1 /* -*- Mode: C; tab-width: 2; indent-tabs-mode: t; c-basic-offset: 2 -*- */
2 /* Copyright 2006-2008 Tim-Philipp Müller <tim centricular net>
3 * Copyright 2005 Jan Schmidt <thaytan@mad.scientist.com>
4 * Copyright 2002,2003 Scott Wheeler <wheeler@kde.org> (portions from taglib)
6 * This library is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Library General Public
8 * License as published by the Free Software Foundation; either
9 * version 2 of the License, or (at your option) any later version.
11 * This library is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Library General Public License for more details.
16 * You should have received a copy of the GNU Library General Public
17 * License along with this library; if not, write to the
18 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 * Boston, MA 02111-1307, USA.
29 #include <gst/tag/tag.h>
30 #include <gst/base/gsttypefindhelper.h>
38 GST_DEBUG_CATEGORY_EXTERN (id3demux_debug);
39 #define GST_CAT_DEFAULT (id3demux_debug)
41 static gboolean parse_comment_frame (ID3TagsWorking * work);
42 static gchar *parse_url_link_frame (ID3TagsWorking * work,
43 const gchar ** tag_name);
44 static GArray *parse_text_identification_frame (ID3TagsWorking * work);
45 static gchar *parse_user_text_identification_frame (ID3TagsWorking * work,
46 const gchar ** tag_name);
47 static gchar *parse_unique_file_identifier (ID3TagsWorking * work,
48 const gchar ** tag_name);
49 static gboolean parse_relative_volume_adjustment_two (ID3TagsWorking * work);
50 static void parse_obsolete_tdat_frame (ID3TagsWorking * work);
51 static gboolean id3v2_tag_to_taglist (ID3TagsWorking * work,
52 const gchar * tag_name, const gchar * tag_str);
53 /* Parse a single string into an array of gchar* */
54 static void parse_split_strings (guint8 encoding, gchar * data, gint data_size,
55 GArray ** out_fields);
56 static void free_tag_strings (GArray * fields);
58 id3v2_genre_fields_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
60 static gboolean parse_picture_frame (ID3TagsWorking * work);
62 #define ID3V2_ENCODING_ISO8859 0x00
63 #define ID3V2_ENCODING_UTF16 0x01
64 #define ID3V2_ENCODING_UTF16BE 0x02
65 #define ID3V2_ENCODING_UTF8 0x03
68 id3demux_id3v2_parse_frame (ID3TagsWorking * work)
70 const gchar *tag_name;
71 gboolean result = FALSE;
73 guint8 *frame_data = work->hdr.frame_data;
74 guint frame_data_size = work->cur_frame_size;
75 gchar *tag_str = NULL;
76 GArray *tag_fields = NULL;
77 guint8 *uu_data = NULL;
80 guint8 *uncompressed_data = NULL;
83 /* Check that the frame id is valid */
84 for (i = 0; i < 5 && work->frame_id[i] != '\0'; i++) {
85 if (!g_ascii_isalnum (work->frame_id[i])) {
86 GST_DEBUG ("Encountered invalid frame_id");
91 /* Can't handle encrypted frames right now (in case we ever do, we'll have
92 * to do the decryption after the un-unsynchronisation and decompression,
94 if (work->frame_flags & ID3V2_FRAME_FORMAT_ENCRYPTION) {
95 GST_WARNING ("Encrypted frames are not supported");
99 tag_name = gst_tag_from_id3_tag (work->frame_id);
100 if (tag_name == NULL &&
101 strncmp (work->frame_id, "RVA2", 4) != 0 &&
102 strncmp (work->frame_id, "TXXX", 4) != 0 &&
103 strncmp (work->frame_id, "TDAT", 4) != 0 &&
104 strncmp (work->frame_id, "UFID", 4) != 0) {
108 if (work->frame_flags & (ID3V2_FRAME_FORMAT_COMPRESSION |
109 ID3V2_FRAME_FORMAT_DATA_LENGTH_INDICATOR)) {
110 if (work->hdr.frame_data_size <= 4)
