert@twinsun.com>
73 Add support for concatenated string literals.
74 * ansi2knr.c (concatlits): New decl.
75 (main): Invoke concatlits to concatenate string literals.
76 (scanstring): Handle backslash-newline correctly. Work with
77 character constants. Fix bug when scanning backwards through
78 backslash-quote. Check for unterminated strings.
79 (convert1): Parse character constants, too.
80 (appendline, concatlits): New functions.
81 * ansi2knr.1: Document this.
83 lpd 1999-08-17 added code to allow preprocessor directives
84 wherever comments are allowed
85 lpd 1999-04-12 added minor fixes from Pavel Roskin
86 <pavel_roskin@geocities.com> for clean compilation with
88 lpd 1999-03-22 added hack to recognize lines consisting of
89 identifier1(identifier2, xxx) as *not* being procedures
90 lpd 1999-02-03 made indentation of preprocessor commands consistent
91 lpd 1999-01-28 fixed two bugs: a '/' in an argument list caused an
92 endless loop; quoted strings within an argument list
94 lpd 1999-01-24 added a check for write errors on the output,
95 suggested by Jim Meyering <meyering@ascend.com>
96 lpd 1998-11-09 added further hack to recognize identifier(void)
98 lpd 1998-10-23 added hack to recognize lines consisting of
99 identifier1(identifier2) as *not* being procedures
100 lpd 1997-12-08 made input_file optional; only closes input and/or
101 output file if not stdin or stdout respectively; prints
102 usage message on stderr rather than stdout; adds
103 --filename switch (changes suggested by
104 <ceder@lysator.liu.se>)
105 lpd 1996-01-21 added code to cope with not HAVE_CONFIG_H and with
106 compilers that don't understand void, as suggested by
108 lpd 1996-01-15 changed to require that the first non-comment token
109 on the line following a function header be a left brace,
110 to reduce sensitivity to macros, as suggested by Tom Lane
112 lpd 1995-06-22 removed #ifndefs whose sole purpose was to define
113 undefined preprocessor symbols as 0; changed all #ifdefs
114 for configuration symbols to #ifs
115 lpd 1995-04-05 changed copyright notice to make it clear that
116 including ansi2knr in a program does not bring the entire
117 program under the GPL
118 lpd 1994-12-18 added conditionals for systems where ctype macros
119 don't handle 8-bit characters properly, suggested by
120 Francois Pinard <pinard@iro.umontreal.ca>;
121 removed --varargs switch (this is now the default)
122 lpd 1994-10-10 removed CONFIG_BROKETS conditional
123 lpd 1994-07-16 added some conditionals to help GNU `configure',
124 suggested by Francois Pinard <pinard@iro.umontreal.ca>;
125 properly erase prototype args in function parameters,
126 contributed by Jim Avera <jima@netcom.com>;
127 correct error in writeblanks (it shouldn't erase EOLs)
128 lpd 1989-xx-xx original version
131 /* Most of the conditionals here are to make ansi2knr work with */
132 /* or without the GNU configure machinery. */
144 For properly autoconfiguring ansi2knr, use AC_CONFIG_HEADER(config.h).
145 This will define HAVE_CONFIG_H and so, activate the following lines.
148 # if STDC_HEADERS || HAVE_STRING_H
151 # include <strings.h>
154 #else /* not HAVE_CONFIG_H */
156 /* Otherwise do it the hard way */
159 # include <strings.h>
162 extern int strlen(), strncmp();
168 #endif /* not HAVE_CONFIG_H */
174 malloc and free should be declared in stdlib.h,
175 but if you've got a K&R compiler, they probably aren't.
181 extern char *malloc();
184 extern char *malloc();
191 /* Define NULL (for *very* old compilers). */
197 * The ctype macros don't always handle 8-bit characters correctly.
198 * Compensate for this here.
201 # undef HAVE_ISASCII /* just in case */
202 # define HAVE_ISASCII 1
205 #if STDC_HEADERS || !HAVE_ISASCII
206 # define is_ascii(c) 1
208 # define is_ascii(c) isascii(c)
211 #define is_space(c) (is_ascii(c) && isspace(c))
212 #define is_alpha(c) (is_ascii(c) && isalpha(c))
213 #define is_alnum(c) (is_ascii(c) && isalnum(c))
215 /* Scanning macros */
216 #define isidchar(ch) (is_alnum(ch) || (ch) == '_')
217 #define isidfirstchar(ch) (is_alpha(ch) || (ch) == '_')
219 /* Forward references */
220 char *ppdirforward();
221 char *ppdirbackward();
228 /* The main program */
236 char *program_name = argv[0];
237 char *output_name = 0;
238 #define bufsize 5000 /* arbitrary size */
243 "Usage: ansi2knr [--filename FILENAME] [INPUT_FILE [OUTPUT_FILE]]\n";
245 * In previous versions, ansi2knr recognized a --varargs switch.
