1 /* GStreamer language codes and names utility functions
2 * Copyright (C) 2009 Tim-Philipp Müller <tim centricular net>
4 * This library is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Library General Public
6 * License as published by the Free Software Foundation; either
7 * version 2 of the License, or (at your option) any later version.
9 * This library 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 GNU
12 * Library General Public License for more details.
14 * You should have received a copy of the GNU Library General Public
15 * License along with this library; if not, write to the
16 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
17 * Boston, MA 02111-1307, USA.
21 * SECTION:gsttaglanguagecodes
22 * @short_description: mappings for ISO-639 language codes and names
23 * @see_also: #GstTagList
27 * Provides helper functions to convert between the various ISO-639 language
28 * codes, and to map language codes to language names.
33 /* FIXME 0.11: maybe switch to ISO-639-2 everywhere incl. GST_TAG_LANGUAGE? */
39 #undef GETTEXT_PACKAGE
40 #define GETTEXT_PACKAGE "iso_639"
42 #define ISO_639_XML_PATH ISO_CODES_PREFIX "/share/xml/iso-codes/iso_639.xml"
43 #define ISO_CODES_LOCALEDIR ISO_CODES_PREFIX "/share/locale"
45 #include <gst/gst-i18n-plugin.h>
52 #include "lang-tables.dat"
54 #ifndef GST_DISABLE_GST_DEBUG
56 #define GST_CAT_DEFAULT ensure_debug_category()
58 static GstDebugCategory *
59 ensure_debug_category (void)
61 static gsize cat_gonce = 0;
63 if (g_once_init_enter (&cat_gonce)) {
66 cat_done = (gsize) _gst_debug_category_new ("tag-langcodes", 0,
67 "GstTag language codes and names");
69 g_once_init_leave (&cat_gonce, cat_done);
72 return (GstDebugCategory *) cat_gonce;
77 #define ensure_debug_category() /* NOOP */
79 #endif /* GST_DISABLE_GST_DEBUG */
81 /* ------------------------------------------------------------------------- */
83 /* Loading and initing */
85 #if defined(HAVE_ISO_CODES)
87 get_val (const gchar ** names, const gchar ** vals, const gchar * name)
89 while (names != NULL && *names != NULL) {
90 if (strcmp (*names, name) == 0)
99 parse_start_element (GMarkupParseContext * ctx, const gchar * element_name,
100 const gchar ** attr_names, const gchar ** attr_vals,
101 gpointer user_data, GError ** error)
103 GHashTable *ht = (GHashTable *) user_data;
104 const gchar *c1, *c2t, *c2b, *name, *tname;
106 if (strcmp (element_name, "iso_639_entry") != 0)
109 c1 = get_val (attr_names, attr_vals, "iso_639_1_code");
111 /* only interested in languages with an ISO 639-1 code for now */
115 c2t = get_val (attr_names, attr_vals, "iso_639_2T_code");
116 c2b = get_val (attr_names, attr_vals, "iso_639_2B_code");
117 name = get_val (attr_names, attr_vals, "name");
119 if (c2t == NULL || c2b == NULL || name == NULL) {
120 GST_WARNING ("broken iso_639.xml entry: c2t=%p, c2b=%p, name=%p", c2t,
125 /* translate language name */
128 /* if no translation was found, it will return the input string, which we
129 * we don't want to put into the hash table because it will be freed again */
130 if (G_UNLIKELY (tname == name))
131 tname = g_intern_string (name);
133 /* now overwrite default/fallback mappings with names in locale language */
134 g_hash_table_replace (ht, (gpointer) g_intern_string (c1), (gpointer) tname);
135 g_hash_table_replace (ht, (gpointer) g_intern_string (c2b), (gpointer) tname);
136 if (strcmp (c2t, c2b) != 0) {
137 g_hash_table_replace (ht, (gpointer) g_intern_string (c2t),
141 GST_LOG ("%s %s %s : %s - %s", c1, c2t, c2b, name, tname);
145 gst_tag_load_iso_639_xml (GHashTable * ht)
153 GST_DEBUG ("binding text domain %s to locale dir %s", GETTEXT_PACKAGE,
154 ISO_CODES_LOCALEDIR);
155 bindtextdomain (GETTEXT_PACKAGE, ISO_CODES_LOCALEDIR);
156 bind_textdomain_codeset (GETTEXT_PACKAGE, "UTF-8");
159 f = g_mapped_file_new (ISO_639_XML_PATH, FALSE, NULL);
161 xml_data = (gchar *) g_mapped_file_get_contents (f);
162 xml_len = g_mapped_file_get_length (f);
164 if (!g_file_get_contents (ISO_639_XML_PATH, &xml_data, &xml_len, &err)) {
165 GST_WARNING ("Could not read %s: %s", ISO_639_XML_PATH, err->message);
171 if (g_utf8_validate (xml_data, xml_len, NULL)) {
172 GMarkupParser xml_parser = { parse_start_element, NULL, NULL, NULL, NULL };
