src/intl/gettext/loadmsgcat.c

   1 /* Load needed message catalogs.
   2    Copyright (C) 1995-1999, 2000, 2001 Free Software Foundation, Inc.
   3
   4    This program is free software; you can redistribute it and/or modify
   5    it under the terms of the GNU General Public License as published by
   6    the Free Software Foundation; either version 2, or (at your option)
   7    any later version.
   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 /* Modified on 2007-07-02 by Kalle Olavi Niemitalo.  */
  19
  20 /* Tell glibc's <string.h> to provide a prototype for mempcpy().
  21    This must come before <config.h> because <config.h> may include
  22    <features.h>, and once <features.h> has been included, it's too late.  */
  23 #ifndef _GNU_SOURCE
  24 #define _GNU_SOURCE    1
  25 #endif
  26
  27 #ifdef HAVE_CONFIG_H
  28 #include "config.h"
  29 #endif
  30
  31 #include <ctype.h>
  32 #include <errno.h>
  33 #include <fcntl.h>
  34 #include <sys/types.h>
  35 #include <sys/stat.h>
  36 #include <stdlib.h>
  37 #include <string.h>
  38
  39 #if defined HAVE_UNISTD_H
  40 #include <unistd.h>
  41 #endif
  42
  43
  44 #if (defined HAVE_MMAP && defined HAVE_MUNMAP && !defined DISALLOW_MMAP)
  45 #include <sys/mman.h>
  46 /* Use a custom macro instead of overloading HAVE_MMAP, because the
  47  * following #include directives may cause "config.h" to be included
  48  * again (bug 960).  It might be good to remove #include "config.h"
  49  * directives from header files and keep them in *.c files only, but
  50  * that seems too risky for the stable elinks-0.11 branch.  */
  51 #undef LOADMSGCAT_USE_MMAP
  52 #define LOADMSGCAT_USE_MMAP     1
  53 #else
  54 #undef LOADMSGCAT_USE_MMAP
  55 #endif
  56
  57 #include "elinks.h"
  58
  59 #include "intl/gettext/gettext.h"
  60 #include "intl/gettext/gettextP.h"
  61 #include "main/main.h"
  62 #include "util/memory.h"
  63 #include "util/string.h"
  64
  65
  66 /* For systems that distinguish between text and binary I/O.
  67    O_BINARY is usually declared in <fcntl.h>. */
  68 #if !defined O_BINARY && defined _O_BINARY
  69   /* For MSC-compatible compilers.  */
  70 #define O_BINARY _O_BINARY
  71 #define O_TEXT _O_TEXT
  72 #endif
  73 #ifdef __BEOS__
  74   /* BeOS 5 has O_BINARY and O_TEXT, but they have no effect.  */
  75 #undef O_BINARY
  76 #undef O_TEXT
  77 #endif
  78 /* On reasonable systems, binary I/O is the default.  */
  79 #ifndef O_BINARY
  80 #define O_BINARY 0
  81 #endif
  82
  83 /* We need a sign, whether a new catalog was loaded, which can be associated
  84    with all translations.  This is important if the translations are
  85    cached by one of GCC's features.  */
  86 int _nl_msg_cat_cntr;
  87
  88
  89 /* For compilers without support for ISO C 99 struct/union initializers:
  90    Initialization at run-time.  */
  91
  92 static struct expression plvar;
  93 static struct expression plone;
  94 static struct expression germanic_plural;
  95
  96 static void
  97 init_germanic_plural(void)
  98 {
  99         if (plone.val.num == 0) {
 100                 plvar.nargs = 0;
 101                 plvar.operation = var;
 102
 103                 plone.nargs = 0;
 104                 plone.operation = num;
 105                 plone.val.num = 1;
 106
 107                 germanic_plural.nargs = 2;
 108                 germanic_plural.operation = not_equal;
 109                 germanic_plural.val.args[0] = &plvar;
 110                 germanic_plural.val.args[1] = &plone;
 111         }
 112 }
 113
 114 #define INIT_GERMANIC_PLURAL() init_germanic_plural ()
 115
 116 /* Initialize the codeset dependent parts of an opened message catalog.
