src/ansi2knr.c

   1 /* Copyright (C) 1989, 1991, 1993, 1994, 1995 Aladdin Enterprises. All rights reserved. */
   2
   3 /* 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 Among other things, the copyright notice and this notice must be preserved
  18 on all copies.
  19
  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.
  27 */
  28
  29 /*
  30  * Usage:
  31         ansi2knr input_file [output_file]
  32  * If no output_file is supplied, output goes to stdout.
  33  * There are no error messages.
  34  *
  35  * ansi2knr recognizes function definitions by seeing a non-keyword
  36  * identifier at the left margin, followed by a left parenthesis,
  37  * with a right parenthesis as the last character on the line.
  38  * It will recognize a multi-line header provided that the last character
  39  * of the last line of the header is a right parenthesis,
  40  * and no intervening line ends with a left or right brace or a semicolon.
  41  * These algorithms ignore whitespace and comments, except that
  42  * the function name must be the first thing on the line.
  43  * The following constructs will confuse it:
  44  *      - Any other construct that starts at the left margin and
  45  *          follows the above syntax (such as a macro or function call).
  46  *      - Macros that tinker with the syntax of the function header.
  47  */
  48
  49 /*
  50  * The original and principal author of ansi2knr is L. Peter Deutsch
  51  * <ghost@aladdin.com>.  Other authors are noted in the change history
  52  * that follows (in reverse chronological order):
  53         lpd 95-06-22 removed #ifndefs whose sole purpose was to define
  54                 undefined preprocessor symbols as 0; changed all #ifdefs
  55                 for configuration symbols to #ifs
  56         lpd 95-04-05 changed copyright notice to make it clear that
  57                 including ansi2knr in a program does not bring the entire
  58                 program under the GPL
  59         lpd 94-12-18 added conditionals for systems where ctype macros
  60                 don't handle 8-bit characters properly, suggested by
  61                 Francois Pinard <pinard@iro.umontreal.ca>;
  62                 removed --varargs switch (this is now the default)
  63         lpd 94-10-10 removed CONFIG_BROKETS conditional
  64         lpd 94-07-16 added some conditionals to help GNU `configure',
  65                 suggested by Francois Pinard <pinard@iro.umontreal.ca>;
  66                 properly erase prototype args in function parameters,
  67                 contributed by Jim Avera <jima@netcom.com>;
  68                 correct error in writeblanks (it shouldn't erase EOLs)
  69         lpd 89-xx-xx original version
  70  */
  71
  72 /* Most of the conditionals here are to make ansi2knr work with */
  73 /* the GNU configure machinery. */
  74
  75 #if HAVE_CONFIG_H
  76 # include <config.h>
  77 #endif
  78
  79 #include <stdio.h>
  80 #include <ctype.h>
  81
  82 #if HAVE_CONFIG_H
  83
  84 /*
  85    For properly autoconfiguring ansi2knr, use AC_CONFIG_HEADER(config.h).
  86    This will define HAVE_CONFIG_H and so, activate the following lines.
  87  */
  88
  89 # if STDC_HEADERS || HAVE_STRING_H
  90 #  include <string.h>
  91 # else
  92 #  include <strings.h>
  93 # endif
  94
  95 #else /* not HAVE_CONFIG_H */
  96
  97 /*
  98    Without AC_CONFIG_HEADER, merely use <string.h> as in the original
  99    Ghostscript distribution.  This loses on older BSD systems.
 100  */
 101
 102 # include <string.h>
 103
 104 #endif /* not HAVE_CONFIG_H */
 105
 106 #if STDC_HEADERS
 107 # include <stdlib.h>
 108 #else
 109 /*
 110    malloc and free should be declared in stdlib.h,
 111    but if you've got a K&R compiler, they probably aren't.
 112  */
 113 char *malloc();
 114 void free();
 115 #endif
 116
 117 /*
 118  * The ctype macros don't always handle 8-bit characters correctly.
 119  * Compensate for this here.
