mksyntax.c

   1 /*
   2  * mksyntax.c - construct shell syntax table for fast char attribute lookup.
   3  */
   4
   5 /* Copyright (C) 2000-2009 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
  10    it under the terms of the GNU General Public License as published by
  11    the Free Software Foundation, either version 3 of the License, or
  12    (at your option) any later version.
  13
  14    Bash is distributed in the hope that it will be useful,
  15    but WITHOUT ANY WARRANTY; without even the implied warranty of
  16    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17    GNU General Public License for more details.
  18
  19    You should have received a copy of the GNU General Public License
  20    along with Bash.  If not, see <http://www.gnu.org/licenses/>.
  21 */
  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         { CBLANK,       "CBLANK" },
  66 };
  67
  68 #define N_WFLAGS        (sizeof (wordflags) / sizeof (wordflags[0]))
  69 #define SYNSIZE         256
  70
  71 int     lsyntax[SYNSIZE];
  72 int     debug;
  73 char    *progname;
  74
  75 char    preamble[] = "\
  76 /*\n\
  77  * This file was generated by mksyntax.  DO NOT EDIT.\n\
  78  */\n\
  79 \n";
  80
  81 char    includes[] = "\
  82 #include \"config.h\"\n\
  83 #include \"stdc.h\"\n\
  84 #include \"syntax.h\"\n\n";
  85
  86 static void
  87 usage()
  88 {
  89   fprintf (stderr, "%s: usage: %s [-d] [-o filename]\n", progname, progname);
  90   exit (2);
  91 }
  92
  93 #ifdef INCLUDE_UNUSED
  94 static int
  95 getcflag (s)
  96      char *s;
  97 {
  98   int i;
  99
 100   for (i = 0; i < N_WFLAGS; i++)
 101     if (strcmp (s, wordflags[i].fstr) == 0)
 102       return wordflags[i].flag;
 103   return -1;
 104 }
 105 #endif
 106
 107 static char *
 108 cdesc (i)
 109      int i;
 110 {
 111   static char xbuf[16];
 112
 113   if (i == ' ')
 114     return "SPC";
 115   else if (ISPRINT (i))
 116     {
 117       xbuf[0] = i;
 118       xbuf[1] = '\0';
 119       return (xbuf);
 120     }
 121   else if (i == CTLESC)
 122     return "CTLESC";
 123   else if (i == CTLNUL)
 124     return "CTLNUL";
 125   else if (i == '\033')         /* ASCII */
 126     return "ESC";
 127
 128   xbuf[0] = '\\';
 129   xbuf[2] = '\0';
 130
 131   switch (i)
 132     {
 133 #ifdef __STDC__
 134     case '\a': xbuf[1] = 'a'; break;
 135     case '\v': xbuf[1] = 'v'; break;
 136 #else
 137     case '\007': xbuf[1] = 'a'; break;
 138     case 0x0B: xbuf[1] = 'v'; break;
 139 #endif
 140     case '\b': xbuf[1] = 'b'; break;
 141     case '\f': xbuf[1] = 'f'; break;
 142     case '\n': xbuf[1] = 'n'; break;
 143     case '\r': xbuf[1] = 'r'; break;
 144     case '\t': xbuf[1] = 't'; break;
 145     default: sprintf (xbuf, "%d", i); break;
 146     }
 147
 148   return xbuf;
 149 }
 150
 151 static char *
 152 getcstr (f)
 153      int f;
 154 {
 155   int i;
 156
 157   for (i = 0; i < N_WFLAGS; i++)
 158     if (f == wordflags[i].flag)
 159       return (wordflags[i].fstr);
 160   return ((char *)NULL);
 161 }
 162
 163 static void
 164 addcstr (str, flag)
 165      char *str;
 166      int flag;
 167 {
 168   char *s, *fstr;
 169   unsigned char uc;
 170
 171   for (s = str; s && *s; s++)
 172     {
 173       uc = *s;
 174
 175       if (debug)
 176         {
 177           fstr = getcstr (flag);
 178           fprintf(stderr, "added %s for character %s\n", fstr, cdesc(uc));
 179         }
 180
 181       lsyntax[uc] |= flag;
 182     }
 183 }
 184
 185 static void
 186 addcchar (c, flag)
 187      unsigned char c;
 188      int flag;
 189 {
 190   char *fstr;
 191
 192   if (debug)
 193     {
 194       fstr = getcstr (flag);
 195       fprintf (stderr, "added %s for character %s\n", fstr, cdesc(c));
 196     }
 197   lsyntax[c] |= flag;
 198 }
 199
 200 static void
 201 addblanks ()
 202 {
 203   register int i;
 204   unsigned char uc;
 205
 206   for (i = 0; i < SYNSIZE; i++)
 207     {
 208       uc = i;
 209       /* Since we don't call setlocale(), this defaults to the "C" locale, and
 210          the default blank characters will be space and tab. */
 211       if (isblank (uc))
 212         lsyntax[uc] |= CBLANK;
 213     }
 214 }
 215
 216 /* load up the correct flag values in lsyntax */
 217 static void
 218 load_lsyntax ()
 219 {
