locale/programs/localedef.c

   1 /* Copyright (C) 1995-2007, 2008 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    This program is free software; you can redistribute it and/or modify
   6    it under the terms of the GNU General Public License as published
   7    by the Free Software Foundation; version 2 of the License, or
   8    (at your option) any later version.
   9
  10    This program 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
  13    GNU General Public License for more details.
  14
  15    You should have received a copy of the GNU General Public License
  16    along with this program; if not, write to the Free Software Foundation,
  17    Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  18
  19 #ifdef HAVE_CONFIG_H
  20 # include <config.h>
  21 #endif
  22
  23 #include <argp.h>
  24 #include <errno.h>
  25 #include <fcntl.h>
  26 #include <libintl.h>
  27 #include <locale.h>
  28 #include <mcheck.h>
  29 #include <stdbool.h>
  30 #include <stdio.h>
  31 #include <stdlib.h>
  32 #include <string.h>
  33 #include <unistd.h>
  34 #include <error.h>
  35 #include <sys/mman.h>
  36 #include <sys/stat.h>
  37
  38 #include "localedef.h"
  39 #include "charmap.h"
  40 #include "locfile.h"
  41
  42 /* Undefine the following line in the production version.  */
  43 /* #define NDEBUG 1 */
  44 #include <assert.h>
  45
  46
  47 /* List of copied locales.  */
  48 struct copy_def_list_t *copy_list;
  49
  50 /* If this is defined be POSIX conform.  */
  51 int posix_conformance;
  52
  53 /* If not zero give a lot more messages.  */
  54 int verbose;
  55
  56 /* If not zero suppress warnings and information messages.  */
  57 int be_quiet;
  58
  59 /* If not zero, produce old-style hash table instead of 3-level access
  60    tables.  */
  61 int oldstyle_tables;
  62
  63 /* If not zero force output even if warning were issued.  */
  64 static int force_output;
  65
  66 /* Prefix for output files.  */
  67 const char *output_prefix;
  68
  69 /* Name of the character map file.  */
  70 static const char *charmap_file;
  71
  72 /* Name of the locale definition file.  */
  73 static const char *input_file;
  74
  75 /* Name of the repertoire map file.  */
  76 const char *repertoire_global;
  77
  78 /* Name of the locale.alias file.  */
  79 const char *alias_file;
  80
  81 /* List of all locales.  */
  82 static struct localedef_t *locales;
  83
  84 /* If true don't add locale data to archive.  */
  85 bool no_archive;
  86
  87 /* If true add named locales to archive.  */
  88 static bool add_to_archive;
  89
  90 /* If true delete named locales from archive.  */
  91 static bool delete_from_archive;
  92
  93 /* If true replace archive content when adding.  */
  94 static bool replace_archive;
  95
  96 /* If true list archive content.  */
  97 static bool list_archive;
  98
  99 /* Maximum number of retries when opening the locale archive.  */
 100 int max_locarchive_open_retry = 10;
 101
 102
 103 /* Name and version of program.  */
 104 static void print_version (FILE *stream, struct argp_state *state);
 105 void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version;
 106
 107 #define OPT_POSIX 301
 108 #define OPT_QUIET 302
 109 #define OPT_OLDSTYLE 303
 110 #define OPT_PREFIX 304
 111 #define OPT_NO_ARCHIVE 305
 112 #define OPT_ADD_TO_ARCHIVE 306
 113 #define OPT_REPLACE 307
 114 #define OPT_DELETE_FROM_ARCHIVE 308
 115 #define OPT_LIST_ARCHIVE 309
 116
 117 /* Definitions of arguments for argp functions.  */
 118 static const struct argp_option options[] =
 119 {
 120   { NULL, 0, NULL, 0, N_("Input Files:") },
 121   { "charmap", 'f', "FILE", 0,
 122     N_("Symbolic character names defined in FILE") },
 123   { "inputfile", 'i', "FILE", 0, N_("Source definitions are found in FILE") },
 124   { "repertoire-map", 'u', "FILE", 0,
 125     N_("FILE contains mapping from symbolic names to UCS4 values") },
 126
 127   { NULL, 0, NULL, 0, N_("Output control:") },
