locale/programs/localedef.c

   1 /* Copyright (C) 1995-2002, 2003, 2004 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 Lesser General Public
   7    License as published by the Free Software Foundation; either
   8    version 2.1 of the 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    Lesser General Public License for more details.
  14
  15    You should have received a copy of the GNU Lesser General Public
  16    License along with the GNU C Library; if not, write to the Free
  17    Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
  18    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 <stdbool.h>
  31 #include <stdio.h>
  32 #include <stdlib.h>
  33 #include <string.h>
  34 #include <unistd.h>
  35 #include <error.h>
  36 #include <sys/mman.h>
  37 #include <sys/stat.h>
  38
  39 #include "localedef.h"
  40 #include "charmap.h"
  41 #include "locfile.h"
  42
  43 /* Undefine the following line in the production version.  */
  44 /* #define NDEBUG 1 */
  45 #include <assert.h>
  46
  47
  48 /* List of copied locales.  */
  49 struct copy_def_list_t *copy_list;
  50
  51 /* If this is defined be POSIX conform.  */
  52 int posix_conformance;
  53
  54 /* If not zero give a lot more messages.  */
  55 int verbose;
  56
  57 /* If not zero suppress warnings and information messages.  */
  58 int be_quiet;
  59
  60 /* If not zero, produce old-style hash table instead of 3-level access
  61    tables.  */
  62 int oldstyle_tables;
  63
  64 /* If not zero force output even if warning were issued.  */
  65 static int force_output;
  66
  67 /* Prefix for output files.  */
  68 const char *output_prefix;
  69
  70 /* Name of the character map file.  */
  71 static const char *charmap_file;
  72
  73 /* Name of the locale definition file.  */
  74 static const char *input_file;
  75
  76 /* Name of the repertoire map file.  */
  77 const char *repertoire_global;
  78
  79 /* Name of the locale.alias file.  */
  80 const char *alias_file;
  81
  82 /* List of all locales.  */
  83 static struct localedef_t *locales;
  84
  85 /* If true don't add locale data to archive.  */
  86 bool no_archive;
  87
  88 /* If true add named locales to archive.  */
  89 static bool add_to_archive;
  90
  91 /* If true delete named locales from archive.  */
  92 static bool delete_from_archive;
  93
  94 /* If true replace archive content when adding.  */
  95 static bool replace_archive;
  96
  97 /* If true list archive content.  */
  98 static bool list_archive;
  99
 100 /* Maximum number of retries when opening the locale archive.  */
 101 int max_locarchive_open_retry = 10;
 102
 103
 104 /* Name and version of program.  */
 105 static void print_version (FILE *stream, struct argp_state *state);
 106 void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version;
 107
 108 #define OPT_POSIX 301
 109 #define OPT_QUIET 302
 110 #define OPT_OLDSTYLE 303
 111 #define OPT_PREFIX 304
 112 #define OPT_NO_ARCHIVE 305
 113 #define OPT_ADD_TO_ARCHIVE 306
 114 #define OPT_REPLACE 307
 115 #define OPT_DELETE_FROM_ARCHIVE 308
 116 #define OPT_LIST_ARCHIVE 309
 117
 118 /* Definitions of arguments for argp functions.  */
 119 static const struct argp_option options[] =
 120 {
 121   { NULL, 0, NULL, 0, N_("Input Files:") },
 122   { "charmap", 'f', "FILE", 0,
 123     N_("Symbolic character names defined in FILE") },
 124   { "inputfile", 'i', "FILE", 0, N_("Source definitions are found in FILE") },
 125   { "repertoire-map", 'u', "FILE", 0,
 126     N_("FILE contains mapping from symbolic names to UCS4 values") },
 127
 128   { NULL, 0, NULL, 0, N_("Output control:") },
 129   { "force", 'c', NULL, 0,
 130     N_("Create output even if warning messages were issued") },
 131   { "old-style", OPT_OLDSTYLE, NULL, 0, N_("Create old-style tables") },
