1 /* Copyright (C) 1989, 2000 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, with a
40 * right parenthesis as the last character on the line, and with a left
41 * brace as the first token on the following line (ignoring possible
42 * intervening comments and/or preprocessor directives), 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 that no intervening
50 * line ends with a left or right brace or a semicolon. These algorithms
51 * ignore whitespace, comments, and preprocessor directives, except that
52 * the function name must be the first thing on the line. The following
53 * 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):
64 lpd 2000-04-12 backs out Eggert's changes because of bugs:
65 - concatlits didn't declare the type of its bufend argument;
66 - concatlits didn't't recognize when it was inside a comment;
67 - scanstring could scan backward past the beginning of the string; when
68 - the check for \ + newline in scanstring was unnecessary.
70 2000-03-05 Paul Eggert <eggert@twinsun.com>
72 Add support for concatenated string literals.
73 * ansi2knr.c (concatlits): New decl.
74 (main): Invoke concatlits to concatenate string literals.
75 (scanstring): Handle backslash-newline correctly. Work with
76 character constants. Fix bug when scanning backwards through
77 backslash-quote. Check for unterminated strings.
78 (convert1): Parse character constants, too.
79 (appendline, concatlits): New functions.
80 * ansi2knr.1: Document this.
82 lpd 1999-08-17 added code to allow preprocessor directives
83 wherever comments are allowed
84 lpd 1999-04-12 added minor fixes from Pavel Roskin
85 <pavel_roskin@geocities.com> for clean compilation with
87 lpd 1999-03-22 added hack to recognize lines consisting of
88 identifier1(identifier2, xxx) as *not* being procedures
89 lpd 1999-02-03 made indentation of preprocessor commands consistent
90 lpd 1999-01-28 fixed two bugs: a '/' in an argument list caused an
91 endless loop; quoted strings within an argument list
93 lpd 1999-01-24 added a check for write errors on the output,
94 suggested by Jim Meyering <meyering@ascend.com>
95 lpd 1998-11-09 added further hack to recognize identifier(void)
97 lpd 1998-10-23 added hack to recognize lines consisting of
98 identifier1(identifier2) as *not* being procedures
99 lpd 1997-12-08 made input_file optional; only closes input and/or
100 output file if not stdin or stdout respectively; prints
101 usage message on stderr rather than stdout; adds
102 --filename switch (changes suggested by
103 <ceder@lysator.liu.se>)
104 lpd 1996-01-21 added code to cope with not HAVE_CONFIG_H and with
105 compilers that don't understand void, as suggested by
107 lpd 1996-01-15 changed to require that the first non-comment token
108 on the line following a function header be a left brace,
109 to reduce sensitivity to macros, as suggested by Tom Lane
111 lpd 1995-06-22 removed #ifndefs whose sole purpose was to define
112 undefined preprocessor symbols as 0; changed all #ifdefs
113 for configuration symbols to #ifs
114 lpd 1995-04-05 changed copyright notice to make it clear that
115 including ansi2knr in a program does not bring the entire
116 program under the GPL
117 lpd 1994-12-18 added conditionals for systems where ctype macros
118 don't handle 8-bit characters properly, suggested by
119 Francois Pinard <pinard@iro.umontreal.ca>;
120 removed --varargs switch (this is now the default)
121 lpd 1994-10-10 removed CONFIG_BROKETS conditional
122 lpd 1994-07-16 added some conditionals to help GNU `configure',
123 suggested by Francois Pinard <pinard@iro.umontreal.ca>;
124 properly erase prototype args in function parameters,
125 contributed by Jim Avera <jima@netcom.com>;
126 correct error in writeblanks (it shouldn't erase EOLs)
127 lpd 1989-xx-xx original version
130 /* Most of the conditionals here are to make ansi2knr work with */
131 /* or without the GNU configure machinery. */
143 For properly autoconfiguring ansi2knr, use AC_CONFIG_HEADER(config.h).
144 This will define HAVE_CONFIG_H and so, activate the following lines.
147 # if STDC_HEADERS || HAVE_STRING_H
150 # include <strings.h>
153 #else /* not HAVE_CONFIG_H */
155 /* Otherwise do it the hard way */
158 # include <strings.h>
161 extern int strlen(), strncmp();
167 #endif /* not HAVE_CONFIG_H */
173 malloc and free should be declared in stdlib.h,
174 but if you've got a K&R compiler, they probably aren't.
180 extern char *malloc();
183 extern char *malloc();
190 /* Define NULL (for *very* old compilers). */
196 * The ctype macros don't always handle 8-bit characters correctly.
197 * Compensate for this here.