 128   { "force", 'c', NULL, 0,
 129     N_("Create output even if warning messages were issued") },
 130   { "old-style", OPT_OLDSTYLE, NULL, 0, N_("Create old-style tables") },
 131   { "prefix", OPT_PREFIX, "PATH", 0, N_("Optional output file prefix") },
 132   { "posix", OPT_POSIX, NULL, 0, N_("Be strictly POSIX conform") },
 133   { "quiet", OPT_QUIET, NULL, 0,
 134     N_("Suppress warnings and information messages") },
 135   { "verbose", 'v', NULL, 0, N_("Print more messages") },
 136   { NULL, 0, NULL, 0, N_("Archive control:") },
 137   { "no-archive", OPT_NO_ARCHIVE, NULL, 0,
 138     N_("Don't add new data to archive") },
 139   { "add-to-archive", OPT_ADD_TO_ARCHIVE, NULL, 0,
 140     N_("Add locales named by parameters to archive") },
 141   { "replace", OPT_REPLACE, NULL, 0, N_("Replace existing archive content") },
 142   { "delete-from-archive", OPT_DELETE_FROM_ARCHIVE, NULL, 0,
 143     N_("Remove locales named by parameters from archive") },
 144   { "list-archive", OPT_LIST_ARCHIVE, NULL, 0, N_("List content of archive") },
 145   { "alias-file", 'A', "FILE", 0,
 146     N_("locale.alias file to consult when making archive")},
 147   { NULL, 0, NULL, 0, NULL }
 148 };
 149
 150 /* Short description of program.  */
 151 static const char doc[] = N_("Compile locale specification");
 152
 153 /* Strings for arguments in help texts.  */
 154 static const char args_doc[] = N_("\
 155 NAME\n\
 156 [--add-to-archive|--delete-from-archive] FILE...\n\
 157 --list-archive [FILE]");
 158
 159 /* Prototype for option handler.  */
 160 static error_t parse_opt (int key, char *arg, struct argp_state *state);
 161
 162 /* Function to print some extra text in the help message.  */
 163 static char *more_help (int key, const char *text, void *input);
 164
 165 /* Data structure to communicate with argp functions.  */
 166 static struct argp argp =
 167 {
 168   options, parse_opt, args_doc, doc, NULL, more_help
 169 };
 170
 171
 172 /* Prototypes for global functions.  */
 173 extern void *xmalloc (size_t __n);
 174
 175 /* Prototypes for local functions.  */
 176 static void error_print (void);
 177 static const char *construct_output_path (char *path);
 178 static const char *normalize_codeset (const char *codeset, size_t name_len);
 179
 180
 181 int
 182 main (int argc, char *argv[])
 183 {
 184   const char *output_path;
 185   int cannot_write_why;
 186   struct charmap_t *charmap;
 187   struct localedef_t global;
 188   int remaining;
 189
 190   /* Set initial values for global variables.  */
 191   copy_list = NULL;
 192   posix_conformance = getenv ("POSIXLY_CORRECT") != NULL;
 193   error_print_progname = error_print;
 194
 195   /* Set locale.  Do not set LC_ALL because the other categories must
 196      not be affected (according to POSIX.2).  */
 197   setlocale (LC_MESSAGES, "");
 198   setlocale (LC_CTYPE, "");
 199
 200   /* Initialize the message catalog.  */
 201   textdomain (_libc_intl_domainname);
 202
 203   /* Parse and process arguments.  */
 204   argp_err_exit_status = 4;
 205   argp_parse (&argp, argc, argv, 0, &remaining, NULL);
 206
 207   /* Handle a few special cases.  */
 208   if (list_archive)
 209     show_archive_content (verbose);
 210   if (add_to_archive)
 211     return add_locales_to_archive (argc - remaining, &argv[remaining],
 212                                    replace_archive);
 213   if (delete_from_archive)
 214     return delete_locales_from_archive (argc - remaining, &argv[remaining]);
 215
 216   /* POSIX.2 requires to be verbose about missing characters in the
 217      character map.  */
 218   verbose |= posix_conformance;
 219
 220   if (argc - remaining != 1)
 221     {
 222       /* We need exactly one non-option parameter.  */
 223       argp_help (&argp, stdout, ARGP_HELP_SEE | ARGP_HELP_EXIT_ERR,
 224                  program_invocation_short_name);
 225       exit (4);
 226     }
 227
 228   /* The parameter describes the output path of the constructed files.
