locale/programs/localedef.c

   1 /* Copyright (C) 1995-2004, 2005, 2006 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 version 2 as
   7    published by the Free Software Foundation.
   8
   9    This program is distributed in the hope that it will be useful,
  10    but WITHOUT ANY WARRANTY; without even the implied warranty of
  11    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  12    GNU General Public License for more details.
  13
  14    You should have received a copy of the GNU General Public License
  15    along with this program; if not, write to the Free Software Foundation,
  16    Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.  */
  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 global functions.  */
 172 extern void *xmalloc (size_t __n);
 173
 174 /* Prototypes for local functions.  */
 175 static void error_print (void);
 176 static const char *construct_output_path (char *path);
 177 static const char *normalize_codeset (const char *codeset, size_t name_len);
 178
 179
 180 int
 181 main (int argc, char *argv[])
 182 {
 183   const char *output_path;
 184   int cannot_write_why;
 185   struct charmap_t *charmap;
 186   struct localedef_t global;
 187   int remaining;
 188
 189   /* Set initial values for global variables.  */
 190   copy_list = NULL;
 191   posix_conformance = getenv ("POSIXLY_CORRECT") != NULL;
 192   error_print_progname = error_print;
 193
 194   /* Set locale.  Do not set LC_ALL because the other categories must
 195      not be affected (according to POSIX.2).  */
 196   setlocale (LC_MESSAGES, "");
 197   setlocale (LC_CTYPE, "");
 198
 199   /* Initialize the message catalog.  */
 200   textdomain (_libc_intl_domainname);
 201
 202   /* Parse and process arguments.  */
 203   argp_err_exit_status = 4;
 204   argp_parse (&argp, argc, argv, 0, &remaining, NULL);
 205
 206   /* Handle a few special cases.  */
 207   if (list_archive)
 208     show_archive_content (verbose);
 209   if (add_to_archive)
 210     return add_locales_to_archive (argc - remaining, &argv[remaining],
 211                                    replace_archive);
 212   if (delete_from_archive)
 213     return delete_locales_from_archive (argc - remaining, &argv[remaining]);
 214
 215   /* POSIX.2 requires to be verbose about missing characters in the
 216      character map.  */
 217   verbose |= posix_conformance;
 218
 219   if (argc - remaining != 1)
 220     {
 221       /* We need exactly one non-option parameter.  */
 222       argp_help (&argp, stdout, ARGP_HELP_SEE | ARGP_HELP_EXIT_ERR,
 223                  program_invocation_short_name);
 224       exit (4);
 225     }
 226
 227   /* The parameter describes the output path of the constructed files.
 228      If the described files cannot be written return a NULL pointer.  */
 229   output_path  = construct_output_path (argv[remaining]);
 230   if (output_path == NULL && ! no_archive)
 231     error (4, errno, _("cannot create directory for output files"));
 232   cannot_write_why = errno;
 233
 234   /* Now that the parameters are processed we have to reset the local
 235      ctype locale.  (P1003.2 4.35.5.2)  */
 236   setlocale (LC_CTYPE, "POSIX");
 237
 238   /* Look whether the system really allows locale definitions.  POSIX
 239      defines error code 3 for this situation so I think it must be
 240      a fatal error (see P1003.2 4.35.8).  */
 241   if (sysconf (_SC_2_LOCALEDEF) < 0)
 242     WITH_CUR_LOCALE (error (3, 0, _("\
 243 FATAL: system does not define `_POSIX2_LOCALEDEF'")));
 244
 245   /* Process charmap file.  */
 246   charmap = charmap_read (charmap_file, verbose, 1, be_quiet, 1);
 247
 248   /* Add the first entry in the locale list.  */
 249   memset (&global, '\0', sizeof (struct localedef_t));
 250   global.name = input_file ?: "/dev/stdin";
 251   global.needed = ALL_LOCALES;
