mksyntax.c

   1 /*
   2  * mksyntax.c - construct shell syntax table for fast char attribute lookup.
   3  */
   4
   5 /* Copyright (C) 2000 Free Software Foundation, Inc.
   6
   7    This file is part of GNU Bash, the Bourne Again SHell.
   8
   9    Bash is free software; you can redistribute it and/or modify it under
  10    the terms of the GNU General Public License as published by the Free
  11    Software Foundation; either version 2, or (at your option) any later
  12    version.
  13
  14    Bash is distributed in the hope that it will be useful, but WITHOUT ANY
  15    WARRANTY; without even the implied warranty of MERCHANTABILITY or
  16    FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License
  17    for more details.
  18
  19    You should have received a copy of the GNU General Public License along
  20    with Bash; see the file COPYING.  If not, write to the Free Software
  21    Foundation, 59 Temple Place, Suite 330, Boston, MA 02111 USA. */
  22
  23 #include "config.h"
  24
  25 #include <stdio.h>
  26 #include "bashansi.h"
  27 #include "chartypes.h"
  28 #include <errno.h>
  29
  30 #ifdef HAVE_UNISTD_H
  31 #  include <unistd.h>
  32 #endif
  33
  34 #include "syntax.h"
  35
  36 extern int optind;
  37 extern char *optarg;
  38
  39 #ifndef errno
  40 extern int errno;
  41 #endif
  42
  43 #ifndef HAVE_STRERROR
  44 extern char *strerror();
  45 #endif
  46
  47 struct wordflag {
  48         int     flag;
  49         char    *fstr;
  50 } wordflags[] = {
  51         { CWORD,        "CWORD" },
  52         { CSHMETA,      "CSHMETA" },
  53         { CSHBRK,       "CSHBRK" },
  54         { CBACKQ,       "CBACKQ" },
  55         { CQUOTE,       "CQUOTE" },
  56         { CSPECL,       "CSPECL" },
  57         { CEXP,         "CEXP" },
  58         { CBSDQUOTE,    "CBSDQUOTE" },
  59         { CBSHDOC,      "CBSHDOC" },
  60         { CGLOB,        "CGLOB" },
  61         { CXGLOB,       "CXGLOB" },
  62         { CXQUOTE,      "CXQUOTE" },
  63         { CSPECVAR,     "CSPECVAR" },
  64         { CSUBSTOP,     "CSUBSTOP" },
  65 };
  66
  67 #define N_WFLAGS        (sizeof (wordflags) / sizeof (wordflags[0]))
  68 #define SYNSIZE         256
  69
  70 int     lsyntax[SYNSIZE];
  71 int     debug;
  72 char    *progname;
  73
  74 char    preamble[] = "\
  75 /*\n\
  76  * This file was generated by mksyntax.  DO NOT EDIT.\n\
  77  */\n\
  78 \n";
  79
  80 char    includes[] = "\
  81 #include \"stdc.h\"\n\
  82 #include \"syntax.h\"\n\n";
  83
  84 static void
  85 usage()
  86 {
  87   fprintf (stderr, "%s: usage: %s [-d] [-o filename]\n", progname, progname);
  88   exit (2);
  89 }
  90
  91 #ifdef INCLUDE_UNUSED
  92 static int
  93 getcflag (s)
  94      char *s;
  95 {
  96   int i;
  97
  98   for (i = 0; i < N_WFLAGS; i++)
  99     if (strcmp (s, wordflags[i].fstr) == 0)
 100       return wordflags[i].flag;
 101   return -1;
 102 }
 103 #endif
 104
 105 static char *
 106 cdesc (i)
 107      int i;
 108 {
 109   static char xbuf[16];
 110
 111   if (i == ' ')
 112     return "SPC";
 113   else if (ISPRINT (i))
 114     {
 115       xbuf[0] = i;
 116       xbuf[1] = '\0';
 117       return (xbuf);
 118     }
 119   else if (i == CTLESC)
 120     return "CTLESC";
 121   else if (i == CTLNUL)
 122     return "CTLNUL";
 123   else if (i == '\033')         /* ASCII */
 124     return "ESC";
 125
 126   xbuf[0] = '\\';
 127   xbuf[2] = '\0';
 128
 129   switch (i)
 130     {
 131     case '\a': xbuf[1] = 'a'; break;
 132     case '\v': xbuf[1] = 'v'; break;
 133     case '\b': xbuf[1] = 'b'; break;
 134     case '\f': xbuf[1] = 'f'; break;
 135     case '\n': xbuf[1] = 'n'; break;
 136     case '\r': xbuf[1] = 'r'; break;
 137     case '\t': xbuf[1] = 't'; break;
 138     default: sprintf (xbuf, "%d", i); break;
 139     }
 140
 141   return xbuf;
 142 }
 143
 144 static char *
 145 getcstr (f)
 146      int f;
 147 {
 148   int i;
 149
 150   for (i = 0; i < N_WFLAGS; i++)
 151     if (f == wordflags[i].flag)
 152       return (wordflags[i].fstr);
 153   return ((char *)NULL);
 154 }
 155
 156 static void
 157 addcstr (str, flag)
 158      char *str;
 159      int flag;
 160 {
 161   char *s, *fstr;
 162   unsigned char uc;
 163
 164   for (s = str; s && *s; s++)
 165     {
 166       uc = *s;
 167
 168       if (debug)
 169         {
 170           fstr = getcstr (flag);
 171           fprintf(stderr, "added %s for character %s\n", fstr, cdesc(uc));
 172         }
 173
 174       lsyntax[uc] |= flag;
 175     }
 176 }
 177
 178 static void
 179 addcchar (c, flag)
