locale/charmap.c

   1 /* Copyright (C) 1995 Free Software Foundation, Inc.
   2
   3 The GNU C Library is free software; you can redistribute it and/or
   4 modify it under the terms of the GNU Library General Public License as
   5 published by the Free Software Foundation; either version 2 of the
   6 License, or (at your option) any later version.
   7
   8 The GNU C Library is distributed in the hope that it will be useful,
   9 but WITHOUT ANY WARRANTY; without even the implied warranty of
  10 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  11 Library General Public License for more details.
  12
  13 You should have received a copy of the GNU Library General Public
  14 License along with the GNU C Library; see the file COPYING.LIB.  If
  15 not, write to the Free Software Foundation, Inc., 675 Mass Ave,
  16 Cambridge, MA 02139, USA.  */
  17
  18 #include <ctype.h>
  19 #include <errno.h>
  20 #include <libintl.h>
  21 #include <limits.h>
  22 #include <stdio.h>
  23 #include <stdlib.h>
  24 #include <string.h>
  25 #include <unistd.h>
  26
  27 #include "localedef.h"
  28 #include "hash.h"
  29
  30 /* Data structure for representing charmap database.  */
  31 struct charmap charmap_data;
  32
  33 /* Line number in charmap file.  */
  34 static unsigned int line_no;
  35
  36 /* Prototypes for local functions.  */
  37 static void read_prolog (FILE *infile);
  38 static unsigned long read_body (FILE *infile);
  39
  40
  41 /* Read complete table of symbolic names for character set from file.  If
  42    this file does not exist or is not readable a default file is tried.
  43    If this also is not readable no character map is defined.  */
  44 void
  45 charmap_read (const char *filename)
  46 {
  47   unsigned long max_char;
  48   long path_max = pathconf (".", _PC_PATH_MAX);
  49   char buf[path_max];
  50   FILE *infile = NULL;
  51
  52   /* Initialize charmap data.  */
  53   charmap_data.codeset_name = NULL;
  54   charmap_data.mb_cur_max = -1;
  55   charmap_data.mb_cur_min = -1;
  56   charmap_data.escape_char = '\\';
  57   charmap_data.comment_char = '#';
  58
  59   if (filename != NULL)
  60     {
  61       strcpy (buf, filename);
  62       infile = fopen (filename, "r");
  63       if (infile == NULL && filename[0] != '/')
  64         {
  65           snprintf (buf, path_max, "%s/%s", CHARMAP_PATH, filename);
  66           infile = fopen (buf, "r");
  67         }
  68     }
  69   if (infile == NULL)
  70     {
  71       if (filename != NULL)
  72         error (0, errno, gettext ("input file `%s' not found"), filename);
  73
  74       snprintf (buf, path_max, "%s/%s", CHARMAP_PATH, DEFAULT_CHARMAP);
  75       infile = fopen (buf, "r");
  76
  77       if (infile == NULL)
  78         error (4, errno, gettext ("input file `%s' not found"), filename);
  79     }
  80
  81   charmap_data.filename = buf;
  82   init_hash (&charmap_data.table, 500);
  83   line_no = 0;
  84
  85   /* Read the prolog of the charmap file.  */
  86   read_prolog (infile);
  87
  88   /* Last works on the charmap tables global data.  */
  89   if (charmap_data.mb_cur_max == -1)
  90     charmap_data.mb_cur_max = 1;
  91   if (charmap_data.mb_cur_min == -1)
  92     charmap_data.mb_cur_min = charmap_data.mb_cur_max;
  93
  94   if ((size_t) charmap_data.mb_cur_max > sizeof (long))
  95     {
  96       error (2, 0, gettext ("program limitation: for now only upto %Zu "
  97                             "bytes per character are allowed"), sizeof (long));
  98     }
  99
 100   /* Now process all entries.  */
 101   max_char = read_body (infile);
 102
 103   /* We don't need the file anymore.  */
 104   fclose (infile);
 105
 106
 107   /* Determine the optimal table size when using the simple modulo hashing
 108      function.  */
 109   if (max_char >= 256)
 110     {
 111       int size;
 112       /* Current best values, initialized to some never reached high value.  */
 113       int best_count = 10000;
 114       int best_size = 10000;
 115       int best_product = best_count * best_size;
 116
 117       /* Give warning.  */
 118       error (-1, 0, gettext ("computing character table size: this may take "
 119                              "a while"));
 120
 121       for (size = 256; size <= best_product; ++size)
 122         {
 123           /* Array with slot counters.  */
 124           int cnt[size];
 125           /* Current character.  */
 126           int ch;
 127           /* Maximal number of characters in any slot.  */
 128           int maxcnt = 0;
 129           /* Product of current size and maximal count.  */
 130           int product = 0;