 220   /* shell metacharacters */
 221   addcstr (shell_meta_chars, CSHMETA);
 222
 223   /* shell word break characters */
 224   addcstr (shell_break_chars, CSHBRK);
 225
 226   addcchar ('`', CBACKQ);
 227
 228   addcstr (shell_quote_chars, CQUOTE);
 229
 230   addcchar (CTLESC, CSPECL);
 231   addcchar (CTLNUL, CSPECL);
 232
 233   addcstr (shell_exp_chars, CEXP);
 234
 235   addcstr (slashify_in_quotes, CBSDQUOTE);
 236   addcstr (slashify_in_here_document, CBSHDOC);
 237
 238   addcstr (shell_glob_chars, CGLOB);
 239
 240 #if defined (EXTENDED_GLOB)
 241   addcstr (ext_glob_chars, CXGLOB);
 242 #endif
 243
 244   addcstr (shell_quote_chars, CXQUOTE);
 245   addcchar ('\\', CXQUOTE);
 246
 247   addcstr ("@*#?-$!", CSPECVAR);        /* omits $0...$9 and $_ */
 248
 249   addcstr ("-=?+", CSUBSTOP);           /* OP in ${paramOPword} */
 250
 251   addblanks ();
 252 }
 253
 254 static void
 255 dump_lflags (fp, ind)
 256      FILE *fp;
 257      int ind;
 258 {
 259   int xflags, first, i;
 260
 261   xflags = lsyntax[ind];
 262   first = 1;
 263
 264   if (xflags == 0)
 265     fputs (wordflags[0].fstr, fp);
 266   else
 267     {
 268       for (i = 1; i < N_WFLAGS; i++)
 269         if (xflags & wordflags[i].flag)
 270           {
 271             if (first)
 272               first = 0;
 273             else
 274               putc ('|', fp);
 275             fputs (wordflags[i].fstr, fp);
 276           }
 277     }
 278 }
 279
 280 static void
 281 wcomment (fp, i)
 282      FILE *fp;
 283      int i;
 284 {
 285   fputs ("\t\t/* ", fp);
 286
 287   fprintf (fp, "%s", cdesc(i));
 288
 289   fputs (" */", fp);
 290 }
 291
 292 static void
 293 dump_lsyntax (fp)
 294      FILE *fp;
 295 {
 296   int i;
 297
 298   fprintf (fp, "int sh_syntabsiz = %d;\n", SYNSIZE);
 299   fprintf (fp, "int sh_syntaxtab[%d] = {\n", SYNSIZE);
 300
 301   for (i = 0; i < SYNSIZE; i++)
 302     {
 303       putc ('\t', fp);
 304       dump_lflags (fp, i);
 305       putc (',', fp);
 306       wcomment (fp, i);
 307       putc ('\n', fp);
 308     }
 309
 310   fprintf (fp, "};\n");
 311 }
 312
 313 int
 314 main(argc, argv)
 315      int argc;
 316      char **argv;
 317 {
 318   int opt, i;
 319   char *filename;
 320   FILE *fp;
 321
 322   if ((progname = strrchr (argv[0], '/')) == 0)
 323     progname = argv[0];
 324   else
 325     progname++;
 326
 327   filename = (char *)NULL;
 328   debug = 0;
 329
 330   while ((opt = getopt (argc, argv, "do:")) != EOF)
 331     {
 332       switch (opt)
 333         {
 334         case 'd':
 335           debug = 1;
 336           break;
 337         case 'o':
 338           filename = optarg;
 339           break;
 340         default:
 341           usage();
 342         }
 343     }
 344
 345   argc -= optind;
 346   argv += optind;
 347
 348   if (filename)
 349     {
 350       fp = fopen (filename, "w");
 351       if (fp == 0)
 352         {
 353           fprintf (stderr, "%s: %s: cannot open: %s\n", progname, filename, strerror(errno));
 354           exit (1);
 355         }
 356     }
 357   else
 358     {
 359       filename = "stdout";
 360       fp = stdout;
 361     }
 362
 363
 364   for (i = 0; i < SYNSIZE; i++)
 365     lsyntax[i] = CWORD;
 366
 367   load_lsyntax ();
 368
 369   fprintf (fp, "%s\n", preamble);
 370   fprintf (fp, "%s\n", includes);
 371
 372   dump_lsyntax (fp);
 373
 374   if (fp != stdout)
 375     fclose (fp);
 376   exit (0);
 377 }
 378
 379
 380 #if !defined (HAVE_STRERROR)
 381
 382 #include <bashtypes.h>
 383 #if defined (HAVE_SYS_PARAM_H)
 384 #  include <sys/param.h>
 385 #endif
 386
 387 #if defined (HAVE_UNISTD_H)
 388 #  include <unistd.h>
 389 #endif
 390
 391 /* Return a string corresponding to the error number E.  From
 392    the ANSI C spec. */
 393 #if defined (strerror)
 394 #  undef strerror
 395 #endif
 396
 397 char *
 398 strerror (e)
 399      int e;
 400 {
 401   static char emsg[40];
 402 #if defined (HAVE_SYS_ERRLIST)
 403   extern int sys_nerr;
 404   extern char *sys_errlist[];
 405
 406   if (e > 0 && e < sys_nerr)
 407     return (sys_errlist[e]);
 408   else
 409 #endif /* HAVE_SYS_ERRLIST */
 410     {
 411       sprintf (emsg, "Unknown system error %d", e);
 412       return (&emsg[0]);
 413     }
 414 }
 415 #endif /* HAVE_STRERROR */