 252   locales = &global;
 253
 254   /* Now read the locale file.  */
 255   if (locfile_read (&global, charmap) != 0)
 256     WITH_CUR_LOCALE (error (4, errno, _("\
 257 cannot open locale definition file `%s'"), input_file));
 258
 259   /* Perhaps we saw some `copy' instructions.  */
 260   while (1)
 261     {
 262       struct localedef_t *runp = locales;
 263
 264       while (runp != NULL && (runp->needed & runp->avail) == runp->needed)
 265         runp = runp->next;
 266
 267       if (runp == NULL)
 268         /* Everything read.  */
 269         break;
 270
 271       if (locfile_read (runp, charmap) != 0)
 272         WITH_CUR_LOCALE (error (4, errno, _("\
 273 cannot open locale definition file `%s'"), runp->name));
 274     }
 275
 276   /* Check the categories we processed in source form.  */
 277   check_all_categories (locales, charmap);
 278
 279   /* We are now able to write the data files.  If warning were given we
 280      do it only if it is explicitly requested (--force).  */
 281   if (error_message_count == 0 || force_output != 0)
 282     {
 283       if (cannot_write_why != 0)
 284         WITH_CUR_LOCALE (error (4, cannot_write_why, _("\
 285 cannot write output files to `%s'"), output_path));
 286       else
 287         write_all_categories (locales, charmap, argv[remaining], output_path);
 288     }
 289   else
 290     WITH_CUR_LOCALE (error (4, 0, _("\
 291 no output file produced because warning were issued")));
 292
 293   /* This exit status is prescribed by POSIX.2 4.35.7.  */
 294   exit (error_message_count != 0);
 295 }
 296
 297
 298 /* Handle program arguments.  */
 299 static error_t
 300 parse_opt (int key, char *arg, struct argp_state *state)
 301 {
 302   switch (key)
 303     {
 304     case OPT_QUIET:
 305       be_quiet = 1;
 306       break;
 307     case OPT_POSIX:
 308       posix_conformance = 1;
 309       break;
 310     case OPT_OLDSTYLE:
 311       oldstyle_tables = 1;
 312       break;
 313     case OPT_PREFIX:
 314       output_prefix = arg;
 315       break;
 316     case OPT_NO_ARCHIVE:
 317       no_archive = true;
 318       break;
 319     case OPT_ADD_TO_ARCHIVE:
 320       add_to_archive = true;
 321       break;
 322     case OPT_REPLACE:
 323       replace_archive = true;
 324       break;
 325     case OPT_DELETE_FROM_ARCHIVE:
 326       delete_from_archive = true;
 327       break;
 328     case OPT_LIST_ARCHIVE:
 329       list_archive = true;
 330       break;
 331     case 'c':
 332       force_output = 1;
 333       break;
 334     case 'f':
 335       charmap_file = arg;
 336       break;
 337     case 'A':
 338       alias_file = arg;
 339       break;
 340     case 'i':
 341       input_file = arg;
 342       break;
 343     case 'u':
 344       repertoire_global = arg;
 345       break;
 346     case 'v':
 347       verbose = 1;
 348       break;
 349     default:
 350       return ARGP_ERR_UNKNOWN;
 351     }
 352   return 0;
 353 }
 354
 355
 356 static char *
 357 more_help (int key, const char *text, void *input)
 358 {
 359   char *cp;
 360
 361   switch (key)
 362     {
 363     case ARGP_KEY_HELP_EXTRA:
 364       /* We print some extra information.  */
 365       if (asprintf (&cp, gettext ("\
 366 System's directory for character maps : %s\n\
 367                        repertoire maps: %s\n\
 368                        locale path    : %s\n\
 369 %s"),
 370                     CHARMAP_PATH, REPERTOIREMAP_PATH, LOCALE_PATH, gettext ("\
 371 For bug reporting instructions, please see:\n\
 372 <http://www.gnu.org/software/libc/bugs.html>.\n")) < 0)
 373         return NULL;
 374       return cp;
 375     default:
 376       break;
 377     }
 378   return (char *) text;
 379 }
 380
 381 /* Print the version information.  */
 382 static void
 383 print_version (FILE *stream, struct argp_state *state)
 384 {