 117    Return the header entry.  */
 118 const unsigned char *
 119 _nl_init_domain_conv(struct loaded_l10nfile *domain_file,
 120                      struct loaded_domain *domain,
 121                      struct binding *domainbinding)
 122 {
 123         /* Find out about the character set the file is encoded with.
 124            This can be found (in textual form) in the entry "".  If this
 125            entry does not exist or if this does not contain the `charset='
 126            information, we will assume the charset matches the one the
 127            current locale and we don't have to perform any conversion.  */
 128         unsigned char *nullentry;
 129         size_t nullentrylen;
 130
 131         /* Preinitialize fields, to avoid recursion during _nl_find_msg.  */
 132         domain->codeset_cntr =
 133                 (domainbinding != NULL ? domainbinding->codeset_cntr : 0);
 134 #if HAVE_ICONV
 135         domain->conv = (iconv_t) - 1;
 136 #endif
 137         domain->conv_tab = NULL;
 138
 139         /* Get the header entry.  */
 140         nullentry = _nl_find_msg(domain_file, domainbinding, "", &nullentrylen);
 141
 142         if (nullentry != NULL) {
 143 #if HAVE_ICONV
 144                 const unsigned char *charsetstr;
 145
 146                 charsetstr = strstr(nullentry, "charset=");
 147                 if (charsetstr != NULL) {
 148                         size_t len;
 149                         unsigned char *charset;
 150                         const unsigned char *outcharset;
 151
 152                         charsetstr += strlen("charset=");
 153                         len = strcspn(charsetstr, " \t\n");
 154
 155                         charset = (unsigned char *) fmem_alloc(len + 1);
 156                         *((unsigned char *) mempcpy(charset, charsetstr, len)) = '\0';
 157
 158                         /* The output charset should normally be determined by the
 159                            locale.  But sometimes the locale is not used or not correctly
 160                            set up, so we provide a possibility for the user to override
 161                            this.  Moreover, the value specified through
 162                            bind_textdomain_codeset overrides both.  */
 163                         if (domainbinding != NULL
 164                             && domainbinding->codeset != NULL)
 165                                 outcharset = domainbinding->codeset;
 166                         else {
 167                                 outcharset = getenv("OUTPUT_CHARSET");
 168                                 if (outcharset == NULL || outcharset[0] == '\0') {
 169                                         extern const unsigned char *elinks_locale_charset(void);
 170
 171                                         outcharset = elinks_locale_charset();
 172                                 }
 173                         }
 174
 175                         /* When using GNU libiconv, we want to use transliteration.  */
 176 #if _LIBICONV_VERSION >= 0x0105
 177                         len = strlen(outcharset);
 178                         {
 179                                 unsigned char *tmp = (unsigned char *) fmem_alloc(len + 10 + 1);
 180
 181                                 memcpy(tmp, outcharset, len);
 182                                 memcpy(tmp + len, "//TRANSLIT", 10 + 1);
 183                                 outcharset = tmp;
 184                         }
 185 #endif
 186                         domain->conv = iconv_open(outcharset, charset);
 187 #if _LIBICONV_VERSION >= 0x0105
 188                         fmem_free((void *) outcharset);
 189 #endif
 190
 191                         fmem_free(charset);
 192                 }
 193 #endif /* HAVE_ICONV */
 194         }
 195
 196         return nullentry;
 197 }
 198
 199 /* Frees the codeset dependent parts of an opened message catalog.  */
 200 void
 201 _nl_free_domain_conv(struct loaded_domain *domain)
 202 {
 203         if (domain->conv_tab != NULL && domain->conv_tab != (unsigned char **) -1)
 204                 free(domain->conv_tab);
 205 #if HAVE_ICONV
 206         if (domain->conv != (iconv_t) - 1)
 207                 iconv_close(domain->conv);
 208 #endif
 209 }
 210
 211 /* We cannot use our memory functions here because of circular library
 212  * dependencies. */
 213
 214 /* This is hacked for ELinks - we want to look up for the translations at the
 215  * correct place even if we are being ran from the source/build tree. */
 216 static struct string *
 217 add_filename_to_string(struct string *str, struct loaded_l10nfile *domain_file)
 218 {
 219         unsigned char *slash = strrchr(program.path, '/');
 220         size_t dirnamelen = (slash ? slash - program.path + 1 : 0);
 221
 222         /* Check if elinks is being run from the source tree. */
 223         if (dirnamelen < 4
