]> Dogcows Code - chaz/tar/blob - src/ansi2knr.c
(dump_regular_file): Fix file padding in case of truncation to zero size.
[chaz/tar] / src / ansi2knr.c
1 /* Copyright (C) 1989, 1997, 1998, 1999 Aladdin Enterprises. All rights reserved. */
2
3 /*$Id: ansi2knr.c $*/
4 /* Convert ANSI C function definitions to K&R ("traditional C") syntax */
5
6 /*
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.
12
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
19 copies.
20
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.
28 */
29
30 /*
31 * Usage:
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.
38 *
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
44 * consisting of only
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.
58 */
59
60 /*
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):
64 lpd 1999-08-17 added code to allow preprocessor directives
65 wherever comments are allowed
66 lpd 1999-04-12 added minor fixes from Pavel Roskin
67 <pavel_roskin@geocities.com> for clean compilation with
68 gcc -W -Wall
69 lpd 1999-03-22 added hack to recognize lines consisting of
70 identifier1(identifier2, xxx) as *not* being procedures
71 lpd 1999-02-03 made indentation of preprocessor commands consistent
72 lpd 1999-01-28 fixed two bugs: a '/' in an argument list caused an
73 endless loop; quoted strings within an argument list
74 confused the parser
75 lpd 1999-01-24 added a check for write errors on the output,
76 suggested by Jim Meyering <meyering@ascend.com>
77 lpd 1998-11-09 added further hack to recognize identifier(void)
78 as being a procedure
79 lpd 1998-10-23 added hack to recognize lines consisting of
80 identifier1(identifier2) as *not* being procedures
81 lpd 1997-12-08 made input_file optional; only closes input and/or
82 output file if not stdin or stdout respectively; prints
83 usage message on stderr rather than stdout; adds
84 --filename switch (changes suggested by
85 <ceder@lysator.liu.se>)
86 lpd 1996-01-21 added code to cope with not HAVE_CONFIG_H and with
87 compilers that don't understand void, as suggested by
88 Tom Lane
89 lpd 1996-01-15 changed to require that the first non-comment token
90 on the line following a function header be a left brace,
91 to reduce sensitivity to macros, as suggested by Tom Lane
92 <tgl@sss.pgh.pa.us>
93 lpd 1995-06-22 removed #ifndefs whose sole purpose was to define
94 undefined preprocessor symbols as 0; changed all #ifdefs
95 for configuration symbols to #ifs
96 lpd 1995-04-05 changed copyright notice to make it clear that
97 including ansi2knr in a program does not bring the entire
98 program under the GPL
99 lpd 1994-12-18 added conditionals for systems where ctype macros
100 don't handle 8-bit characters properly, suggested by
101 Francois Pinard <pinard@iro.umontreal.ca>;
102 removed --varargs switch (this is now the default)
103 lpd 1994-10-10 removed CONFIG_BROKETS conditional
104 lpd 1994-07-16 added some conditionals to help GNU `configure',
105 suggested by Francois Pinard <pinard@iro.umontreal.ca>;
106 properly erase prototype args in function parameters,
107 contributed by Jim Avera <jima@netcom.com>;
108 correct error in writeblanks (it shouldn't erase EOLs)
109 lpd 1989-xx-xx original version
110 */
111
112 /* Most of the conditionals here are to make ansi2knr work with */
113 /* or without the GNU configure machinery. */
114
115 #if HAVE_CONFIG_H
116 # include <config.h>
117 #endif
118
119 #include <stdio.h>
120 #include <ctype.h>
121
122 #if HAVE_CONFIG_H
123
124 /*
125 For properly autoconfiguring ansi2knr, use AC_CONFIG_HEADER(config.h).
126 This will define HAVE_CONFIG_H and so, activate the following lines.
127 */
128
129 # if STDC_HEADERS || HAVE_STRING_H
130 # include <string.h>
131 # else
132 # include <strings.h>
133 # endif
134
135 #else /* not HAVE_CONFIG_H */
136
137 /* Otherwise do it the hard way */
138
139 # ifdef BSD
140 # include <strings.h>
141 # else
142 # ifdef VMS
143 extern int strlen(), strncmp();
144 # else
145 # include <string.h>
146 # endif
147 # endif
148
149 #endif /* not HAVE_CONFIG_H */
150
151 #if STDC_HEADERS
152 # include <stdlib.h>
153 #else
154 /*
155 malloc and free should be declared in stdlib.h,
156 but if you've got a K&R compiler, they probably aren't.
