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.
37 #undef GETTEXT_PACKAGE
38 #define GETTEXT_PACKAGE "iso_639"
40 #define ISO_639_XML_PATH ISO_CODES_PREFIX "/share/xml/iso-codes/iso_639.xml"
41 #define ISO_CODES_LOCALEDIR ISO_CODES_PREFIX "/share/locale"
43 #include <gst/gst-i18n-plugin.h>
50 #include "lang-tables.dat"
52 #ifndef GST_DISABLE_GST_DEBUG
54 #define GST_CAT_DEFAULT ensure_debug_category()
56 static GstDebugCategory *
57 ensure_debug_category (void)
59 static gsize cat_gonce = 0;
61 if (g_once_init_enter (&cat_gonce)) {
64 cat_done = (gsize) _gst_debug_category_new ("tag-langcodes", 0,
65 "GstTag language codes and names");
67 g_once_init_leave (&cat_gonce, cat_done);
70 return (GstDebugCategory *) cat_gonce;
75 #define ensure_debug_category() /* NOOP */
77 #endif /* GST_DISABLE_GST_DEBUG */
79 /* ------------------------------------------------------------------------- */
81 /* Loading and initing */
83 #if defined(HAVE_ISO_CODES)
85 get_val (const gchar ** names, const gchar ** vals, const gchar * name)
87 while (names != NULL && *names != NULL) {
88 if (strcmp (*names, name) == 0)
97 parse_start_element (GMarkupParseContext * ctx, const gchar * element_name,
98 const gchar ** attr_names, const gchar ** attr_vals,
99 gpointer user_data, GError ** error)
101 GHashTable *ht = (GHashTable *) user_data;
102 const gchar *c1, *c2t, *c2b, *name, *tname;
104 if (strcmp (element_name, "iso_639_entry") != 0)
107 c1 = get_val (attr_names, attr_vals, "iso_639_1_code");
109 /* only interested in languages with an ISO 639-1 code for now */
113 c2t = get_val (attr_names, attr_vals, "iso_639_2T_code");
114 c2b = get_val (attr_names, attr_vals, "iso_639_2B_code");
115 name = get_val (attr_names, attr_vals, "name");
117 if (c2t == NULL || c2b == NULL || name == NULL) {
118 GST_WARNING ("broken iso_639.xml entry: c2t=%p, c2b=%p, name=%p", c2t,
123 /* translate language name */
126 /* if no translation was found, it will return the input string, which we
127 * we don't want to put into the hash table because it will be freed again */
128 if (G_UNLIKELY (tname == name))
129 tname = g_intern_string (name);
131 /* now overwrite default/fallback mappings with names in locale language */
132 g_hash_table_replace (ht, (gpointer) g_intern_string (c1), (gpointer) tname);
133 g_hash_table_replace (ht, (gpointer) g_intern_string (c2b), (gpointer) tname);
134 if (strcmp (c2t, c2b) != 0) {
135 g_hash_table_replace (ht, (gpointer) g_intern_string (c2t),
139 GST_LOG ("%s %s %s : %s - %s", c1, c2t, c2b, name, tname);
143 gst_tag_load_iso_639_xml (GHashTable * ht)
151 GST_DEBUG ("binding text domain %s to locale dir %s", GETTEXT_PACKAGE,
152 ISO_CODES_LOCALEDIR);
153 bindtextdomain (GETTEXT_PACKAGE, ISO_CODES_LOCALEDIR);
154 bind_textdomain_codeset (GETTEXT_PACKAGE, "UTF-8");
157 f = g_mapped_file_new (ISO_639_XML_PATH, FALSE, NULL);
159 xml_data = (gchar *) g_mapped_file_get_contents (f);
160 xml_len = g_mapped_file_get_length (f);
162 if (!g_file_get_contents (ISO_639_XML_PATH, &xml_data, &xml_len, &err)) {
163 GST_WARNING ("Could not read %s: %s", ISO_639_XML_PATH, err->message);
169 if (g_utf8_validate (xml_data, xml_len, NULL)) {
170 GMarkupParser xml_parser = { parse_start_element, NULL, NULL, NULL, NULL };
171 GMarkupParseContext *ctx;
173 ctx = g_markup_parse_context_new (&xml_parser, 0, ht, NULL);
174 if (!g_markup_parse_context_parse (ctx, xml_data, xml_len, &err)) {
175 GST_WARNING ("Parsing iso_639.xml failed: %s", err->message);
178 g_markup_parse_context_free (ctx);
180 GST_WARNING ("iso_639.xml file is not valid UTF-8");
181 GST_MEMDUMP ("iso_639.xml file", (guint8 *) xml_data, xml_len);
184 /* ... and clean up */
186 g_mapped_file_unref (f);
190 #endif /* HAVE_ISO_CODES */
193 gst_tag_get_iso_639_ht (void)
195 static gsize once_val = 0;
198 if (g_once_init_enter (&once_val)) {
202 GST_MEMDUMP ("iso 639 language names (internal default/fallback)",
203 (guint8 *) iso_639_names, sizeof (iso_639_names));
205 /* maps code -> language name; all strings are either interned strings
206 * or const static strings from lang-table.c */
