glib/gcharset.c

   1 /* gcharset.c - Charset information
   2  *
   3  * Copyright (C) 2011 Red Hat, Inc.
   4  *
   5  * This 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  * This 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 this library; if not, see <http://www.gnu.org/licenses/>.
  17  */
  18
  19 #include "config.h"
  20
  21 #include "gcharsetprivate.h"
  22
  23 #include "garray.h"
  24 #include "genviron.h"
  25 #include "ghash.h"
  26 #include "gmessages.h"
  27 #include "gstrfuncs.h"
  28 #include "gthread.h"
  29 #ifdef G_OS_WIN32
  30 #include "gwin32.h"
  31 #endif
  32
  33 #include "libcharset/libcharset.h"
  34
  35 #include <string.h>
  36 #include <stdio.h>
  37
  38 G_LOCK_DEFINE_STATIC (aliases);
  39
  40 static GHashTable *
  41 get_alias_hash (void)
  42 {
  43   static GHashTable *alias_hash = NULL;
  44   const char *aliases;
  45
  46   G_LOCK (aliases);
  47
  48   if (!alias_hash)
  49     {
  50       alias_hash = g_hash_table_new (g_str_hash, g_str_equal);
  51
  52       aliases = _g_locale_get_charset_aliases ();
  53       while (*aliases != '\0')
  54         {
  55           const char *canonical;
  56           const char *alias;
  57           const char **alias_array;
  58           int count = 0;
  59
  60           alias = aliases;
  61           aliases += strlen (aliases) + 1;
  62           canonical = aliases;
  63           aliases += strlen (aliases) + 1;
  64
  65           alias_array = g_hash_table_lookup (alias_hash, canonical);
  66           if (alias_array)
  67             {
  68               while (alias_array[count])
  69                 count++;
  70             }
  71
  72           alias_array = g_renew (const char *, alias_array, count + 2);
  73           alias_array[count] = alias;
  74           alias_array[count + 1] = NULL;
  75
  76           g_hash_table_insert (alias_hash, (char *)canonical, alias_array);
  77         }
  78     }
  79
  80   G_UNLOCK (aliases);
  81
  82   return alias_hash;
  83 }
  84
  85 /* As an abuse of the alias table, the following routines gets
  86  * the charsets that are aliases for the canonical name.
  87  */
  88 const char **
  89 _g_charset_get_aliases (const char *canonical_name)
  90 {
  91   GHashTable *alias_hash = get_alias_hash ();
  92
  93   return g_hash_table_lookup (alias_hash, canonical_name);
  94 }
  95
  96 static gboolean
  97 g_utf8_get_charset_internal (const char  *raw_data,
  98                              const char **a)
  99 {
 100   const char *charset = g_getenv ("CHARSET");
 101
 102   if (charset && *charset)
 103     {
 104       *a = charset;
 105
 106       if (charset && strstr (charset, "UTF-8"))
 107         return TRUE;
 108       else
 109         return FALSE;
 110     }
 111
 112   /* The libcharset code tries to be thread-safe without
 113    * a lock, but has a memory leak and a missing memory
 114    * barrier, so we lock for it
 115    */
 116   G_LOCK (aliases);
 117   charset = _g_locale_charset_unalias (raw_data);
 118   G_UNLOCK (aliases);
 119
 120   if (charset && *charset)
 121     {
 122       *a = charset;
 123
 124       if (charset && strstr (charset, "UTF-8"))
 125         return TRUE;
 126       else
 127         return FALSE;
 128     }
 129
 130   /* Assume this for compatibility at present.  */
 131   *a = "US-ASCII";
 132
 133   return FALSE;
 134 }
 135
 136 typedef struct _GCharsetCache GCharsetCache;
 137
 138 struct _GCharsetCache {
 139   gboolean is_utf8;
 140   gchar *raw;
 141   gchar *charset;
 142 };
 143
 144 static void
 145 charset_cache_free (gpointer data)
 146 {
 147   GCharsetCache *cache = data;
 148   g_free (cache->raw);
 149   g_free (cache->charset);
 150   g_free (cache);
 151 }
 152
 153 /**
 154  * g_get_charset:
 155  * @charset: (out) (optional) (transfer none): return location for character set
 156  *   name, or %NULL.
 157  *
 158  * Obtains the character set for the [current locale][setlocale]; you
 159  * might use this character set as an argument to g_convert(), to convert
 160  * from the current locale's encoding to some other encoding. (Frequently
 161  * g_locale_to_utf8() and g_locale_from_utf8() are nice shortcuts, though.)
 162  *
 163  * On Windows the character set returned by this function is the
 164  * so-called system default ANSI code-page. That is the character set
 165  * used by the "narrow" versions of C library and Win32 functions that
 166  * handle file names. It might be different from the character set
 167  * used by the C library's current locale.