246 * If this switch was supplied, ansi2knr would attempt to convert
247 * a ... argument to va_alist and va_dcl; if this switch was not
248 * supplied, ansi2knr would simply drop any such arguments.
249 * Now, ansi2knr always does this conversion, and we only
250 * check for this switch for backward compatibility.
252 int convert_varargs = 1;
255 while ( argc > 1 && argv[1][0] == '-' ) {
256 if ( !strcmp(argv[1], "--varargs") ) {
262 if ( !strcmp(argv[1], "--filename") && argc > 2 ) {
268 fprintf(stderr, "%s: Unrecognized switch: %s\n", program_name,
270 fprintf(stderr, usage);
276 fprintf(stderr, usage);
279 output_name = argv[2];
280 out = fopen(output_name, "w");
282 fprintf(stderr, "%s: Cannot open output file %s\n",
283 program_name, output_name);
288 in = fopen(argv[1], "r");
290 fprintf(stderr, "%s: Cannot open input file %s\n",
291 program_name, argv[1]);
301 fprintf(out, "#line 1 \"%s\"\n", filename);
302 buf = malloc(bufsize);
305 fprintf(stderr, "Unable to allocate read buffer!\n");
309 while ( fgets(line, (unsigned)(buf + bufsize - line), in) != NULL )
311 test: line += strlen(line);
312 switch ( test1(buf) )
314 case 2: /* a function header */
315 convert1(buf, out, 1, convert_varargs);
317 case 1: /* a function */
318 /* Check for a { at the start of the next line. */
320 f: if ( line >= buf + (bufsize - 1) ) /* overflow check */
322 if ( fgets(line, (unsigned)(buf + bufsize - line), in) == NULL )
324 switch ( *skipspace(ppdirforward(more), 1) )
327 /* Definitely a function header. */
328 convert1(buf, out, 0, convert_varargs);
332 /* The next line was blank or a comment: */
333 /* keep scanning for a non-comment. */
334 line += strlen(line);
337 /* buf isn't a function header, but */
345 case -1: /* maybe the start of a function */
346 if ( line != buf + (bufsize - 1) ) /* overflow check */
349 default: /* not a function */
359 output_error = ferror(out);
360 output_error |= fclose(out);
361 } else { /* out == stdout */
363 output_error = ferror(out);
365 if ( output_error ) {
366 fprintf(stderr, "%s: error writing to %s\n", program_name,
367 (output_name ? output_name : "stdout"));
376 * Skip forward or backward over one or more preprocessor directives.
382 for (; *p == '#'; ++p) {
383 for (; *p != '\r' && *p != '\n'; ++p)
386 if (*p == '\r' && p[1] == '\n')
392 ppdirbackward(p, limit)
399 if (*np == '\n' && np[-1] == '\r')
401 for (; np > limit && np[-1] != '\r' && np[-1] != '\n'; --np)
410 * Skip over whitespace, comments, and preprocessor directives,
411 * in either direction.
416 int dir; /* 1 for forward, -1 for backward */
419 while ( is_space(*p) )
421 if ( !(*p == '/' && p[dir] == '*') )
424 while ( !(*p == '*' && p[dir] == '/') ) {
426 return p; /* multi-line comment?? */
434 /* Scan over a quoted string, in either direction. */
440 for (p += dir; ; p += dir)
441 if (*p == '"' && p[-dir] != '\\')
446 * Write blanks over part of a string.
447 * Don't overwrite end-of-line characters.
450 writeblanks(start, end)
454 for ( p = start; p < end; p++ )
455 if ( *p != '\r' && *p != '\n' )
461 * Test whether the string in buf is a function definition.
462 * The string may contain and/or end with a newline.
464 * 0 - definitely not a function definition;
465 * 1 - definitely a function definition;
466 * 2 - definitely a function prototype (NOT USED);
467 * -1 - may be the beginning of a function definition,
468 * append another line and look again.
469 * The reason we don't attempt to convert function prototypes is that
470 * Ghostscript's declaration-generating macros look too much like
471 * prototypes, and confuse the algorithms.
