lib/ephy-langs.c

   1 /*
   2  *  Copyright © 2003, 2004 Christian Persch
   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
  16  *  Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  17  *
  18  *  $Id: ephy-langs.c 6952 2007-03-11 19:42:02Z chpe $
  19  */
  20
  21 #include "config.h"
  22
  23 #include "ephy-langs.h"
  24
  25 #include "ephy-debug.h"
  26
  27 #include <glib/gi18n.h>
  28
  29 #include <string.h>
  30
  31 #include <libxml/xmlreader.h>
  32
  33 /* If you add more language codes here, remember to also
  34  * add them to data/generate-font-schemas.py .
  35  */
  36 static const EphyFontsLanguageInfo font_languages [] =
  37 {
  38         /* Translators: The text before the "|" is context to help you decide on
  39          * the correct translation. You MUST OMIT it in the translated string. */
  40         { N_("select fonts for|Arabic"),                                "ar" },
  41         /* Translators: The text before the "|" is context to help you decide on
  42          * the correct translation. You MUST OMIT it in the translated string. */
  43         { N_("select fonts for|Baltic"),                                "x-baltic" },
  44         /* Translators: The text before the "|" is context to help you decide on
  45          * the correct translation. You MUST OMIT it in the translated string. */
  46         { N_("select fonts for|Central European"),                      "x-central-euro" },
  47         /* Translators: The text before the "|" is context to help you decide on
  48          * the correct translation. You MUST OMIT it in the translated string. */
  49         { N_("select fonts for|Cyrillic"),                              "x-cyrillic" },
  50         /* Translators: The text before the "|" is context to help you decide on
  51          * the correct translation. You MUST OMIT it in the translated string. */
  52         { N_("select fonts for|Devanagari"),                            "x-devanagari" },
  53         /* Translators: The text before the "|" is context to help you decide on
  54          * the correct translation. You MUST OMIT it in the translated string. */
  55         { N_("select fonts for|Greek"),                                 "el" },
  56         /* Translators: The text before the "|" is context to help you decide on
  57          * the correct translation. You MUST OMIT it in the translated string. */
  58         { N_("select fonts for|Hebrew"),                                "he" },
  59         /* Translators: The text before the "|" is context to help you decide on
  60          * the correct translation. You MUST OMIT it in the translated string. */
  61         { N_("select fonts for|Japanese"),                              "ja" },
  62         /* Translators: The text before the "|" is context to help you decide on
  63          * the correct translation. You MUST OMIT it in the translated string. */
  64         { N_("select fonts for|Korean"),                                "ko" },
  65         /* Translators: The text before the "|" is context to help you decide on
  66          * the correct translation. You MUST OMIT it in the translated string. */
  67         { N_("select fonts for|Simplified Chinese"),                    "zh-CN" },
  68         /* Translators: The text before the "|" is context to help you decide on
  69          * the correct translation. You MUST OMIT it in the translated string. */
  70         { N_("select fonts for|Tamil"),                                 "x-tamil" },
  71         /* Translators: The text before the "|" is context to help you decide on
  72          * the correct translation. You MUST OMIT it in the translated string. */
  73         { N_("select fonts for|Thai"),                                  "th" },
  74         /* Translators: The text before the "|" is context to help you decide on
  75          * the correct translation. You MUST OMIT it in the translated string. */
  76         { N_("select fonts for|Traditional Chinese"),                   "zh-TW" },
  77         /* Translators: The text before the "|" is context to help you decide on
  78          * the correct translation. You MUST OMIT it in the translated string. */
  79         { N_("select fonts for|Traditional Chinese (Hong Kong)"),       "zh-HK" },
  80         /* Translators: The text before the "|" is context to help you decide on
  81          * the correct translation. You MUST OMIT it in the translated string. */
  82         { N_("select fonts for|Turkish"),                               "tr" },
