1 /* Copyright (C) 1989, 1997, 1998, 1999 Aladdin Enterprises. All rights reserved. */
3 /* Convert ANSI C function definitions to K&R ("traditional C") syntax */
6 ansi2knr is distributed in the hope that it will be useful, but WITHOUT ANY
7 WARRANTY. No author or distributor accepts responsibility to anyone for the
8 consequences of using it or for whether it serves any particular purpose or
9 works at all, unless he says so in writing. Refer to the GNU General Public
10 License (the "GPL") for full details.
12 Everyone is granted permission to copy, modify and redistribute ansi2knr,
13 but only under the conditions described in the GPL. A copy of this license
14 is supposed to have been given to you along with ansi2knr so you can know
15 your rights and responsibilities. It should be in a file named COPYLEFT,
16 or, if there is no file named COPYLEFT, a file named COPYING. Among other
17 things, the copyright notice and this notice must be preserved on all
20 We explicitly state here what we believe is already implied by the GPL: if
21 the ansi2knr program is distributed as a separate set of sources and a
22 separate executable file which are aggregated on a storage medium together
23 with another program, this in itself does not bring the other program under
24 the GPL, nor does the mere fact that such a program or the procedures for
25 constructing it invoke the ansi2knr executable bring any other part of the
26 program under the GPL.
31 ansi2knr [--filename FILENAME] [INPUT_FILE [OUTPUT_FILE]]
32 * --filename provides the file name for the #line directive in the output,
33 * overriding input_file (if present).
34 * If no input_file is supplied, input is read from stdin.
35 * If no output_file is supplied, output goes to stdout.
36 * There are no error messages.
38 * ansi2knr recognizes function definitions by seeing a non-keyword
39 * identifier at the left margin, followed by a left parenthesis,
40 * with a right parenthesis as the last character on the line,
41 * and with a left brace as the first token on the following line
42 * (ignoring possible intervening comments), except that a line
44 * identifier1(identifier2)
45 * will not be considered a function definition unless identifier2 is
46 * the word "void", and a line consisting of
47 * identifier1(identifier2, <<arbitrary>>)
48 * will not be considered a function definition.
49 * ansi2knr will recognize a multi-line header provided
50 * that no intervening line ends with a left or right brace or a semicolon.
51 * These algorithms ignore whitespace and comments, except that
52 * the function name must be the first thing on the line.
53 * The following constructs will confuse it:
54 * - Any other construct that starts at the left margin and
55 * follows the above syntax (such as a macro or function call).
56 * - Some macros that tinker with the syntax of function headers.
60 * The original and principal author of ansi2knr is L. Peter Deutsch
61 * <ghost@aladdin.com>. Other authors are noted in the change history
62 * that follows (in reverse chronological order):
63 lpd 1999-04-12 added minor fixes from Pavel Roskin
64 <pavel_roskin@geocities.com> for clean compilation with
66 lpd 1999-03-22 added hack to recognize lines consisting of
67 identifier1(identifier2, xxx) as *not* being procedures
68 lpd 1999-02-03 made indentation of preprocessor commands consistent
69 lpd 1999-01-28 fixed two bugs: a '/' in an argument list caused an
70 endless loop; quoted strings within an argument list
72 lpd 1999-01-24 added a check for write errors on the output,