 168  *
 169  * The return value is %TRUE if the locale's encoding is UTF-8, in that
 170  * case you can perhaps avoid calling g_convert().
 171  *
 172  * The string returned in @charset is not allocated, and should not be
 173  * freed.
 174  *
 175  * Returns: %TRUE if the returned charset is UTF-8
 176  */
 177 gboolean
 178 g_get_charset (const char **charset)
 179 {
 180   static GPrivate cache_private = G_PRIVATE_INIT (charset_cache_free);
 181   GCharsetCache *cache = g_private_get (&cache_private);
 182   const gchar *raw;
 183
 184   if (!cache)
 185     {
 186       cache = g_new0 (GCharsetCache, 1);
 187       g_private_set (&cache_private, cache);
 188     }
 189
 190   G_LOCK (aliases);
 191   raw = _g_locale_charset_raw ();
 192   G_UNLOCK (aliases);
 193
 194   if (!(cache->raw && strcmp (cache->raw, raw) == 0))
 195     {
 196       const gchar *new_charset;
 197
 198       g_free (cache->raw);
 199       g_free (cache->charset);
 200       cache->raw = g_strdup (raw);
 201       cache->is_utf8 = g_utf8_get_charset_internal (raw, &new_charset);
 202       cache->charset = g_strdup (new_charset);
 203     }
 204
 205   if (charset)
 206     *charset = cache->charset;
 207
 208   return cache->is_utf8;
 209 }
 210
 211 /**
 212  * g_get_codeset:
 213  *
 214  * Gets the character set for the current locale.
 215  *
 216  * Returns: a newly allocated string containing the name
 217  *     of the character set. This string must be freed with g_free().
 218  */
 219 gchar *
 220 g_get_codeset (void)
 221 {
 222   const gchar *charset;
 223
 224   g_get_charset (&charset);
 225
 226   return g_strdup (charset);
 227 }
 228
 229 #ifndef G_OS_WIN32
 230
 231 /* read an alias file for the locales */
 232 static void
 233 read_aliases (gchar      *file,
 234               GHashTable *alias_table)
 235 {
 236   FILE *fp;
 237   char buf[256];
 238
 239   fp = fopen (file,"r");
 240   if (!fp)
 241     return;
 242   while (fgets (buf, 256, fp))
 243     {
 244       char *p, *q;
 245
 246       g_strstrip (buf);
 247
 248       /* Line is a comment */
 249       if ((buf[0] == '#') || (buf[0] == '\0'))
 250         continue;
 251
 252       /* Reads first column */
 253       for (p = buf, q = NULL; *p; p++) {
 254         if ((*p == '\t') || (*p == ' ') || (*p == ':')) {
 255           *p = '\0';
 256           q = p+1;
 257           while ((*q == '\t') || (*q == ' ')) {
 258             q++;
 259           }
 260           break;
 261         }
 262       }
 263       /* The line only had one column */
 264       if (!q || *q == '\0')
 265         continue;
 266
 267       /* Read second column */
 268       for (p = q; *p; p++) {
 269         if ((*p == '\t') || (*p == ' ')) {
 270           *p = '\0';
 271           break;
 272         }
 273       }
 274
 275       /* Add to alias table if necessary */
 276       if (!g_hash_table_lookup (alias_table, buf)) {
 277         g_hash_table_insert (alias_table, g_strdup (buf), g_strdup (q));
 278       }
 279     }
 280   fclose (fp);
 281 }
 282
 283 #endif
 284
 285 static char *
 286 unalias_lang (char *lang)
 287 {
 288 #ifndef G_OS_WIN32
 289   static GHashTable *alias_table = NULL;
 290   char *p;
 291   int i;
 292
 293   if (g_once_init_enter (&alias_table))
 294     {
 295       GHashTable *table = g_hash_table_new (g_str_hash, g_str_equal);
 296       read_aliases ("/usr/share/locale/locale.alias", table);
 297       g_once_init_leave (&alias_table, table);
 298     }
 299
 300   i = 0;
 301   while ((p = g_hash_table_lookup (alias_table, lang)) && (strcmp (p, lang) != 0))
 302     {
 303       lang = p;
 304       if (i++ == 30)
 305         {
 306           static gboolean said_before = FALSE;
 307           if (!said_before)
 308             g_warning ("Too many alias levels for a locale, "
 309                        "may indicate a loop");
 310           said_before = TRUE;
 311           return lang;
 312         }
 313     }
 314 #endif
 315   return lang;
 316 }
 317
 318 /* Mask for components of locale spec. The ordering here is from
 319  * least significant to most significant
 320  */
 321 enum
 322 {
 323   COMPONENT_CODESET =   1 << 0,
 324   COMPONENT_TERRITORY = 1 << 1,
 325   COMPONENT_MODIFIER =  1 << 2
 326 };
 327
 328 /* Break an X/Open style locale specification into components
 329  */
 330 static guint