200 # undef HAVE_ISASCII /* just in case */
201 # define HAVE_ISASCII 1
204 #if STDC_HEADERS || !HAVE_ISASCII
205 # define is_ascii(c) 1
207 # define is_ascii(c) isascii(c)
210 #define is_space(c) (is_ascii(c) && isspace(c))
211 #define is_alpha(c) (is_ascii(c) && isalpha(c))
212 #define is_alnum(c) (is_ascii(c) && isalnum(c))
214 /* Scanning macros */
215 #define isidchar(ch) (is_alnum(ch) || (ch) == '_')
216 #define isidfirstchar(ch) (is_alpha(ch) || (ch) == '_')
218 /* Forward references */
219 char *ppdirforward();
220 char *ppdirbackward();
227 /* The main program */
235 char *program_name = argv[0];
236 char *output_name = 0;
237 #define bufsize 5000 /* arbitrary size */
242 "Usage: ansi2knr [--filename FILENAME] [INPUT_FILE [OUTPUT_FILE]]\n";
244 * In previous versions, ansi2knr recognized a --varargs switch.
245 * If this switch was supplied, ansi2knr would attempt to convert
246 * a ... argument to va_alist and va_dcl; if this switch was not
247 * supplied, ansi2knr would simply drop any such arguments.
248 * Now, ansi2knr always does this conversion, and we only
249 * check for this switch for backward compatibility.
251 int convert_varargs = 1;
254 while ( argc > 1 && argv[1][0] == '-' ) {
255 if ( !strcmp(argv[1], "--varargs") ) {
261 if ( !strcmp(argv[1], "--filename") && argc > 2 ) {
267 fprintf(stderr, "%s: Unrecognized switch: %s\n", program_name,
269 fprintf(stderr, usage);
275 fprintf(stderr, usage);
278 output_name = argv[2];
279 out = fopen(output_name, "w");
281 fprintf(stderr, "%s: Cannot open output file %s\n",
282 program_name, output_name);
287 in = fopen(argv[1], "r");
289 fprintf(stderr, "%s: Cannot open input file %s\n",
290 program_name, argv[1]);
300 fprintf(out, "#line 1 \"%s\"\n", filename);
301 buf = malloc(bufsize);
304 fprintf(stderr, "Unable to allocate read buffer!\n");
308 while ( fgets(line, (unsigned)(buf + bufsize - line), in) != NULL )
310 test: line += strlen(line);
311 switch ( test1(buf) )
313 case 2: /* a function header */
314 convert1(buf, out, 1, convert_varargs);
316 case 1: /* a function */
317 /* Check for a { at the start of the next line. */
319 f: if ( line >= buf + (bufsize - 1) ) /* overflow check */
321 if ( fgets(line, (unsigned)(buf + bufsize - line), in) == NULL )
323 switch ( *skipspace(ppdirforward(more), 1) )
326 /* Definitely a function header. */
327 convert1(buf, out, 0, convert_varargs);
331 /* The next line was blank or a comment: */
332 /* keep scanning for a non-comment. */
333 line += strlen(line);
336 /* buf isn't a function header, but */
344 case -1: /* maybe the start of a function */
345 if ( line != buf + (bufsize - 1) ) /* overflow check */
348 default: /* not a function */
358 output_error = ferror(out);
359 output_error |= fclose(out);
360 } else { /* out == stdout */
362 output_error = ferror(out);
364 if ( output_error ) {
365 fprintf(stderr, "%s: error writing to %s\n", program_name,
366 (output_name ? output_name : "stdout"));
375 * Skip forward or backward over one or more preprocessor directives.
381 for (; *p == '#'; ++p) {
382 for (; *p != '\r' && *p != '\n'; ++p)
385 if (*p == '\r' && p[1] == '\n')
391 ppdirbackward(p, limit)
398 if (*np == '\n' && np[-1] == '\r')
400 for (; np > limit && np[-1] != '\r' && np[-1] != '\n'; --np)
409 * Skip over whitespace, comments, and preprocessor directives,
410 * in either direction.
415 int dir; /* 1 for forward, -1 for backward */
418 while ( is_space(*p) )
420 if ( !(*p == '/' && p[dir] == '*') )
423 while ( !(*p == '*' && p[dir] == '/') ) {
425 return p; /* multi-line comment?? */
433 /* Scan over a quoted string, in either direction. */
439 for (p += dir; ; p += dir)
440 if (*p == '"' && p[-dir] != '\\')
445 * Write blanks over part of a string.
446 * Don't overwrite end-of-line characters.
449 writeblanks(start, end)
453 for ( p = start; p < end; p++ )
454 if ( *p != '\r' && *p != '\n' )
460 * Test whether the string in buf is a function definition.
461 * The string may contain and/or end with a newline.
463 * 0 - definitely not a function definition;
464 * 1 - definitely a function definition;
465 * 2 - definitely a function prototype (NOT USED);
466 * -1 - may be the beginning of a function definition,
467 * append another line and look again.
468 * The reason we don't attempt to convert function prototypes is that
469 * Ghostscript's declaration-generating macros look too much like
470 * prototypes, and confuse the algorithms.