173 GMarkupParseContext *ctx;
175 ctx = g_markup_parse_context_new (&xml_parser, 0, ht, NULL);
176 if (!g_markup_parse_context_parse (ctx, xml_data, xml_len, &err)) {
177 GST_WARNING ("Parsing iso_639.xml failed: %s", err->message);
180 g_markup_parse_context_free (ctx);
182 GST_WARNING ("iso_639.xml file is not valid UTF-8");
183 GST_MEMDUMP ("iso_639.xml file", (guint8 *) xml_data, xml_len);
186 /* ... and clean up */
188 g_mapped_file_unref (f);
192 #endif /* HAVE_ISO_CODES */
195 gst_tag_get_iso_639_ht (void)
197 static gsize once_val = 0;
200 if (g_once_init_enter (&once_val)) {
204 GST_MEMDUMP ("iso 639 language names (internal default/fallback)",
205 (guint8 *) iso_639_names, sizeof (iso_639_names));
207 /* maps code -> language name; all strings are either interned strings
208 * or const static strings from lang-table.c */
209 ht = g_hash_table_new (g_str_hash, g_str_equal);
211 /* set up default/fallback mappings */
212 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
213 GST_LOG ("%3d %s %s %c%c 0x%04x %s", i, iso_639_codes[i].iso_639_1,
214 iso_639_codes[i].iso_639_2,
215 ((iso_639_codes[i].flags & ISO_639_FLAG_2B)) ? 'B' : '.',
216 ((iso_639_codes[i].flags & ISO_639_FLAG_2T)) ? 'T' : '.',
217 iso_639_codes[i].name_offset,
218 iso_639_names + iso_639_codes[i].name_offset);
220 #ifdef HAVE_ISO_CODES
221 /* intern these in order to minimise allocations when interning strings
222 * read from the xml file later */
223 g_intern_static_string (iso_639_codes[i].iso_639_1);
224 g_intern_static_string (iso_639_codes[i].iso_639_2);
225 g_intern_static_string (iso_639_names + iso_639_codes[i].name_offset);
228 /* and add default mapping (these strings are always valid) */
229 g_hash_table_insert (ht, (gpointer) iso_639_codes[i].iso_639_1,
230 (gpointer) (iso_639_names + iso_639_codes[i].name_offset));
231 g_hash_table_insert (ht, (gpointer) iso_639_codes[i].iso_639_2,
232 (gpointer) (iso_639_names + iso_639_codes[i].name_offset));
235 #ifdef HAVE_ISO_CODES
237 GstClockTime ts = gst_util_get_timestamp ();
239 gst_tag_load_iso_639_xml (ht);
241 ts = gst_util_get_timestamp () - ts;
242 GST_INFO ("iso_639.xml loading took %.2gms", (double) ts / GST_MSECOND);
245 GST_INFO ("iso-codes disabled or not available");
248 done_val = (gsize) ht;
249 g_once_init_leave (&once_val, done_val);
252 return (GHashTable *) once_val;
255 /* ------------------------------------------------------------------------- */
258 qsort_strcmp_func (const void *p1, const void *p2)
260 return strcmp (*(char *const *) p1, *(char *const *) p2);
264 * gst_tag_get_language_codes:
266 * Returns a list of known language codes (in form of two-letter ISO-639-1
267 * codes). This is useful for UIs to build a list of available languages for
268 * tagging purposes (e.g. to tag an audio track appropriately in a video or
271 * Returns: NULL-terminated string array with two-letter language codes. Free
272 * with g_strfreev() when no longer needed.
277 gst_tag_get_language_codes (void)
285 ensure_debug_category ();
287 ht = gst_tag_get_iso_639_ht ();
289 /* we have at least two keys for each language (-1 code and -2 code) */
290 codes = g_new (gchar *, (g_hash_table_size (ht) / 2) + 1);
293 g_hash_table_iter_init (&iter, ht);
294 while (g_hash_table_iter_next (&iter, &key, NULL)) {
295 const gchar *lang_code = key;
297 if (strlen (lang_code) == 2) {
298 codes[i] = g_strdup (lang_code);
304 /* be nice and sort the list */
305 qsort (&codes[0], i, sizeof (gchar *), qsort_strcmp_func);
311 * gst_tag_get_language_name:
312 * @language_code: two or three-letter ISO-639 language code
314 * Returns the name of the language given an ISO-639 language code, such
315 * as often found in a GST_TAG_LANGUAGE tag. The name will be translated
316 * according to the current locale (if the library was built against the
317 * iso-codes package, otherwise the English name will be returned).
319 * Language codes are case-sensitive and expected to be lower case.
321 * Returns: language name in UTF-8 format, or NULL if @language_code could
322 * not be mapped to a language name. The returned string must not be
323 * modified and does not need to freed; it will stay valid until the
324 * application is terminated.