 331 explode_locale (const gchar *locale,
 332                 gchar      **language,
 333                 gchar      **territory,
 334                 gchar      **codeset,
 335                 gchar      **modifier)
 336 {
 337   const gchar *uscore_pos;
 338   const gchar *at_pos;
 339   const gchar *dot_pos;
 340
 341   guint mask = 0;
 342
 343   uscore_pos = strchr (locale, '_');
 344   dot_pos = strchr (uscore_pos ? uscore_pos : locale, '.');
 345   at_pos = strchr (dot_pos ? dot_pos : (uscore_pos ? uscore_pos : locale), '@');
 346
 347   if (at_pos)
 348     {
 349       mask |= COMPONENT_MODIFIER;
 350       *modifier = g_strdup (at_pos);
 351     }
 352   else
 353     at_pos = locale + strlen (locale);
 354
 355   if (dot_pos)
 356     {
 357       mask |= COMPONENT_CODESET;
 358       *codeset = g_strndup (dot_pos, at_pos - dot_pos);
 359     }
 360   else
 361     dot_pos = at_pos;
 362
 363   if (uscore_pos)
 364     {
 365       mask |= COMPONENT_TERRITORY;
 366       *territory = g_strndup (uscore_pos, dot_pos - uscore_pos);
 367     }
 368   else
 369     uscore_pos = dot_pos;
 370
 371   *language = g_strndup (locale, uscore_pos - locale);
 372
 373   return mask;
 374 }
 375
 376 /*
 377  * Compute all interesting variants for a given locale name -
 378  * by stripping off different components of the value.
 379  *
 380  * For simplicity, we assume that the locale is in
 381  * X/Open format: language[_territory][.codeset][@modifier]
 382  *
 383  * TODO: Extend this to handle the CEN format (see the GNUlibc docs)
 384  *       as well. We could just copy the code from glibc wholesale
 385  *       but it is big, ugly, and complicated, so I'm reluctant
 386  *       to do so when this should handle 99% of the time...
 387  */
 388 static void
 389 append_locale_variants (GPtrArray *array,
 390                         const gchar *locale)
 391 {
 392   gchar *language = NULL;
 393   gchar *territory = NULL;
 394   gchar *codeset = NULL;
 395   gchar *modifier = NULL;
 396
 397   guint mask;
 398   guint i, j;
 399
 400   g_return_if_fail (locale != NULL);
 401
 402   mask = explode_locale (locale, &language, &territory, &codeset, &modifier);
 403
 404   /* Iterate through all possible combinations, from least attractive
 405    * to most attractive.
 406    */
 407   for (j = 0; j <= mask; ++j)
 408     {
 409       i = mask - j;
 410
 411       if ((i & ~mask) == 0)
 412         {
 413           gchar *val = g_strconcat (language,
 414                                     (i & COMPONENT_TERRITORY) ? territory : "",
 415                                     (i & COMPONENT_CODESET) ? codeset : "",
 416                                     (i & COMPONENT_MODIFIER) ? modifier : "",
 417                                     NULL);
 418           g_ptr_array_add (array, val);
 419         }
 420     }
 421
 422   g_free (language);
 423   if (mask & COMPONENT_CODESET)
 424     g_free (codeset);
 425   if (mask & COMPONENT_TERRITORY)
 426     g_free (territory);
 427   if (mask & COMPONENT_MODIFIER)
 428     g_free (modifier);
 429 }
 430
 431 /**
 432  * g_get_locale_variants:
 433  * @locale: a locale identifier
 434  *
 435  * Returns a list of derived variants of @locale, which can be used to
 436  * e.g. construct locale-dependent filenames or search paths. The returned
 437  * list is sorted from most desirable to least desirable.
 438  * This function handles territory, charset and extra locale modifiers.
 439  *
 440  * For example, if @locale is "fr_BE", then the returned list
 441  * is "fr_BE", "fr".
 442  *
 443  * If you need the list of variants for the current locale,
 444  * use g_get_language_names().
 445  *
 446  * Returns: (transfer full) (array zero-terminated=1) (element-type utf8): a newly
 447  *   allocated array of newly allocated strings with the locale variants. Free with
 448  *   g_strfreev().
 449  *
 450  * Since: 2.28
 451  */
 452 gchar **
 453 g_get_locale_variants (const gchar *locale)
 454 {
 455   GPtrArray *array;
 456
 457   g_return_val_if_fail (locale != NULL, NULL);
 458
 459   array = g_ptr_array_sized_new (8);
 460   append_locale_variants (array, locale);
 461   g_ptr_array_add (array, NULL);
 462
 463   return (gchar **) g_ptr_array_free (array, FALSE);
 464 }
 465
 466 /* The following is (partly) taken from the gettext package.