 229      If the described files cannot be written return a NULL pointer.  */
 230   output_path  = construct_output_path (argv[remaining]);
 231   if (output_path == NULL && ! no_archive)
 232     error (4, errno, _("cannot create directory for output files"));
 233   cannot_write_why = errno;
 234
 235   /* Now that the parameters are processed we have to reset the local
 236      ctype locale.  (P1003.2 4.35.5.2)  */
 237   setlocale (LC_CTYPE, "POSIX");
 238
 239   /* Look whether the system really allows locale definitions.  POSIX
 240      defines error code 3 for this situation so I think it must be
 241      a fatal error (see P1003.2 4.35.8).  */
 242   if (sysconf (_SC_2_LOCALEDEF) < 0)
 243     WITH_CUR_LOCALE (error (3, 0, _("\
 244 FATAL: system does not define `_POSIX2_LOCALEDEF'")));
 245
 246   /* Process charmap file.  */
 247   charmap = charmap_read (charmap_file, verbose, 1, be_quiet, 1);
 248
 249   /* Add the first entry in the locale list.  */
 250   memset (&global, '\0', sizeof (struct localedef_t));
 251   global.name = input_file ?: "/dev/stdin";
 252   global.needed = ALL_LOCALES;
 253   locales = &global;
 254
 255   /* Now read the locale file.  */
 256   if (locfile_read (&global, charmap) != 0)
 257     WITH_CUR_LOCALE (error (4, errno, _("\
 258 cannot open locale definition file `%s'"), input_file));
 259
 260   /* Perhaps we saw some `copy' instructions.  */
 261   while (1)
 262     {
 263       struct localedef_t *runp = locales;
 264
 265       while (runp != NULL && (runp->needed & runp->avail) == runp->needed)
 266         runp = runp->next;
 267
 268       if (runp == NULL)
 269         /* Everything read.  */
 270         break;
 271
 272       if (locfile_read (runp, charmap) != 0)
 273         WITH_CUR_LOCALE (error (4, errno, _("\
 274 cannot open locale definition file `%s'"), runp->name));
 275     }
 276
 277   /* Check the categories we processed in source form.  */
 278   check_all_categories (locales, charmap);
 279
 280   /* We are now able to write the data files.  If warning were given we
 281      do it only if it is explicitly requested (--force).  */
 282   if (error_message_count == 0 || force_output != 0)
 283     {
 284       if (cannot_write_why != 0)
 285         WITH_CUR_LOCALE (error (4, cannot_write_why, _("\
 286 cannot write output files to `%s'"), output_path));
 287       else
 288         write_all_categories (locales, charmap, argv[remaining], output_path);
 289     }
 290   else
 291     WITH_CUR_LOCALE (error (4, 0, _("\
 292 no output file produced because warnings were issued")));
 293
 294   /* This exit status is prescribed by POSIX.2 4.35.7.  */
 295   exit (error_message_count != 0);
 296 }
 297
 298
 299 /* Handle program arguments.  */
 300 static error_t
 301 parse_opt (int key, char *arg, struct argp_state *state)
 302 {
 303   switch (key)
 304     {
 305     case OPT_QUIET:
 306       be_quiet = 1;
 307       break;
 308     case OPT_POSIX:
 309       posix_conformance = 1;
 310       break;
 311     case OPT_OLDSTYLE:
 312       oldstyle_tables = 1;
 313       break;
 314     case OPT_PREFIX:
 315       output_prefix = arg;
 316       break;
 317     case OPT_NO_ARCHIVE:
 318       no_archive = true;
 319       break;
 320     case OPT_ADD_TO_ARCHIVE:
 321       add_to_archive = true;
 322       break;
 323     case OPT_REPLACE:
 324       replace_archive = true;
 325       break;
 326     case OPT_DELETE_FROM_ARCHIVE:
 327       delete_from_archive = true;
 328       break;
 329     case OPT_LIST_ARCHIVE:
 330       list_archive = true;
 331       break;
 332     case 'c':
 333       force_output = 1;
 334       break;
 335     case 'f':
 336       charmap_file = arg;
 337       break;
 338     case 'A':
 339       alias_file = arg;
 340       break;
 341     case 'i':
 342       input_file = arg;
 343       break;
 344     case 'u':
 345       repertoire_global = arg;
 346       break;
 347     case 'v':
 348       verbose = 1;
 349       break;
 350     default:
 351       return ARGP_ERR_UNKNOWN;
 352     }
 353   return 0;
 354 }
 355
 356
 357 static char *
 358 more_help (int key, const char *text, void *input)
 359 {
 360   char *cp;
 361
 362   switch (key)
 363     {