 132   { "prefix", OPT_PREFIX, "PATH", 0, N_("Optional output file prefix") },
 133   { "posix", OPT_POSIX, NULL, 0, N_("Be strictly POSIX conform") },
 134   { "quiet", OPT_QUIET, NULL, 0,
 135     N_("Suppress warnings and information messages") },
 136   { "verbose", 'v', NULL, 0, N_("Print more messages") },
 137   { NULL, 0, NULL, 0, N_("Archive control:") },
 138   { "no-archive", OPT_NO_ARCHIVE, NULL, 0,
 139     N_("Don't add new data to archive") },
 140   { "add-to-archive", OPT_ADD_TO_ARCHIVE, NULL, 0,
 141     N_("Add locales named by parameters to archive") },
 142   { "replace", OPT_REPLACE, NULL, 0, N_("Replace existing archive content") },
 143   { "delete-from-archive", OPT_DELETE_FROM_ARCHIVE, NULL, 0,
 144     N_("Remove locales named by parameters from archive") },
 145   { "list-archive", OPT_LIST_ARCHIVE, NULL, 0, N_("List content of archive") },
 146   { "alias-file", 'A', "FILE", 0,
 147     N_("locale.alias file to consult when making archive")},
 148   { NULL, 0, NULL, 0, NULL }
 149 };
 150
 151 /* Short description of program.  */
 152 static const char doc[] = N_("Compile locale specification");
 153
 154 /* Strings for arguments in help texts.  */
 155 static const char args_doc[] = N_("\
 156 NAME\n\
 157 [--add-to-archive|--delete-from-archive] FILE...\n\
 158 --list-archive [FILE]");
 159
 160 /* Prototype for option handler.  */
 161 static error_t parse_opt (int key, char *arg, struct argp_state *state);
 162
 163 /* Function to print some extra text in the help message.  */
 164 static char *more_help (int key, const char *text, void *input);
 165
 166 /* Data structure to communicate with argp functions.  */
 167 static struct argp argp =
 168 {
 169   options, parse_opt, args_doc, doc, NULL, more_help
 170 };
 171
 172
 173 /* Prototypes for global functions.  */
 174 extern void *xmalloc (size_t __n);
 175
 176 /* Prototypes for local functions.  */
 177 static void error_print (void);
 178 static const char *construct_output_path (char *path);
 179 static const char *normalize_codeset (const char *codeset, size_t name_len);
 180
 181
 182 int
 183 main (int argc, char *argv[])
 184 {
 185   const char *output_path;
 186   int cannot_write_why;
 187   struct charmap_t *charmap;
 188   struct localedef_t global;
 189   int remaining;
 190
 191   /* Set initial values for global variables.  */
 192   copy_list = NULL;
 193   posix_conformance = getenv ("POSIXLY_CORRECT") != NULL;
 194   error_print_progname = error_print;
 195
 196   /* Set locale.  Do not set LC_ALL because the other categories must
 197      not be affected (according to POSIX.2).  */
 198   setlocale (LC_MESSAGES, "");
 199   setlocale (LC_CTYPE, "");
 200
 201   /* Initialize the message catalog.  */
 202   textdomain (_libc_intl_domainname);
 203
 204   /* Parse and process arguments.  */
 205   argp_err_exit_status = 4;
 206   argp_parse (&argp, argc, argv, 0, &remaining, NULL);
 207
 208   /* Handle a few special cases.  */
 209   if (list_archive)
 210     show_archive_content (verbose);
 211   if (add_to_archive)
 212     return add_locales_to_archive (argc - remaining, &argv[remaining],
 213                                    replace_archive);
 214   if (delete_from_archive)
 215     return delete_locales_from_archive (argc - remaining, &argv[remaining]);
 216
 217   /* POSIX.2 requires to be verbose about missing characters in the
 218      character map.  */
 219   verbose |= posix_conformance;
 220
 221   if (argc - remaining != 1)
 222     {
 223       /* We need exactly one non-option parameter.  */
 224       argp_help (&argp, stdout, ARGP_HELP_SEE | ARGP_HELP_EXIT_ERR,
 225                  program_invocation_short_name);
 226       exit (4);
 227     }
 228
 229   /* The parameter describes the output path of the constructed files.
