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 /* FIXME: remove once we depend on GLib >= 2.22 */
55 #if !GLIB_CHECK_VERSION (2, 22, 0)
56 #define g_mapped_file_unref g_mapped_file_free
59 #ifndef GST_DISABLE_GST_DEBUG
61 #define GST_CAT_DEFAULT ensure_debug_category()
63 static GstDebugCategory *
64 ensure_debug_category (void)
66 static gsize cat_gonce = 0;
68 if (g_once_init_enter (&cat_gonce)) {
71 cat_done = (gsize) _gst_debug_category_new ("tag-langcodes", 0,
72 "GstTag language codes and names");
74 g_once_init_leave (&cat_gonce, cat_done);
77 return (GstDebugCategory *) cat_gonce;
82 #define ensure_debug_category() /* NOOP */
84 #endif /* GST_DISABLE_GST_DEBUG */
86 /* ------------------------------------------------------------------------- */
88 /* Loading and initing */
90 #if defined(HAVE_ISO_CODES)
92 get_val (const gchar ** names, const gchar ** vals, const gchar * name)
94 while (names != NULL && *names != NULL) {
95 if (strcmp (*names, name) == 0)
104 parse_start_element (GMarkupParseContext * ctx, const gchar * element_name,
105 const gchar ** attr_names, const gchar ** attr_vals,
106 gpointer user_data, GError ** error)
108 GHashTable *ht = (GHashTable *) user_data;
109 const gchar *c1, *c2t, *c2b, *name, *tname;
111 if (strcmp (element_name, "iso_639_entry") != 0)
114 c1 = get_val (attr_names, attr_vals, "iso_639_1_code");
116 /* only interested in languages with an ISO 639-1 code for now */
120 c2t = get_val (attr_names, attr_vals, "iso_639_2T_code");
121 c2b = get_val (attr_names, attr_vals, "iso_639_2B_code");
122 name = get_val (attr_names, attr_vals, "name");
124 if (c2t == NULL || c2b == NULL || name == NULL) {
125 GST_WARNING ("broken iso_639.xml entry: c2t=%p, c2b=%p, name=%p", c2t,
130 /* translate language name */
133 /* if no translation was found, it will return the input string, which we
134 * we don't want to put into the hash table because it will be freed again */
135 if (G_UNLIKELY (tname == name))
136 tname = g_intern_string (name);
138 /* now overwrite default/fallback mappings with names in locale language */
139 g_hash_table_replace (ht, (gpointer) g_intern_string (c1), (gpointer) tname);
140 g_hash_table_replace (ht, (gpointer) g_intern_string (c2b), (gpointer) tname);
141 if (strcmp (c2t, c2b) != 0) {
142 g_hash_table_replace (ht, (gpointer) g_intern_string (c2t),
146 GST_LOG ("%s %s %s : %s - %s", c1, c2t, c2b, name, tname);
150 gst_tag_load_iso_639_xml (GHashTable * ht)
158 GST_DEBUG ("binding text domain %s to locale dir %s", GETTEXT_PACKAGE,
159 ISO_CODES_LOCALEDIR);
160 bindtextdomain (GETTEXT_PACKAGE, ISO_CODES_LOCALEDIR);
161 bind_textdomain_codeset (GETTEXT_PACKAGE, "UTF-8");
164 f = g_mapped_file_new (ISO_639_XML_PATH, FALSE, NULL);
166 xml_data = (gchar *) g_mapped_file_get_contents (f);
167 xml_len = g_mapped_file_get_length (f);
169 if (!g_file_get_contents (ISO_639_XML_PATH, &xml_data, &xml_len, &err)) {
170 GST_WARNING ("Could not read %s: %s", ISO_639_XML_PATH, err->message);
176 if (g_utf8_validate (xml_data, xml_len, NULL)) {
177 GMarkupParser xml_parser = { parse_start_element, NULL, NULL, NULL, NULL };
178 GMarkupParseContext *ctx;
180 ctx = g_markup_parse_context_new (&xml_parser, 0, ht, NULL);
181 if (!g_markup_parse_context_parse (ctx, xml_data, xml_len, &err)) {
182 GST_WARNING ("Parsing iso_639.xml failed: %s", err->message);
185 g_markup_parse_context_free (ctx);
187 GST_WARNING ("iso_639.xml file is not valid UTF-8");
188 GST_MEMDUMP ("iso_639.xml file", (guint8 *) xml_data, xml_len);
191 /* ... and clean up */
193 g_mapped_file_unref (f);
197 #endif /* HAVE_ISO_CODES */
200 gst_tag_get_iso_639_ht (void)
202 static gsize once_val = 0;
205 if (g_once_init_enter (&once_val)) {
209 GST_MEMDUMP ("iso 639 language names (internal default/fallback)",
210 (guint8 *) iso_639_names, sizeof (iso_639_names));
212 /* maps code -> language name; all strings are either interned strings
213 * or const static strings from lang-table.c */
214 ht = g_hash_table_new (g_str_hash, g_str_equal);