207 ht = g_hash_table_new (g_str_hash, g_str_equal);
209 /* set up default/fallback mappings */
210 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
211 GST_LOG ("%3d %s %s %c%c 0x%04x %s", i, iso_639_codes[i].iso_639_1,
212 iso_639_codes[i].iso_639_2,
213 ((iso_639_codes[i].flags & ISO_639_FLAG_2B)) ? 'B' : '.',
214 ((iso_639_codes[i].flags & ISO_639_FLAG_2T)) ? 'T' : '.',
215 iso_639_codes[i].name_offset,
216 iso_639_names + iso_639_codes[i].name_offset);
218 #ifdef HAVE_ISO_CODES
219 /* intern these in order to minimise allocations when interning strings
220 * read from the xml file later */
221 g_intern_static_string (iso_639_codes[i].iso_639_1);
222 g_intern_static_string (iso_639_codes[i].iso_639_2);
223 g_intern_static_string (iso_639_names + iso_639_codes[i].name_offset);
226 /* and add default mapping (these strings are always valid) */
227 g_hash_table_insert (ht, (gpointer) iso_639_codes[i].iso_639_1,
228 (gpointer) (iso_639_names + iso_639_codes[i].name_offset));
229 g_hash_table_insert (ht, (gpointer) iso_639_codes[i].iso_639_2,
230 (gpointer) (iso_639_names + iso_639_codes[i].name_offset));
233 #ifdef HAVE_ISO_CODES
235 GstClockTime ts = gst_util_get_timestamp ();
237 gst_tag_load_iso_639_xml (ht);
239 ts = gst_util_get_timestamp () - ts;
240 GST_INFO ("iso_639.xml loading took %.2gms", (double) ts / GST_MSECOND);
243 GST_INFO ("iso-codes disabled or not available");
246 done_val = (gsize) ht;
247 g_once_init_leave (&once_val, done_val);
250 return (GHashTable *) once_val;
253 /* ------------------------------------------------------------------------- */
256 qsort_strcmp_func (const void *p1, const void *p2)
258 return strcmp (*(char *const *) p1, *(char *const *) p2);
262 * gst_tag_get_language_codes:
264 * Returns a list of known language codes (in form of two-letter ISO-639-1
265 * codes). This is useful for UIs to build a list of available languages for
266 * tagging purposes (e.g. to tag an audio track appropriately in a video or
269 * Returns: (transfer full): NULL-terminated string array with two-letter
270 * language codes. Free with g_strfreev() when no longer needed.
273 gst_tag_get_language_codes (void)
281 ensure_debug_category ();
283 ht = gst_tag_get_iso_639_ht ();
285 /* we have at least two keys for each language (-1 code and -2 code) */
286 codes = g_new (gchar *, (g_hash_table_size (ht) / 2) + 1);
289 g_hash_table_iter_init (&iter, ht);
290 while (g_hash_table_iter_next (&iter, &key, NULL)) {
291 const gchar *lang_code = key;
293 if (strlen (lang_code) == 2) {
294 codes[i] = g_strdup (lang_code);
300 /* be nice and sort the list */
301 qsort (&codes[0], i, sizeof (gchar *), qsort_strcmp_func);
307 * gst_tag_get_language_name:
308 * @language_code: two or three-letter ISO-639 language code
310 * Returns the name of the language given an ISO-639 language code as
311 * found in a GST_TAG_LANGUAGE_CODE tag. The name will be translated
312 * according to the current locale (if the library was built against the
313 * iso-codes package, otherwise the English name will be returned).
315 * Language codes are case-sensitive and expected to be lower case.
317 * Returns: language name in UTF-8 format, or NULL if @language_code could
318 * not be mapped to a language name. The returned string must not be
319 * modified and does not need to freed; it will stay valid until the
320 * application is terminated.
323 gst_tag_get_language_name (const gchar * language_code)
325 const gchar *lang_name;
328 g_return_val_if_fail (language_code != NULL, NULL);
330 ensure_debug_category ();
332 ht = gst_tag_get_iso_639_ht ();
334 lang_name = g_hash_table_lookup (ht, (gpointer) language_code);
335 GST_LOG ("%s -> %s", language_code, GST_STR_NULL (lang_name));
341 * gst_tag_get_language_code_iso_639_1:
342 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
344 * Returns two-letter ISO-639-1 language code given a three-letter ISO-639-2
345 * language code or two-letter ISO-639-1 language code (both are accepted for
348 * Language codes are case-sensitive and expected to be lower case.