 364     case ARGP_KEY_HELP_EXTRA:
 365       /* We print some extra information.  */
 366       if (asprintf (&cp, gettext ("\
 367 System's directory for character maps : %s\n\
 368                        repertoire maps: %s\n\
 369                        locale path    : %s\n\
 370 %s"),
 371                     CHARMAP_PATH, REPERTOIREMAP_PATH, LOCALE_PATH, gettext ("\
 372 For bug reporting instructions, please see:\n\
 373 <http://www.gnu.org/software/libc/bugs.html>.\n")) < 0)
 374         return NULL;
 375       return cp;
 376     default:
 377       break;
 378     }
 379   return (char *) text;
 380 }
 381
 382 /* Print the version information.  */
 383 static void
 384 print_version (FILE *stream, struct argp_state *state)
 385 {
 386   fprintf (stream, "localedef (GNU %s) %s\n", PACKAGE, VERSION);
 387   fprintf (stream, gettext ("\
 388 Copyright (C) %s Free Software Foundation, Inc.\n\
 389 This is free software; see the source for copying conditions.  There is NO\n\
 390 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 391 "), "2008");
 392   fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
 393 }
 394
 395
 396 /* The address of this function will be assigned to the hook in the error
 397    functions.  */
 398 static void
 399 error_print (void)
 400 {
 401 }
 402
 403
 404 /* The parameter to localedef describes the output path.  If it does
 405    contain a '/' character it is a relative path.  Otherwise it names the
 406    locale this definition is for.  */
 407 static const char *
 408 construct_output_path (char *path)
 409 {
 410   const char *normal = NULL;
 411   char *result;
 412   char *endp;
 413
 414   if (strchr (path, '/') == NULL)
 415     {
 416       /* This is a system path.  First examine whether the locale name
 417          contains a reference to the codeset.  This should be
 418          normalized.  */
 419       char *startp;
 420       size_t n;
 421
 422       startp = path;
 423       /* We must be prepared for finding a CEN name or a location of
 424          the introducing `.' where it is not possible anymore.  */
 425       while (*startp != '\0' && *startp != '@' && *startp != '.')
 426         ++startp;
 427       if (*startp == '.')
 428         {
 429           /* We found a codeset specification.  Now find the end.  */
 430           endp = ++startp;
 431           while (*endp != '\0' && *endp != '@')
 432             ++endp;
 433
 434           if (endp > startp)
 435             normal = normalize_codeset (startp, endp - startp);
 436         }
 437       else
 438         /* This is to keep gcc quiet.  */
 439         endp = NULL;
 440
 441       /* We put an additional '\0' at the end of the string because at
 442          the end of the function we need another byte for the trailing
 443          '/'.  */
 444       if (normal == NULL)
 445         n = asprintf (&result, "%s%s/%s%c",
 446                       output_prefix ?: "", LOCALEDIR, path, '\0');
 447       else
 448         n = asprintf (&result, "%s%s/%.*s%s%s%c",
 449                       output_prefix ?: "", LOCALEDIR,
 450                       (int) (startp - path), path, normal, endp, '\0');
 451
 452       if (n < 0)
 453         return NULL;
 454
 455       endp = result + n - 1;
 456     }
 457   else
 458     {
 459       /* This is a user path.  Please note the additional byte in the
 460          memory allocation.  */
 461       size_t len = strlen (path) + 1;
 462       result = xmalloc (len + 1);
 463       endp = mempcpy (result, path, len) - 1;
 464
 465       /* If the user specified an output path we cannot add the output
 466          to the archive.  */
 467       no_archive = true;
 468     }
 469
 470   errno = 0;
 471
 472   if (no_archive && euidaccess (result, W_OK) == -1)
 473     /* Perhaps the directory does not exist now.  Try to create it.  */
 474     if (errno == ENOENT)
 475       {
 476         errno = 0;
 477         if (mkdir (result, 0777) < 0)
 478           return NULL;
 479       }
 480
 481   *endp++ = '/';
 482   *endp = '\0';
 483
 484   return result;
 485 }
 486
 487
 488 /* Normalize codeset name.  There is no standard for the codeset
 489    names.  Normalization allows the user to use any of the common