 385   fprintf (stream, "localedef (GNU %s) %s\n", PACKAGE, VERSION);
 386   fprintf (stream, gettext ("\
 387 Copyright (C) %s Free Software Foundation, Inc.\n\
 388 This is free software; see the source for copying conditions.  There is NO\n\
 389 warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.\n\
 390 "), "2006");
 391   fprintf (stream, gettext ("Written by %s.\n"), "Ulrich Drepper");
 392 }
 393
 394
 395 /* The address of this function will be assigned to the hook in the error
 396    functions.  */
 397 static void
 398 error_print (void)
 399 {
 400 }
 401
 402
 403 /* The parameter to localedef describes the output path.  If it does
 404    contain a '/' character it is a relative path.  Otherwise it names the
 405    locale this definition is for.  */
 406 static const char *
 407 construct_output_path (char *path)
 408 {
 409   const char *normal = NULL;
 410   char *result;
 411   char *endp;
 412
 413   if (strchr (path, '/') == NULL)
 414     {
 415       /* This is a system path.  First examine whether the locale name
 416          contains a reference to the codeset.  This should be
 417          normalized.  */
 418       char *startp;
 419       size_t n;
 420
 421       startp = path;
 422       /* We must be prepared for finding a CEN name or a location of
 423          the introducing `.' where it is not possible anymore.  */
 424       while (*startp != '\0' && *startp != '@' && *startp != '.')
 425         ++startp;
 426       if (*startp == '.')
 427         {
 428           /* We found a codeset specification.  Now find the end.  */
 429           endp = ++startp;
 430           while (*endp != '\0' && *endp != '@')
 431             ++endp;
 432
 433           if (endp > startp)
 434             normal = normalize_codeset (startp, endp - startp);
 435         }
 436       else
 437         /* This is to keep gcc quiet.  */
 438         endp = NULL;
 439
 440       /* We put an additional '\0' at the end of the string because at
 441          the end of the function we need another byte for the trailing
 442          '/'.  */
 443       if (normal == NULL)
 444         n = asprintf (&result, "%s%s/%s%c",
 445                       output_prefix ?: "", LOCALEDIR, path, '\0');
 446       else
 447         n = asprintf (&result, "%s%s/%.*s%s%s%c",
 448                       output_prefix ?: "", LOCALEDIR,
 449                       (int) (startp - path), path, normal, endp, '\0');
 450
 451       if (n < 0)
 452         return NULL;
 453
 454       endp = result + n - 1;
 455     }
 456   else
 457     {
 458       /* This is a user path.  Please note the additional byte in the
 459          memory allocation.  */
 460       size_t len = strlen (path) + 1;
 461       result = xmalloc (len + 1);
 462       endp = mempcpy (result, path, len) - 1;
 463
 464       /* If the user specified an output path we cannot add the output
 465          to the archive.  */
 466       no_archive = true;
 467     }
 468
 469   errno = 0;
 470
 471   if (no_archive && euidaccess (result, W_OK) == -1)
 472     /* Perhaps the directory does not exist now.  Try to create it.  */
 473     if (errno == ENOENT)
 474       {
 475         errno = 0;
 476         if (mkdir (result, 0777) < 0)
 477           return NULL;
 478       }
 479
 480   *endp++ = '/';
 481   *endp = '\0';
 482
 483   return result;
 484 }
 485
 486
 487 /* Normalize codeset name.  There is no standard for the codeset
 488    names.  Normalization allows the user to use any of the common
 489    names.  */
 490 static const char *
 491 normalize_codeset (codeset, name_len)
 492      const char *codeset;
 493      size_t name_len;
 494 {
 495   int len = 0;
 496   int only_digit = 1;
 497   char *retval;
 498   char *wp;
 499   size_t cnt;
 500
 501   for (cnt = 0; cnt < name_len; ++cnt)
 502     if (isalnum (codeset[cnt]))
 503       {
