locale/programs/localedef.c

   1 /* Copyright (C) 1995,1996,1997,1998,1999,2000 Free Software Foundation, Inc.
   2    This file is part of the GNU C Library.
   3    Contributed by Ulrich Drepper <drepper@cygnus.com>, 1995.
   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 <argp.h>
  25 #include <errno.h>
  26 #include <fcntl.h>
  27 #include <libintl.h>
  28 #include <locale.h>
  29 #include <mcheck.h>
  30 #include <stdio.h>
  31 #include <stdlib.h>
  32 #include <string.h>
  33 #include <unistd.h>
  34 #include <sys/mman.h>
  35 #include <sys/stat.h>
  36
  37 #include "error.h"
  38 #include "charmap.h"
  39 #include "locfile.h"
  40
  41 /* Undefine the following line in the production version.  */
  42 /* #define NDEBUG 1 */
  43 #include <assert.h>
  44
  45
  46 /* List of copied locales.  */
  47 struct copy_def_list_t *copy_list;
  48
  49 /* If this is defined be POSIX conform.  */
  50 int posix_conformance;
  51
  52 /* If not zero give a lot more messages.  */
  53 int verbose;
  54
  55 /* If not zero suppress warnings and information messages.  */
  56 int be_quiet;
  57
  58 /* If not zero, produce old-style hash table instead of 3-level access tables.  */
  59 int oldstyle_tables;
  60
  61 /* If not zero force output even if warning were issued.  */
  62 static int force_output;
  63
  64 /* Name of the character map file.  */
  65 static const char *charmap_file;
  66
  67 /* Name of the locale definition file.  */
  68 static const char *input_file;
  69
  70 /* Name of the repertoire map file.  */
  71 const char *repertoire_global;
  72
  73 /* List of all locales.  */
  74 static struct localedef_t *locales;
  75
  76
  77 /* Name and version of program.  */
  78 static void print_version (FILE *stream, struct argp_state *state);
  79 void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version;
  80
  81 #define OPT_POSIX 1
  82 #define OPT_QUIET 2
  83 #define OPT_OLDSTYLE 3
  84
  85 /* Definitions of arguments for argp functions.  */
  86 static const struct argp_option options[] =
  87 {
  88   { NULL, 0, NULL, 0, N_("Input Files:") },
  89   { "charmap", 'f', "FILE", 0,
  90     N_("Symbolic character names defined in FILE") },
  91   { "inputfile", 'i', "FILE", 0, N_("Source definitions are found in FILE") },
  92   { "repertoire-map", 'u', "FILE", 0,
  93     N_("FILE contains mapping from symbolic names to UCS4 values") },
  94
  95   { NULL, 0, NULL, 0, N_("Output control:") },
  96   { "force", 'c', NULL, 0,
  97     N_("Create output even if warning messages were issued") },
  98   { "old-style", OPT_OLDSTYLE, NULL, 0, N_("Create old-style tables") },
  99   { "posix", OPT_POSIX, NULL, 0, N_("Be strictly POSIX conform") },
 100   { "quiet", OPT_QUIET, NULL, 0,
 101     N_("Suppress warnings and information messages") },
 102   { "verbose", 'v', NULL, 0, N_("Print more messages") },
 103   { NULL, 0, NULL, 0, NULL }
 104 };
 105
 106 /* Short description of program.  */
 107 static const char doc[] = N_("Compile locale specification");
 108
 109 /* Strings for arguments in help texts.  */
 110 static const char args_doc[] = N_("NAME");
 111
 112 /* Prototype for option handler.  */
 113 static error_t parse_opt (int key, char *arg, struct argp_state *state);
 114
 115 /* Function to print some extra text in the help message.  */
 116 static char *more_help (int key, const char *text, void *input);
 117
 118 /* Data structure to communicate with argp functions.  */
 119 static struct argp argp =
 120 {
 121   options, parse_opt, args_doc, doc, NULL, more_help
 122 };
 123
 124
 125 /* Prototypes for global functions.  */
 126 extern void *xmalloc (size_t __n);
 127
 128 /* Prototypes for local functions.  */
 129 static void error_print (void);
 130 static const char *construct_output_path (char *path);
 131 static const char *normalize_codeset (const char *codeset, size_t name_len);
 132
 133
 134 int
 135 main (int argc, char *argv[])
 136 {
 137   const char *output_path;
 138   int cannot_write_why;
 139   struct charmap_t *charmap;
 140   struct localedef_t global;
 141   int remaining;
 142
 143   /* Set initial values for global variables.  */
