locale/programs/charmap.c

   1 /* Copyright (C) 1996, 1997 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3    Contributed by Ulrich Drepper <drepper@gnu.ai.mit.edu>, 1996.
   4
   5    The GNU C Library is free software; you can redistribute it and/or
   6    modify it under the terms of the GNU Library General Public License as
   7    published by the Free Software Foundation; either version 2 of the
   8    License, or (at your option) any later version.
   9
  10    The GNU C Library is distributed in the hope that it will be useful,
  11    but WITHOUT ANY WARRANTY; without even the implied warranty of
  12    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  13    Library General Public License for more details.
  14
  15    You should have received a copy of the GNU Library General Public
  16    License along with the GNU C Library; see the file COPYING.LIB.  If not,
  17    write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
  18    Boston, MA 02111-1307, USA.  */
  19
  20 #ifdef HAVE_CONFIG_H
  21 # include <config.h>
  22 #endif
  23
  24 #include <ctype.h>
  25 #include <errno.h>
  26 #include <libintl.h>
  27 #include <obstack.h>
  28 #include <stdlib.h>
  29 #include <string.h>
  30 #include <unistd.h>
  31
  32 #include "error.h"
  33 #include "linereader.h"
  34 #include "charset.h"
  35
  36
  37 /* Uncomment following line for production version.  */
  38 /* define NDEBUG 1 */
  39 #include <assert.h>
  40
  41
  42 /* Define the lookup function.  */
  43 #include "charmap-kw.h"
  44
  45
  46 void *xmalloc (size_t __n);
  47
  48 /* Prototypes for local functions.  */
  49 static struct charset_t *parse_charmap (const char *filename);
  50 static void new_width (struct linereader *cmfile, struct charset_t *result,
  51                        const char *from, const char *to,
  52                        unsigned long int width);
  53
  54
  55 struct charset_t *
  56 charmap_read (const char *filename)
  57 {
  58   const char *pathnfile;
  59   struct charset_t *result = NULL;
  60
  61   if (filename != NULL)
  62     {
  63       if (euidaccess (filename, R_OK) >= 0)
  64         pathnfile = filename;
  65       else
  66         {
  67           char *cp = xmalloc (strlen (filename) + sizeof CHARMAP_PATH + 1);
  68           stpcpy (stpcpy (stpcpy (cp, CHARMAP_PATH), "/"), filename);
  69
  70           pathnfile = (const char *) cp;
  71         }
  72
  73       result = parse_charmap (pathnfile);
  74
  75       if (result == NULL)
  76         error (0, errno, _("character map file `%s' not found"), filename);
  77     }
  78
  79   if (result == NULL)
  80     {
  81       pathnfile = CHARMAP_PATH "/" DEFAULT_CHARMAP;
  82
  83       result = parse_charmap (pathnfile);
  84
  85       if (result == NULL)
  86         error (4, errno, _("default character map file `%s' not found"),
  87                DEFAULT_CHARMAP);
  88     }
  89
  90   return result;
  91 }
  92
  93
  94 static struct charset_t *
  95 parse_charmap (const char *filename)
  96 {
  97   struct linereader *cmfile;
  98   struct charset_t *result;
  99   int state;
 100   enum token_t expected_tok = tok_error;
 101   const char *expected_str = NULL;
 102   char *from_name = NULL;
 103   char *to_name = NULL;
 104
 105   /* Determine path.  */
 106   cmfile = lr_open (filename, charmap_hash);
 107   if (cmfile == NULL)
 108     {
 109       if (strchr (filename, '/') == NULL)
 110         {
 111           /* Look in the systems charmap directory.  */
 112           char *buf = xmalloc (strlen (filename) + 1 + sizeof (CHARMAP_PATH));
 113
 114           stpcpy (stpcpy (stpcpy (buf, CHARMAP_PATH), "/"), filename);
 115           cmfile = lr_open (buf, charmap_hash);
 116
 117           if (cmfile == NULL)
 118             free (buf);
 119         }
 120
 121       if (cmfile == NULL)
 122         return NULL;
 123     }
 124
 125   /* Allocate room for result.  */
 126   result = (struct charset_t *) xmalloc (sizeof (struct charset_t));
 127   memset (result, '\0', sizeof (struct charset_t));
 128   /* The default DEFAULT_WIDTH is 1.  */
 129   result->width_default = 1;
 130
 131 #define obstack_chunk_alloc malloc
 132 #define obstack_chunk_free free
 133   obstack_init (&result->mem_pool);
 134
 135   if (init_hash (&result->char_table, 256))
 136     {
 137       free (result);
 138       return NULL;
 139     }
 140
