locale/programs/localedef.c

   1 /* Copyright (C) 1995-2014 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, see <http://www.gnu.org/licenses/>.  */
  17
  18 #ifdef HAVE_CONFIG_H
  19 # include <config.h>
  20 #endif
  21
  22 #include <argp.h>
  23 #include <errno.h>
  24 #include <fcntl.h>
  25 #include <libintl.h>
  26 #include <locale.h>
  27 #include <mcheck.h>
  28 #include <stdbool.h>
  29 #include <stdio.h>
  30 #include <stdlib.h>
  31 #include <string.h>
  32 #include <unistd.h>
  33 #include <error.h>
  34 #include <sys/mman.h>
  35 #include <sys/stat.h>
  36
  37 #include "localedef.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
  59    tables.  */
  60 int oldstyle_tables;
  61
  62 /* If not zero force output even if warning were issued.  */
  63 static int force_output;
  64
  65 /* Prefix for output files.  */
  66 const char *output_prefix;
  67
  68 /* Name of the character map file.  */
  69 static const char *charmap_file;
  70
  71 /* Name of the locale definition file.  */
  72 static const char *input_file;
  73
  74 /* Name of the repertoire map file.  */
  75 const char *repertoire_global;
  76
  77 /* Name of the locale.alias file.  */
  78 const char *alias_file;
  79
  80 /* List of all locales.  */
  81 static struct localedef_t *locales;
  82
  83 /* If true don't add locale data to archive.  */
  84 bool no_archive;
  85
  86 /* If true add named locales to archive.  */
  87 static bool add_to_archive;
  88
  89 /* If true delete named locales from archive.  */
  90 static bool delete_from_archive;
  91
  92 /* If true replace archive content when adding.  */
  93 static bool replace_archive;
  94
  95 /* If true list archive content.  */
  96 static bool list_archive;
  97
  98 /* Maximum number of retries when opening the locale archive.  */
  99 int max_locarchive_open_retry = 10;
 100
 101
 102 /* Name and version of program.  */
 103 static void print_version (FILE *stream, struct argp_state *state);
 104 void (*argp_program_version_hook) (FILE *, struct argp_state *) = print_version;
 105
 106 #define OPT_POSIX 301
 107 #define OPT_QUIET 302
 108 #define OPT_OLDSTYLE 303
 109 #define OPT_PREFIX 304
 110 #define OPT_NO_ARCHIVE 305
 111 #define OPT_ADD_TO_ARCHIVE 306
 112 #define OPT_REPLACE 307
 113 #define OPT_DELETE_FROM_ARCHIVE 308
 114 #define OPT_LIST_ARCHIVE 309
 115 #define OPT_LITTLE_ENDIAN 400
 116 #define OPT_BIG_ENDIAN 401
 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', N_("FILE"), 0,
 123     N_("Symbolic character names defined in FILE") },
 124   { "inputfile", 'i', N_("FILE"), 0,
 125     N_("Source definitions are found in FILE") },
 126   { "repertoire-map", 'u', N_("FILE"), 0,
 127     N_("FILE contains mapping from symbolic names to UCS4 values") },
 128
 129   { NULL, 0, NULL, 0, N_("Output control:") },
 130   { "force", 'c', NULL, 0,
 131     N_("Create output even if warning messages were issued") },
 132   { "old-style", OPT_OLDSTYLE, NULL, 0, N_("Create old-style tables") },
 133   { "prefix", OPT_PREFIX, N_("PATH"), 0, N_("Optional output file prefix") },
 134   { "posix", OPT_POSIX, NULL, 0, N_("Strictly conform to POSIX") },
 135   { "quiet", OPT_QUIET, NULL, 0,
 136     N_("Suppress warnings and information messages") },
 137   { "verbose", 'v', NULL, 0, N_("Print more messages") },
 138   { NULL, 0, NULL, 0, N_("Archive control:") },
 139   { "no-archive", OPT_NO_ARCHIVE, NULL, 0,
 140     N_("Don't add new data to archive") },