 144   copy_list = NULL;
 145   posix_conformance = getenv ("POSIXLY_CORRECT") != NULL;
 146   error_print_progname = error_print;
 147
 148   /* Set locale.  Do not set LC_ALL because the other categories must
 149      not be affected (according to POSIX.2).  */
 150   setlocale (LC_MESSAGES, "");
 151   setlocale (LC_CTYPE, "");
 152
 153   /* Initialize the message catalog.  */
 154   textdomain (_libc_intl_domainname);
 155
 156   /* Parse and process arguments.  */
 157   argp_err_exit_status = 4;
 158   argp_parse (&argp, argc, argv, 0, &remaining, NULL);
 159
 160   /* POSIX.2 requires to be verbose about missing characters in the
 161      character map.  */
 162   verbose |= posix_conformance;
 163
 164   if (argc - remaining != 1)
 165     {
 166       /* We need exactly one non-option parameter.  */
 167       argp_help (&argp, stdout, ARGP_HELP_SEE | ARGP_HELP_EXIT_ERR,
 168                  program_invocation_short_name);
 169       exit (4);
 170     }
 171
 172   /* The parameter describes the output path of the constructed files.
 173      If the described files cannot be written return a NULL pointer.  */
 174   output_path  = construct_output_path (argv[remaining]);
 175   cannot_write_why = errno;
 176
 177   /* Now that the parameters are processed we have to reset the local
 178      ctype locale.  (P1003.2 4.35.5.2)  */
 179   setlocale (LC_CTYPE, "POSIX");
 180
 181   /* Look whether the system really allows locale definitions.  POSIX
 182      defines error code 3 for this situation so I think it must be
 183      a fatal error (see P1003.2 4.35.8).  */
 184   if (sysconf (_SC_2_LOCALEDEF) < 0)
 185     error (3, 0, _("FATAL: system does not define `_POSIX2_LOCALEDEF'"));
 186
 187   /* Process charmap file.  */
 188   charmap = charmap_read (charmap_file);
 189
 190   /* Add the first entry in the locale list.  */
 191   memset (&global, '\0', sizeof (struct localedef_t));
 192   global.name = input_file;
 193   global.needed = ALL_LOCALES;
 194   locales = &global;
 195
 196   /* Now read the locale file.  */
 197   if (locfile_read (&global, charmap) != 0)
 198     error (4, errno, _("cannot open locale definition file `%s'"), input_file);
 199
 200   /* Perhaps we saw some `copy' instructions.  */
 201   while (1)
 202     {
 203       struct localedef_t *runp = locales;
 204
 205       while (runp != NULL && (runp->needed & runp->avail) == runp->needed)
 206         runp = runp->next;
 207
 208       if (runp == NULL)
 209         /* Everything read.  */
 210         break;
 211
 212       if (locfile_read (runp, charmap) != 0)
 213         error (4, errno, _("cannot open locale definition file `%s'"),
 214                runp->name);
 215     }
 216
 217   /* Check the categories we processed in source form.  */
 218   check_all_categories (locales, charmap);
 219
 220   /* We are now able to write the data files.  If warning were given we
 221      do it only if it is explicitly requested (--force).  */
 222   if (error_message_count == 0 || force_output != 0)
 223     {
 224       if (cannot_write_why != 0)
 225         error (4, cannot_write_why, _("cannot write output files to `%s'"),
 226                output_path);
 227       else
 228         write_all_categories (locales, charmap, output_path);
 229     }
 230   else
 231     error (4, 0, _("no output file produced because warning were issued"));
 232
 233   /* This exit status is prescribed by POSIX.2 4.35.7.  */
 234   exit (error_message_count != 0);
 235 }
 236
 237
 238 /* Handle program arguments.  */
 239 static error_t
 240 parse_opt (int key, char *arg, struct argp_state *state)
 241 {
 242   switch (key)
 243     {
 244     case OPT_QUIET:
 245       be_quiet = 1;
 246       break;
 247     case OPT_POSIX:
 248       posix_conformance = 1;
 249       break;
 250     case OPT_OLDSTYLE:
 251       oldstyle_tables = 1;
 252       break;
 253     case 'c':
 254       force_output = 1;
 255       break;
 256     case 'f':
 257       charmap_file = arg;
 258       break;
 259     case 'i':
 260       input_file = arg;
 261       break;
 262     case 'u':