 467    Copyright (C) 1995, 1996, 1997, 1998 Free Software Foundation, Inc.  */
 468
 469 static const gchar *
 470 guess_category_value (const gchar *category_name)
 471 {
 472   const gchar *retval;
 473
 474   /* The highest priority value is the 'LANGUAGE' environment
 475      variable.  This is a GNU extension.  */
 476   retval = g_getenv ("LANGUAGE");
 477   if ((retval != NULL) && (retval[0] != '\0'))
 478     return retval;
 479
 480   /* 'LANGUAGE' is not set.  So we have to proceed with the POSIX
 481      methods of looking to 'LC_ALL', 'LC_xxx', and 'LANG'.  On some
 482      systems this can be done by the 'setlocale' function itself.  */
 483
 484   /* Setting of LC_ALL overwrites all other.  */
 485   retval = g_getenv ("LC_ALL");
 486   if ((retval != NULL) && (retval[0] != '\0'))
 487     return retval;
 488
 489   /* Next comes the name of the desired category.  */
 490   retval = g_getenv (category_name);
 491   if ((retval != NULL) && (retval[0] != '\0'))
 492     return retval;
 493
 494   /* Last possibility is the LANG environment variable.  */
 495   retval = g_getenv ("LANG");
 496   if ((retval != NULL) && (retval[0] != '\0'))
 497     return retval;
 498
 499 #ifdef G_PLATFORM_WIN32
 500   /* g_win32_getlocale() first checks for LC_ALL, LC_MESSAGES and
 501    * LANG, which we already did above. Oh well. The main point of
 502    * calling g_win32_getlocale() is to get the thread's locale as used
 503    * by Windows and the Microsoft C runtime (in the "English_United
 504    * States" format) translated into the Unixish format.
 505    */
 506   {
 507     char *locale = g_win32_getlocale ();
 508     retval = g_intern_string (locale);
 509     g_free (locale);
 510     return retval;
 511   }
 512 #endif
 513
 514   return NULL;
 515 }
 516
 517 typedef struct _GLanguageNamesCache GLanguageNamesCache;
 518
 519 struct _GLanguageNamesCache {
 520   gchar *languages;
 521   gchar **language_names;
 522 };
 523
 524 static void
 525 language_names_cache_free (gpointer data)
 526 {
 527   GLanguageNamesCache *cache = data;
 528   g_free (cache->languages);
 529   g_strfreev (cache->language_names);
 530   g_free (cache);
 531 }
 532
 533 /**
 534  * g_get_language_names:
 535  *
 536  * Computes a list of applicable locale names, which can be used to
 537  * e.g. construct locale-dependent filenames or search paths. The returned
 538  * list is sorted from most desirable to least desirable and always contains
 539  * the default locale "C".
 540  *
 541  * For example, if LANGUAGE=de:en_US, then the returned list is
 542  * "de", "en_US", "en", "C".
 543  *
 544  * This function consults the environment variables `LANGUAGE`, `LC_ALL`,
 545  * `LC_MESSAGES` and `LANG` to find the list of locales specified by the
 546  * user.
 547  *
 548  * Returns: (array zero-terminated=1) (transfer none): a %NULL-terminated array of strings owned by GLib
 549  *    that must not be modified or freed.
 550  *
 551  * Since: 2.6
 552  **/
 553 const gchar * const *
 554 g_get_language_names (void)
 555 {
 556   static GPrivate cache_private = G_PRIVATE_INIT (language_names_cache_free);
 557   GLanguageNamesCache *cache = g_private_get (&cache_private);
 558   const gchar *value;
 559
 560   if (!cache)
 561     {
 562       cache = g_new0 (GLanguageNamesCache, 1);
 563       g_private_set (&cache_private, cache);
 564     }
 565
 566   value = guess_category_value ("LC_MESSAGES");
 567   if (!value)
 568     value = "C";
 569
 570   if (!(cache->languages && strcmp (cache->languages, value) == 0))
 571     {
 572       GPtrArray *array;
 573       gchar **alist, **a;
 574
 575       g_free (cache->languages);
 576       g_strfreev (cache->language_names);
 577       cache->languages = g_strdup (value);
 578
 579       array = g_ptr_array_sized_new (8);
 580
 581       alist = g_strsplit (value, ":", 0);
 582       for (a = alist; *a; a++)
 583         append_locale_variants (array, unalias_lang (*a));
 584       g_strfreev (alist);
 585       g_ptr_array_add (array, g_strdup ("C"));
 586       g_ptr_array_add (array, NULL);
 587
 588       cache->language_names = (gchar **) g_ptr_array_free (array, FALSE);
 589     }
 590
 591   return (const gchar * const *) cache->language_names;
 592 }