locale/programs/localedef.c

   1 /* Copyright (C) 1995-2011, 2012 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
 116 /* Definitions of arguments for argp functions.  */
 117 static const struct argp_option options[] =
 118 {
 119   { NULL, 0, NULL, 0, N_("Input Files:") },
 120   { "charmap", 'f', "FILE", 0,
 121     N_("Symbolic character names defined in FILE") },
 122   { "inputfile", 'i', "FILE", 0, N_("Source definitions are found in FILE") },
 123   { "repertoire-map", 'u', "FILE", 0,
 124     N_("FILE contains mapping from symbolic names to UCS4 values") },
 125
 126   { NULL, 0, NULL, 0, N_("Output control:") },
 127   { "force", 'c', NULL, 0,
 128     N_("Create output even if warning messages were issued") },
 129   { "old-style", OPT_OLDSTYLE, NULL, 0, N_("Create old-style tables") },
 130   { "prefix", OPT_PREFIX, "PATH", 0, N_("Optional output file prefix") },
 131   { "posix", OPT_POSIX, NULL, 0, N_("Be strictly POSIX conform") },
 132   { "quiet", OPT_QUIET, NULL, 0,
 133     N_("Suppress warnings and information messages") },
 134   { "verbose", 'v', NULL, 0, N_("Print more messages") },
 135   { NULL, 0, NULL, 0, N_("Archive control:") },
 136   { "no-archive", OPT_NO_ARCHIVE, NULL, 0,
 137     N_("Don't add new data to archive") },
 138   { "add-to-archive", OPT_ADD_TO_ARCHIVE, NULL, 0,
 139     N_("Add locales named by parameters to archive") },
 140   { "replace", OPT_REPLACE, NULL, 0, N_("Replace existing archive content") },
 141   { "delete-from-archive", OPT_DELETE_FROM_ARCHIVE, NULL, 0,
 142     N_("Remove locales named by parameters from archive") },
 143   { "list-archive", OPT_LIST_ARCHIVE, NULL, 0, N_("List content of archive") },
 144   { "alias-file", 'A', "FILE", 0,
 145     N_("locale.alias file to consult when making archive")},
 146   { NULL, 0, NULL, 0, NULL }
 147 };
 148
 149 /* Short description of program.  */
 150 static const char doc[] = N_("Compile locale specification");
 151
 152 /* Strings for arguments in help texts.  */
 153 static const char args_doc[] = N_("\
 154 NAME\n\
 155 [--add-to-archive|--delete-from-archive] FILE...\n\
 156 --list-archive [FILE]");
 157
 158 /* Prototype for option handler.  */
 159 static error_t parse_opt (int key, char *arg, struct argp_state *state);
 160
 161 /* Function to print some extra text in the help message.  */
 162 static char *more_help (int key, const char *text, void *input);
 163
 164 /* Data structure to communicate with argp functions.  */
 165 static struct argp argp =
 166 {
 167   options, parse_opt, args_doc, doc, NULL, more_help
 168 };
 169
 170
 171 /* Prototypes for local functions.  */
 172 static void error_print (void);
 173 static const char *construct_output_path (char *path);
 174 static const char *normalize_codeset (const char *codeset, size_t name_len);
 175
 176
 177 int
 178 main (int argc, char *argv[])
 179 {
 180   const char *output_path;
 181   int cannot_write_why;
 182   struct charmap_t *charmap;
 183   struct localedef_t global;
 184   int remaining;
 185
 186   /* Set initial values for global variables.  */
 187   copy_list = NULL;
 188   posix_conformance = getenv ("POSIXLY_CORRECT") != NULL;
 189   error_print_progname = error_print;
 190
 191   /* Set locale.  Do not set LC_ALL because the other categories must
 192      not be affected (according to POSIX.2).  */
 193   setlocale (LC_MESSAGES, "");
 194   setlocale (LC_CTYPE, "");
 195
 196   /* Initialize the message catalog.  */
 197   textdomain (_libc_intl_domainname);
 198
 199   /* Parse and process arguments.  */
 200   argp_err_exit_status = 4;
 201   argp_parse (&argp, argc, argv, 0, &remaining, NULL);
 202
 203   /* Handle a few special cases.  */
 204   if (list_archive)
 205     show_archive_content (verbose);
 206   if (add_to_archive)
 207     return add_locales_to_archive (argc - remaining, &argv[remaining],
 208                                    replace_archive);
 209   if (delete_from_archive)
 210     return delete_locales_from_archive (argc - remaining, &argv[remaining]);
 211
 212   /* POSIX.2 requires to be verbose about missing characters in the
 213      character map.  */
 214   verbose |= posix_conformance;
 215
 216   if (argc - remaining != 1)
 217     {
 218       /* We need exactly one non-option parameter.  */
 219       argp_help (&argp, stdout, ARGP_HELP_SEE | ARGP_HELP_EXIT_ERR,
 220                  program_invocation_short_name);
 221       exit (4);
 222     }
 223
 224   /* The parameter describes the output path of the constructed files.