 131           /* Iteration pointer through hashing table.  */
 132           char *ptr = NULL;
 133
 134           /* Initializes counters to zero.  */
 135           memset(cnt, 0, size * sizeof (int));
 136
 137           /* Iterate through whole hashing table.  */
 138           while (product < best_product
 139                  && iterate_table (&charmap_data.table, (void **) &ptr,
 140                                    (void **) &ch))
 141             {
 142               /* Increment slot counter.  */
 143               ++cnt[ch % size];
 144               /* Test for current maximum.  */
 145               if (cnt[ch % size] > maxcnt)
 146                 {
 147                   maxcnt = cnt[ch % size];
 148                   product = maxcnt * size;
 149                 }
 150             }
 151
 152           if (product < best_product)
 153             {
 154               best_count = maxcnt;
 155               best_size = size;
 156               best_product = best_count * best_size;
 157             }
 158         }
 159
 160       charmap_data.hash_size = best_size;
 161       charmap_data.hash_layers = best_count;
 162     }
 163   else
 164     {
 165       charmap_data.hash_size = 256;
 166       charmap_data.hash_layers = 1;
 167     }
 168 }
 169
 170
 171 #define SYNTAX_ERROR                                                         \
 172   do { error (0, 0, gettext ("%s:%u: syntax error in charmap file"),         \
 173               charmap_data.filename, line_no);                               \
 174        goto end_of_loop; } while (0)
 175
 176 /* Read the prolog of the charmap file until the line containing `CHARMAP'.
 177    All possible entries are processed.  */
 178 static void
 179 read_prolog (FILE *infile)
 180 {
 181   size_t bufsize = sysconf (_SC_LINE_MAX);
 182   char buf[bufsize];
 183
 184   while (1)
 185     {
 186       char *cp = buf;
 187       char len;
 188
 189       /* Read the next line.  */
 190       fgets (buf, bufsize, infile);
 191       len = strlen (buf);
 192
 193       /* On EOF simply return.  */
 194       if (len == 0 || buf[len - 1] != '\n')
 195         error (4, 0, gettext ("%s: unexpected end of file in charmap"),
 196                charmap_data.filename);
 197
 198       /* This is the next line.  */
 199       ++line_no;
 200
 201       /* Comments and empty lines are ignored.  */
 202       if (len == 1 || buf[0] == charmap_data.comment_char)
 203         continue;
 204
 205       buf[len - 1] = '\0';
 206
 207       /* Throw away leading white spaces.  This is not defined in POSIX.2
 208          so don't do it if conformance is requested.  */
 209       if (!posix_conformance)
 210         while (isspace (*cp))
 211           ++cp;
 212
 213       /* If `CHARMAP' is read the prolog is over.  */
 214       if (strncmp (cp, "CHARMAP", 7) == 0
 215           && (!posix_conformance || cp[7] == '\0'))
 216         return;
 217
 218       /* Now it can be only one of special symbols defining the charmap
 219          parameters.  All are beginning with '<'.  */
 220       if (*cp != '<')
 221         SYNTAX_ERROR;
 222
 223       ++cp;
 224       if (strncmp (cp, "code_set_name>", 14) == 0)
 225         {
 226           char *startp;
 227
 228 #define cp_to_arg(no,pred)                                                    \
 229           cp += no;                                                           \
 230           while (isspace (*cp))                                               \
 231             ++cp;                                                             \
 232           if (*cp == '\0' || !pred (*cp))                                     \
 233             SYNTAX_ERROR;
 234
 235           cp_to_arg (14,isgraph)
 236
 237           if (charmap_data.codeset_name != NULL)
 238             {
 239               error (0, 0, gettext ("%s:%u: duplicate code set name "
 240                                     "specification"),
 241                      charmap_data.filename, line_no);
 242               free (charmap_data.codeset_name);
 243             }
 244
 245           startp = cp;
 246           while (*cp != '\0' && isgraph (*cp) && !isspace (*cp))
 247             ++cp;
 248
 249           charmap_data.codeset_name = (char *) xmalloc (cp - startp + 1);
 250           strncpy (startp, startp, cp - startp);
 251         }
 252       else if (strncmp (cp, "mb_cur_max>", 11) == 0)
 253         {
 254           int new_val;
 255           cp_to_arg (11,isdigit)
 256
 257           if (charmap_data.mb_cur_max != -1)
 258             error (0, 0,