 224             || strncmp(program.path + dirnamelen - 4, "src", 3))
 225                 return NULL;
 226
 227         if ((dirnamelen && !add_bytes_to_string(str, program.path, dirnamelen))
 228             || !add_to_string(str, "../po/")
 229             || !add_bytes_to_string(str,
 230                                     (unsigned char *) domain_file->langdirname,
 231                                     domain_file->langdirnamelen)
 232             || !add_to_string(str, ".gmo"))
 233                 return NULL;
 234
 235         return str;
 236 }
 237
 238 /* Load the message catalogs specified by FILENAME.  If it is no valid
 239    message catalog do nothing.  */
 240 void
 241 _nl_load_domain(struct loaded_l10nfile *domain_file,
 242                 struct binding *domainbinding)
 243 {
 244         int fd = -1;
 245         size_t size;
 246         struct stat st;
 247         struct mo_file_header *data = (struct mo_file_header *) -1;
 248         int use_mmap = 0;
 249         struct loaded_domain *domain;
 250         const unsigned char *nullentry;
 251
 252         domain_file->decided = 1;
 253         domain_file->data = NULL;
 254
 255         /* Note that it would be useless to store domainbinding in domain_file
 256            because domainbinding might be == NULL now but != NULL later (after
 257            a call to bind_textdomain_codeset).  */
 258
 259         {
 260                 struct string filename;
 261
 262                 if (init_string(&filename)
 263                     && add_filename_to_string(&filename, domain_file)) {
 264                         fd = open(filename.source, O_RDONLY | O_BINARY);
 265                 }
 266
 267                 done_string(&filename);
 268
 269                 if (fd != -1)
 270                         goto source_success;
 271         }
 272
 273         /* If the record does not represent a valid locale the FILENAME
 274            might be NULL.  This can happen when according to the given
 275            specification the locale file name is different for XPG and CEN
 276            syntax.  */
 277         if (domain_file->filename == NULL)
 278                 return;
 279
 280         /* Try to open the addressed file.  */
 281         fd = open(domain_file->filename, O_RDONLY | O_BINARY);
 282         if (fd == -1)
 283                 return;
 284
 285 source_success:
 286
 287         /* We must know about the size of the file.  */
 288         if (fstat(fd, &st) != 0
 289             || (size = (size_t) st.st_size) != st.st_size
 290             || (size < sizeof(struct mo_file_header))) {
 291                 /* Something went wrong.  */
 292                 close(fd);
 293                 return;
 294         }
 295 #ifdef LOADMSGCAT_USE_MMAP
 296         /* Now we are ready to load the file.  If mmap() is available we try
 297            this first.  If not available or it failed we try to load it.  */
 298         data = (struct mo_file_header *) mmap(NULL, size, PROT_READ,
 299                                               MAP_PRIVATE, fd, 0);
 300
 301         if (data != (struct mo_file_header *) -1) {
 302                 /* mmap() call was successful.  */
 303                 close(fd);
 304                 use_mmap = 1;
 305         }
 306 #endif
 307
 308         /* If the data is not yet available (i.e. mmap'ed) we try to load
 309            it manually.  */
 310         if (data == (struct mo_file_header *) -1) {
 311                 size_t to_read;
 312                 unsigned char *read_ptr;
 313
 314                 data = (struct mo_file_header *) malloc(size);
 315                 if (data == NULL)
 316                         return;
 317
 318                 to_read = size;
 319                 read_ptr = (unsigned char *) data;
 320                 do {
 321                         ssize_t nb = safe_read(fd, read_ptr, to_read);
 322
 323                         if (nb <= 0) {
 324                                 close(fd);
 325                                 return;
 326                         }
 327                         read_ptr += nb;
 328                         to_read -= nb;
 329                 } while (to_read > 0);
 330
 331                 close(fd);
 332         }
 333
 334         /* Using the magic number we can test whether it really is a message
 335            catalog file.  */
 336         if (data->magic != _MAGIC && data->magic != _MAGIC_SWAPPED) {
 337                 /* The magic number is wrong: not a message catalog file.  */
 338 #ifdef LOADMSGCAT_USE_MMAP
 339                 if (use_mmap)
 340                         munmap((void *) data, size);
 341                 else
 342 #endif
 343                         free(data);
 344                 return;
 345         }
 346
 347         domain = (struct loaded_domain *) malloc(sizeof(struct loaded_domain));
 348         if (domain == NULL)