 141   { "add-to-archive", OPT_ADD_TO_ARCHIVE, NULL, 0,
 142     N_("Add locales named by parameters to archive") },
 143   { "replace", OPT_REPLACE, NULL, 0, N_("Replace existing archive content") },
 144   { "delete-from-archive", OPT_DELETE_FROM_ARCHIVE, NULL, 0,
 145     N_("Remove locales named by parameters from archive") },
 146   { "list-archive", OPT_LIST_ARCHIVE, NULL, 0, N_("List content of archive") },
 147   { "alias-file", 'A', N_("FILE"), 0,
 148     N_("locale.alias file to consult when making archive")},
 149   { "little-endian", OPT_LITTLE_ENDIAN, NULL, 0,
 150     N_("Generate little-endian output") },
 151   { "big-endian", OPT_BIG_ENDIAN, NULL, 0,
 152     N_("Generate big-endian output") },
 153   { NULL, 0, NULL, 0, NULL }
 154 };
 155
 156 /* Short description of program.  */
 157 static const char doc[] = N_("Compile locale specification");
 158
 159 /* Strings for arguments in help texts.  */
 160 static const char args_doc[] = N_("\
 161 NAME\n\
 162 [--add-to-archive|--delete-from-archive] FILE...\n\
 163 --list-archive [FILE]");
 164
 165 /* Prototype for option handler.  */
 166 static error_t parse_opt (int key, char *arg, struct argp_state *state);
 167
 168 /* Function to print some extra text in the help message.  */
 169 static char *more_help (int key, const char *text, void *input);
 170
 171 /* Data structure to communicate with argp functions.  */
 172 static struct argp argp =
 173 {
 174   options, parse_opt, args_doc, doc, NULL, more_help
 175 };
 176
 177
 178 /* Prototypes for local functions.  */
 179 static void error_print (void);
 180 static const char *construct_output_path (char *path);
 181 static const char *normalize_codeset (const char *codeset, size_t name_len);
 182
 183
 184 int
 185 main (int argc, char *argv[])
 186 {
 187   const char *output_path;
 188   int cannot_write_why;
 189   struct charmap_t *charmap;
 190   struct localedef_t global;
 191   int remaining;
 192
 193   /* Set initial values for global variables.  */
 194   copy_list = NULL;
 195   posix_conformance = getenv ("POSIXLY_CORRECT") != NULL;
 196   error_print_progname = error_print;
 197
 198   /* Set locale.  Do not set LC_ALL because the other categories must
 199      not be affected (according to POSIX.2).  */
 200   setlocale (LC_MESSAGES, "");
 201   setlocale (LC_CTYPE, "");
 202
 203   /* Initialize the message catalog.  */
 204   textdomain (_libc_intl_domainname);
 205
 206   /* Parse and process arguments.  */
 207   argp_err_exit_status = 4;
 208   argp_parse (&argp, argc, argv, 0, &remaining, NULL);
 209
 210   /* Handle a few special cases.  */
 211   if (list_archive)
 212     show_archive_content (remaining > 1 ? argv[remaining] : NULL, verbose);
 213   if (add_to_archive)
 214     return add_locales_to_archive (argc - remaining, &argv[remaining],
 215                                    replace_archive);
 216   if (delete_from_archive)
 217     return delete_locales_from_archive (argc - remaining, &argv[remaining]);
 218
 219   /* POSIX.2 requires to be verbose about missing characters in the
 220      character map.  */
 221   verbose |= posix_conformance;
 222
 223   if (argc - remaining != 1)
 224     {
 225       /* We need exactly one non-option parameter.  */
 226       argp_help (&argp, stdout, ARGP_HELP_SEE | ARGP_HELP_EXIT_ERR,
 227                  program_invocation_short_name);
 228       exit (4);
 229     }
 230
 231   /* The parameter describes the output path of the constructed files.