  83         /* Translators: The text before the "|" is context to help you decide on
  84          * the correct translation. You MUST OMIT it in the translated string. */
  85         { N_("select fonts for|Armenian"),                              "x-armn" },
  86         /* Translators: The text before the "|" is context to help you decide on
  87          * the correct translation. You MUST OMIT it in the translated string. */
  88         { N_("select fonts for|Bengali"),                               "x-beng" },
  89         /* Translators: The text before the "|" is context to help you decide on
  90          * the correct translation. You MUST OMIT it in the translated string. */
  91         { N_("select fonts for|Unified Canadian Syllabics"),            "x-cans" },
  92         /* Translators: The text before the "|" is context to help you decide on
  93          * the correct translation. You MUST OMIT it in the translated string. */
  94         { N_("select fonts for|Ethiopic"),                              "x-ethi" },
  95         /* Translators: The text before the "|" is context to help you decide on
  96          * the correct translation. You MUST OMIT it in the translated string. */
  97         { N_("select fonts for|Georgian"),                              "x-geor" },
  98         /* Translators: The text before the "|" is context to help you decide on
  99          * the correct translation. You MUST OMIT it in the translated string. */
 100         { N_("select fonts for|Gujarati"),                              "x-gujr" },
 101         /* Translators: The text before the "|" is context to help you decide on
 102          * the correct translation. You MUST OMIT it in the translated string. */
 103         { N_("select fonts for|Gurmukhi"),                              "x-guru" },
 104         /* Translators: The text before the "|" is context to help you decide on
 105          * the correct translation. You MUST OMIT it in the translated string. */
 106         { N_("select fonts for|Khmer"),                                 "x-khmr" },
 107         /* Translators: The text before the "|" is context to help you decide on
 108          * the correct translation. You MUST OMIT it in the translated string. */
 109         { N_("select fonts for|Malayalam"),                             "x-mlym" },
 110         /* Translators: The text before the "|" is context to help you decide on
 111          * the correct translation. You MUST OMIT it in the translated string. */
 112         { N_("select fonts for|Western"),                               "x-western" },
 113         /* Translators: The text before the "|" is context to help you decide on
 114          * the correct translation. You MUST OMIT it in the translated string. */
 115         { N_("select fonts for|Other Scripts"),                         "x-unicode" }
 116 };
 117
 118 const EphyFontsLanguageInfo *
 119 ephy_font_languages (void)
 120 {
 121         return font_languages;
 122 }
 123
 124 guint
 125 ephy_font_n_languages (void)
 126 {
 127         return G_N_ELEMENTS (font_languages);
 128 }
 129
 130 /* sanitise the languages list according to the rules for HTTP accept-language
 131  * in RFC 2616, Sect. 14.4
 132  */
 133 void
 134 ephy_langs_sanitise (GArray *array)
 135 {
 136         char *lang1, *lang2;
 137         int i, j;
 138
 139         /* if we have 'xy-ab' in list but not 'xy', append 'xy' */
 140         for (i = 0; i < array->len; i++)
 141         {
 142                 gboolean found = FALSE;
 143                 char *dash, *prefix;
 144
 145                 lang1 = (char *) g_array_index (array,char *, i);
 146
 147                 dash = strchr (lang1, '-');
 148                 if (dash == NULL) continue;
 149
 150                 for (j = i + 1; j < array->len; j++)
 151                 {
 152                         lang2 = (char *) g_array_index (array, char *, j);
 153                         if (strchr (lang2, '-') == NULL &&
 154                             g_str_has_prefix (lang1, lang2))
 155                         {
 156                                 found = TRUE;
 157                         }
 158                 }
 159
 160                 if (found == FALSE)
 161                 {
 162                         prefix = g_strndup (lang1, dash - lang1);
 163                         g_array_append_val (array, prefix);
 164                 }
 165         }
 166
 167         /* uniquify */
 168         for (i = 0; i < (int) array->len - 1; i++)
 169         {
 170                 for (j = (int) array->len - 1; j > i; j--)
 171                 {