 141   /* We use a state machine to describe the charmap description file
 142      format.  */
 143   state = 1;
 144   while (1)
 145     {
 146       /* What's on?  */
 147       struct token *now = lr_token (cmfile, NULL);
 148       enum token_t nowtok = now->tok;
 149       struct token *arg;
 150
 151       if (nowtok == tok_eof)
 152         break;
 153
 154       switch (state)
 155         {
 156         case 1:
 157           /* The beginning.  We expect the special declarations, EOL or
 158              `CHARMAP'.  */
 159           if (nowtok == tok_eol)
 160             /* Ignore empty lines.  */
 161             continue;
 162
 163           if (nowtok == tok_charmap)
 164             {
 165               from_name = NULL;
 166               to_name = NULL;
 167
 168               /* We have to set up the real work.  Fill in some
 169                  default values.  */
 170               if (result->mb_cur_max == 0)
 171                 result->mb_cur_max = 1;
 172               if (result->mb_cur_min == 0)
 173                 result->mb_cur_min = result->mb_cur_max;
 174               if (result->mb_cur_min > result->mb_cur_max)
 175                 {
 176                   error (0, 0, _("\
 177 %s: <mb_cur_max> must be greater than <mb_cur_min>\n"),
 178                          cmfile->fname);
 179
 180                   result->mb_cur_min = result->mb_cur_max;
 181                 }
 182
 183               lr_ignore_rest (cmfile, 1);
 184
 185               state = 2;
 186               continue;
 187             }
 188
 189           if (nowtok != tok_code_set_name && nowtok != tok_mb_cur_max
 190               && nowtok != tok_mb_cur_min && nowtok != tok_escape_char
 191               && nowtok != tok_comment_char && nowtok != tok_g0esc
 192               && nowtok != tok_g1esc && nowtok != tok_g2esc
 193               && nowtok != tok_g3esc)
 194             {
 195               lr_error (cmfile, _("syntax error in prolog: %s"),
 196                         _("illegal definition"));
 197
 198               lr_ignore_rest (cmfile, 0);
 199               continue;
 200             }
 201
 202           /* We know that we need an argument.  */
 203           arg = lr_token (cmfile, NULL);
 204
 205           switch (nowtok)
 206             {
 207             case tok_code_set_name:
 208               if (arg->tok != tok_ident)
 209                 {
 210                 badarg:
 211                   lr_error (cmfile, _("syntax error in prolog: %s"),
 212                             _("bad argument"));
 213
 214                   lr_ignore_rest (cmfile, 0);
 215                   continue;
 216                 }
 217
 218               result->code_set_name = obstack_copy0 (&result->mem_pool,
 219                                                      arg->val.str.start,
 220                                                      arg->val.str.len);
 221
 222               lr_ignore_rest (cmfile, 1);
 223               continue;
 224
 225             case tok_mb_cur_max:
 226             case tok_mb_cur_min:
 227               if (arg->tok != tok_number)
 228                 goto badarg;
 229
 230               if (arg->val.num < 1 || arg->val.num > 4)
 231                 {
 232                   lr_error (cmfile,
 233                             _("value for <%s> must lie between 1 and 4"),
 234                             nowtok == tok_mb_cur_min ? "mb_cur_min"
 235                                                      : "mb_cur_max");
 236
 237                   lr_ignore_rest (cmfile, 0);
 238                   continue;
 239                 }
 240               if ((nowtok == tok_mb_cur_max && result->mb_cur_min != 0
 241                    && (int) arg->val.num < result->mb_cur_min)
 242                   || (nowtok == tok_mb_cur_min && result->mb_cur_max != 0
 243                       && (int) arg->val.num > result->mb_cur_max))
 244                 {
 245                   lr_error (cmfile, _("\
 246 value of <mb_cur_max> must be greater than the value of <mb_cur_min>"));
 247
 248                   lr_ignore_rest (cmfile, 0);
 249                   continue;
 250                 }
 251
 252               if (nowtok == tok_mb_cur_max)
 253                 result->mb_cur_max = arg->val.num;
 254               else
 255                 result->mb_cur_min = arg->val.num;
 256
 257               lr_ignore_rest (cmfile, 1);
 258               continue;