 232      If the described files cannot be written return a NULL pointer.  */
 233   output_path  = construct_output_path (argv[remaining]);
 234   if (output_path == NULL && ! no_archive)
 235     error (4, errno, _("cannot create directory for output files"));
 236   cannot_write_why = errno;
 237
 238   /* Now that the parameters are processed we have to reset the local
 239      ctype locale.  (P1003.2 4.35.5.2)  */
 240   setlocale (LC_CTYPE, "POSIX");
 241
 242   /* Look whether the system really allows locale definitions.  POSIX
 243      defines error code 3 for this situation so I think it must be
 244      a fatal error (see P1003.2 4.35.8).  */
 245   if (sysconf (_SC_2_LOCALEDEF) < 0)
 246     WITH_CUR_LOCALE (error (3, 0, _("\
 247 FATAL: system does not define `_POSIX2_LOCALEDEF'")));
 248
 249   /* Process charmap file.  */
 250   charmap = charmap_read (charmap_file, verbose, 1, be_quiet, 1);
 251
 252   /* Add the first entry in the locale list.  */
 253   memset (&global, '\0', sizeof (struct localedef_t));
 254   global.name = input_file ?: "/dev/stdin";
 255   global.needed = ALL_LOCALES;
 256   locales = &global;
 257
 258   /* Now read the locale file.  */
 259   if (locfile_read (&global, charmap) != 0)
 260     WITH_CUR_LOCALE (error (4, errno, _("\
 261 cannot open locale definition file `%s'"), input_file));
 262
 263   /* Perhaps we saw some `copy' instructions.  */
 264   while (1)
 265     {
 266       struct localedef_t *runp = locales;
 267
 268       while (runp != NULL && (runp->needed & runp->avail) == runp->needed)
 269         runp = runp->next;
 270
 271       if (runp == NULL)
 272         /* Everything read.  */
 273         break;
 274
 275       if (locfile_read (runp, charmap) != 0)
 276         WITH_CUR_LOCALE (error (4, errno, _("\
 277 cannot open locale definition file `%s'"), runp->name));
 278     }
 279
 280   /* Check the categories we processed in source form.  */
 281   check_all_categories (locales, charmap);
 282
 283   /* We are now able to write the data files.  If warning were given we
 284      do it only if it is explicitly requested (--force).  */
 285   if (error_message_count == 0 || force_output != 0)
 286     {
 287       if (cannot_write_why != 0)
 288         WITH_CUR_LOCALE (error (4, cannot_write_why, _("\
 289 cannot write output files to `%s'"), output_path));
 290       else
 291         write_all_categories (locales, charmap, argv[remaining], output_path);
 292     }
 293   else
 294     WITH_CUR_LOCALE (error (4, 0, _("\
 295 no output file produced because warnings were issued")));
 296
 297   /* This exit status is prescribed by POSIX.2 4.35.7.  */
 298   exit (error_message_count != 0);
 299 }
 300
 301
 302 /* Handle program arguments.  */
 303 static error_t
 304 parse_opt (int key, char *arg, struct argp_state *state)
 305 {
 306   switch (key)
 307     {
 308     case OPT_QUIET:
 309       be_quiet = 1;
 310       break;
 311     case OPT_POSIX:
 312       posix_conformance = 1;
 313       break;
 314     case OPT_OLDSTYLE:
 315       oldstyle_tables = 1;
 316       break;
 317     case OPT_PREFIX:
 318       output_prefix = arg;
 319       break;
 320     case OPT_NO_ARCHIVE:
 321       no_archive = true;
 322       break;
 323     case OPT_ADD_TO_ARCHIVE:
 324       add_to_archive = true;
 325       break;
 326     case OPT_REPLACE:
 327       replace_archive = true;
 328       break;
 329     case OPT_DELETE_FROM_ARCHIVE:
 330       delete_from_archive = true;
 331       break;
 332     case OPT_LIST_ARCHIVE:
 333       list_archive = true;
 334       break;
 335     case OPT_LITTLE_ENDIAN:
 336       set_big_endian (false);
 337       break;
 338     case OPT_BIG_ENDIAN:
 339       set_big_endian (true);
 340       break;
 341     case 'c':
 342       force_output = 1;
 343       break;
 344     case 'f':
 345       charmap_file = arg;
 346       break;
 347     case 'A':
 348       alias_file = arg;
 349       break;
 350     case 'i':
 351       input_file = arg;
 352       break;
 353     case 'u':
 354       repertoire_global = arg;
 355       break;
 356     case 'v':
 357       verbose = 1;
 358       break;
 359     default:
 360       return ARGP_ERR_UNKNOWN;
 361     }
 362   return 0;
 363 }
 364
 365
 366 static char *
 367 more_help (int key, const char *text, void *input)