 172                         lang1 = (char *) g_array_index (array,char *, i);
 173                         lang2 = (char *) g_array_index (array, char *, j);
 174
 175                         if (strcmp (lang1, lang2) == 0)
 176                         {
 177                                 g_array_remove_index (array, j);
 178                                 g_free (lang2);
 179                         }
 180                 }
 181         }
 182
 183         /* move 'xy' code behind all 'xy-ab' codes */
 184         for (i = (int) array->len - 2; i >= 0; i--)
 185         {
 186                 for (j = (int) array->len - 1; j > i; j--)
 187                 {
 188                         lang1 = (char *) g_array_index (array, char *, i);
 189                         lang2 = (char *) g_array_index (array, char *, j);
 190
 191                         if (strchr (lang1, '-') == NULL &&
 192                             strchr (lang2, '-') != NULL &&
 193                             g_str_has_prefix (lang2, lang1))
 194                         {
 195                                 g_array_insert_val (array, j + 1, lang1);
 196                                 g_array_remove_index (array, i);
 197                                 break;
 198                         }
 199                 }
 200         }
 201 }
 202
 203 void
 204 ephy_langs_append_languages (GArray *array)
 205 {
 206         const char * const * languages;
 207         char *lang;
 208         int i;
 209
 210         languages = g_get_language_names ();
 211         g_return_if_fail (languages != NULL);
 212
 213         /* FIXME: maybe just use the first, instead of all of them? */
 214         for (i = 0; languages[i] != NULL; i++)
 215         {
 216
 217                 if (strstr (languages[i], ".") == 0 &&
 218                     strstr (languages[i], "@") == 0 &&
 219                     strcmp (languages[i], "C") != 0)
 220                 {
 221                         /* change to lowercase and '_' to '-' */
 222                         lang = g_strdelimit (g_ascii_strdown
 223                                                 (languages[i], -1), "_", '-');
 224
 225                         g_array_append_val (array, lang);
 226                 }
 227         }
 228
 229         /* Fallback: add "en" if list is empty */
 230         if (array->len == 0)
 231         {
 232                 lang = g_strdup ("en");
 233                 g_array_append_val (array, lang);
 234         }
 235 }
 236
 237 char **
 238 ephy_langs_get_languages (void)
 239 {
 240         GArray *array;
 241
 242         array = g_array_new (TRUE, FALSE, sizeof (char *));
 243
 244         ephy_langs_append_languages (array);
 245
 246         ephy_langs_sanitise (array);
 247
 248         return (char **) g_array_free (array, FALSE);
 249 }
 250
 251 #define ISOCODESLOCALEDIR       ISO_CODES_PREFIX "/share/locale"
 252
 253 static void
 254 ephy_langs_bind_iso_domains (void)
 255 {
 256 #ifdef ENABLE_NLS
 257         static gboolean bound = FALSE;
 258
 259         if (bound == FALSE)
 260         {
 261                 bindtextdomain (ISO_639_DOMAIN, ISOCODESLOCALEDIR);
 262                 bind_textdomain_codeset (ISO_639_DOMAIN, "UTF-8");
 263
 264                 bindtextdomain(ISO_3166_DOMAIN, ISOCODESLOCALEDIR);
 265                 bind_textdomain_codeset (ISO_3166_DOMAIN, "UTF-8");
 266
 267                 bound = TRUE;
 268         }
 269 #endif
 270 }
 271
 272 static void
 273 read_iso_639_entry (xmlTextReaderPtr reader,
 274                     GHashTable *table)
 275 {
 276         xmlChar *code, *name;
 277
 278         code = xmlTextReaderGetAttribute (reader, (const xmlChar *) "iso_639_1_code");
 279         name = xmlTextReaderGetAttribute (reader, (const xmlChar *) "name");
 280
 281         /* Get iso-639-2 code */
 282         if (code == NULL || code[0] == '\0')
 283         {
 284                 xmlFree (code);
 285                 /* FIXME: use the 2T or 2B code? */
 286                 code = xmlTextReaderGetAttribute (reader, (const xmlChar *) "iso_639_2T_code");
 287         }
 288
 289         if (code != NULL && code[0] != '\0' && name != NULL && name[0] != '\0')
 290         {
 291                 g_hash_table_insert (table, code, name);
 292         }
 293         else
 294         {
 295                 xmlFree (code);
 296                 xmlFree (name);
 297         }
 298 }
 299
 300 static void
 301 read_iso_3166_entry (xmlTextReaderPtr reader,
 302                      GHashTable *table)
 303 {
 304         xmlChar *code, *name;
 305