 259
 260             case tok_escape_char:
 261             case tok_comment_char:
 262               if (arg->tok != tok_ident)
 263                 goto badarg;
 264
 265               if (arg->val.str.len != 1)
 266                 {
 267                   lr_error (cmfile, _("\
 268 argument to <%s> must be a single character"),
 269                             nowtok == tok_escape_char ? "escape_char"
 270                                                       : "comment_char");
 271
 272                   lr_ignore_rest (cmfile, 0);
 273                   continue;
 274                 }
 275
 276               if (nowtok == tok_escape_char)
 277                 cmfile->escape_char = *arg->val.str.start;
 278               else
 279                 cmfile->comment_char = *arg->val.str.start;
 280
 281               lr_ignore_rest (cmfile, 1);
 282               continue;
 283
 284             case tok_g0esc:
 285             case tok_g1esc:
 286             case tok_g2esc:
 287             case tok_g3esc:
 288               lr_ignore_rest (cmfile, 0); /* XXX */
 289               continue;
 290
 291             default:
 292               /* Cannot happen.  */
 293               assert (! "Should not happen");
 294             }
 295           break;
 296
 297         case 2:
 298           /* We have seen `CHARMAP' and now are in the body.  Each line
 299              must have the format "%s %s %s\n" or "%s...%s %s %s\n".  */
 300           if (nowtok == tok_eol)
 301             /* Ignore empty lines.  */
 302             continue;
 303
 304           if (nowtok == tok_end)
 305             {
 306               expected_tok = tok_charmap;
 307               expected_str = "CHARMAP";
 308               state = 90;
 309               continue;
 310             }
 311
 312           if (nowtok != tok_bsymbol)
 313             {
 314               lr_error (cmfile, _("syntax error in %s definition: %s"),
 315                         "CHARMAP", _("no symbolic name given"));
 316
 317               lr_ignore_rest (cmfile, 0);
 318               continue;
 319             }
 320
 321           /* If the previous line was not completely correct free the
 322              used memory.  */
 323           if (from_name != NULL)
 324             obstack_free (&result->mem_pool, from_name);
 325
 326           from_name = (char *) obstack_copy0 (&result->mem_pool,
 327                                               now->val.str.start,
 328                                               now->val.str.len);
 329           to_name = NULL;
 330
 331           state = 3;
 332           continue;
 333
 334         case 3:
 335           /* We have two possibilities: We can see an ellipsis or an
 336              encoding value.  */
 337           if (nowtok == tok_ellipsis)
 338             {
 339               state = 4;
 340               continue;
 341             }
 342           /* FALLTHROUGH */
 343
 344         case 5:
 345           if (nowtok != tok_charcode && nowtok != tok_ucs2
 346               && nowtok != tok_ucs4)
 347             {
 348               lr_error (cmfile, _("syntax error in %s definition: %s"),
 349                         "CHARMAP", _("illegal encoding given"));
 350
 351               lr_ignore_rest (cmfile, 0);
 352
 353               state = 2;
 354               continue;
 355             }
 356
 357           if (nowtok == tok_charcode)
 358             /* Write char value in table.  */
 359             charset_new_char (cmfile, result, now->val.charcode.nbytes,
 360                               now->val.charcode.val, from_name, to_name);
 361           else
 362             /* Determine ISO 10646 value and write into table.  */
 363             charset_new_unicode (cmfile, result, now->val.charcode.nbytes,
 364                                  now->val.charcode.val, from_name, to_name);
 365
 366           /* Ignore trailing comment silently.  */
 367           lr_ignore_rest (cmfile, 0);
 368
 369           from_name = NULL;
 370           to_name = NULL;
 371
 372           state = 2;
 373           continue;
 374
 375         case 4:
 376           if (nowtok != tok_bsymbol)
 377             {
 378               lr_error (cmfile, _("syntax error in %s definition: %s"),
 379                         "CHARMAP",
 380                         _("no symbolic name given for end of range"));
 381