 225      If the described files cannot be written return a NULL pointer.  */
 226   output_path  = construct_output_path (argv[remaining]);
 227   if (output_path == NULL && ! no_archive)
 228     error (4, errno, _("cannot create directory for output files"));
 229   cannot_write_why = errno;
 230
 231   /* Now that the parameters are processed we have to reset the local
 232      ctype locale.  (P1003.2 4.35.5.2)  */
 233   setlocale (LC_CTYPE, "POSIX");
 234
 235   /* Look whether the system really allows locale definitions.  POSIX
 236      defines error code 3 for this situation so I think it must be
 237      a fatal error (see P1003.2 4.35.8).  */
 238   if (sysconf (_SC_2_LOCALEDEF) < 0)
 239     WITH_CUR_LOCALE (error (3, 0, _("\
 240 FATAL: system does not define `_POSIX2_LOCALEDEF'")));
 241
 242   /* Process charmap file.  */
 243   charmap = charmap_read (charmap_file, verbose, 1, be_quiet, 1);
 244
 245   /* Add the first entry in the locale list.  */
 246   memset (&global, '\0', sizeof (struct localedef_t));
 247   global.name = input_file ?: "/dev/stdin";
 248   global.needed = ALL_LOCALES;
 249   locales = &global;
 250
 251   /* Now read the locale file.  */
 252   if (locfile_read (&global, charmap) != 0)
 253     WITH_CUR_LOCALE (error (4, errno, _("\
 254 cannot open locale definition file `%s'"), input_file));
 255
 256   /* Perhaps we saw some `copy' instructions.  */
 257   while (1)
 258     {
 259       struct localedef_t *runp = locales;
 260
 261       while (runp != NULL && (runp->needed & runp->avail) == runp->needed)
 262         runp = runp->next;
 263
 264       if (runp == NULL)
 265         /* Everything read.  */
 266         break;
 267
 268       if (locfile_read (runp, charmap) != 0)
 269         WITH_CUR_LOCALE (error (4, errno, _("\
 270 cannot open locale definition file `%s'"), runp->name));
 271     }
 272
 273   /* Check the categories we processed in source form.  */
 274   check_all_categories (locales, charmap);
 275
 276   /* We are now able to write the data files.  If warning were given we
 277      do it only if it is explicitly requested (--force).  */
 278   if (error_message_count == 0 || force_output != 0)
 279     {
 280       if (cannot_write_why != 0)
 281         WITH_CUR_LOCALE (error (4, cannot_write_why, _("\
 282 cannot write output files to `%s'"), output_path));
 283       else
 284         write_all_categories (locales, charmap, argv[remaining], output_path);
 285     }
 286   else
 287     WITH_CUR_LOCALE (error (4, 0, _("\
 288 no output file produced because warnings were issued")));
 289
 290   /* This exit status is prescribed by POSIX.2 4.35.7.  */
 291   exit (error_message_count != 0);
 292 }
 293
 294
 295 /* Handle program arguments.  */
 296 static error_t
 297 parse_opt (int key, char *arg, struct argp_state *state)
 298 {
 299   switch (key)
 300     {
 301     case OPT_QUIET:
 302       be_quiet = 1;
 303       break;
 304     case OPT_POSIX:
 305       posix_conformance = 1;
 306       break;
 307     case OPT_OLDSTYLE:
 308       oldstyle_tables = 1;
 309       break;
 310     case OPT_PREFIX:
 311       output_prefix = arg;
 312       break;
 313     case OPT_NO_ARCHIVE:
 314       no_archive = true;
 315       break;
 316     case OPT_ADD_TO_ARCHIVE:
 317       add_to_archive = true;
 318       break;
 319     case OPT_REPLACE:
 320       replace_archive = true;
 321       break;
 322     case OPT_DELETE_FROM_ARCHIVE:
 323       delete_from_archive = true;
 324       break;
 325     case OPT_LIST_ARCHIVE:
 326       list_archive = true;
 327       break;
 328     case 'c':
 329       force_output = 1;
 330       break;
 331     case 'f':
 332       charmap_file = arg;
 333       break;
 334     case 'A':
 335       alias_file = arg;
 336       break;
 337     case 'i':
 338       input_file = arg;
 339       break;
 340     case 'u':
 341       repertoire_global = arg;
 342       break;
 343     case 'v':
 344       verbose = 1;
 345       break;
 346     default:
 347       return ARGP_ERR_UNKNOWN;
 348     }
 349   return 0;
 350 }
 351
 352
 353 static char *
 354 more_help (int key, const char *text, void *input)
 355 {
 356   char *cp;
 357
 358   switch (key)
 359     {