 306         code = xmlTextReaderGetAttribute (reader, (const xmlChar *) "alpha_2_code");
 307         name = xmlTextReaderGetAttribute (reader, (const xmlChar *) "name");
 308
 309         if (code != NULL && code[0] != '\0' && name != NULL && name[0] != '\0')
 310         {
 311                 char *lcode;
 312
 313                 lcode = g_ascii_strdown ((char *) code, -1);
 314                 xmlFree (code);
 315
 316                 g_hash_table_insert (table, lcode, name);
 317         }
 318         else
 319         {
 320                 xmlFree (code);
 321                 xmlFree (name);
 322         }
 323
 324 }
 325
 326 typedef enum
 327 {
 328         STATE_START,
 329         STATE_STOP,
 330         STATE_ENTRIES,
 331 } ParserState;
 332
 333 static void
 334 load_iso_entries (int iso,
 335                   GFunc read_entry_func,
 336                   gpointer user_data)
 337 {
 338         xmlTextReaderPtr reader;
 339         ParserState state = STATE_START;
 340         xmlChar iso_entries[32], iso_entry[32];
 341         char *filename;
 342         int ret = -1;
 343
 344         LOG ("Loading ISO-%d codes", iso);
 345
 346         START_PROFILER ("Loading ISO codes")
 347
 348         filename = g_strdup_printf (ISO_CODES_PREFIX "/share/xml/iso-codes/iso_%d.xml", iso);
 349         reader = xmlNewTextReaderFilename (filename);
 350         if (reader == NULL) goto out;
 351
 352         xmlStrPrintf (iso_entries, sizeof (iso_entries), (const xmlChar *)"iso_%d_entries", iso);
 353         xmlStrPrintf (iso_entry, sizeof (iso_entry), (const xmlChar *)"iso_%d_entry", iso);
 354
 355         ret = xmlTextReaderRead (reader);
 356
 357         while (ret == 1)
 358         {
 359                 const xmlChar *tag;
 360                 xmlReaderTypes type;
 361
 362                 tag = xmlTextReaderConstName (reader);
 363                 type = xmlTextReaderNodeType (reader);
 364
 365                 if (state == STATE_ENTRIES &&
 366                     type == XML_READER_TYPE_ELEMENT &&
 367                     xmlStrEqual (tag, iso_entry))
 368                 {
 369                         read_entry_func (reader, user_data);
 370                 }
 371                 else if (state == STATE_START &&
 372                          type == XML_READER_TYPE_ELEMENT &&
 373                          xmlStrEqual (tag, iso_entries))
 374                 {
 375                         state = STATE_ENTRIES;
 376                 }
 377                 else if (state == STATE_ENTRIES &&
 378                          type == XML_READER_TYPE_END_ELEMENT &&
 379                          xmlStrEqual (tag, iso_entries))
 380                 {
 381                         state = STATE_STOP;
 382                 }
 383                 else if (type == XML_READER_TYPE_SIGNIFICANT_WHITESPACE ||
 384                          type == XML_READER_TYPE_WHITESPACE ||
 385                          type == XML_READER_TYPE_TEXT ||
 386                          type == XML_READER_TYPE_COMMENT)
 387                 {
 388                         /* eat it */
 389                 }
 390                 else
 391                 {
 392                         /* ignore it */
 393                 }
 394
 395                 ret = xmlTextReaderRead (reader);
 396         }
 397
 398         xmlFreeTextReader (reader);
 399
 400 out:
 401         if (ret < 0 || state != STATE_STOP)
 402         {
 403                 g_warning ("Failed to load ISO-%d codes from %s!\n",
 404                            iso, filename);
 405         }
 406
 407         g_free (filename);
 408
 409         STOP_PROFILER ("Loading ISO codes")
 410 }
 411
 412 GHashTable *
 413 ephy_langs_iso_639_table (void)
 414 {
 415         GHashTable *table;
 416
 417         ephy_langs_bind_iso_domains ();
 418         table = g_hash_table_new_full (g_str_hash, g_str_equal,
 419                                        (GDestroyNotify) xmlFree,
 420                                        (GDestroyNotify) xmlFree);
 421
 422         load_iso_entries (639, (GFunc) read_iso_639_entry, table);
 423
 424         return table;
 425 }
 426
 427 GHashTable *
 428 ephy_langs_iso_3166_table (void)
 429 {
 430         GHashTable *table;
 431
 432         ephy_langs_bind_iso_domains ();
 433         table = g_hash_table_new_full (g_str_hash, g_str_equal,
 434                                        (GDestroyNotify) g_free,
 435                                        (GDestroyNotify) xmlFree);
 436
 437         load_iso_entries (3166, (GFunc) read_iso_3166_entry, table);
 438
 439         return table;
 440 }