 490    names.  */
 491 static const char *
 492 normalize_codeset (codeset, name_len)
 493      const char *codeset;
 494      size_t name_len;
 495 {
 496   int len = 0;
 497   int only_digit = 1;
 498   char *retval;
 499   char *wp;
 500   size_t cnt;
 501
 502   for (cnt = 0; cnt < name_len; ++cnt)
 503     if (isalnum (codeset[cnt]))
 504       {
 505         ++len;
 506
 507         if (isalpha (codeset[cnt]))
 508           only_digit = 0;
 509       }
 510
 511   retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
 512
 513   if (retval != NULL)
 514     {
 515       if (only_digit)
 516         wp = stpcpy (retval, "iso");
 517       else
 518         wp = retval;
 519
 520       for (cnt = 0; cnt < name_len; ++cnt)
 521         if (isalpha (codeset[cnt]))
 522           *wp++ = tolower (codeset[cnt]);
 523         else if (isdigit (codeset[cnt]))
 524           *wp++ = codeset[cnt];
 525
 526       *wp = '\0';
 527     }
 528
 529   return (const char *) retval;
 530 }
 531
 532
 533 struct localedef_t *
 534 add_to_readlist (int category, const char *name, const char *repertoire_name,
 535                  int generate, struct localedef_t *copy_locale)
 536 {
 537   struct localedef_t *runp = locales;
 538
 539   while (runp != NULL && strcmp (name, runp->name) != 0)
 540     runp = runp->next;
 541
 542   if (runp == NULL)
 543     {
 544       /* Add a new entry at the end.  */
 545       struct localedef_t *newp;
 546
 547       assert (generate == 1);
 548
 549       newp = xcalloc (1, sizeof (struct localedef_t));
 550       newp->name = name;
 551       newp->repertoire_name = repertoire_name;
 552
 553       if (locales == NULL)
 554         runp = locales = newp;
 555       else
 556         {
 557           runp = locales;
 558           while (runp->next != NULL)
 559             runp = runp->next;
 560           runp = runp->next = newp;
 561         }
 562     }
 563
 564   if (generate
 565       && (runp->needed & (1 << category)) != 0
 566       && (runp->avail & (1 << category)) == 0)
 567     WITH_CUR_LOCALE (error (5, 0, _("\
 568 circular dependencies between locale definitions")));
 569
 570   if (copy_locale != NULL)
 571     {
 572       if (runp->categories[category].generic != NULL)
 573         WITH_CUR_LOCALE (error (5, 0, _("\
 574 cannot add already read locale `%s' a second time"), name));
 575       else
 576         runp->categories[category].generic =
 577           copy_locale->categories[category].generic;
 578     }
 579
 580   runp->needed |= 1 << category;
 581
 582   return runp;
 583 }
 584
 585
 586 struct localedef_t *
 587 find_locale (int category, const char *name, const char *repertoire_name,
 588              const struct charmap_t *charmap)
 589 {
 590   struct localedef_t *result;
 591
 592   /* Find the locale, but do not generate it since this would be a bug.  */
 593   result = add_to_readlist (category, name, repertoire_name, 0, NULL);
 594
 595   assert (result != NULL);
 596
 597   if ((result->avail & (1 << category)) == 0
 598       && locfile_read (result, charmap) != 0)
 599     WITH_CUR_LOCALE (error (4, errno, _("\
 600 cannot open locale definition file `%s'"), result->name));
 601
 602   return result;
 603 }
 604
 605
 606 struct localedef_t *
 607 load_locale (int category, const char *name, const char *repertoire_name,
 608              const struct charmap_t *charmap, struct localedef_t *copy_locale)
 609 {
 610   struct localedef_t *result;
 611
 612   /* Generate the locale if it does not exist.  */
 613   result = add_to_readlist (category, name, repertoire_name, 1, copy_locale);
 614
 615   assert (result != NULL);
 616
 617   if ((result->avail & (1 << category)) == 0
 618       && locfile_read (result, charmap) != 0)
 619     WITH_CUR_LOCALE (error (4, errno, _("\
 620 cannot open locale definition file `%s'"), result->name));
 621
 622   return result;
 623 }
 624
 625 static void
 626 turn_on_mcheck (void)
 627 {
 628   /* Enable `malloc' debugging.  */
 629   mcheck (NULL);
 630   /* Use the following line for a more thorough but much slower testing.  */
 631   /* mcheck_pedantic (NULL); */
 632 }
 633
 634 void (*__malloc_initialize_hook) (void) = turn_on_mcheck;