73 suggested by Jim Meyering <meyering@ascend.com>
74 lpd 1998-11-09 added further hack to recognize identifier(void)
76 lpd 1998-10-23 added hack to recognize lines consisting of
77 identifier1(identifier2) as *not* being procedures
78 lpd 1997-12-08 made input_file optional; only closes input and/or
79 output file if not stdin or stdout respectively; prints
80 usage message on stderr rather than stdout; adds
81 --filename switch (changes suggested by
82 <ceder@lysator.liu.se>)
83 lpd 1996-01-21 added code to cope with not HAVE_CONFIG_H and with
84 compilers that don't understand void, as suggested by
86 lpd 1996-01-15 changed to require that the first non-comment token
87 on the line following a function header be a left brace,
88 to reduce sensitivity to macros, as suggested by Tom Lane
90 lpd 1995-06-22 removed #ifndefs whose sole purpose was to define
91 undefined preprocessor symbols as 0; changed all #ifdefs
92 for configuration symbols to #ifs
93 lpd 1995-04-05 changed copyright notice to make it clear that
94 including ansi2knr in a program does not bring the entire
96 lpd 1994-12-18 added conditionals for systems where ctype macros
97 don't handle 8-bit characters properly, suggested by
98 Francois Pinard <pinard@iro.umontreal.ca>;
99 removed --varargs switch (this is now the default)
100 lpd 1994-10-10 removed CONFIG_BROKETS conditional
101 lpd 1994-07-16 added some conditionals to help GNU `configure',
102 suggested by Francois Pinard <pinard@iro.umontreal.ca>;
103 properly erase prototype args in function parameters,
104 contributed by Jim Avera <jima@netcom.com>;
105 correct error in writeblanks (it shouldn't erase EOLs)
106 lpd 1989-xx-xx original version
109 /* Most of the conditionals here are to make ansi2knr work with */
110 /* or without the GNU configure machinery. */
122 For properly autoconfiguring ansi2knr, use AC_CONFIG_HEADER(config.h).
123 This will define HAVE_CONFIG_H and so, activate the following lines.
126 # if STDC_HEADERS || HAVE_STRING_H
129 # include <strings.h>
132 #else /* not HAVE_CONFIG_H */
134 /* Otherwise do it the hard way */
137 # include <strings.h>
140 extern int strlen(), strncmp();
146 #endif /* not HAVE_CONFIG_H */
152 malloc and free should be declared in stdlib.h,
153 but if you've got a K&R compiler, they probably aren't.
159 extern char *malloc();
162 extern char *malloc();
169 /* Define NULL (for *very* old compilers). */
175 * The ctype macros don't always handle 8-bit characters correctly.
176 * Compensate for this here.
179 # undef HAVE_ISASCII /* just in case */
180 # define HAVE_ISASCII 1
183 #if STDC_HEADERS || !HAVE_ISASCII
184 # define is_ascii(c) 1
186 # define is_ascii(c) isascii(c)
189 #define is_space(c) (is_ascii(c) && isspace(c))
190 #define is_alpha(c) (is_ascii(c) && isalpha(c))
191 #define is_alnum(c) (is_ascii(c) && isalnum(c))
193 /* Scanning macros */
194 #define isidchar(ch) (is_alnum(ch) || (ch) == '_')
195 #define isidfirstchar(ch) (is_alpha(ch) || (ch) == '_')
197 /* Forward references */
204 /* The main program */
212 char *program_name = argv[0];
213 char *output_name = 0;
214 #define bufsize 5000 /* arbitrary size */
219 "Usage: ansi2knr [--filename FILENAME] [INPUT_FILE [OUTPUT_FILE]]\n";
221 * In previous versions, ansi2knr recognized a --varargs switch.
222 * If this switch was supplied, ansi2knr would attempt to convert
223 * a ... argument to va_alist and va_dcl; if this switch was not
224 * supplied, ansi2knr would simply drop any such arguments.
225 * Now, ansi2knr always does this conversion, and we only
226 * check for this switch for backward compatibility.