112 if (ID3V2_VER_MAJOR (work->hdr.version) == 3) {
113 work->parse_size = GST_READ_UINT32_BE (frame_data);
115 work->parse_size = read_synch_uint (frame_data, 4);
118 frame_data_size -= 4;
119 GST_LOG ("Un-unsynced data size %d (of %d)", work->parse_size,
121 if (work->parse_size > frame_data_size) {
122 GST_WARNING ("ID3v2 frame %s data has invalid size %d (>%d)",
123 work->frame_id, work->parse_size, frame_data_size);
128 /* in v2.3 the frame sizes are not syncsafe, so the entire tag had to be
129 * unsynced. In v2.4 the frame sizes are syncsafe so it's just the frame
130 * data that needs un-unsyncing, but not the frame headers. */
131 if (ID3V2_VER_MAJOR (work->hdr.version) == 4) {
132 if ((work->hdr.flags & ID3V2_HDR_FLAG_UNSYNC) != 0 ||
133 ((work->frame_flags & ID3V2_FRAME_FORMAT_UNSYNCHRONISATION) != 0)) {
134 GST_DEBUG ("Un-unsyncing frame %s", work->frame_id);
135 uu_data = id3demux_ununsync_data (frame_data, &frame_data_size);
136 frame_data = uu_data;
137 GST_MEMDUMP ("ID3v2 frame (un-unsyced)", frame_data, frame_data_size);
141 work->parse_size = frame_data_size;
143 if (work->frame_flags & ID3V2_FRAME_FORMAT_COMPRESSION) {
145 uLongf destSize = work->parse_size;
148 uncompressed_data = g_malloc (work->parse_size);
150 dest = (Bytef *) uncompressed_data;
151 src = (Bytef *) frame_data;
153 if (uncompress (dest, &destSize, src, frame_data_size) != Z_OK) {
154 g_free (uncompressed_data);
158 if (destSize != work->parse_size) {
160 ("Decompressing ID3v2 frame %s did not produce expected size %d bytes (got %lu)",
161 tag_name, work->parse_size, destSize);
162 g_free (uncompressed_data);
166 work->parse_data = uncompressed_data;
168 GST_WARNING ("Compressed ID3v2 tag frame could not be decompressed"
169 " because gstid3demux was compiled without zlib support");
174 work->parse_data = frame_data;
177 if (work->frame_id[0] == 'T') {
178 if (strcmp (work->frame_id, "TDAT") == 0) {
179 parse_obsolete_tdat_frame (work);
181 } else if (strcmp (work->frame_id, "TXXX") == 0) {
182 /* Handle user text frame */
183 tag_str = parse_user_text_identification_frame (work, &tag_name);
185 /* Text identification frame */
186 tag_fields = parse_text_identification_frame (work);
188 } else if (work->frame_id[0] == 'W' && strcmp (work->frame_id, "WXXX") != 0) {
189 /* URL link frame: ISO-8859-1 encoded, one frame per tag */
190 tag_str = parse_url_link_frame (work, &tag_name);
191 } else if (!strcmp (work->frame_id, "COMM")) {
193 result = parse_comment_frame (work);
194 } else if (!strcmp (work->frame_id, "APIC")) {
195 /* Attached picture */
196 result = parse_picture_frame (work);
197 } else if (!strcmp (work->frame_id, "RVA2")) {
198 /* Relative volume */
199 result = parse_relative_volume_adjustment_two (work);
200 } else if (!strcmp (work->frame_id, "UFID")) {
201 /* Unique file identifier */
202 tag_str = parse_unique_file_identifier (work, &tag_name);
205 if (work->frame_flags & ID3V2_FRAME_FORMAT_COMPRESSION) {
206 g_free (uncompressed_data);
207 uncompressed_data = NULL;
208 work->parse_data = frame_data;
212 if (tag_str != NULL) {
213 /* g_print ("Tag %s value %s\n", tag_name, tag_str); */
214 result = id3v2_tag_to_taglist (work, tag_name, tag_str);
217 if (tag_fields != NULL) {
218 if (strcmp (work->frame_id, "TCON") == 0) {
219 /* Genre strings need special treatment */
220 result |= id3v2_genre_fields_to_taglist (work, tag_name, tag_fields);