157 */
158 # ifdef MSDOS
159 # include <malloc.h>
160 # else
161 # ifdef VMS
162 extern char *malloc();
163 extern void free();
164 # else
165 extern char *malloc();
166 extern int free();
167 # endif
168 # endif
169
170 #endif
171
172 /* Define NULL (for *very* old compilers). */
173 #ifndef NULL
174 # define NULL (0)
175 #endif
176
177 /*
178 * The ctype macros don't always handle 8-bit characters correctly.
179 * Compensate for this here.
180 */
181 #ifdef isascii
182 # undef HAVE_ISASCII /* just in case */
183 # define HAVE_ISASCII 1
184 #else
185 #endif
186 #if STDC_HEADERS || !HAVE_ISASCII
187 # define is_ascii(c) 1
188 #else
189 # define is_ascii(c) isascii(c)
190 #endif
191
192 #define is_space(c) (is_ascii(c) && isspace(c))
193 #define is_alpha(c) (is_ascii(c) && isalpha(c))
194 #define is_alnum(c) (is_ascii(c) && isalnum(c))
195
196 /* Scanning macros */
197 #define isidchar(ch) (is_alnum(ch) || (ch) == '_')
198 #define isidfirstchar(ch) (is_alpha(ch) || (ch) == '_')
199
200 /* Forward references */
201 char *ppdirforward();
202 char *ppdirbackward();
203 char *skipspace();
204 char *scanstring();
205 int writeblanks();
206 int test1();
207 int convert1();
208
209 /* The main program */
210 int
211 main(argc, argv)
212 int argc;
213 char *argv[];
214 { FILE *in = stdin;
215 FILE *out = stdout;
216 char *filename = 0;
217 char *program_name = argv[0];
218 char *output_name = 0;
219 #define bufsize 5000 /* arbitrary size */
220 char *buf;
221 char *line;
222 char *more;
223 char *usage =
224 "Usage: ansi2knr [--filename FILENAME] [INPUT_FILE [OUTPUT_FILE]]\n";
225 /*
226 * In previous versions, ansi2knr recognized a --varargs switch.
227 * If this switch was supplied, ansi2knr would attempt to convert
228 * a ... argument to va_alist and va_dcl; if this switch was not
229 * supplied, ansi2knr would simply drop any such arguments.
230 * Now, ansi2knr always does this conversion, and we only
231 * check for this switch for backward compatibility.
232 */
233 int convert_varargs = 1;
234 int output_error;
235
236 while ( argc > 1 && argv[1][0] == '-' ) {
237 if ( !strcmp(argv[1], "--varargs") ) {
238 convert_varargs = 1;
239 argc--;
240 argv++;
241 continue;
242 }
243 if ( !strcmp(argv[1], "--filename") && argc > 2 ) {
244 filename = argv[2];
245 argc -= 2;
246 argv += 2;
247 continue;
248 }
249 fprintf(stderr, "%s: Unrecognized switch: %s\n", program_name,
250 argv[1]);
251 fprintf(stderr, usage);
252 exit(1);
253 }
254 switch ( argc )
255 {
256 default:
257 fprintf(stderr, usage);
258 exit(0);
259 case 3:
260 output_name = argv[2];
261 out = fopen(output_name, "w");
262 if ( out == NULL ) {
263 fprintf(stderr, "%s: Cannot open output file %s\n",
264 program_name, output_name);
265 exit(1);
266 }
267 /* falls through */
268 case 2:
269 in = fopen(argv[1], "r");
270 if ( in == NULL ) {
271 fprintf(stderr, "%s: Cannot open input file %s\n",
272 program_name, argv[1]);
273 exit(1);
274 }
275 if ( filename == 0 )
276 filename = argv[1];
277 /* falls through */
278 case 1:
279 break;
280 }
281 if ( filename )
282 fprintf(out, "#line 1 \"%s\"\n", filename);
283 buf = malloc(bufsize);
284 if ( buf == NULL )
285 {
286 fprintf(stderr, "Unable to allocate read buffer!\n");
287 exit(1);
288 }
289 line = buf;
290 while ( fgets(line, (unsigned)(buf + bufsize - line), in) != NULL )
291 {
292 test: line += strlen(line);