 120  */
 121 #ifdef isascii
 122 #  undef HAVE_ISASCII           /* just in case */
 123 #  define HAVE_ISASCII 1
 124 #else
 125 #endif
 126 #if STDC_HEADERS || !HAVE_ISASCII
 127 #  define is_ascii(c) 1
 128 #else
 129 #  define is_ascii(c) isascii(c)
 130 #endif
 131
 132 #define is_space(c) (is_ascii(c) && isspace(c))
 133 #define is_alpha(c) (is_ascii(c) && isalpha(c))
 134 #define is_alnum(c) (is_ascii(c) && isalnum(c))
 135
 136 /* Scanning macros */
 137 #define isidchar(ch) (is_alnum(ch) || (ch) == '_')
 138 #define isidfirstchar(ch) (is_alpha(ch) || (ch) == '_')
 139
 140 /* Forward references */
 141 char *skipspace();
 142 void writeblanks();
 143 int test1();
 144 int convert1();
 145
 146 /* The main program */
 147 int
 148 main(argc, argv)
 149     int argc;
 150     char *argv[];
 151 {       FILE *in, *out;
 152 #define bufsize 5000                    /* arbitrary size */
 153         char *buf;
 154         char *line;
 155         /*
 156          * In previous versions, ansi2knr recognized a --varargs switch.
 157          * If this switch was supplied, ansi2knr would attempt to convert
 158          * a ... argument to va_alist and va_dcl; if this switch was not
 159          * supplied, ansi2knr would simply drop any such arguments.
 160          * Now, ansi2knr always does this conversion, and we only
 161          * check for this switch for backward compatibility.
 162          */
 163         int convert_varargs = 1;
 164
 165         if ( argc > 1 && argv[1][0] == '-' )
 166           {     if ( !strcmp(argv[1], "--varargs") )
 167                   {     convert_varargs = 1;
 168                         argc--;
 169                         argv++;
 170                   }
 171                 else
 172                   {     fprintf(stderr, "Unrecognized switch: %s\n", argv[1]);
 173                         exit(1);
 174                   }
 175           }
 176         switch ( argc )
 177            {
 178         default:
 179                 printf("Usage: ansi2knr input_file [output_file]\n");
 180                 exit(0);
 181         case 2:
 182                 out = stdout;
 183                 break;
 184         case 3:
 185                 out = fopen(argv[2], "w");
 186                 if ( out == NULL )
 187                    {    fprintf(stderr, "Cannot open output file %s\n", argv[2]);
 188                         exit(1);
 189                    }
 190            }
 191         in = fopen(argv[1], "r");
 192         if ( in == NULL )
 193            {    fprintf(stderr, "Cannot open input file %s\n", argv[1]);
 194                 exit(1);
 195            }
 196         fprintf(out, "#line 1 \"%s\"\n", argv[1]);
 197         buf = malloc(bufsize);
 198         line = buf;
 199         while ( fgets(line, (unsigned)(buf + bufsize - line), in) != NULL )
 200            {    switch ( test1(buf) )
 201                    {
 202                 case 2:                 /* a function header */
 203                         convert1(buf, out, 1, convert_varargs);
 204                         break;
 205                 case 1:                 /* a function */
 206                         convert1(buf, out, 0, convert_varargs);
 207                         break;
 208                 case -1:                /* maybe the start of a function */
 209                         line = buf + strlen(buf);
 210                         if ( line != buf + (bufsize - 1) ) /* overflow check */
 211                                 continue;
 212                         /* falls through */
 213                 default:                /* not a function */
 214                         fputs(buf, out);
 215                         break;
 216                    }
 217                 line = buf;
 218            }
 219         if ( line != buf ) fputs(buf, out);
 220         free(buf);
 221         fclose(out);
 222         fclose(in);
 223         return 0;
 224 }
 225
 226 /* Skip over space and comments, in either direction. */
 227 char *
 228 skipspace(p, dir)
 229     register char *p;
 230     register int dir;                   /* 1 for forward, -1 for backward */
 231 {       for ( ; ; )
 232            {    while ( is_space(*p) ) p += dir;
 233                 if ( !(*p == '/' && p[dir] == '*') ) break;
 234                 p += dir;  p += dir;
 235                 while ( !(*p == '*' && p[dir] == '/') )
 236                    {    if ( *p == 0 ) return p;        /* multi-line comment?? */
 237                         p += dir;
 238                    }
 239                 p += dir;  p += dir;
 240            }
 241         return p;
 242 }
 243
 244 /*
 245  * Write blanks over part of a string.
 246  * Don't overwrite end-of-line characters.