224 for (t = 0; t < tag_fields->len; t++) {
225 tag_str = g_array_index (tag_fields, gchar *, t);
226 if (tag_str != NULL && tag_str[0] != '\0')
227 result |= id3v2_tag_to_taglist (work, tag_name, tag_str);
230 free_tag_strings (tag_fields);
239 parse_comment_frame (ID3TagsWorking * work)
244 GArray *fields = NULL;
245 gchar *description, *text;
247 if (work->parse_size < 6)
250 encoding = work->parse_data[0];
251 language[0] = g_ascii_tolower (work->parse_data[1]);
252 language[1] = g_ascii_tolower (work->parse_data[2]);
253 language[2] = g_ascii_tolower (work->parse_data[3]);
256 parse_split_strings (encoding, (gchar *) work->parse_data + 4,
257 work->parse_size - 4, &fields);
259 if (fields == NULL || fields->len < 2) {
260 GST_WARNING ("Failed to decode comment frame");
263 description = g_array_index (fields, gchar *, 0);
264 text = g_array_index (fields, gchar *, 1);
266 if (!g_utf8_validate (text, -1, NULL)) {
267 GST_WARNING ("Converted string is not valid utf-8");
271 /* skip our own dummy descriptions (from id3v2mux) */
272 if (strlen (description) > 0 && g_utf8_validate (description, -1, NULL) &&
273 sscanf (description, "c%u", &dummy) != 1) {
276 /* must be either an ISO-639-1 or ISO-639-2 language code */
277 if (language[0] != '\0' &&
278 g_ascii_isalpha (language[0]) &&
279 g_ascii_isalpha (language[1]) &&
280 (g_ascii_isalpha (language[2]) || language[2] == '\0')) {
281 const gchar *lang_code;
283 /* prefer two-letter ISO 639-1 code if we have a mapping */
284 lang_code = gst_tag_get_language_code (language);
285 s = g_strdup_printf ("%s[%s]=%s", description,
286 (lang_code) ? lang_code : language, text);
288 s = g_strdup_printf ("%s=%s", description, text);
290 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
291 GST_TAG_EXTENDED_COMMENT, s, NULL);
293 } else if (text != NULL && *text != '\0') {
294 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
295 GST_TAG_COMMENT, text, NULL);
300 free_tag_strings (fields);
305 GST_WARNING ("failed to parse COMM frame");
306 free_tag_strings (fields);
312 parse_text_identification_frame (ID3TagsWorking * work)
315 GArray *fields = NULL;
317 if (work->parse_size < 2)
320 encoding = work->parse_data[0];
321 parse_split_strings (encoding, (gchar *) work->parse_data + 1,
322 work->parse_size - 1, &fields);
324 if (fields->len > 0) {
325 GST_LOG ("Read %d fields from Text ID frame of size %d with encoding %d"
326 ". First is '%s'", fields->len, work->parse_size - 1, encoding,
327 g_array_index (fields, gchar *, 0));
329 GST_LOG ("Read 0 fields from Text ID frame of size %d with encoding %d",
330 work->parse_size - 1, encoding);
338 link_is_known_license (const gchar * url)
340 return g_str_has_prefix (url, "http://creativecommons.org/licenses/");
344 parse_url_link_frame (ID3TagsWorking * work, const gchar ** tag_name)
347 gchar *nul, *data, *link;
351 if (work->parse_size == 0)
354 data = (gchar *) work->parse_data;
355 /* if there's more data then the string is long, we only want to parse the
356 * data up to the terminating zero to g_convert and ignore the rest, as
358 nul = memchr (data, '\0', work->parse_size);
360 len = (gsize) (nul - data);
362 len = work->parse_size;
365 link = g_convert (data, len, "UTF-8", "ISO-8859-1", NULL, NULL, NULL);
367 if (link == NULL || !gst_uri_is_valid (link)) {
368 GST_DEBUG ("Invalid URI in %s frame: %s", work->frame_id,
369 GST_STR_NULL (link));