 259                    gettext ("%s:%u: duplicate definition of mb_cur_max"),
 260                    charmap_data.filename, line_no);
 261
 262           new_val = (int) strtol (cp, &cp, posix_conformance ? 10 : 0);
 263           if (new_val < 1)
 264             error (0, 0, gettext ("%s:%u: illegal value for mb_cur_max: %d"),
 265                    charmap_data.filename, line_no, new_val);
 266           else
 267             charmap_data.mb_cur_max = new_val;
 268         }
 269       else if (strncmp (cp, "mb_cur_min>", 11) == 0)
 270         {
 271           int new_val;
 272           cp_to_arg (11,isdigit)
 273
 274           if (charmap_data.mb_cur_max != -1)
 275             error (0, 0,
 276                    gettext ("%s:%u: duplicate definition of mb_cur_min"),
 277                    charmap_data.filename, line_no);
 278
 279           new_val = (int) strtol (cp, &cp, posix_conformance ? 10 : 0);
 280           if (new_val < 1)
 281             error (0, 0, gettext ("%s:%u: illegal value for mb_cur_min: %d"),
 282                    charmap_data.filename, line_no, new_val);
 283           else
 284             charmap_data.mb_cur_min = new_val;
 285         }
 286       else if (strncmp (cp, "escape_char>", 12) == 0)
 287         {
 288           cp_to_arg (12, isgraph)
 289           charmap_data.escape_char = *cp;
 290         }
 291       else if (strncmp (cp, "comment_char>", 13) == 0)
 292         {
 293           cp_to_arg (13, isgraph)
 294           charmap_data.comment_char = *cp;
 295         }
 296       else
 297         SYNTAX_ERROR;
 298       end_of_loop:
 299     }
 300 }
 301 #undef cp_to_arg
 302
 303
 304 static unsigned long
 305 read_body (FILE *infile)
 306 {
 307   unsigned long max_char = 0;
 308   size_t bufsize = sysconf (_SC_LINE_MAX);
 309   char buf[bufsize];
 310   char name_str[bufsize / 2];
 311   char code_str[bufsize / 2];
 312
 313   while (1)
 314     {
 315       char *cp = buf;
 316       size_t len;
 317
 318       /* Read the next line.  */
 319       fgets (buf, bufsize, infile);
 320       len = strlen (buf);
 321
 322       /* On EOF simply return.  */
 323       if (len == 0)
 324         error (0, 0, gettext ("%s: `END CHARMAP' is missing"),
 325                charmap_data.filename);
 326
 327       /* This is the next line.  */
 328       ++line_no;
 329
 330       if (len == bufsize - 1)
 331         {
 332           error (0, 0, gettext ("%s:%u: line too long;  use `getconf "
 333                                 "LINE_MAX' to get the current maximum line"
 334                                 "length"), charmap_data.filename, line_no);
 335           do
 336             {
 337               fgets (buf, bufsize, infile);
 338               len = strlen (buf);
 339             }
 340           while (len == bufsize - 1);
 341           continue;
 342         }
 343
 344       /* Comments and empty lines are ignored.  */
 345       if (len == 1 || buf[0] == charmap_data.comment_char)
 346         continue;
 347
 348       buf[len - 1] = '\0';
 349
 350       /* Throw away leading white spaces.  This is not defined in POSIX.2
 351          so don't do it if conformance is requested.  */
 352       if (!posix_conformance)
 353         while (isspace (*cp))
 354           ++cp;
 355
 356       if (*cp == '<')
 357         {
 358           char *end1p, *end2p, *start2p;
 359           size_t cnt = 0;
 360           unsigned long char_value = 0;
 361
 362           if (sscanf (cp + 1, "%s %s", name_str, code_str) != 2)
 363             SYNTAX_ERROR;
 364
 365           end1p = cp = name_str;
 366           while (*cp != '\0' && *cp != '>')
 367             {
 368               if (*cp == charmap_data.escape_char)
 369                 if (*++cp == '\0')
 370                   SYNTAX_ERROR;
 371               *end1p++ = *cp++;
 372             }
 373           if (*cp == '\0')
 374             /* No final '>'.  Make error condition.  */
 375             end1p = name_str;
 376           else
 377             ++cp;
 378
 379           *end1p = '\0';
 380
 381           if (*cp == '.' && *++cp == '.' && *++cp == '.' && *++cp == '<')
 382             {
 383               /* This might be the alternate form.  */
 384               start2p = end2p = ++cp;
 385               while (*cp != '\0' && *cp != '>')
 386                 {
 387                   if (*cp == charmap_data.escape_char)
 388                     if (*++cp == '\0')
 389                       SYNTAX_ERROR;
 390                   *end2p = *cp++;
 391                 }
 392               if (*cp == '\0')
 393                 /* NO final '>'.  Make error condition.  */