228 int convert_varargs = 1;
231 while ( argc > 1 && argv[1][0] == '-' ) {
232 if ( !strcmp(argv[1], "--varargs") ) {
238 if ( !strcmp(argv[1], "--filename") && argc > 2 ) {
244 fprintf(stderr, "%s: Unrecognized switch: %s\n", program_name,
246 fprintf(stderr, usage);
252 fprintf(stderr, usage);
255 output_name = argv[2];
256 out = fopen(output_name, "w");
258 fprintf(stderr, "%s: Cannot open output file %s\n",
259 program_name, output_name);
264 in = fopen(argv[1], "r");
266 fprintf(stderr, "%s: Cannot open input file %s\n",
267 program_name, argv[1]);
277 fprintf(out, "#line 1 \"%s\"\n", filename);
278 buf = malloc(bufsize);
281 fprintf(stderr, "Unable to allocate read buffer!\n");
285 while ( fgets(line, (unsigned)(buf + bufsize - line), in) != NULL )
287 test: line += strlen(line);
288 switch ( test1(buf) )
290 case 2: /* a function header */
291 convert1(buf, out, 1, convert_varargs);
293 case 1: /* a function */
294 /* Check for a { at the start of the next line. */
296 f: if ( line >= buf + (bufsize - 1) ) /* overflow check */
298 if ( fgets(line, (unsigned)(buf + bufsize - line), in) == NULL )
300 switch ( *skipspace(more, 1) )
303 /* Definitely a function header. */
304 convert1(buf, out, 0, convert_varargs);
308 /* The next line was blank or a comment: */
309 /* keep scanning for a non-comment. */
310 line += strlen(line);
313 /* buf isn't a function header, but */
321 case -1: /* maybe the start of a function */
322 if ( line != buf + (bufsize - 1) ) /* overflow check */
325 default: /* not a function */
335 output_error = ferror(out);
336 output_error |= fclose(out);
337 } else { /* out == stdout */
339 output_error = ferror(out);
341 if ( output_error ) {
342 fprintf(stderr, "%s: error writing to %s\n", program_name,
343 (output_name ? output_name : "stdout"));
351 /* Skip over whitespace and comments, in either direction. */
355 register int dir; /* 1 for forward, -1 for backward */
357 { while ( is_space(*p) )
359 if ( !(*p == '/' && p[dir] == '*') )
362 while ( !(*p == '*' && p[dir] == '/') )
364 return p; /* multi-line comment?? */
372 /* Scan over a quoted string, in either direction. */
378 for (p += dir; ; p += dir)
379 if (*p == '"' && p[-dir] != '\\')
384 * Write blanks over part of a string.
385 * Don't overwrite end-of-line characters.
388 writeblanks(start, end)
392 for ( p = start; p < end; p++ )
393 if ( *p != '\r' && *p != '\n' )
399 * Test whether the string in buf is a function definition.
400 * The string may contain and/or end with a newline.
402 * 0 - definitely not a function definition;
403 * 1 - definitely a function definition;
404 * 2 - definitely a function prototype (NOT USED);
405 * -1 - may be the beginning of a function definition,
406 * append another line and look again.
407 * The reason we don't attempt to convert function prototypes is that
408 * Ghostscript's declaration-generating macros look too much like
409 * prototypes, and confuse the algorithms.
414 { register char *p = buf;
419 if ( !isidfirstchar(*p) )
420 return 0; /* no name at left margin */
421 bend = skipspace(buf + strlen(buf) - 1, -1);
424 case ';': contin = 0 /*2*/; break;
425 case ')': contin = 1; break;
426 case '{': return 0; /* not a function */
427 case '}': return 0; /* not a function */
428 default: contin = -1;
430 while ( isidchar(*p) )
435 return 0; /* not a function */
438 return 0; /* no parameters */
439 /* Check that the apparent function name isn't a keyword. */
440 /* We only need to check for keywords that could be followed */
441 /* by a left parenthesis (which, unfortunately, is most of them). */
442 { static char *words[] =
443 { "asm", "auto", "case", "char", "const", "double",
444 "extern", "float", "for", "if", "int", "long",
445 "register", "return", "short", "signed", "sizeof",
446 "static", "switch", "typedef", "unsigned",
447 "void", "volatile", "while", 0
451 unsigned len = endfn - buf;
453 while ( (kp = *key) != 0 )
454 { if ( strlen(kp) == len && !strncmp(kp, buf, len) )
455 return 0; /* name is a keyword */
463 * Check for identifier1(identifier2) and not
464 * identifier1(void), or identifier1(identifier2, xxxx).