 230      If the described files cannot be written return a NULL pointer.  */
 231   output_path  = construct_output_path (argv[remaining]);
 232   if (output_path == NULL && ! no_archive)
 233     error (4, errno, _("cannot create directory for output files"));
 234   cannot_write_why = errno;
 235
 236   /* Now that the parameters are processed we have to reset the local
 237      ctype locale.  (P1003.2 4.35.5.2)  */
 238   setlocale (LC_CTYPE, "POSIX");
 239
 240   /* Look whether the system really allows locale definitions.  POSIX
 241      defines error code 3 for this situation so I think it must be
 242      a fatal error (see P1003.2 4.35.8).  */
 243   if (sysconf (_SC_2_LOCALEDEF) < 0)
 244     WITH_CUR_LOCALE (error (3, 0, _("\
 245 FATAL: system does not define `_POSIX2_LOCALEDEF'")));
 246
 247   /* Process charmap file.  */
 248   charmap = charmap_read (charmap_file, verbose, be_quiet, 1);
 249
 250   /* Add the first entry in the locale list.  */
 251   memset (&global, '\0', sizeof (struct localedef_t));
 252   global.name = input_file ?: "/dev/stdin";
 253   global.needed = ALL_LOCALES;
 254   locales = &global;
 255
 256   /* Now read the locale file.  */
 257   if (locfile_read (&global, charmap) != 0)
 258     WITH_CUR_LOCALE (error (4, errno, _("\
 259 cannot open locale definition file `%s'"), input_file));
 260
 261   /* Perhaps we saw some `copy' instructions.  */
 262   while (1)
 263     {
 264       struct localedef_t *runp = locales;
 265
 266       while (runp != NULL && (runp->needed & runp->avail) == runp->needed)
 267         runp = runp->next;
 268
 269       if (runp == NULL)
 270         /* Everything read.  */
 271         break;
 272
 273       if (locfile_read (runp, charmap) != 0)
 274         WITH_CUR_LOCALE (error (4, errno, _("\
 275 cannot open locale definition file `%s'"), runp->name));
 276     }
 277
 278   /* Check the categories we processed in source form.  */
 279   check_all_categories (locales, charmap);
 280
 281   /* We are now able to write the data files.  If warning were given we
 282      do it only if it is explicitly requested (--force).  */
 283   if (error_message_count == 0 || force_output != 0)
 284     {
 285       if (cannot_write_why != 0)
 286         WITH_CUR_LOCALE (error (4, cannot_write_why, _("\
 287 cannot write output files to `%s'"), output_path));
 288       else
 289         write_all_categories (locales, charmap, argv[remaining], output_path);
 290     }
 291   else
 292     WITH_CUR_LOCALE (error (4, 0, _("\
 293 no output file produced because warning were issued")));
 294
 295   /* This exit status is prescribed by POSIX.2 4.35.7.  */
 296   exit (error_message_count != 0);
 297 }
 298
 299
 300 /* Handle program arguments.  */
 301 static error_t
 302 parse_opt (int key, char *arg, struct argp_state *state)
 303 {
 304   switch (key)
 305     {
 306     case OPT_QUIET:
 307       be_quiet = 1;
 308       break;
 309     case OPT_POSIX:
 310       posix_conformance = 1;
 311       break;
 312     case OPT_OLDSTYLE:
 313       oldstyle_tables = 1;
 314       break;
 315     case OPT_PREFIX:
 316       output_prefix = arg;
 317       break;
 318     case OPT_NO_ARCHIVE:
 319       no_archive = true;
 320       break;
 321     case OPT_ADD_TO_ARCHIVE:
 322       add_to_archive = true;
 323       break;
 324     case OPT_REPLACE:
 325       replace_archive = true;
 326       break;
 327     case OPT_DELETE_FROM_ARCHIVE:
 328       delete_from_archive = true;
 329       break;
 330     case OPT_LIST_ARCHIVE:
 331       list_archive = true;
 332       break;
 333     case 'c':
 334       force_output = 1;
 335       break;
 336     case 'f':
 337       charmap_file = arg;
 338       break;
 339     case 'A':
 340       alias_file = arg;
 341       break;
 342     case 'i':
 343       input_file = arg;
 344       break;
 345     case 'u':
 346       repertoire_global = arg;
 347       break;
 348     case 'v':
 349       verbose = 1;
 350       break;
 351     default:
 352       return ARGP_ERR_UNKNOWN;
 353     }
 354   return 0;
 355 }
 356
 357
 358 static char *
 359 more_help (int key, const char *text, void *input)
 360 {
 361   char *cp;
 362
 363   switch (key)
 364     {