 504         ++len;
 505
 506         if (isalpha (codeset[cnt]))
 507           only_digit = 0;
 508       }
 509
 510   retval = (char *) malloc ((only_digit ? 3 : 0) + len + 1);
 511
 512   if (retval != NULL)
 513     {
 514       if (only_digit)
 515         wp = stpcpy (retval, "iso");
 516       else
 517         wp = retval;
 518
 519       for (cnt = 0; cnt < name_len; ++cnt)
 520         if (isalpha (codeset[cnt]))
 521           *wp++ = tolower (codeset[cnt]);
 522         else if (isdigit (codeset[cnt]))
 523           *wp++ = codeset[cnt];
 524
 525       *wp = '\0';
 526     }
 527
 528   return (const char *) retval;
 529 }
 530
 531
 532 struct localedef_t *
 533 add_to_readlist (int category, const char *name, const char *repertoire_name,
 534                  int generate, struct localedef_t *copy_locale)
 535 {
 536   struct localedef_t *runp = locales;
 537
 538   while (runp != NULL && strcmp (name, runp->name) != 0)
 539     runp = runp->next;
 540
 541   if (runp == NULL)
 542     {
 543       /* Add a new entry at the end.  */
 544       struct localedef_t *newp;
 545
 546       assert (generate == 1);
 547
 548       newp = xcalloc (1, sizeof (struct localedef_t));
 549       newp->name = name;
 550       newp->repertoire_name = repertoire_name;
 551
 552       if (locales == NULL)
 553         runp = locales = newp;
 554       else
 555         {
 556           runp = locales;
 557           while (runp->next != NULL)
 558             runp = runp->next;
 559           runp = runp->next = newp;
 560         }
 561     }
 562
 563   if (generate
 564       && (runp->needed & (1 << category)) != 0
 565       && (runp->avail & (1 << category)) == 0)
 566     WITH_CUR_LOCALE (error (5, 0, _("\
 567 circular dependencies between locale definitions")));
 568
 569   if (copy_locale != NULL)
 570     {
 571       if (runp->categories[category].generic != NULL)
 572         WITH_CUR_LOCALE (error (5, 0, _("\
 573 cannot add already read locale `%s' a second time"), name));
 574       else
 575         runp->categories[category].generic =
 576           copy_locale->categories[category].generic;
 577     }
 578
 579   runp->needed |= 1 << category;
 580
 581   return runp;
 582 }
 583
 584
 585 struct localedef_t *
 586 find_locale (int category, const char *name, const char *repertoire_name,
 587              const struct charmap_t *charmap)
 588 {
 589   struct localedef_t *result;
 590
 591   /* Find the locale, but do not generate it since this would be a bug.  */
 592   result = add_to_readlist (category, name, repertoire_name, 0, NULL);
 593
 594   assert (result != NULL);
 595
 596   if ((result->avail & (1 << category)) == 0
 597       && locfile_read (result, charmap) != 0)
 598     WITH_CUR_LOCALE (error (4, errno, _("\
 599 cannot open locale definition file `%s'"), result->name));
 600
 601   return result;
 602 }
 603
 604
 605 struct localedef_t *
 606 load_locale (int category, const char *name, const char *repertoire_name,
 607              const struct charmap_t *charmap, struct localedef_t *copy_locale)
 608 {
 609   struct localedef_t *result;
 610
 611   /* Generate the locale if it does not exist.  */
 612   result = add_to_readlist (category, name, repertoire_name, 1, copy_locale);
 613
 614   assert (result != NULL);
 615
 616   if ((result->avail & (1 << category)) == 0
 617       && locfile_read (result, charmap) != 0)
 618     WITH_CUR_LOCALE (error (4, errno, _("\
 619 cannot open locale definition file `%s'"), result->name));
 620
 621   return result;
 622 }
 623
 624 static void
 625 turn_on_mcheck (void)
 626 {
 627   /* Enable `malloc' debugging.  */
 628   mcheck (NULL);
 629   /* Use the following line for a more thorough but much slower testing.  */
 630   /* mcheck_pedantic (NULL); */
 631 }
 632
 633 void (*__malloc_initialize_hook) (void) = turn_on_mcheck;