480 if ( !isidfirstchar(*p) )
481 return 0; /* no name at left margin */
482 bend = skipspace(ppdirbackward(buf + strlen(buf) - 1, buf), -1);
485 case ';': contin = 0 /*2*/; break;
486 case ')': contin = 1; break;
487 case '{': return 0; /* not a function */
488 case '}': return 0; /* not a function */
489 default: contin = -1;
491 while ( isidchar(*p) )
496 return 0; /* not a function */
499 return 0; /* no parameters */
500 /* Check that the apparent function name isn't a keyword. */
501 /* We only need to check for keywords that could be followed */
502 /* by a left parenthesis (which, unfortunately, is most of them). */
503 { static char *words[] =
504 { "asm", "auto", "case", "char", "const", "double",
505 "extern", "float", "for", "if", "int", "long",
506 "register", "return", "short", "signed", "sizeof",
507 "static", "switch", "typedef", "unsigned",
508 "void", "volatile", "while", 0
512 unsigned len = endfn - buf;
514 while ( (kp = *key) != 0 )
515 { if ( strlen(kp) == len && !strncmp(kp, buf, len) )
516 return 0; /* name is a keyword */
524 * Check for identifier1(identifier2) and not
525 * identifier1(void), or identifier1(identifier2, xxxx).
528 while ( isidchar(*p) )
533 (*p == ')' && (len != 4 || strncmp(id, "void", 4)))
535 return 0; /* not a function */
538 * If the last significant character was a ), we need to count
539 * parentheses, because it might be part of a formal parameter
540 * that is a procedure.
545 for (p = skipspace(buf, 1); *p; p = skipspace(p + 1, 1))
546 level += (*p == '(' ? 1 : *p == ')' ? -1 : 0);
553 /* Convert a recognized function definition or header to K&R syntax. */
555 convert1(buf, out, header, convert_varargs)
558 int header; /* Boolean */
559 int convert_varargs; /* Boolean */
563 * The breaks table contains pointers to the beginning and end
567 unsigned num_breaks = 2; /* for testing */
573 /* Pre-ANSI implementations don't agree on whether strchr */
574 /* is called strchr or index, so we open-code it here. */
575 for ( endfn = buf; *(endfn++) != '('; )
578 breaks = (char **)malloc(sizeof(char *) * num_breaks * 2);
579 if ( breaks == NULL )
580 { /* Couldn't allocate break table, give up */
581 fprintf(stderr, "Unable to allocate break table!\n");
585 btop = breaks + num_breaks * 2 - 2;
587 /* Parse the argument list */
595 { /* Filled up break table. */
596 /* Allocate a bigger one and start over. */
597 free((char *)breaks);
602 /* Find the end of the argument */
603 for ( ; end == NULL; p++ )
607 if ( !level ) end = p;
610 if ( !level ) lp = p;
614 if ( --level < 0 ) end = p;
619 p = skipspace(p, 1) - 1;
622 p = scanstring(p, 1) - 1;
628 /* Erase any embedded prototype parameters. */
630 writeblanks(lp + 1, rp);
631 p--; /* back up over terminator */
632 /* Find the name being declared. */
633 /* This is complicated because of procedure and */
634 /* array modifiers. */
636 { p = skipspace(p - 1, -1);
639 case ']': /* skip array dimension(s) */
640 case ')': /* skip procedure args OR name */
652 if (p > buf && p[-1] == '*')
653 p = skipspace(p, -1) + 1;
656 p = scanstring(p, -1) + 1;
661 if ( *p == '(' && *skipspace(p + 1, 1) == '*' )
662 { /* We found the name being declared */
663 while ( !isidfirstchar(*p) )
664 p = skipspace(p, 1) + 1;
672 found: if ( *p == '.' && p[-1] == '.' && p[-2] == '.' )
673 { if ( convert_varargs )
674 { *bp++ = "va_alist";
679 if ( bp == breaks + 1 ) /* sole argument */
680 writeblanks(breaks[0], p);
682 writeblanks(bp[-1] - 1, p);
687 { while ( isidchar(*p) ) p--;
692 while ( *p++ == ',' );
694 /* Make a special check for 'void' arglist */
695 if ( bp == breaks+2 )
696 { p = skipspace(breaks[0], 1);
697 if ( !strncmp(p, "void", 4) )
698 { p = skipspace(p+4, 1);
699 if ( p == breaks[2] - 1 )
700 { bp = breaks; /* yup, pretend arglist is empty */
701 writeblanks(breaks[0], p + 1);
705 /* Put out the function name and left parenthesis. */
707 while ( p != endfn ) putc(*p, out), p++;
708 /* Put out the declaration. */
711 for ( p = breaks[0]; *p; p++ )
712 if ( *p == '\r' || *p == '\n' )
716 { for ( ap = breaks+1; ap < bp; ap += 2 )
718 while ( isidchar(*p) )
724 /* Put out the argument declarations */
725 for ( ap = breaks+2; ap <= bp; ap += 2 )
729 fputs(breaks[0], out); /* any prior args */
730 fputs("va_dcl", out); /* the final arg */
734 fputs(breaks[0], out);
736 free((char *)breaks);