293 switch ( test1(buf) )
294 {
295 case 2: /* a function header */
296 convert1(buf, out, 1, convert_varargs);
297 break;
298 case 1: /* a function */
299 /* Check for a { at the start of the next line. */
300 more = ++line;
301 f: if ( line >= buf + (bufsize - 1) ) /* overflow check */
302 goto wl;
303 if ( fgets(line, (unsigned)(buf + bufsize - line), in) == NULL )
304 goto wl;
305 switch ( *skipspace(ppdirforward(more), 1) )
306 {
307 case '{':
308 /* Definitely a function header. */
309 convert1(buf, out, 0, convert_varargs);
310 fputs(more, out);
311 break;
312 case 0:
313 /* The next line was blank or a comment: */
314 /* keep scanning for a non-comment. */
315 line += strlen(line);
316 goto f;
317 default:
318 /* buf isn't a function header, but */
319 /* more might be. */
320 fputs(buf, out);
321 strcpy(buf, more);
322 line = buf;
323 goto test;
324 }
325 break;
326 case -1: /* maybe the start of a function */
327 if ( line != buf + (bufsize - 1) ) /* overflow check */
328 continue;
329 /* falls through */
330 default: /* not a function */
331 wl: fputs(buf, out);
332 break;
333 }
334 line = buf;
335 }
336 if ( line != buf )
337 fputs(buf, out);
338 free(buf);
339 if ( output_name ) {
340 output_error = ferror(out);
341 output_error |= fclose(out);
342 } else { /* out == stdout */
343 fflush(out);
344 output_error = ferror(out);
345 }
346 if ( output_error ) {
347 fprintf(stderr, "%s: error writing to %s\n", program_name,
348 (output_name ? output_name : "stdout"));
349 exit(1);
350 }
351 if ( in != stdin )
352 fclose(in);
353 return 0;
354 }
355
356 /*
357 * Skip forward or backward over one or more preprocessor directives.
358 */
359 char *
360 ppdirforward(p)
361 char *p;
362 {
363 for (; *p == '#'; ++p) {
364 for (; *p != '\r' && *p != '\n'; ++p)
365 if (*p == 0)
366 return p;
367 if (*p == '\r' && p[1] == '\n')
368 ++p;
369 }
370 return p;
371 }
372 char *
373 ppdirbackward(p, limit)
374 char *p;
375 char *limit;
376 {
377 char *np = p;
378
379 for (;; p = --np) {
380 if (*np == '\n' && np[-1] == '\r')
381 --np;
382 for (; np > limit && np[-1] != '\r' && np[-1] != '\n'; --np)
383 if (np[-1] == 0)
384 return np;
385 if (*np != '#')
386 return p;
387 }
388 }
389
390 /*
391 * Skip over whitespace, comments, and preprocessor directives,
392 * in either direction.
393 */
394 char *
395 skipspace(p, dir)
396 char *p;
397 int dir; /* 1 for forward, -1 for backward */
398 {
399 for ( ; ; ) {
400 while ( is_space(*p) )
401 p += dir;
402 if ( !(*p == '/' && p[dir] == '*') )
403 break;
404 p += dir; p += dir;
405 while ( !(*p == '*' && p[dir] == '/') ) {
406 if ( *p == 0 )
407 return p; /* multi-line comment?? */
408 p += dir;
409 }
410 p += dir; p += dir;
411 }
412 return p;
413 }
414
415 /* Scan over a quoted string, in either direction. */
416 char *
417 scanstring(p, dir)
418 char *p;
419 int dir;
420 {
421 for (p += dir; ; p += dir)
422 if (*p == '"' && p[-dir] != '\\')
423 return p + dir;
424 }
425
426 /*
427 * Write blanks over part of a string.
428 * Don't overwrite end-of-line characters.
429 */
430 int
431 writeblanks(start, end)
432 char *start;
433 char *end;
434 { char *p;
435 for ( p = start; p < end; p++ )
436 if ( *p != '\r' && *p != '\n' )
437 *p = ' ';
438 return 0;
439 }
440
441 /*
442 * Test whether the string in buf is a function definition.
443 * The string may contain and/or end with a newline.