 394                 end2p = start2p;
 395               else
 396                 ++cp;
 397             }
 398           else
 399             start2p = end2p = NULL;
 400
 401
 402           if (end1p == name_str || (start2p != NULL && start2p != end2p)
 403               || *cp != '\0'
 404               || *code_str != charmap_data.escape_char)
 405             SYNTAX_ERROR;
 406
 407           cp = code_str;
 408           do
 409             {
 410               char *begin;
 411               long val;
 412
 413               switch (*++cp)
 414                 {
 415                 case 'd':
 416                   val = strtol ((begin = cp + 1), &cp, 10);
 417                   break;
 418                 case 'x':
 419                   val = strtol ((begin = cp + 1), &cp, 16);
 420                   break;
 421                 default:
 422                   val = strtol ((begin = cp), &cp, 8);
 423                   break;
 424                 }
 425               if (begin == cp)
 426                 SYNTAX_ERROR;
 427
 428               if (posix_conformance && cp - begin < 2)
 429                 error (0, 0, gettext ("%s:%u: byte constant has less than "
 430                                       "two digits"),
 431                        charmap_data.filename, line_no);
 432
 433               if (val < 0 || val > 255)
 434                 {
 435                   error (0, 0, gettext ("%s:%u: character encoding must be "
 436                                         "given in 8-bit bytes"),
 437                          charmap_data.filename, line_no);
 438                   goto end_of_loop;
 439                 }
 440
 441               if (cnt < (size_t) charmap_data.mb_cur_max)
 442                 {
 443                   if (cnt < sizeof (long))  /* FIXME */
 444                     char_value = (char_value << 8) | val;
 445                 }
 446               else
 447                 {
 448                   error (0, 0, gettext ("%s:%u: number of bytes in character "
 449                                         "definition exceeds `mb_cur_max'"),
 450                          charmap_data.filename, line_no);
 451                   break;
 452                 }
 453               ++cnt;
 454             }
 455           while (*cp == charmap_data.escape_char);
 456
 457           /* Ignore the rest of the line (comment).  */
 458           if (end2p == NULL)
 459             {
 460               if (insert_entry (&charmap_data.table, name_str,
 461                                 end1p - name_str, (void *) char_value))
 462                 error (0, 0, gettext ("%s:%u: duplicate entry"),
 463                        charmap_data.filename, line_no);
 464
 465               max_char = MAX (max_char, char_value);
 466             }
 467           else
 468             {
 469               char *en1, *en2, *start1p;
 470               long n1, n2, n;
 471
 472               start1p = name_str;
 473
 474               while (*start1p == *start2p && !isdigit (*start1p)
 475                      && start1p < end1p)
 476                   ++start1p, ++start2p;
 477
 478               n1 = strtol (start1p, &en1, 10);
 479               n2 = strtol (start2p, &en2, 10);
 480
 481               if (en1 - start1p != en2 - start2p || en1 != end1p
 482                   || en2 != end2p)
 483                 SYNTAX_ERROR;
 484
 485               if (n1 > n2)
 486                 error (0, 0, gettext ("%s:%u: starting character is bigger "
 487                                       "than last"),
 488                        charmap_data.filename, line_no);
 489
 490               n = n1;
 491               while (n <= n2)
 492                 {
 493                   snprintf(start1p, en1 - start1p, "%0*d", en1 - start1p, n);
 494
 495                   if (insert_entry (&charmap_data.table, name_str,
 496                                     en1 - name_str,
 497                                     (void *) (char_value + n - n1)))
 498                     error (0, 0, gettext ("%s:%u: duplicate entry"),
 499                            charmap_data.filename, line_no);
 500
 501                   max_char = MAX (max_char, char_value + n - n1);
 502                   ++n;
 503                 }
 504             }
 505         }
 506       else
 507         {
 508           if (strncmp (cp, "END CHARMAP", 11) == 0)
 509             return max_char;
 510
 511           SYNTAX_ERROR;
 512         }
 513       end_of_loop:
 514     }
 515
 516   return max_char;
 517 }
 518
 519 /*
 520  * Local Variables:
 521  *  mode:c
 522  *  c-basic-offset:2
 523  * End:
 524  */