374 /* we don't know if it's a link to a page that explains the copyright
375 * situation, or a link that points to/represents a license, the ID3 spec
376 * does not separate those two things; for now only put known license URIs
377 * into GST_TAG_LICENSE_URI and everything else into GST_TAG_COPYRIGHT_URI */
378 if (strcmp (work->frame_id, "WCOP") == 0) {
379 if (link_is_known_license (link))
380 *tag_name = GST_TAG_LICENSE_URI;
382 *tag_name = GST_TAG_COPYRIGHT_URI;
383 } else if (strcmp (work->frame_id, "WOAF") == 0) {
384 /* can't be bothered to create a CONTACT_URI tag for this, so let's just
385 * put into into GST_TAG_CONTACT, which is where it ends up when reading
386 * the info from vorbis comments as well */
387 *tag_name = GST_TAG_CONTACT;
395 parse_user_text_identification_frame (ID3TagsWorking * work,
396 const gchar ** tag_name)
400 GArray *fields = NULL;
404 if (work->parse_size < 2)
407 encoding = work->parse_data[0];
409 parse_split_strings (encoding, (gchar *) work->parse_data + 1,
410 work->parse_size - 1, &fields);
415 if (fields->len != 2) {
416 GST_WARNING ("Expected 2 fields in TXXX frame, but got %d", fields->len);
417 free_tag_strings (fields);
422 gst_tag_from_id3_user_tag ("TXXX", g_array_index (fields, gchar *, 0));
424 GST_LOG ("TXXX frame of size %d. Mapped descriptor '%s' to GStreamer tag %s",
425 work->parse_size - 1, g_array_index (fields, gchar *, 0),
426 GST_STR_NULL (*tag_name));
429 ret = g_strdup (g_array_index (fields, gchar *, 1));
430 /* GST_LOG ("%s = %s", *tag_name, GST_STR_NULL (ret)); */
435 free_tag_strings (fields);
440 parse_id_string (ID3TagsWorking * work, gchar ** p_str, gint * p_len,
445 if (work->parse_size < 2)
448 for (len = 0; len < work->parse_size - 1; ++len) {
449 if (work->parse_data[len] == '\0')
453 datalen = work->parse_size - (len + 1);
454 if (len == 0 || datalen <= 0)
457 *p_str = g_strndup ((gchar *) work->parse_data, len);
459 *p_datalen = datalen;
465 parse_unique_file_identifier (ID3TagsWorking * work, const gchar ** tag_name)
468 gchar *owner_id, *data, *ret = NULL;
470 GST_LOG ("parsing UFID frame of size %d", work->parse_size);
472 if (!parse_id_string (work, &owner_id, &len, &datalen))
475 data = (gchar *) work->parse_data + len + 1;
476 GST_LOG ("UFID owner ID: %s (+ %d bytes of data)", owner_id, datalen);
478 if (strcmp (owner_id, "http://musicbrainz.org") == 0 &&
479 g_utf8_validate (data, datalen, NULL)) {
480 *tag_name = GST_TAG_MUSICBRAINZ_TRACKID;
481 ret = g_strndup (data, datalen);
483 GST_INFO ("Unknown UFID owner ID: %s", owner_id);
490 /* parse data and return length of the next string in the given encoding,
491 * including the NUL terminator */
493 scan_encoded_string (guint8 encoding, gchar * data, gint data_size)
498 case ID3V2_ENCODING_ISO8859:
499 case ID3V2_ENCODING_UTF8:
500 for (i = 0; i < data_size; ++i) {
505 case ID3V2_ENCODING_UTF16:
506 case ID3V2_ENCODING_UTF16BE:
507 /* we don't care about BOMs here and treat them as part of the string */
508 /* Find '\0\0' terminator */
509 for (i = 0; i < data_size - 1; i += 2) {
510 if (data[i] == '\0' && data[i + 1] == '\0')
522 parse_picture_frame (ID3TagsWorking * work)
524 guint8 txt_encoding, pic_type;
525 gchar *mime_str = NULL;
528 GST_LOG ("APIC frame (ID3v2.%u)", ID3V2_VER_MAJOR (work->hdr.version));
530 if (work->parse_size < 1 + 1 + 1 + 1 + 1)
531 goto not_enough_data;
533 txt_encoding = work->parse_data[0];