 382               lr_ignore_rest (cmfile, 0);
 383               continue;
 384             }
 385
 386           /* If the previous line was not completely correct free the
 387              used memory.  */
 388           to_name = (char *) obstack_copy0 (&result->mem_pool,
 389                                             cmfile->token.val.str.start,
 390                                             cmfile->token.val.str.len);
 391
 392           state = 3;
 393           continue;
 394
 395         case 90:
 396           if (nowtok != expected_tok)
 397             lr_error (cmfile, _("\
 398 `%1$s' definition does not end with `END %1$s'"), expected_str);
 399
 400           lr_ignore_rest (cmfile, nowtok == expected_tok);
 401           state = 91;
 402           continue;
 403
 404         case 91:
 405           /* Waiting for WIDTH... */
 406           if (nowtok == tok_eol)
 407             /* Ignore empty lines.  */
 408             continue;
 409
 410           if (nowtok == tok_width_default)
 411             {
 412               state = 92;
 413               continue;
 414             }
 415
 416           if (nowtok == tok_width)
 417             {
 418               lr_ignore_rest (cmfile, 1);
 419               state = 93;
 420               continue;
 421             }
 422
 423           if (nowtok == tok_width_variable)
 424             {
 425               lr_ignore_rest (cmfile, 1);
 426               state = 98;
 427               continue;
 428             }
 429
 430           lr_error (cmfile, _("\
 431 only WIDTH definitions are allowed to follow the CHARMAP definition"));
 432
 433           lr_ignore_rest (cmfile, 0);
 434           continue;
 435
 436         case 92:
 437           if (nowtok != tok_number)
 438             lr_error (cmfile, _("value for %s must be an integer"),
 439                       "WIDTH_DEFAULT");
 440           else
 441             result->width_default = now->val.num;
 442
 443           lr_ignore_rest (cmfile, nowtok == tok_number);
 444
 445           state = 91;
 446           continue;
 447
 448         case 93:
 449           /* We now expect `END WIDTH' or lines of the format "%s %d\n" or
 450              "%s...%s %d\n".  */
 451           if (nowtok == tok_eol)
 452             /* ignore empty lines.  */
 453             continue;
 454
 455           if (nowtok == tok_end)
 456             {
 457               expected_tok = tok_width;
 458               expected_str = "WIDTH";
 459               state = 90;
 460               continue;
 461             }
 462
 463           if (nowtok != tok_bsymbol)
 464             {
 465               lr_error (cmfile, _("syntax error in %s definition: %s"),
 466                         "WIDTH", _("no symbolic name given"));
 467
 468               lr_ignore_rest (cmfile, 0);
 469               continue;
 470             }
 471
 472           if (from_name != NULL)
 473             obstack_free (&result->mem_pool, from_name);
 474
 475           from_name = (char *) obstack_copy0 (&result->mem_pool,
 476                                               now->val.str.start,
 477                                               now->val.str.len);
 478           to_name = NULL;
 479
 480           state = 94;
 481           continue;
 482
 483         case 94:
 484           if (nowtok == tok_ellipsis)
 485             {
 486               state = 95;
 487               continue;
 488             }
 489
 490         case 96:
 491           if (nowtok != tok_number)
 492             lr_error (cmfile, _("value for %s must be an integer"),
 493                       "WIDTH");
 494           else
 495             {
 496               /* Store width for chars.  */
 497               new_width (cmfile, result, from_name, to_name, now->val.num);
 498
 499               from_name = NULL;
 500               to_name = NULL;
 501             }
 502
 503           lr_ignore_rest (cmfile, nowtok == tok_number);
 504
 505           state = 93;
 506           continue;
 507
 508         case 95:
 509           if (nowtok != tok_bsymbol)
 510             {
 511               lr_error (cmfile, _("syntax error in %s definition: %s"),
 512                         "WIDTH", _("no symbolic name given for end of range"));
 513
 514               lr_ignore_rest (cmfile, 0);
 515
 516               state = 93;
 517               continue;
 518             }
 519
 520           to_name = (char *) obstack_copy0 (&result->mem_pool,