444 * Return as follows:
445 * 0 - definitely not a function definition;
446 * 1 - definitely a function definition;
447 * 2 - definitely a function prototype (NOT USED);
448 * -1 - may be the beginning of a function definition,
449 * append another line and look again.
450 * The reason we don't attempt to convert function prototypes is that
451 * Ghostscript's declaration-generating macros look too much like
452 * prototypes, and confuse the algorithms.
453 */
454 int
455 test1(buf)
456 char *buf;
457 { char *p = buf;
458 char *bend;
459 char *endfn;
460 int contin;
461
462 if ( !isidfirstchar(*p) )
463 return 0; /* no name at left margin */
464 bend = skipspace(ppdirbackward(buf + strlen(buf) - 1, buf), -1);
465 switch ( *bend )
466 {
467 case ';': contin = 0 /*2*/; break;
468 case ')': contin = 1; break;
469 case '{': return 0; /* not a function */
470 case '}': return 0; /* not a function */
471 default: contin = -1;
472 }
473 while ( isidchar(*p) )
474 p++;
475 endfn = p;
476 p = skipspace(p, 1);
477 if ( *p++ != '(' )
478 return 0; /* not a function */
479 p = skipspace(p, 1);
480 if ( *p == ')' )
481 return 0; /* no parameters */
482 /* Check that the apparent function name isn't a keyword. */
483 /* We only need to check for keywords that could be followed */
484 /* by a left parenthesis (which, unfortunately, is most of them). */
485 { static char *words[] =
486 { "asm", "auto", "case", "char", "const", "double",
487 "extern", "float", "for", "if", "int", "long",
488 "register", "return", "short", "signed", "sizeof",
489 "static", "switch", "typedef", "unsigned",
490 "void", "volatile", "while", 0
491 };
492 char **key = words;
493 char *kp;
494 unsigned len = endfn - buf;
495
496 while ( (kp = *key) != 0 )
497 { if ( strlen(kp) == len && !strncmp(kp, buf, len) )
498 return 0; /* name is a keyword */
499 key++;
500 }
501 }
502 {
503 char *id = p;
504 int len;
505 /*
506 * Check for identifier1(identifier2) and not
507 * identifier1(void), or identifier1(identifier2, xxxx).
508 */
509
510 while ( isidchar(*p) )
511 p++;
512 len = p - id;
513 p = skipspace(p, 1);
514 if (*p == ',' ||
515 (*p == ')' && (len != 4 || strncmp(id, "void", 4)))
516 )
517 return 0; /* not a function */
518 }
519 /*
520 * If the last significant character was a ), we need to count
521 * parentheses, because it might be part of a formal parameter
522 * that is a procedure.
523 */
524 if (contin > 0) {
525 int level = 0;
526
527 for (p = skipspace(buf, 1); *p; p = skipspace(p + 1, 1))
528 level += (*p == '(' ? 1 : *p == ')' ? -1 : 0);
529 if (level > 0)
530 contin = -1;
531 }
532 return contin;
533 }
534
535 /* Convert a recognized function definition or header to K&R syntax. */
536 int
537 convert1(buf, out, header, convert_varargs)
538 char *buf;
539 FILE *out;
540 int header; /* Boolean */
541 int convert_varargs; /* Boolean */
542 { char *endfn;
543 char *p;
544 /*
545 * The breaks table contains pointers to the beginning and end
546 * of each argument.