 360     case ARGP_KEY_HELP_EXTRA:
 361       /* We print some extra information.  */
 362       if (asprintf (&cp, gettext ("\
 363 System's directory for character maps : %s\n\
 364                        repertoire maps: %s\n\
 365                        locale path    : %s\n\
 366 %s"),
 367                     CHARMAP_PATH, REPERTOIREMAP_PATH, LOCALE_PATH, gettext ("\
 368 For bug reporting instructions, please see:\n\
 369 <http://www.gnu.org/software/libc/bugs.html>.\n")) < 0)
 370         return NULL;
 371       return cp;
 372     default:
 373       break;
 374     }
 375   return (char *) text;
 376 }
 377
 378 /* Print the version information.  */
 379 static void
 380 print_version (FILE *stream, struct argp_state *state)
 381 {
 382   fprintf (stream, "localedef (GNU %s) %s\n", PACKAGE, VERSION);
 383   fprintf (stream, gettext ("\
 384 Copyright (C) %s Free Software Foundation, Inc.\n\
 385 This is free software; see the source for copying conditions.  There is NO\n\
 386 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 387 "), "2012");
 388   fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
 389 }
 390
 391
 392 /* The address of this function will be assigned to the hook in the error
 393    functions.  */
 394 static void
 395 error_print (void)
 396 {
 397 }
 398
 399
 400 /* The parameter to localedef describes the output path.  If it does
 401    contain a '/' character it is a relative path.  Otherwise it names the
 402    locale this definition is for.  */
 403 static const char *
 404 construct_output_path (char *path)
 405 {
 406   const char *normal = NULL;
 407   char *result;
 408   char *endp;
 409
 410   if (strchr (path, '/') == NULL)
 411     {
 412       /* This is a system path.  First examine whether the locale name
 413          contains a reference to the codeset.  This should be
 414          normalized.  */
 415       char *startp;
 416
 417       startp = path;
 418       /* We must be prepared for finding a CEN name or a location of
 419          the introducing `.' where it is not possible anymore.  */
 420       while (*startp != '\0' && *startp != '@' && *startp != '.')
 421         ++startp;
 422       if (*startp == '.')
 423         {
 424           /* We found a codeset specification.  Now find the end.  */
 425           endp = ++startp;
 426           while (*endp != '\0' && *endp != '@')
 427             ++endp;
 428
 429           if (endp > startp)
 430             normal = normalize_codeset (startp, endp - startp);
 431         }
 432       else
 433         /* This is to keep gcc quiet.  */
 434         endp = NULL;
 435
 436       /* We put an additional '\0' at the end of the string because at
 437          the end of the function we need another byte for the trailing
 438          '/'.  */
 439       ssize_t n;
 440       if (normal == NULL)
 441         n = asprintf (&result, "%s%s/%s%c",
 442                       output_prefix ?: "", LOCALEDIR, path, '\0');
 443       else
 444         n = asprintf (&result, "%s%s/%.*s%s%s%c",
 445                       output_prefix ?: "", LOCALEDIR,
 446                       (int) (startp - path), path, normal, endp, '\0');
 447
 448       if (n < 0)
 449         return NULL;
 450
 451       endp = result + n - 1;
 452     }
 453   else
 454     {
 455       /* This is a user path.  Please note the additional byte in the
 456          memory allocation.  */
 457       size_t len = strlen (path) + 1;
 458       result = xmalloc (len + 1);
 459       endp = mempcpy (result, path, len) - 1;
 460
 461       /* If the user specified an output path we cannot add the output
 462          to the archive.  */
 463       no_archive = true;
 464     }
 465
 466   errno = 0;
 467
 468   if (no_archive && euidaccess (result, W_OK) == -1)
 469     /* Perhaps the directory does not exist now.  Try to create it.  */
 470     if (errno == ENOENT)
 471       {
 472         errno = 0;
 473         if (mkdir (result, 0777) < 0)
 474           return NULL;
 475       }
 476
 477   *endp++ = '/';
 478   *endp = '\0';
 479
 480   return result;
 481 }
 482
 483
 484 /* Normalize codeset name.  There is no standard for the codeset
 485    names.  Normalization allows the user to use any of the common