216 /* set up default/fallback mappings */
217 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
218 GST_LOG ("%3d %s %s %c%c 0x%04x %s", i, iso_639_codes[i].iso_639_1,
219 iso_639_codes[i].iso_639_2,
220 ((iso_639_codes[i].flags & ISO_639_FLAG_2B)) ? 'B' : '.',
221 ((iso_639_codes[i].flags & ISO_639_FLAG_2T)) ? 'T' : '.',
222 iso_639_codes[i].name_offset,
223 iso_639_names + iso_639_codes[i].name_offset);
225 #ifdef HAVE_ISO_CODES
226 /* intern these in order to minimise allocations when interning strings
227 * read from the xml file later */
228 g_intern_static_string (iso_639_codes[i].iso_639_1);
229 g_intern_static_string (iso_639_codes[i].iso_639_2);
230 g_intern_static_string (iso_639_names + iso_639_codes[i].name_offset);
233 /* and add default mapping (these strings are always valid) */
234 g_hash_table_insert (ht, (gpointer) iso_639_codes[i].iso_639_1,
235 (gpointer) (iso_639_names + iso_639_codes[i].name_offset));
236 g_hash_table_insert (ht, (gpointer) iso_639_codes[i].iso_639_2,
237 (gpointer) (iso_639_names + iso_639_codes[i].name_offset));
240 #ifdef HAVE_ISO_CODES
242 GstClockTime ts = gst_util_get_timestamp ();
244 gst_tag_load_iso_639_xml (ht);
246 ts = gst_util_get_timestamp () - ts;
247 GST_INFO ("iso_639.xml loading took %.2gms", (double) ts / GST_MSECOND);
250 GST_INFO ("iso-codes disabled or not available");
253 done_val = (gsize) ht;
254 g_once_init_leave (&once_val, done_val);
257 return (GHashTable *) once_val;
260 /* ------------------------------------------------------------------------- */
263 qsort_strcmp_func (const void *p1, const void *p2)
265 return strcmp (*(char *const *) p1, *(char *const *) p2);
269 * gst_tag_get_language_codes:
271 * Returns a list of known language codes (in form of two-letter ISO-639-1
272 * codes). This is useful for UIs to build a list of available languages for
273 * tagging purposes (e.g. to tag an audio track appropriately in a video or
276 * Returns: NULL-terminated string array with two-letter language codes. Free
277 * with g_strfreev() when no longer needed.
282 gst_tag_get_language_codes (void)
290 ensure_debug_category ();
292 ht = gst_tag_get_iso_639_ht ();
294 /* we have at least two keys for each language (-1 code and -2 code) */
295 codes = g_new (gchar *, (g_hash_table_size (ht) / 2) + 1);
298 g_hash_table_iter_init (&iter, ht);
299 while (g_hash_table_iter_next (&iter, &key, NULL)) {
300 const gchar *lang_code = key;
302 if (strlen (lang_code) == 2) {
303 codes[i] = g_strdup (lang_code);
309 /* be nice and sort the list */
310 qsort (&codes[0], i, sizeof (gchar *), qsort_strcmp_func);
316 * gst_tag_get_language_name:
317 * @language_code: two or three-letter ISO-639 language code
319 * Returns the name of the language given an ISO-639 language code, such
320 * as often found in a GST_TAG_LANGUAGE tag. The name will be translated
321 * according to the current locale (if the library was built against the
322 * iso-codes package, otherwise the English name will be returned).
324 * Language codes are case-sensitive and expected to be lower case.
326 * Returns: language name in UTF-8 format, or NULL if @language_code could
327 * not be mapped to a language name. The returned string must not be
328 * modified and does not need to freed; it will stay valid until the
329 * application is terminated.
334 gst_tag_get_language_name (const gchar * language_code)
336 const gchar *lang_name;
339 g_return_val_if_fail (language_code != NULL, NULL);
341 ensure_debug_category ();
343 ht = gst_tag_get_iso_639_ht ();
345 lang_name = g_hash_table_lookup (ht, (gpointer) language_code);
346 GST_LOG ("%s -> %s", language_code, GST_STR_NULL (lang_name));
352 * gst_tag_get_language_code_iso_639_1:
353 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
355 * Returns two-letter ISO-639-1 language code given a three-letter ISO-639-2
356 * language code or two-letter ISO-639-1 language code (both are accepted for
359 * Language codes are case-sensitive and expected to be lower case.