 521                                             now->val.str.start,
 522                                             now->val.str.len);
 523
 524           state = 96;
 525           continue;
 526
 527         case 98:
 528           /* We now expect `END WIDTH_VARIABLE' or lines of the format
 529              "%s\n" or "%s...%s\n".  */
 530           if (nowtok == tok_eol)
 531             /* ignore empty lines.  */
 532             continue;
 533
 534           if (nowtok == tok_end)
 535             {
 536               expected_tok = tok_width_variable;
 537               expected_str = "WIDTH_VARIABLE";
 538               state = 90;
 539               continue;
 540             }
 541
 542           if (nowtok != tok_bsymbol)
 543             {
 544               lr_error (cmfile, _("syntax error in %s definition: %s"),
 545                         "WIDTH_VARIABLE", _("no symbolic name given"));
 546
 547               lr_ignore_rest (cmfile, 0);
 548
 549               continue;
 550             }
 551
 552           if (from_name != NULL)
 553             obstack_free (&result->mem_pool, from_name);
 554
 555           from_name = (char *) obstack_copy0 (&result->mem_pool,
 556                                               now->val.str.start,
 557                                               now->val.str.len);
 558           to_name = NULL;
 559
 560           state = 99;
 561           continue;
 562
 563         case 99:
 564           if (nowtok == tok_ellipsis)
 565             state = 100;
 566
 567           /* Store info.  */
 568           from_name = NULL;
 569
 570           /* Warn */
 571           state = 98;
 572           continue;
 573
 574         case 100:
 575           if (nowtok != tok_bsymbol)
 576             lr_error (cmfile, _("syntax error in %s definition: %s"),
 577                       "WIDTH_VARIABLE",
 578                       _("no symbolic name given for end of range"));
 579           else
 580             {
 581               to_name = (char *) obstack_copy0 (&result->mem_pool,
 582                                                 now->val.str.start,
 583                                                 now->val.str.len);
 584               /* XXX Enter value into table.  */
 585             }
 586
 587           lr_ignore_rest (cmfile, nowtok == tok_bsymbol);
 588
 589           state = 98;
 590           continue;
 591
 592         default:
 593           error (5, 0, _("%s: error in state machine"), __FILE__);
 594           /* NOTREACHED */
 595         }
 596       break;
 597     }
 598
 599   if (state != 91)
 600     error (0, 0, _("%s: premature end of file"), cmfile->fname);
 601
 602   lr_close (cmfile);
 603
 604   return result;
 605 }
 606
 607
 608 static void
 609 new_width (struct linereader *cmfile, struct charset_t *result,
 610            const char *from, const char *to, unsigned long int width)
 611 {
 612   unsigned int from_val, to_val;
 613
 614   from_val = charset_find_value (result, from, strlen (from));
 615   if ((wchar_t) from_val == ILLEGAL_CHAR_VALUE)
 616     {
 617       lr_error (cmfile, _("unknown character `%s'"), from);
 618       return;
 619     }
 620
 621   if (to == NULL)
 622     to_val = from_val;
 623   else
 624     {
 625       to_val = charset_find_value (result, to, strlen (to));
 626       if ((wchar_t) to_val == ILLEGAL_CHAR_VALUE)
 627         {
 628           lr_error (cmfile, _("unknown character `%s'"), to);
 629           return;
 630         }
 631     }
 632
 633   if (result->nwidth_rules >= result->nwidth_rules_max)
 634     {
 635       size_t new_size = result->nwidth_rules + 32;
 636       struct width_rule *new_rules =
 637         (struct width_rule *) obstack_alloc (&result->mem_pool,
 638                                              (new_size
 639                                               * sizeof (struct width_rule)));
 640
 641       memcpy (new_rules, result->width_rules,
 642               result->nwidth_rules_max * sizeof (struct width_rule));
 643
 644       result->width_rules = new_rules;
 645       result->nwidth_rules_max = new_size;
 646     }
 647
 648   result->width_rules[result->nwidth_rules].from = from_val;
 649   result->width_rules[result->nwidth_rules].to = to_val;
 650   result->width_rules[result->nwidth_rules].width = (unsigned int) width;
 651   ++result->nwidth_rules;
 652 }