467 while ( isidchar(*p) )
472 (*p == ')' && (len != 4 || strncmp(id, "void", 4)))
474 return 0; /* not a function */
477 * If the last significant character was a ), we need to count
478 * parentheses, because it might be part of a formal parameter
479 * that is a procedure.
484 for (p = skipspace(buf, 1); *p; p = skipspace(p + 1, 1))
485 level += (*p == '(' ? 1 : *p == ')' ? -1 : 0);
492 /* Convert a recognized function definition or header to K&R syntax. */
494 convert1(buf, out, header, convert_varargs)
497 int header; /* Boolean */
498 int convert_varargs; /* Boolean */
502 * The breaks table contains pointers to the beginning and end
506 unsigned num_breaks = 2; /* for testing */
512 /* Pre-ANSI implementations don't agree on whether strchr */
513 /* is called strchr or index, so we open-code it here. */
514 for ( endfn = buf; *(endfn++) != '('; )
517 breaks = (char **)malloc(sizeof(char *) * num_breaks * 2);
518 if ( breaks == NULL )
519 { /* Couldn't allocate break table, give up */
520 fprintf(stderr, "Unable to allocate break table!\n");
524 btop = breaks + num_breaks * 2 - 2;
526 /* Parse the argument list */
534 { /* Filled up break table. */
535 /* Allocate a bigger one and start over. */
536 free((char *)breaks);
541 /* Find the end of the argument */
542 for ( ; end == NULL; p++ )
546 if ( !level ) end = p;
549 if ( !level ) lp = p;
553 if ( --level < 0 ) end = p;
558 p = skipspace(p, 1) - 1;
561 p = scanstring(p, 1) - 1;
567 /* Erase any embedded prototype parameters. */
569 writeblanks(lp + 1, rp);
570 p--; /* back up over terminator */
571 /* Find the name being declared. */
572 /* This is complicated because of procedure and */
573 /* array modifiers. */
575 { p = skipspace(p - 1, -1);
578 case ']': /* skip array dimension(s) */
579 case ')': /* skip procedure args OR name */
591 if (p > buf && p[-1] == '*')
592 p = skipspace(p, -1) + 1;
595 p = scanstring(p, -1) + 1;
600 if ( *p == '(' && *skipspace(p + 1, 1) == '*' )
601 { /* We found the name being declared */
602 while ( !isidfirstchar(*p) )
603 p = skipspace(p, 1) + 1;
611 found: if ( *p == '.' && p[-1] == '.' && p[-2] == '.' )
612 { if ( convert_varargs )
613 { *bp++ = "va_alist";
618 if ( bp == breaks + 1 ) /* sole argument */
619 writeblanks(breaks[0], p);
621 writeblanks(bp[-1] - 1, p);
626 { while ( isidchar(*p) ) p--;
631 while ( *p++ == ',' );
633 /* Make a special check for 'void' arglist */
634 if ( bp == breaks+2 )
635 { p = skipspace(breaks[0], 1);
636 if ( !strncmp(p, "void", 4) )
637 { p = skipspace(p+4, 1);
638 if ( p == breaks[2] - 1 )
639 { bp = breaks; /* yup, pretend arglist is empty */
640 writeblanks(breaks[0], p + 1);
644 /* Put out the function name and left parenthesis. */
646 while ( p != endfn ) putc(*p, out), p++;
647 /* Put out the declaration. */
650 for ( p = breaks[0]; *p; p++ )
651 if ( *p == '\r' || *p == '\n' )
655 { for ( ap = breaks+1; ap < bp; ap += 2 )
657 while ( isidchar(*p) )
663 /* Put out the argument declarations */
664 for ( ap = breaks+2; ap <= bp; ap += 2 )
668 fputs(breaks[0], out); /* any prior args */
669 fputs("va_dcl", out); /* the final arg */
673 fputs(breaks[0], out);
675 free((char *)breaks);