 486    names.  */
 487 static const char *
 488 normalize_codeset (codeset, name_len)
 489      const char *codeset;
 490      size_t name_len;
 491 {
 492   int len = 0;
 493   int only_digit = 1;
 494   char *retval;
 495   char *wp;
 496   size_t cnt;
 497
 498   for (cnt = 0; cnt < name_len; ++cnt)
 499     if (isalnum (codeset[cnt]))
 500       {
 501         ++len;
 502
 503         if (isalpha (codeset[cnt]))
 504           only_digit = 0;
 505       }
 506
 507   retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
 508
 509   if (retval != NULL)
 510     {
 511       if (only_digit)
 512         wp = stpcpy (retval, "iso");
 513       else
 514         wp = retval;
 515
 516       for (cnt = 0; cnt < name_len; ++cnt)
 517         if (isalpha (codeset[cnt]))
 518           *wp++ = tolower (codeset[cnt]);
 519         else if (isdigit (codeset[cnt]))
 520           *wp++ = codeset[cnt];
 521
 522       *wp = '\0';
 523     }
 524
 525   return (const char *) retval;
 526 }
 527
 528
 529 struct localedef_t *
 530 add_to_readlist (int category, const char *name, const char *repertoire_name,
 531                  int generate, struct localedef_t *copy_locale)
 532 {
 533   struct localedef_t *runp = locales;
 534
 535   while (runp != NULL && strcmp (name, runp->name) != 0)
 536     runp = runp->next;
 537
 538   if (runp == NULL)
 539     {
 540       /* Add a new entry at the end.  */
 541       struct localedef_t *newp;
 542
 543       assert (generate == 1);
 544
 545       newp = xcalloc (1, sizeof (struct localedef_t));
 546       newp->name = name;
 547       newp->repertoire_name = repertoire_name;
 548
 549       if (locales == NULL)
 550         runp = locales = newp;
 551       else
 552         {
 553           runp = locales;
 554           while (runp->next != NULL)
 555             runp = runp->next;
 556           runp = runp->next = newp;
 557         }
 558     }
 559
 560   if (generate
 561       && (runp->needed & (1 << category)) != 0
 562       && (runp->avail & (1 << category)) == 0)
 563     WITH_CUR_LOCALE (error (5, 0, _("\
 564 circular dependencies between locale definitions")));
 565
 566   if (copy_locale != NULL)
 567     {
 568       if (runp->categories[category].generic != NULL)
 569         WITH_CUR_LOCALE (error (5, 0, _("\
 570 cannot add already read locale `%s' a second time"), name));
 571       else
 572         runp->categories[category].generic =
 573           copy_locale->categories[category].generic;
 574     }
 575
 576   runp->needed |= 1 << category;
 577
 578   return runp;
 579 }
 580
 581
 582 struct localedef_t *
 583 find_locale (int category, const char *name, const char *repertoire_name,
 584              const struct charmap_t *charmap)
 585 {
 586   struct localedef_t *result;
 587
 588   /* Find the locale, but do not generate it since this would be a bug.  */
 589   result = add_to_readlist (category, name, repertoire_name, 0, NULL);
 590
 591   assert (result != NULL);
 592
 593   if ((result->avail & (1 << category)) == 0
 594       && locfile_read (result, charmap) != 0)
 595     WITH_CUR_LOCALE (error (4, errno, _("\
 596 cannot open locale definition file `%s'"), result->name));
 597
 598   return result;
 599 }
 600
 601
 602 struct localedef_t *
 603 load_locale (int category, const char *name, const char *repertoire_name,
 604              const struct charmap_t *charmap, struct localedef_t *copy_locale)
 605 {
 606   struct localedef_t *result;
 607
 608   /* Generate the locale if it does not exist.  */
 609   result = add_to_readlist (category, name, repertoire_name, 1, copy_locale);
 610
 611   assert (result != NULL);
 612
 613   if ((result->avail & (1 << category)) == 0
 614       && locfile_read (result, charmap) != 0)
 615     WITH_CUR_LOCALE (error (4, errno, _("\
 616 cannot open locale definition file `%s'"), result->name));
 617
 618   return result;
 619 }
 620
 621 static void
 622 turn_on_mcheck (void)
 623 {
 624   /* Enable `malloc' debugging.  */
 625   mcheck (NULL);
 626   /* Use the following line for a more thorough but much slower testing.  */
 627   /* mcheck_pedantic (NULL); */
 628 }
 629
 630 void (*__malloc_initialize_hook) (void) = turn_on_mcheck;