547 */
548 char **breaks;
549 unsigned num_breaks = 2; /* for testing */
550 char **btop;
551 char **bp;
552 char **ap;
553 char *vararg = 0;
554
555 /* Pre-ANSI implementations don't agree on whether strchr */
556 /* is called strchr or index, so we open-code it here. */
557 for ( endfn = buf; *(endfn++) != '('; )
558 ;
559 top: p = endfn;
560 breaks = (char **)malloc(sizeof(char *) * num_breaks * 2);
561 if ( breaks == NULL )
562 { /* Couldn't allocate break table, give up */
563 fprintf(stderr, "Unable to allocate break table!\n");
564 fputs(buf, out);
565 return -1;
566 }
567 btop = breaks + num_breaks * 2 - 2;
568 bp = breaks;
569 /* Parse the argument list */
570 do
571 { int level = 0;
572 char *lp = NULL;
573 char *rp = NULL;
574 char *end = NULL;
575
576 if ( bp >= btop )
577 { /* Filled up break table. */
578 /* Allocate a bigger one and start over. */
579 free((char *)breaks);
580 num_breaks <<= 1;
581 goto top;
582 }
583 *bp++ = p;
584 /* Find the end of the argument */
585 for ( ; end == NULL; p++ )
586 { switch(*p)
587 {
588 case ',':
589 if ( !level ) end = p;
590 break;
591 case '(':
592 if ( !level ) lp = p;
593 level++;
594 break;
595 case ')':
596 if ( --level < 0 ) end = p;
597 else rp = p;
598 break;
599 case '/':
600 if (p[1] == '*')
601 p = skipspace(p, 1) - 1;
602 break;
603 case '"':
604 p = scanstring(p, 1) - 1;
605 break;
606 default:
607 ;
608 }
609 }
610 /* Erase any embedded prototype parameters. */
611 if ( lp && rp )
612 writeblanks(lp + 1, rp);
613 p--; /* back up over terminator */
614 /* Find the name being declared. */
615 /* This is complicated because of procedure and */
616 /* array modifiers. */
617 for ( ; ; )
618 { p = skipspace(p - 1, -1);
619 switch ( *p )
620 {
621 case ']': /* skip array dimension(s) */
622 case ')': /* skip procedure args OR name */
623 { int level = 1;
624 while ( level )
625 switch ( *--p )
626 {
627 case ']': case ')':
628 level++;
629 break;
630 case '[': case '(':
631 level--;
632 break;
633 case '/':
634 if (p > buf && p[-1] == '*')
635 p = skipspace(p, -1) + 1;
636 break;
637 case '"':
638 p = scanstring(p, -1) + 1;
639 break;
640 default: ;
641 }
642 }
643 if ( *p == '(' && *skipspace(p + 1, 1) == '*' )
644 { /* We found the name being declared */
645 while ( !isidfirstchar(*p) )
646 p = skipspace(p, 1) + 1;
647 goto found;
648 }
649 break;
650 default:
651 goto found;
652 }
653 }
654 found: if ( *p == '.' && p[-1] == '.' && p[-2] == '.' )
655 { if ( convert_varargs )
656 { *bp++ = "va_alist";
657 vararg = p-2;
658 }
659 else
660 { p++;
661 if ( bp == breaks + 1 ) /* sole argument */
662 writeblanks(breaks[0], p);
663 else
664 writeblanks(bp[-1] - 1, p);
665 bp--;
666 }
667 }
668 else
669 { while ( isidchar(*p) ) p--;
670 *bp++ = p+1;
671 }
672 p = end;
673 }
674 while ( *p++ == ',' );
675 *bp = p;
676 /* Make a special check for 'void' arglist */
677 if ( bp == breaks+2 )
678 { p = skipspace(breaks[0], 1);
679 if ( !strncmp(p, "void", 4) )
680 { p = skipspace(p+4, 1);
681 if ( p == breaks[2] - 1 )
682 { bp = breaks; /* yup, pretend arglist is empty */
683 writeblanks(breaks[0], p + 1);
684 }
685 }
686 }
687 /* Put out the function name and left parenthesis. */
688 p = buf;
689 while ( p != endfn ) putc(*p, out), p++;
690 /* Put out the declaration. */
691 if ( header )
692 { fputs(");", out);
693 for ( p = breaks[0]; *p; p++ )
694 if ( *p == '\r' || *p == '\n' )
695 putc(*p, out);
696 }
697 else
698 { for ( ap = breaks+1; ap < bp; ap += 2 )
699 { p = *ap;
700 while ( isidchar(*p) )
701 putc(*p, out), p++;
702 if ( ap < bp - 1 )
703 fputs(", ", out);
704 }
705 fputs(") ", out);
706 /* Put out the argument declarations */
707 for ( ap = breaks+2; ap <= bp; ap += 2 )
708 (*ap)[-1] = ';';
709 if ( vararg != 0 )
710 { *vararg = 0;
711 fputs(breaks[0], out); /* any prior args */
712 fputs("va_dcl", out); /* the final arg */
713 fputs(bp[0], out);
714 }
715 else
716 fputs(breaks[0], out);
717 }
718 free((char *)breaks);
719 return 0;
720 }
This page took 0.068324 seconds and 4 git commands to generate.