481 if ( !isidfirstchar(*p) )
482 return 0; /* no name at left margin */
483 bend = skipspace(ppdirbackward(buf + strlen(buf) - 1, buf), -1);
486 case ';': contin = 0 /*2*/; break;
487 case ')': contin = 1; break;
488 case '{': return 0; /* not a function */
489 case '}': return 0; /* not a function */
490 default: contin = -1;
492 while ( isidchar(*p) )
497 return 0; /* not a function */
500 return 0; /* no parameters */
501 /* Check that the apparent function name isn't a keyword. */
502 /* We only need to check for keywords that could be followed */
503 /* by a left parenthesis (which, unfortunately, is most of them). */
504 { static char *words[] =
505 { "asm", "auto", "case", "char", "const", "double",
506 "extern", "float", "for", "if", "int", "long",
507 "register", "return", "short", "signed", "sizeof",
508 "static", "switch", "typedef", "unsigned",
509 "void", "volatile", "while", 0
513 unsigned len = endfn - buf;
515 while ( (kp = *key) != 0 )
516 { if ( strlen(kp) == len && !strncmp(kp, buf, len) )
517 return 0; /* name is a keyword */
525 * Check for identifier1(identifier2) and not
526 * identifier1(void), or identifier1(identifier2, xxxx).
529 while ( isidchar(*p) )
534 (*p == ')' && (len != 4 || strncmp(id, "void", 4)))
536 return 0; /* not a function */
539 * If the last significant character was a ), we need to count
540 * parentheses, because it might be part of a formal parameter
541 * that is a procedure.
546 for (p = skipspace(buf, 1); *p; p = skipspace(p + 1, 1))
547 level += (*p == '(' ? 1 : *p == ')' ? -1 : 0);
554 /* Convert a recognized function definition or header to K&R syntax. */
556 convert1(buf, out, header, convert_varargs)
559 int header; /* Boolean */
560 int convert_varargs; /* Boolean */
564 * The breaks table contains pointers to the beginning and end
568 unsigned num_breaks = 2; /* for testing */
574 /* Pre-ANSI implementations don't agree on whether strchr */
575 /* is called strchr or index, so we open-code it here. */
576 for ( endfn = buf; *(endfn++) != '('; )
579 breaks = (char **)malloc(sizeof(char *) * num_breaks * 2);
580 if ( breaks == NULL )
581 { /* Couldn't allocate break table, give up */
582 fprintf(stderr, "Unable to allocate break table!\n");
586 btop = breaks + num_breaks * 2 - 2;
588 /* Parse the argument list */
596 { /* Filled up break table. */
597 /* Allocate a bigger one and start over. */
598 free((char *)breaks);
603 /* Find the end of the argument */
604 for ( ; end == NULL; p++ )
608 if ( !level ) end = p;
611 if ( !level ) lp = p;
615 if ( --level < 0 ) end = p;
620 p = skipspace(p, 1) - 1;
623 p = scanstring(p, 1) - 1;
629 /* Erase any embedded prototype parameters. */
631 writeblanks(lp + 1, rp);
632 p--; /* back up over terminator */
633 /* Find the name being declared. */
634 /* This is complicated because of procedure and */
635 /* array modifiers. */
637 { p = skipspace(p - 1, -1);
640 case ']': /* skip array dimension(s) */
641 case ')': /* skip procedure args OR name */
653 if (p > buf && p[-1] == '*')
654 p = skipspace(p, -1) + 1;
657 p = scanstring(p, -1) + 1;
662 if ( *p == '(' && *skipspace(p + 1, 1) == '*' )
663 { /* We found the name being declared */
664 while ( !isidfirstchar(*p) )
665 p = skipspace(p, 1) + 1;
673 found: if ( *p == '.' && p[-1] == '.' && p[-2] == '.' )
674 { if ( convert_varargs )
675 { *bp++ = "va_alist";
680 if ( bp == breaks + 1 ) /* sole argument */
681 writeblanks(breaks[0], p);
683 writeblanks(bp[-1] - 1, p);
688 { while ( isidchar(*p) ) p--;
693 while ( *p++ == ',' );
695 /* Make a special check for 'void' arglist */
696 if ( bp == breaks+2 )
697 { p = skipspace(breaks[0], 1);
698 if ( !strncmp(p, "void", 4) )
699 { p = skipspace(p+4, 1);
700 if ( p == breaks[2] - 1 )
701 { bp = breaks; /* yup, pretend arglist is empty */
702 writeblanks(breaks[0], p + 1);
706 /* Put out the function name and left parenthesis. */
708 while ( p != endfn ) putc(*p, out), p++;
709 /* Put out the declaration. */
712 for ( p = breaks[0]; *p; p++ )
713 if ( *p == '\r' || *p == '\n' )
717 { for ( ap = breaks+1; ap < bp; ap += 2 )
719 while ( isidchar(*p) )
725 /* Put out the argument declarations */
726 for ( ap = breaks+2; ap <= bp; ap += 2 )
730 fputs(breaks[0], out); /* any prior args */
731 fputs("va_dcl", out); /* the final arg */
735 fputs(breaks[0], out);
737 free((char *)breaks);