 349                 return;
 350         domain_file->data = domain;
 351
 352         domain->data = (unsigned char *) data;
 353         domain->use_mmap = use_mmap;
 354         domain->mmap_size = size;
 355         domain->must_swap = data->magic != _MAGIC;
 356
 357         /* Fill in the information about the available tables.  */
 358         switch (W(domain->must_swap, data->revision)) {
 359                 case 0:
 360                         domain->nstrings = W(domain->must_swap, data->nstrings);
 361                         domain->orig_tab = (struct string_desc *)
 362                                 ((unsigned char *) data +
 363                                  W(domain->must_swap, data->orig_tab_offset));
 364                         domain->trans_tab = (struct string_desc *)
 365                                 ((unsigned char *) data +
 366                                  W(domain->must_swap, data->trans_tab_offset));
 367                         domain->hash_size =
 368                                 W(domain->must_swap, data->hash_tab_size);
 369                         domain->hash_tab = (nls_uint32 *)
 370                                 ((unsigned char *) data +
 371                                  W(domain->must_swap, data->hash_tab_offset));
 372                         break;
 373 default:
 374                         /* This is an invalid revision.  */
 375 #ifdef LOADMSGCAT_USE_MMAP
 376                         if (use_mmap)
 377                                 munmap((void *) data, size);
 378                         else
 379 #endif
 380                                 free(data);
 381                         free(domain);
 382                         domain_file->data = NULL;
 383                         return;
 384         }
 385
 386         /* Now initialize the character set converter from the character set
 387            the file is encoded with (found in the header entry) to the domain's
 388            specified character set or the locale's character set.  */
 389         nullentry = _nl_init_domain_conv(domain_file, domain, domainbinding);
 390
 391         /* Also look for a plural specification.  */
 392         if (nullentry != NULL) {
 393                 const unsigned char *plural;
 394                 const unsigned char *nplurals;
 395
 396                 plural = strstr(nullentry, "plural=");
 397                 nplurals = strstr(nullentry, "nplurals=");
 398                 if (plural == NULL || nplurals == NULL)
 399                         goto no_plural;
 400                 else {
 401                         /* First get the number.  */
 402                         unsigned char *endp;
 403                         unsigned long int n;
 404                         struct parse_args args;
 405
 406                         nplurals += 9;
 407                         skip_space(nplurals);
 408
 409                         for (endp = (unsigned char *) nplurals, n = 0;
 410                              isdigit(*endp); endp++)
 411                                 n = n * 10 + (*endp - '0');
 412
 413                         domain->nplurals = n;
 414                         if (nplurals == endp)
 415                                 goto no_plural;
 416
 417                         /* Due to the restrictions bison imposes onto the interface of the
 418                            scanner function we have to put the input string and the result
 419                            passed up from the parser into the same structure which address
 420                            is passed down to the parser.  */
 421                         plural += 7;
 422                         args.cp = plural;
 423                         if (gettext__parse(&args) != 0)
 424                                 goto no_plural;
 425                         domain->plural = args.res;
 426                 }
 427         } else {
 428                 /* By default we are using the Germanic form: singular form only
 429                    for `one', the plural form otherwise.  Yes, this is also what
 430                    English is using since English is a Germanic language.  */
 431 no_plural:
 432                 INIT_GERMANIC_PLURAL();
 433                 domain->plural = &germanic_plural;
 434                 domain->nplurals = 2;
 435         }
 436 }
 437
 438 #if 0
 439 void
 440 _nl_unload_domain(struct loaded_domain *domain)
 441 {
 442         if (domain->plural != &germanic_plural)
 443                 __gettext_free_exp(domain->plural);
 444
 445         _nl_free_domain_conv(domain);
 446
 447 #ifdef _POSIX_MAPPED_FILES
 448         if (domain->use_mmap)
 449                 munmap((void *) domain->data, domain->mmap_size);
 450         else
 451 #endif  /* _POSIX_MAPPED_FILES */
 452                 free((void *) domain->data);
 453
 454         free(domain);
 455 }
 456 #endif