329 gst_tag_get_language_name (const gchar * language_code)
331 const gchar *lang_name;
334 g_return_val_if_fail (language_code != NULL, NULL);
336 ensure_debug_category ();
338 ht = gst_tag_get_iso_639_ht ();
340 lang_name = g_hash_table_lookup (ht, (gpointer) language_code);
341 GST_LOG ("%s -> %s", language_code, GST_STR_NULL (lang_name));
347 * gst_tag_get_language_code_iso_639_1:
348 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
350 * Returns two-letter ISO-639-1 language code given a three-letter ISO-639-2
351 * language code or two-letter ISO-639-1 language code (both are accepted for
354 * Language codes are case-sensitive and expected to be lower case.
356 * Returns: two-letter ISO-639-1 language code string that maps to @lang_code,
357 * or NULL if no mapping is known. The returned string must not be
363 gst_tag_get_language_code_iso_639_1 (const gchar * lang_code)
365 const gchar *c = NULL;
368 g_return_val_if_fail (lang_code != NULL, NULL);
370 ensure_debug_category ();
372 /* FIXME: we are being a bit inconsistent here in the sense that will only
373 * map the language codes from our static table. Theoretically the iso-codes
374 * XML file might have had additional codes that are now in the hash table.
375 * We keep it simple for now and don't waste memory on additional tables. */
376 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
377 /* we check both codes here, so function can be used in a more versatile
378 * way, to convert a language tag to a two-letter language code and/or
379 * verify an existing code */
380 if (strcmp (lang_code, iso_639_codes[i].iso_639_1) == 0 ||
381 strcmp (lang_code, iso_639_codes[i].iso_639_2) == 0) {
382 c = iso_639_codes[i].iso_639_1;
387 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));
393 gst_tag_get_language_code_iso_639_2X (const gchar * lang_code, guint8 flags)
397 /* FIXME: we are being a bit inconsistent here in the sense that we will only
398 * map the language codes from our static table. Theoretically the iso-codes
399 * XML file might have had additional codes that are now in the hash table.
400 * We keep it simple for now and don't waste memory on additional tables.
401 * Also, we currently only parse the iso_639.xml file if language names or
402 * a list of all codes is requested, and it'd be nice to keep it like that. */
403 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
404 /* we check both codes here, so function can be used in a more versatile
405 * way, to convert a language tag to a three-letter language code and/or
406 * verify an existing code */
407 if (strcmp (lang_code, iso_639_codes[i].iso_639_1) == 0 ||
408 strcmp (lang_code, iso_639_codes[i].iso_639_2) == 0) {
409 if ((iso_639_codes[i].flags & flags) == flags) {
410 return iso_639_codes[i].iso_639_2;
411 } else if (i > 0 && (iso_639_codes[i - 1].flags & flags) == flags &&
412 iso_639_codes[i].name_offset == iso_639_codes[i - 1].name_offset) {
413 return iso_639_codes[i - 1].iso_639_2;
414 } else if ((i + 1) < G_N_ELEMENTS (iso_639_codes) &&
415 (iso_639_codes[i + 1].flags & flags) == flags &&
416 iso_639_codes[i].name_offset == iso_639_codes[i + 1].name_offset) {
417 return iso_639_codes[i + 1].iso_639_2;
425 * gst_tag_get_language_code_iso_639_2T:
426 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
428 * Returns three-letter ISO-639-2 "terminological" language code given a
429 * two-letter ISO-639-1 language code or a three-letter ISO-639-2 language
430 * code (both are accepted for convenience).
432 * The "terminological" code is derived from the local name of the language
433 * (e.g. "deu" for German instead of "ger"). In most scenarios, the
434 * "terminological" codes are prefered over the "bibliographic" ones.
436 * Language codes are case-sensitive and expected to be lower case.
438 * Returns: three-letter ISO-639-2 language code string that maps to @lang_code,
439 * or NULL if no mapping is known. The returned string must not be
445 gst_tag_get_language_code_iso_639_2T (const gchar * lang_code)
449 g_return_val_if_fail (lang_code != NULL, NULL);
451 ensure_debug_category ();
453 c = gst_tag_get_language_code_iso_639_2X (lang_code, ISO_639_FLAG_2T);
455 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));
461 * gst_tag_get_language_code_iso_639_2B:
462 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
464 * Returns three-letter ISO-639-2 "bibliographic" language code given a
465 * two-letter ISO-639-1 language code or a three-letter ISO-639-2 language
466 * code (both are accepted for convenience).
468 * The "bibliographic" code is derived from the English name of the language
469 * (e.g. "ger" for German instead of "de" or "deu"). In most scenarios, the
470 * "terminological" codes are prefered.
472 * Language codes are case-sensitive and expected to be lower case.
474 * Returns: three-letter ISO-639-2 language code string that maps to @lang_code,
475 * or NULL if no mapping is known. The returned string must not be
481 gst_tag_get_language_code_iso_639_2B (const gchar * lang_code)
485 g_return_val_if_fail (lang_code != NULL, NULL);
487 ensure_debug_category ();
489 c = gst_tag_get_language_code_iso_639_2X (lang_code, ISO_639_FLAG_2B);
491 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));