 180      unsigned char c;
 181      int flag;
 182 {
 183   char *fstr;
 184
 185   if (debug)
 186     {
 187       fstr = getcstr (flag);
 188       fprintf (stderr, "added %s for character %s\n", fstr, cdesc(c));
 189     }
 190   lsyntax[c] |= flag;
 191 }
 192
 193 /* load up the correct flag values in lsyntax */
 194 static void
 195 load_lsyntax ()
 196 {
 197   /* shell metacharacters */
 198   addcstr (shell_meta_chars, CSHMETA);
 199
 200   /* shell word break characters */
 201   addcstr (shell_break_chars, CSHBRK);
 202
 203   addcchar ('`', CBACKQ);
 204
 205   addcstr (shell_quote_chars, CQUOTE);
 206
 207   addcchar (CTLESC, CSPECL);
 208   addcchar (CTLNUL, CSPECL);
 209
 210   addcstr (shell_exp_chars, CEXP);
 211
 212   addcstr (slashify_in_quotes, CBSDQUOTE);
 213   addcstr (slashify_in_here_document, CBSHDOC);
 214
 215   addcstr (shell_glob_chars, CGLOB);
 216
 217 #if defined (EXTENDED_GLOB)
 218   addcstr (ext_glob_chars, CXGLOB);
 219 #endif
 220
 221   addcstr (shell_quote_chars, CXQUOTE);
 222   addcchar ('\\', CXQUOTE);
 223
 224   addcstr ("@*#?-$!", CSPECVAR);        /* omits $0...$9 and $_ */
 225
 226   addcstr ("-=?+", CSUBSTOP);           /* OP in ${paramOPword} */
 227 }
 228
 229 static void
 230 dump_lflags (fp, ind)
 231      FILE *fp;
 232      int ind;
 233 {
 234   int xflags, first, i;
 235
 236   xflags = lsyntax[ind];
 237   first = 1;
 238
 239   if (xflags == 0)
 240     fputs (wordflags[0].fstr, fp);
 241   else
 242     {
 243       for (i = 1; i < N_WFLAGS; i++)
 244         if (xflags & wordflags[i].flag)
 245           {
 246             if (first)
 247               first = 0;
 248             else
 249               putc ('|', fp);
 250             fputs (wordflags[i].fstr, fp);
 251           }
 252     }
 253 }
 254
 255 static void
 256 wcomment (fp, i)
 257      FILE *fp;
 258      int i;
 259 {
 260   fputs ("\t\t/* ", fp);
 261
 262   fprintf (fp, "%s", cdesc(i));
 263
 264   fputs (" */", fp);
 265 }
 266
 267 static void
 268 dump_lsyntax (fp)
 269      FILE *fp;
 270 {
 271   int i;
 272
 273   fprintf (fp, "const int sh_syntaxtab[%d] = {\n", SYNSIZE);
 274
 275   for (i = 0; i < SYNSIZE; i++)
 276     {
 277       putc ('\t', fp);
 278       dump_lflags (fp, i);
 279       putc (',', fp);
 280       wcomment (fp, i);
 281       putc ('\n', fp);
 282     }
 283
 284   fprintf (fp, "};\n");
 285 }
 286
 287 int
 288 main(argc, argv)
 289      int argc;
 290      char **argv;
 291 {
 292   int opt, i;
 293   char *filename;
 294   FILE *fp;
 295
 296   if ((progname = strrchr (argv[0], '/')) == 0)
 297     progname = argv[0];
 298   else
 299     progname++;
 300
 301   filename = (char *)NULL;
 302   debug = 0;
 303
 304   while ((opt = getopt (argc, argv, "do:")) != EOF)
 305     {
 306       switch (opt)
 307         {
 308         case 'd':
 309           debug = 1;
 310           break;
 311         case 'o':
 312           filename = optarg;
 313           break;
 314         default:
 315           usage();
 316         }
 317     }
 318
 319   argc -= optind;
 320   argv += optind;
 321
 322   if (filename)
 323     {
 324       fp = fopen (filename, "w");
 325       if (fp == 0)
 326         {
 327           fprintf (stderr, "%s: %s: cannot open: %s\n", progname, filename, strerror(errno));
 328           exit (1);
 329         }
 330     }
 331   else
 332     {
 333       filename = "stdout";
 334       fp = stdout;
 335     }
 336
 337
 338   for (i = 0; i < SYNSIZE; i++)
 339     lsyntax[i] = CWORD;
 340
 341   load_lsyntax ();
 342
 343   fprintf (fp, "%s\n", preamble);
 344   fprintf (fp, "%s\n", includes);
 345
 346   dump_lsyntax (fp);
 347
 348   if (fp != stdout)
 349     fclose (fp);
 350   exit (0);
 351 }
 352
 353
 354 #if !defined (HAVE_STRERROR)
 355
 356 #include <bashtypes.h>
 357 #ifndef _MINIX
 358 #  include <sys/param.h>
 359 #endif
 360
 361 #if defined (HAVE_UNISTD_H)
 362 #  include <unistd.h>
 363 #endif
 364
 365 /* Return a string corresponding to the error number E.  From
 366    the ANSI C spec. */
 367 #if defined (strerror)
 368 #  undef strerror
 369 #endif
 370
 371 char *
 372 strerror (e)
 373      int e;
 374 {
 375   static char emsg[40];
 376 #if defined (HAVE_SYS_ERRLIST)
 377   extern int sys_nerr;
 378   extern char *sys_errlist[];
 379
 380   if (e > 0 && e < sys_nerr)
 381     return (sys_errlist[e]);
 382   else
 383 #endif /* HAVE_SYS_ERRLIST */
 384     {
 385       sprintf (emsg, "Unknown system error %d", e);
 386       return (&emsg[0]);
 387     }
 388 }
 389 #endif /* HAVE_STRERROR */