 365     case ARGP_KEY_HELP_EXTRA:
 366       /* We print some extra information.  */
 367       if (asprintf (&cp, gettext ("\
 368 System's directory for character maps : %s\n\
 369                        repertoire maps: %s\n\
 370                        locale path    : %s\n\
 371 %s"),
 372                     CHARMAP_PATH, REPERTOIREMAP_PATH, LOCALE_PATH, gettext ("\
 373 For bug reporting instructions, please see:\n\
 374 <http://www.gnu.org/software/libc/bugs.html>.\n")) < 0)
 375         return NULL;
 376       return cp;
 377     default:
 378       break;
 379     }
 380   return (char *) text;
 381 }
 382
 383 /* Print the version information.  */
 384 static void
 385 print_version (FILE *stream, struct argp_state *state)
 386 {
 387   fprintf (stream, "localedef (GNU %s) %s\n", PACKAGE, VERSION);
 388   fprintf (stream, gettext ("\
 389 Copyright (C) %s Free Software Foundation, Inc.\n\
 390 This is free software; see the source for copying conditions.  There is NO\n\
 391 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 392 "), "2004");
 393   fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
 394 }
 395
 396
 397 /* The address of this function will be assigned to the hook in the error
 398    functions.  */
 399 static void
 400 error_print (void)
 401 {
 402 }
 403
 404
 405 /* The parameter to localedef describes the output path.  If it does
 406    contain a '/' character it is a relative path.  Otherwise it names the
 407    locale this definition is for.  */
 408 static const char *
 409 construct_output_path (char *path)
 410 {
 411   const char *normal = NULL;
 412   char *result;
 413   char *endp;
 414
 415   if (strchr (path, '/') == NULL)
 416     {
 417       /* This is a system path.  First examine whether the locale name
 418          contains a reference to the codeset.  This should be
 419          normalized.  */
 420       char *startp;
 421       size_t n;
 422
 423       startp = path;
 424       /* We must be prepared for finding a CEN name or a location of
 425          the introducing `.' where it is not possible anymore.  */
 426       while (*startp != '\0' && *startp != '@' && *startp != '.')
 427         ++startp;
 428       if (*startp == '.')
 429         {
 430           /* We found a codeset specification.  Now find the end.  */
 431           endp = ++startp;
 432           while (*endp != '\0' && *endp != '@')
 433             ++endp;
 434
 435           if (endp > startp)
 436             normal = normalize_codeset (startp, endp - startp);
 437         }
 438       else
 439         /* This is to keep gcc quiet.  */
 440         endp = NULL;
 441
 442       /* We put an additional '\0' at the end of the string because at
 443          the end of the function we need another byte for the trailing
 444          '/'.  */
 445       if (normal == NULL)
 446         n = asprintf (&result, "%s%s/%s%c",
 447                       output_prefix ?: "", LOCALEDIR, path, '\0');
 448       else
 449         n = asprintf (&result, "%s%s/%.*s%s%s%c",
 450                       output_prefix ?: "", LOCALEDIR,
 451                       (int) (startp - path), path, normal, endp, '\0');
 452
 453       if (n < 0)
 454         return NULL;
 455
 456       endp = result + n - 1;
 457     }
 458   else
 459     {
 460       /* This is a user path.  Please note the additional byte in the
 461          memory allocation.  */
 462       size_t len = strlen (path) + 1;
 463       result = xmalloc (len + 1);
 464       endp = mempcpy (result, path, len) - 1;
 465
 466       /* If the user specified an output path we cannot add the output
 467          to the archive.  */
 468       no_archive = true;
 469     }
 470
 471   errno = 0;
 472
 473   if (no_archive && euidaccess (result, W_OK) == -1)
 474     /* Perhaps the directory does not exist now.  Try to create it.  */
 475     if (errno == ENOENT)
 476       {
 477         errno = 0;
 478         if (mkdir (result, 0777) < 0)
 479           return NULL;
 480       }
 481
 482   *endp++ = '/';
 483   *endp = '\0';
 484
 485   return result;
 486 }
 487
 488
 489 /* Normalize codeset name.  There is no standard for the codeset
 490    names.  Normalization allows the user to use any of the common