 247  */
 248 void
 249 writeblanks(start, end)
 250     char *start;
 251     char *end;
 252 {       char *p;
 253         for ( p = start; p < end; p++ )
 254           if ( *p != '\r' && *p != '\n' ) *p = ' ';
 255 }
 256
 257 /*
 258  * Test whether the string in buf is a function definition.
 259  * The string may contain and/or end with a newline.
 260  * Return as follows:
 261  *      0 - definitely not a function definition;
 262  *      1 - definitely a function definition;
 263  *      2 - definitely a function prototype (NOT USED);
 264  *      -1 - may be the beginning of a function definition,
 265  *              append another line and look again.
 266  * The reason we don't attempt to convert function prototypes is that
 267  * Ghostscript's declaration-generating macros look too much like
 268  * prototypes, and confuse the algorithms.
 269  */
 270 int
 271 test1(buf)
 272     char *buf;
 273 {       register char *p = buf;
 274         char *bend;
 275         char *endfn;
 276         int contin;
 277         if ( !isidfirstchar(*p) )
 278                 return 0;               /* no name at left margin */
 279         bend = skipspace(buf + strlen(buf) - 1, -1);
 280         switch ( *bend )
 281            {
 282         case ';': contin = 0 /*2*/; break;
 283         case ')': contin = 1; break;
 284         case '{': return 0;             /* not a function */
 285         case '}': return 0;             /* not a function */
 286         default: contin = -1;
 287            }
 288         while ( isidchar(*p) ) p++;
 289         endfn = p;
 290         p = skipspace(p, 1);
 291         if ( *p++ != '(' )
 292                 return 0;               /* not a function */
 293         p = skipspace(p, 1);
 294         if ( *p == ')' )
 295                 return 0;               /* no parameters */
 296         /* Check that the apparent function name isn't a keyword. */
 297         /* We only need to check for keywords that could be followed */
 298         /* by a left parenthesis (which, unfortunately, is most of them). */
 299            {    static char *words[] =
 300                    {    "asm", "auto", "case", "char", "const", "double",
 301                         "extern", "float", "for", "if", "int", "long",
 302                         "register", "return", "short", "signed", "sizeof",
 303                         "static", "switch", "typedef", "unsigned",
 304                         "void", "volatile", "while", 0
 305                    };
 306                 char **key = words;
 307                 char *kp;
 308                 int len = endfn - buf;
 309                 while ( (kp = *key) != 0 )
 310                    {    if ( strlen(kp) == len && !strncmp(kp, buf, len) )
 311                                 return 0;       /* name is a keyword */
 312                         key++;
 313                    }
 314            }
 315         return contin;
 316 }
 317
 318 /* Convert a recognized function definition or header to K&R syntax. */
 319 int
 320 convert1(buf, out, header, convert_varargs)
 321     char *buf;
 322     FILE *out;
 323     int header;                 /* Boolean */
 324     int convert_varargs;        /* Boolean */
 325 {       char *endfn;
 326         register char *p;
 327         char **breaks;
 328         unsigned num_breaks = 2;        /* for testing */
 329         char **btop;
 330         char **bp;
 331         char **ap;
 332         char *vararg = 0;
 333         /* Pre-ANSI implementations don't agree on whether strchr */
 334         /* is called strchr or index, so we open-code it here. */
 335         for ( endfn = buf; *(endfn++) != '('; ) ;
 336 top:    p = endfn;
 337         breaks = (char **)malloc(sizeof(char *) * num_breaks * 2);
 338         if ( breaks == 0 )
 339            {    /* Couldn't allocate break table, give up */
 340                 fprintf(stderr, "Unable to allocate break table!\n");
 341                 fputs(buf, out);
 342                 return -1;
 343            }
 344         btop = breaks + num_breaks * 2 - 2;
 345         bp = breaks;
 346         /* Parse the argument list */
 347         do
 348            {    int level = 0;
 349                 char *lp = NULL;
 350                 char *rp;
 351                 char *end = NULL;
 352                 if ( bp >= btop )
 353                    {    /* Filled up break table. */
 354                         /* Allocate a bigger one and start over. */
 355                         free((char *)breaks);
 356                         num_breaks <<= 1;
 357                         goto top;
 358                    }
 359                 *bp++ = p;
 360                 /* Find the end of the argument */
 361                 for ( ; end == NULL; p++ )
 362                    {    switch(*p)
 363                            {
 364                         case ',':
 365                                 if ( !level ) end = p;
 366                                 break;
 367                         case '(':
 368                                 if ( !level ) lp = p;
 369                                 level++;
 370                                 break;