361 * Returns: two-letter ISO-639-1 language code string that maps to @lang_code,
362 * or NULL if no mapping is known. The returned string must not be
368 gst_tag_get_language_code_iso_639_1 (const gchar * lang_code)
370 const gchar *c = NULL;
373 g_return_val_if_fail (lang_code != NULL, NULL);
375 ensure_debug_category ();
377 /* FIXME: we are being a bit inconsistent here in the sense that will only
378 * map the language codes from our static table. Theoretically the iso-codes
379 * XML file might have had additional codes that are now in the hash table.
380 * We keep it simple for now and don't waste memory on additional tables. */
381 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
382 /* we check both codes here, so function can be used in a more versatile
383 * way, to convert a language tag to a two-letter language code and/or
384 * verify an existing code */
385 if (strcmp (lang_code, iso_639_codes[i].iso_639_1) == 0 ||
386 strcmp (lang_code, iso_639_codes[i].iso_639_2) == 0) {
387 c = iso_639_codes[i].iso_639_1;
392 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));
398 gst_tag_get_language_code_iso_639_2X (const gchar * lang_code, guint8 flags)
402 /* FIXME: we are being a bit inconsistent here in the sense that we will only
403 * map the language codes from our static table. Theoretically the iso-codes
404 * XML file might have had additional codes that are now in the hash table.
405 * We keep it simple for now and don't waste memory on additional tables.
406 * Also, we currently only parse the iso_639.xml file if language names or
407 * a list of all codes is requested, and it'd be nice to keep it like that. */
408 for (i = 0; i < G_N_ELEMENTS (iso_639_codes); ++i) {
409 /* we check both codes here, so function can be used in a more versatile
410 * way, to convert a language tag to a three-letter language code and/or
411 * verify an existing code */
412 if (strcmp (lang_code, iso_639_codes[i].iso_639_1) == 0 ||
413 strcmp (lang_code, iso_639_codes[i].iso_639_2) == 0) {
414 if ((iso_639_codes[i].flags & flags) == flags) {
415 return iso_639_codes[i].iso_639_2;
416 } else if (i > 0 && (iso_639_codes[i - 1].flags & flags) == flags &&
417 iso_639_codes[i].name_offset == iso_639_codes[i - 1].name_offset) {
418 return iso_639_codes[i - 1].iso_639_2;
419 } else if (i < G_N_ELEMENTS (iso_639_codes) &&
420 (iso_639_codes[i + 1].flags & flags) == flags &&
421 iso_639_codes[i].name_offset == iso_639_codes[i + 1].name_offset) {
422 return iso_639_codes[i + 1].iso_639_2;
430 * gst_tag_get_language_code_iso_639_2T:
431 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
433 * Returns three-letter ISO-639-2 "terminological" language code given a
434 * two-letter ISO-639-1 language code or a three-letter ISO-639-2 language
435 * code (both are accepted for convenience).
437 * The "terminological" code is derived from the local name of the language
438 * (e.g. "deu" for German instead of "ger"). In most scenarios, the
439 * "terminological" codes are prefered over the "bibliographic" ones.
441 * Language codes are case-sensitive and expected to be lower case.
443 * Returns: three-letter ISO-639-2 language code string that maps to @lang_code,
444 * or NULL if no mapping is known. The returned string must not be
450 gst_tag_get_language_code_iso_639_2T (const gchar * lang_code)
454 g_return_val_if_fail (lang_code != NULL, NULL);
456 ensure_debug_category ();
458 c = gst_tag_get_language_code_iso_639_2X (lang_code, ISO_639_FLAG_2T);
460 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));
466 * gst_tag_get_language_code_iso_639_2B:
467 * @lang_code: ISO-639 language code (e.g. "deu" or "ger" or "de")
469 * Returns three-letter ISO-639-2 "bibliographic" language code given a
470 * two-letter ISO-639-1 language code or a three-letter ISO-639-2 language
471 * code (both are accepted for convenience).
473 * The "bibliographic" code is derived from the English name of the language
474 * (e.g. "ger" for German instead of "de" or "deu"). In most scenarios, the
475 * "terminological" codes are prefered.
477 * Language codes are case-sensitive and expected to be lower case.
479 * Returns: three-letter ISO-639-2 language code string that maps to @lang_code,
480 * or NULL if no mapping is known. The returned string must not be
486 gst_tag_get_language_code_iso_639_2B (const gchar * lang_code)
490 g_return_val_if_fail (lang_code != NULL, NULL);
492 ensure_debug_category ();
494 c = gst_tag_get_language_code_iso_639_2X (lang_code, ISO_639_FLAG_2B);
496 GST_LOG ("%s -> %s", lang_code, GST_STR_NULL (c));