 368 {
 369   char *cp;
 370   char *tp;
 371
 372   switch (key)
 373     {
 374     case ARGP_KEY_HELP_EXTRA:
 375       /* We print some extra information.  */
 376       if (asprintf (&tp, gettext ("\
 377 For bug reporting instructions, please see:\n\
 378 %s.\n"), REPORT_BUGS_TO) < 0)
 379         return NULL;
 380       if (asprintf (&cp, gettext ("\
 381 System's directory for character maps : %s\n\
 382                        repertoire maps: %s\n\
 383                        locale path    : %s\n\
 384 %s"),
 385                     CHARMAP_PATH, REPERTOIREMAP_PATH, LOCALE_PATH, tp) < 0)
 386         {
 387           free (tp);
 388           return NULL;
 389         }
 390       return cp;
 391     default:
 392       break;
 393     }
 394   return (char *) text;
 395 }
 396
 397 /* Print the version information.  */
 398 static void
 399 print_version (FILE *stream, struct argp_state *state)
 400 {
 401   fprintf (stream, "localedef %s%s\n", PKGVERSION, VERSION);
 402   fprintf (stream, gettext ("\
 403 Copyright (C) %s Free Software Foundation, Inc.\n\
 404 This is free software; see the source for copying conditions.  There is NO\n\
 405 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 406 "), "2013");
 407   fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
 408 }
 409
 410
 411 /* The address of this function will be assigned to the hook in the error
 412    functions.  */
 413 static void
 414 error_print (void)
 415 {
 416 }
 417
 418
 419 /* The parameter to localedef describes the output path.  If it does
 420    contain a '/' character it is a relative path.  Otherwise it names the
 421    locale this definition is for.  */
 422 static const char *
 423 construct_output_path (char *path)
 424 {
 425   const char *normal = NULL;
 426   char *result;
 427   char *endp;
 428
 429   if (strchr (path, '/') == NULL)
 430     {
 431       /* This is a system path.  First examine whether the locale name
 432          contains a reference to the codeset.  This should be
 433          normalized.  */
 434       char *startp;
 435
 436       startp = path;
 437       /* We must be prepared for finding a CEN name or a location of
 438          the introducing `.' where it is not possible anymore.  */
 439       while (*startp != '\0' && *startp != '@' && *startp != '.')
 440         ++startp;
 441       if (*startp == '.')
 442         {
 443           /* We found a codeset specification.  Now find the end.  */
 444           endp = ++startp;
 445           while (*endp != '\0' && *endp != '@')
 446             ++endp;
 447
 448           if (endp > startp)
 449             normal = normalize_codeset (startp, endp - startp);
 450         }
 451       else
 452         /* This is to keep gcc quiet.  */
 453         endp = NULL;
 454
 455       /* We put an additional '\0' at the end of the string because at
 456          the end of the function we need another byte for the trailing
 457          '/'.  */
 458       ssize_t n;
 459       if (normal == NULL)
 460         n = asprintf (&result, "%s%s/%s%c",
 461                       output_prefix ?: "", LOCALEDIR, path, '\0');
 462       else
 463         n = asprintf (&result, "%s%s/%.*s%s%s%c",
 464                       output_prefix ?: "", LOCALEDIR,
 465                       (int) (startp - path), path, normal, endp, '\0');
 466
 467       if (n < 0)
 468         return NULL;
 469
 470       endp = result + n - 1;
 471     }
 472   else
 473     {
 474       /* This is a user path.  Please note the additional byte in the
 475          memory allocation.  */
 476       size_t len = strlen (path) + 1;
 477       result = xmalloc (len + 1);
 478       endp = mempcpy (result, path, len) - 1;
 479
 480       /* If the user specified an output path we cannot add the output
 481          to the archive.  */
 482       no_archive = true;
 483     }
 484
 485   errno = 0;
 486
 487   if (no_archive && euidaccess (result, W_OK) == -1)
 488     /* Perhaps the directory does not exist now.  Try to create it.  */
 489     if (errno == ENOENT)
 490       {
 491         errno = 0;
 492         if (mkdir (result, 0777) < 0)
 493           return NULL;
 494       }
 495
 496   *endp++ = '/';
 497   *endp = '\0';
 498
 499   return result;
 500 }
 501
 502
 503 /* Normalize codeset name.  There is no standard for the codeset