 263       repertoire_global = arg;
 264       break;
 265     case 'v':
 266       verbose = 1;
 267       break;
 268     default:
 269       return ARGP_ERR_UNKNOWN;
 270     }
 271   return 0;
 272 }
 273
 274
 275 static char *
 276 more_help (int key, const char *text, void *input)
 277 {
 278   char *cp;
 279
 280   switch (key)
 281     {
 282     case ARGP_KEY_HELP_EXTRA:
 283       /* We print some extra information.  */
 284       asprintf (&cp, gettext ("\
 285 System's directory for character maps : %s\n\
 286                        repertoire maps: %s\n\
 287                        locale path    : %s\n\
 288 %s"),
 289                 CHARMAP_PATH, REPERTOIREMAP_PATH, LOCALE_PATH, gettext ("\
 290 Report bugs using the `glibcbug' script to <bugs@gnu.org>.\n"));
 291       return cp;
 292     default:
 293       break;
 294     }
 295   return (char *) text;
 296 }
 297
 298 /* Print the version information.  */
 299 static void
 300 print_version (FILE *stream, struct argp_state *state)
 301 {
 302   fprintf (stream, "localedef (GNU %s) %s\n", PACKAGE, VERSION);
 303   fprintf (stream, gettext ("\
 304 Copyright (C) %s Free Software Foundation, Inc.\n\
 305 This is free software; see the source for copying conditions.  There is NO\n\
 306 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 307 "), "2000");
 308   fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
 309 }
 310
 311
 312 /* The address of this function will be assigned to the hook in the error
 313    functions.  */
 314 static void
 315 error_print (void)
 316 {
 317 }
 318
 319
 320 /* The parameter to localedef describes the output path.  If it does
 321    contain a '/' character it is a relative path.  Otherwise it names the
 322    locale this definition is for.  */
 323 static const char *
 324 construct_output_path (char *path)
 325 {
 326   const char *normal = NULL;
 327   char *result;
 328   char *endp;
 329
 330   if (strchr (path, '/') == NULL)
 331     {
 332       /* This is a system path.  First examine whether the locale name
 333          contains a reference to the codeset.  This should be
 334          normalized.  */
 335       char *startp;
 336       size_t n;
 337
 338       startp = path;
 339       /* We must be prepared for finding a CEN name or a location of
 340          the introducing `.' where it is not possible anymore.  */
 341       while (*startp != '\0' && *startp != '@' && *startp != '.'
 342              && *startp != '+' && *startp != ',')
 343         ++startp;
 344       if (*startp == '.')
 345         {
 346           /* We found a codeset specification.  Now find the end.  */
 347           endp = ++startp;
 348           while (*endp != '\0' && *endp != '@')
 349             ++endp;
 350
 351           if (endp > startp)
 352             normal = normalize_codeset (startp, endp - startp);
 353         }
 354       else
 355         /* This is to keep gcc quiet.  */
 356         endp = NULL;
 357
 358       /* We put an additional '\0' at the end of the string because at
 359          the end of the function we need another byte for the trailing
 360          '/'.  */
 361       if (normal == NULL)
 362         n = asprintf (&result, "%s/%s%c", LOCALEDIR, path, '\0');
 363       else
 364         n = asprintf (&result, "%s/%.*s%s%s%c", LOCALEDIR,
 365                       (int) (startp - path), path, normal, endp, '\0');
 366
 367       endp = result + n;
 368     }
 369   else
 370     {
 371       /* This is a user path.  Please note the additional byte in the
 372          memory allocation.  */
 373       size_t len = strlen (path) + 1;
 374       result = xmalloc (len + 1);
 375       endp = mempcpy (result, path, len) - 1;
 376     }
 377
 378   errno = 0;
 379
 380   if (euidaccess (result, W_OK) == -1)
 381     /* Perhaps the directory does not exist now.  Try to create it.  */
 382     if (errno == ENOENT)
 383       {
 384         errno = 0;
 385         mkdir (result, 0777);
 386       }
 387
 388   *endp++ = '/';
 389   *endp = '\0';
 390
 391   return result;
 392 }
 393
 394
 395 /* Normalize codeset name.  There is no standard for the codeset