 371                         case ')':
 372                                 if ( --level < 0 ) end = p;
 373                                 else rp = p;
 374                                 break;
 375                         case '/':
 376                                 p = skipspace(p, 1) - 1;
 377                                 break;
 378                         default:
 379                                 ;
 380                            }
 381                    }
 382                 /* Erase any embedded prototype parameters. */
 383                 if ( lp )
 384                   writeblanks(lp + 1, rp);
 385                 p--;                    /* back up over terminator */
 386                 /* Find the name being declared. */
 387                 /* This is complicated because of procedure and */
 388                 /* array modifiers. */
 389                 for ( ; ; )
 390                    {    p = skipspace(p - 1, -1);
 391                         switch ( *p )
 392                            {
 393                         case ']':       /* skip array dimension(s) */
 394                         case ')':       /* skip procedure args OR name */
 395                            {    int level = 1;
 396                                 while ( level )
 397                                  switch ( *--p )
 398                                    {
 399                                 case ']': case ')': level++; break;
 400                                 case '[': case '(': level--; break;
 401                                 case '/': p = skipspace(p, -1) + 1; break;
 402                                 default: ;
 403                                    }
 404                            }
 405                                 if ( *p == '(' && *skipspace(p + 1, 1) == '*' )
 406                                    {    /* We found the name being declared */
 407                                         while ( !isidfirstchar(*p) )
 408                                                 p = skipspace(p, 1) + 1;
 409                                         goto found;
 410                                    }
 411                                 break;
 412                         default: goto found;
 413                            }
 414                    }
 415 found:          if ( *p == '.' && p[-1] == '.' && p[-2] == '.' )
 416                   {     if ( convert_varargs )
 417                           {     *bp++ = "va_alist";
 418                                 vararg = p-2;
 419                           }
 420                         else
 421                           {     p++;
 422                                 if ( bp == breaks + 1 ) /* sole argument */
 423                                   writeblanks(breaks[0], p);
 424                                 else
 425                                   writeblanks(bp[-1] - 1, p);
 426                                 bp--;
 427                           }
 428                    }
 429                 else
 430                    {    while ( isidchar(*p) ) p--;
 431                         *bp++ = p+1;
 432                    }
 433                 p = end;
 434            }
 435         while ( *p++ == ',' );
 436         *bp = p;
 437         /* Make a special check for 'void' arglist */
 438         if ( bp == breaks+2 )
 439            {    p = skipspace(breaks[0], 1);
 440                 if ( !strncmp(p, "void", 4) )
 441                    {    p = skipspace(p+4, 1);
 442                         if ( p == breaks[2] - 1 )
 443                            {    bp = breaks;    /* yup, pretend arglist is empty */
 444                                 writeblanks(breaks[0], p + 1);
 445                            }
 446                    }
 447            }
 448         /* Put out the function name and left parenthesis. */
 449         p = buf;
 450         while ( p != endfn ) putc(*p, out), p++;
 451         /* Put out the declaration. */
 452         if ( header )
 453           {     fputs(");", out);
 454                 for ( p = breaks[0]; *p; p++ )
 455                   if ( *p == '\r' || *p == '\n' )
 456                     putc(*p, out);
 457           }
 458         else
 459           {     for ( ap = breaks+1; ap < bp; ap += 2 )
 460                   {     p = *ap;
 461                         while ( isidchar(*p) )
 462                           putc(*p, out), p++;
 463                         if ( ap < bp - 1 )
 464                           fputs(", ", out);
 465                   }
 466                 fputs(")  ", out);
 467                 /* Put out the argument declarations */
 468                 for ( ap = breaks+2; ap <= bp; ap += 2 )
 469                   (*ap)[-1] = ';';
 470                 if ( vararg != 0 )
 471                   {     *vararg = 0;
 472                         fputs(breaks[0], out);          /* any prior args */
 473                         fputs("va_dcl", out);           /* the final arg */
 474                         fputs(bp[0], out);
 475                   }
 476                 else
 477                   fputs(breaks[0], out);
 478           }
 479         free((char *)breaks);
 480         return 0;
 481 }