350 * Returns: two-letter ISO-639-1 language code string that maps to @lang_code,
351 * or NULL if no mapping is known. The returned string must not be
355 gst_tag_get_language_code_iso_639_1 (const gchar * lang_code)
357 const gchar *c = NULL;
360 g_return_val_if_fail (lang_code != NULL, NULL);
362 ensure_debug_category ();
364 /* FIXME: we are being a bit inconsistent here in the sense that will only
365 * map the language codes from our static table. Theoretically the iso-codes
366 * XML file might have had additional codes that are now in the hash table.
367 * We keep it simple for now and don't waste memory on additional tables. */
368 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
369 /* we check both codes here, so function can be used in a more versatile
370 * way, to convert a language tag to a two-letter language code and/or
371 * verify an existing code */
372 if (strcmp (lang_code, iso_639_codes[i].iso_639_1) == 0 ||
373 strcmp (lang_code, iso_639_codes[i].iso_639_2) == 0) {
374 c = iso_639_codes[i].iso_639_1;
379 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));
385 gst_tag_get_language_code_iso_639_2X (const gchar * lang_code, guint8 flags)
389 /* FIXME: we are being a bit inconsistent here in the sense that we will only
390 * map the language codes from our static table. Theoretically the iso-codes
391 * XML file might have had additional codes that are now in the hash table.
392 * We keep it simple for now and don't waste memory on additional tables.
393 * Also, we currently only parse the iso_639.xml file if language names or
394 * a list of all codes is requested, and it'd be nice to keep it like that. */
395 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
396 /* we check both codes here, so function can be used in a more versatile
397 * way, to convert a language tag to a three-letter language code and/or
398 * verify an existing code */
399 if (strcmp (lang_code, iso_639_codes[i].iso_639_1) == 0 ||
400 strcmp (lang_code, iso_639_codes[i].iso_639_2) == 0) {
401 if ((iso_639_codes[i].flags & flags) == flags) {
402 return iso_639_codes[i].iso_639_2;
403 } else if (i > 0 && (iso_639_codes[i - 1].flags & flags) == flags &&
404 iso_639_codes[i].name_offset == iso_639_codes[i - 1].name_offset) {
405 return iso_639_codes[i - 1].iso_639_2;
406 } else if ((i + 1) < G_N_ELEMENTS (iso_639_codes) &&
407 (iso_639_codes[i + 1].flags & flags) == flags &&
408 iso_639_codes[i].name_offset == iso_639_codes[i + 1].name_offset) {
409 return iso_639_codes[i + 1].iso_639_2;
417 * gst_tag_get_language_code_iso_639_2T:
418 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
420 * Returns three-letter ISO-639-2 "terminological" language code given a
421 * two-letter ISO-639-1 language code or a three-letter ISO-639-2 language
422 * code (both are accepted for convenience).
424 * The "terminological" code is derived from the local name of the language
425 * (e.g. "deu" for German instead of "ger"). In most scenarios, the
426 * "terminological" codes are prefered over the "bibliographic" ones.
428 * Language codes are case-sensitive and expected to be lower case.
430 * Returns: three-letter ISO-639-2 language code string that maps to @lang_code,
431 * or NULL if no mapping is known. The returned string must not be
435 gst_tag_get_language_code_iso_639_2T (const gchar * lang_code)
439 g_return_val_if_fail (lang_code != NULL, NULL);
441 ensure_debug_category ();
443 c = gst_tag_get_language_code_iso_639_2X (lang_code, ISO_639_FLAG_2T);
445 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));
451 * gst_tag_get_language_code_iso_639_2B:
452 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
454 * Returns three-letter ISO-639-2 "bibliographic" language code given a
455 * two-letter ISO-639-1 language code or a three-letter ISO-639-2 language
456 * code (both are accepted for convenience).
458 * The "bibliographic" code is derived from the English name of the language
459 * (e.g. "ger" for German instead of "de" or "deu"). In most scenarios, the
460 * "terminological" codes are prefered.
462 * Language codes are case-sensitive and expected to be lower case.
464 * Returns: three-letter ISO-639-2 language code string that maps to @lang_code,
465 * or NULL if no mapping is known. The returned string must not be
469 gst_tag_get_language_code_iso_639_2B (const gchar * lang_code)
473 g_return_val_if_fail (lang_code != NULL, NULL);
475 ensure_debug_category ();
477 c = gst_tag_get_language_code_iso_639_2X (lang_code, ISO_639_FLAG_2B);
479 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));
485 * gst_tag_check_language_code:
486 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
488 * Check if a given string contains a known ISO 639 language code.
490 * This is useful in situations where it's not clear whether a given
491 * string is a language code (which should be put into a #GST_TAG_LANGUAGE_CODE
492 * tag) or a free-form language name descriptor (which should be put into a
493 * #GST_TAG_LANGUAGE_NAME tag instead).
495 * Returns: TRUE if the two- or three-letter language code in @lang_code
496 * is a valid ISO-639 language code.
499 gst_tag_check_language_code (const gchar * lang_code)
501 return (gst_tag_get_language_code_iso_639_1 (lang_code) != NULL);