 396    names.  Normalization allows the user to use any of the common
 397    names.  */
 398 static const char *
 399 normalize_codeset (codeset, name_len)
 400      const char *codeset;
 401      size_t name_len;
 402 {
 403   int len = 0;
 404   int only_digit = 1;
 405   char *retval;
 406   char *wp;
 407   size_t cnt;
 408
 409   for (cnt = 0; cnt < name_len; ++cnt)
 410     if (isalnum (codeset[cnt]))
 411       {
 412         ++len;
 413
 414         if (isalpha (codeset[cnt]))
 415           only_digit = 0;
 416       }
 417
 418   retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
 419
 420   if (retval != NULL)
 421     {
 422       if (only_digit)
 423         wp = stpcpy (retval, "iso");
 424       else
 425         wp = retval;
 426
 427       for (cnt = 0; cnt < name_len; ++cnt)
 428         if (isalpha (codeset[cnt]))
 429           *wp++ = tolower (codeset[cnt]);
 430         else if (isdigit (codeset[cnt]))
 431           *wp++ = codeset[cnt];
 432
 433       *wp = '\0';
 434     }
 435
 436   return (const char *) retval;
 437 }
 438
 439
 440 struct localedef_t *
 441 add_to_readlist (int locale, const char *name, const char *repertoire_name,
 442                  int generate, struct localedef_t *copy_locale)
 443 {
 444   struct localedef_t *runp = locales;
 445
 446   while (runp != NULL && strcmp (name, runp->name) != 0)
 447     runp = runp->next;
 448
 449   if (runp == NULL)
 450     {
 451       /* Add a new entry at the end.  */
 452       struct localedef_t *newp;
 453
 454       assert (generate == 1);
 455
 456       newp = xcalloc (1, sizeof (struct localedef_t));
 457       newp->name = name;
 458       newp->repertoire_name = repertoire_name;
 459
 460       if (locales == NULL)
 461         runp = locales = newp;
 462       else
 463         {
 464           runp = locales;
 465           while (runp->next != NULL)
 466             runp = runp->next;
 467           runp = runp->next = newp;
 468         }
 469     }
 470
 471   if (generate && (runp->needed & (1 << locale)) != 0)
 472     error (5, 0, _("circular dependencies between locale definitions"));
 473
 474   if (copy_locale != NULL)
 475     {
 476       if (runp->categories[locale].generic != NULL)
 477         error (5, 0, _("cannot add already read locale `%s' a second time"),
 478                name);
 479       else
 480         runp->categories[locale].generic =
 481           copy_locale->categories[locale].generic;
 482     }
 483
 484   runp->needed |= 1 << locale;
 485
 486   return runp;
 487 }
 488
 489
 490 struct localedef_t *
 491 find_locale (int locale, const char *name, const char *repertoire_name,
 492              struct charmap_t *charmap)
 493 {
 494   struct localedef_t *result;
 495
 496   /* Find the locale, but do not generate it since this would be a bug.  */
 497   result = add_to_readlist (locale, name, repertoire_name, 0, NULL);
 498
 499   assert (result != NULL);
 500
 501   if ((result->avail & (1 << locale)) == 0
 502       && locfile_read (result, charmap) != 0)
 503     error (4, errno, _("cannot open locale definition file `%s'"),
 504            result->name);
 505
 506   return result;
 507 }
 508
 509
 510 struct localedef_t *
 511 load_locale (int locale, const char *name, const char *repertoire_name,
 512              struct charmap_t *charmap, struct localedef_t *copy_locale)
 513 {
 514   struct localedef_t *result;
 515
 516   /* Generate the locale if it does not exist.  */
 517   result = add_to_readlist (locale, name, repertoire_name, 1, copy_locale);
 518
 519   assert (result != NULL);
 520
 521   if ((result->avail & (1 << locale)) == 0
 522       && locfile_read (result, charmap) != 0)
 523     error (4, errno, _("cannot open locale definition file `%s'"),
 524            result->name);
 525
 526   return result;
 527 }
 528
 529 static void
 530 turn_on_mcheck (void)
 531 {
 532   /* Enable `malloc' debugging.  */
 533   mcheck (NULL);
 534   /* Use the following line for a more thorough but much slower testing.  */
 535   /* mcheck_pedantic (NULL); */
 536 }
 537
 538 void (*__malloc_initialize_hook) (void) = turn_on_mcheck;