 491    names.  */
 492 static const char *
 493 normalize_codeset (codeset, name_len)
 494      const char *codeset;
 495      size_t name_len;
 496 {
 497   int len = 0;
 498   int only_digit = 1;
 499   char *retval;
 500   char *wp;
 501   size_t cnt;
 502
 503   for (cnt = 0; cnt < name_len; ++cnt)
 504     if (isalnum (codeset[cnt]))
 505       {
 506         ++len;
 507
 508         if (isalpha (codeset[cnt]))
 509           only_digit = 0;
 510       }
 511
 512   retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
 513
 514   if (retval != NULL)
 515     {
 516       if (only_digit)
 517         wp = stpcpy (retval, "iso");
 518       else
 519         wp = retval;
 520
 521       for (cnt = 0; cnt < name_len; ++cnt)
 522         if (isalpha (codeset[cnt]))
 523           *wp++ = tolower (codeset[cnt]);
 524         else if (isdigit (codeset[cnt]))
 525           *wp++ = codeset[cnt];
 526
 527       *wp = '\0';
 528     }
 529
 530   return (const char *) retval;
 531 }
 532
 533
 534 struct localedef_t *
 535 add_to_readlist (int locale, const char *name, const char *repertoire_name,
 536                  int generate, struct localedef_t *copy_locale)
 537 {
 538   struct localedef_t *runp = locales;
 539
 540   while (runp != NULL && strcmp (name, runp->name) != 0)
 541     runp = runp->next;
 542
 543   if (runp == NULL)
 544     {
 545       /* Add a new entry at the end.  */
 546       struct localedef_t *newp;
 547
 548       assert (generate == 1);
 549
 550       newp = xcalloc (1, sizeof (struct localedef_t));
 551       newp->name = name;
 552       newp->repertoire_name = repertoire_name;
 553
 554       if (locales == NULL)
 555         runp = locales = newp;
 556       else
 557         {
 558           runp = locales;
 559           while (runp->next != NULL)
 560             runp = runp->next;
 561           runp = runp->next = newp;
 562         }
 563     }
 564
 565   if (generate
 566       && (runp->needed & (1 << locale)) != 0
 567       && (runp->avail & (1 << locale)) == 0)
 568     WITH_CUR_LOCALE (error (5, 0, _("\
 569 circular dependencies between locale definitions")));
 570
 571   if (copy_locale != NULL)
 572     {
 573       if (runp->categories[locale].generic != NULL)
 574         WITH_CUR_LOCALE (error (5, 0, _("\
 575 cannot add already read locale `%s' a second time"), name));
 576       else
 577         runp->categories[locale].generic =
 578           copy_locale->categories[locale].generic;
 579     }
 580
 581   runp->needed |= 1 << locale;
 582
 583   return runp;
 584 }
 585
 586
 587 struct localedef_t *
 588 find_locale (int locale, const char *name, const char *repertoire_name,
 589              const struct charmap_t *charmap)
 590 {
 591   struct localedef_t *result;
 592
 593   /* Find the locale, but do not generate it since this would be a bug.  */
 594   result = add_to_readlist (locale, name, repertoire_name, 0, NULL);
 595
 596   assert (result != NULL);
 597
 598   if ((result->avail & (1 << locale)) == 0
 599       && locfile_read (result, charmap) != 0)
 600     WITH_CUR_LOCALE (error (4, errno, _("\
 601 cannot open locale definition file `%s'"), result->name));
 602
 603   return result;
 604 }
 605
 606
 607 struct localedef_t *
 608 load_locale (int locale, const char *name, const char *repertoire_name,
 609              const struct charmap_t *charmap, struct localedef_t *copy_locale)
 610 {
 611   struct localedef_t *result;
 612
 613   /* Generate the locale if it does not exist.  */
 614   result = add_to_readlist (locale, name, repertoire_name, 1, copy_locale);
 615
 616   assert (result != NULL);
 617
 618   if ((result->avail & (1 << locale)) == 0
 619       && locfile_read (result, charmap) != 0)
 620     WITH_CUR_LOCALE (error (4, errno, _("\
 621 cannot open locale definition file `%s'"), result->name));
 622
 623   return result;
 624 }
 625
 626 static void
 627 turn_on_mcheck (void)
 628 {
 629   /* Enable `malloc' debugging.  */
 630   mcheck (NULL);
 631   /* Use the following line for a more thorough but much slower testing.  */
 632   /* mcheck_pedantic (NULL); */
 633 }
 634
 635 void (*__malloc_initialize_hook) (void) = turn_on_mcheck;