 504    names.  Normalization allows the user to use any of the common
 505    names.  */
 506 static const char *
 507 normalize_codeset (codeset, name_len)
 508      const char *codeset;
 509      size_t name_len;
 510 {
 511   int len = 0;
 512   int only_digit = 1;
 513   char *retval;
 514   char *wp;
 515   size_t cnt;
 516
 517   for (cnt = 0; cnt < name_len; ++cnt)
 518     if (isalnum (codeset[cnt]))
 519       {
 520         ++len;
 521
 522         if (isalpha (codeset[cnt]))
 523           only_digit = 0;
 524       }
 525
 526   retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
 527
 528   if (retval != NULL)
 529     {
 530       if (only_digit)
 531         wp = stpcpy (retval, "iso");
 532       else
 533         wp = retval;
 534
 535       for (cnt = 0; cnt < name_len; ++cnt)
 536         if (isalpha (codeset[cnt]))
 537           *wp++ = tolower (codeset[cnt]);
 538         else if (isdigit (codeset[cnt]))
 539           *wp++ = codeset[cnt];
 540
 541       *wp = '\0';
 542     }
 543
 544   return (const char *) retval;
 545 }
 546
 547
 548 struct localedef_t *
 549 add_to_readlist (int category, const char *name, const char *repertoire_name,
 550                  int generate, struct localedef_t *copy_locale)
 551 {
 552   struct localedef_t *runp = locales;
 553
 554   while (runp != NULL && strcmp (name, runp->name) != 0)
 555     runp = runp->next;
 556
 557   if (runp == NULL)
 558     {
 559       /* Add a new entry at the end.  */
 560       struct localedef_t *newp;
 561
 562       assert (generate == 1);
 563
 564       newp = xcalloc (1, sizeof (struct localedef_t));
 565       newp->name = name;
 566       newp->repertoire_name = repertoire_name;
 567
 568       if (locales == NULL)
 569         runp = locales = newp;
 570       else
 571         {
 572           runp = locales;
 573           while (runp->next != NULL)
 574             runp = runp->next;
 575           runp = runp->next = newp;
 576         }
 577     }
 578
 579   if (generate
 580       && (runp->needed & (1 << category)) != 0
 581       && (runp->avail & (1 << category)) == 0)
 582     WITH_CUR_LOCALE (error (5, 0, _("\
 583 circular dependencies between locale definitions")));
 584
 585   if (copy_locale != NULL)
 586     {
 587       if (runp->categories[category].generic != NULL)
 588         WITH_CUR_LOCALE (error (5, 0, _("\
 589 cannot add already read locale `%s' a second time"), name));
 590       else
 591         runp->categories[category].generic =
 592           copy_locale->categories[category].generic;
 593     }
 594
 595   runp->needed |= 1 << category;
 596
 597   return runp;
 598 }
 599
 600
 601 struct localedef_t *
 602 find_locale (int category, const char *name, const char *repertoire_name,
 603              const struct charmap_t *charmap)
 604 {
 605   struct localedef_t *result;
 606
 607   /* Find the locale, but do not generate it since this would be a bug.  */
 608   result = add_to_readlist (category, name, repertoire_name, 0, NULL);
 609
 610   assert (result != NULL);
 611
 612   if ((result->avail & (1 << category)) == 0
 613       && locfile_read (result, charmap) != 0)
 614     WITH_CUR_LOCALE (error (4, errno, _("\
 615 cannot open locale definition file `%s'"), result->name));
 616
 617   return result;
 618 }
 619
 620
 621 struct localedef_t *
 622 load_locale (int category, const char *name, const char *repertoire_name,
 623              const struct charmap_t *charmap, struct localedef_t *copy_locale)
 624 {
 625   struct localedef_t *result;
 626
 627   /* Generate the locale if it does not exist.  */
 628   result = add_to_readlist (category, name, repertoire_name, 1, copy_locale);
 629
 630   assert (result != NULL);
 631
 632   if ((result->avail & (1 << category)) == 0
 633       && locfile_read (result, charmap) != 0)
 634     WITH_CUR_LOCALE (error (4, errno, _("\
 635 cannot open locale definition file `%s'"), result->name));
 636
 637   return result;
 638 }
 639
 640 static void
 641 turn_on_mcheck (void)
 642 {
 643   /* Enable `malloc' debugging.  */
 644   mcheck (NULL);
 645   /* Use the following line for a more thorough but much slower testing.  */
 646   /* mcheck_pedantic (NULL); */
 647 }
 648
 649 void (*__malloc_initialize_hook) (void) = turn_on_mcheck;