537 /* Read image format; in early ID3v2 versions this is a fixed-length
538 * 3-character string without terminator; in later versions (>= 2.3.0)
539 * this is a NUL-terminated string of variable length */
540 if (ID3V2_VER_MAJOR (work->hdr.version) < 3) {
541 if (work->parse_size < 3)
542 goto not_enough_data;
544 mime_str = g_strndup ((gchar *) work->parse_data, 3);
547 if (!parse_id_string (work, &mime_str, &len, &datalen))
549 ++len; /* for string terminator */
552 if (work->parse_size < len + 1 + 1 + 1)
553 goto not_enough_data;
555 work->parse_data += len;
556 work->parse_size -= len;
558 /* Read image type */
559 pic_type = work->parse_data[0];
563 GST_LOG ("APIC frame mime type : %s", GST_STR_NULL (mime_str));
564 GST_LOG ("APIC frame picture type : 0x%02x", (guint) pic_type);
566 if (work->parse_size < 1 + 1)
567 goto not_enough_data;
569 len = scan_encoded_string (txt_encoding, (gchar *) work->parse_data,
575 /* just skip the description string ... */
576 GST_LOG ("Skipping description string (%d bytes in original coding)", len);
578 if (work->parse_size < len + 1)
579 goto not_enough_data;
581 work->parse_data += len;
582 work->parse_size -= len;
584 GST_DEBUG ("image data is %u bytes", work->parse_size);
586 if (work->parse_size <= 0)
587 goto not_enough_data;
589 if (!gst_tag_list_add_id3_image (work->tags, (guint8 *) work->parse_data,
590 work->parse_size, pic_type)) {
599 GST_DEBUG ("not enough data, skipping APIC frame");
600 /* fall through to error */
604 GST_DEBUG ("problem parsing APIC frame, skipping");
610 #define ID3V2_RVA2_CHANNEL_MASTER 1
613 parse_relative_volume_adjustment_two (ID3TagsWorking * work)
615 const gchar *gain_tag_name = NULL;
616 const gchar *peak_tag_name = NULL;
617 gdouble gain_dB, peak_val;
619 guint8 *data, chan, peak_bits;
621 gint len, datalen, i;
623 if (!parse_id_string (work, &id, &len, &datalen))
626 if (datalen < (1 + 2 + 1)) {
627 GST_WARNING ("broken RVA2 frame, data size only %d bytes", datalen);
632 data = work->parse_data + len + 1;
633 chan = GST_READ_UINT8 (data);
634 gain_dB = (gdouble) ((gint16) GST_READ_UINT16_BE (data + 1)) / 512.0;
635 /* The meaning of the peak value is not defined in the ID3v2 spec. However,
636 * the first/only implementation of this seems to have been in XMMS, and
637 * other libs (like mutagen) seem to follow that implementation as well:
638 * see http://bugs.xmms.org/attachment.cgi?id=113&action=view */
639 peak_bits = GST_READ_UINT8 (data + 1 + 2);
640 if (peak_bits > 64) {
641 GST_WARNING ("silly peak precision of %d bits, ignoring", (gint) peak_bits);
645 datalen -= 1 + 2 + 1;
646 if (peak_bits == 16) {
647 peak = GST_READ_UINT16_BE (data);
650 for (i = 0; i < (GST_ROUND_UP_8 (peak_bits) / 8) && datalen > 0; ++i) {
652 peak |= GST_READ_UINT8 (data);
658 peak = peak << (64 - GST_ROUND_UP_8 (peak_bits));
660 gst_guint64_to_gdouble (peak) / gst_util_guint64_to_gdouble (G_MAXINT64);
661 GST_LOG ("RVA2 frame: id=%s, chan=%u, adj=%.2fdB, peak_bits=%u, peak=%.2f",
662 id, chan, gain_dB, (guint) peak_bits, peak_val);
664 if (chan == ID3V2_RVA2_CHANNEL_MASTER && strcmp (id, "track") == 0) {
665 gain_tag_name = GST_TAG_TRACK_GAIN;
666 peak_tag_name = GST_TAG_TRACK_PEAK;
667 } else if (chan == ID3V2_RVA2_CHANNEL_MASTER && strcmp (id, "album") == 0) {
668 gain_tag_name = GST_TAG_ALBUM_GAIN;
669 peak_tag_name = GST_TAG_ALBUM_PEAK;
671 GST_INFO ("Unhandled RVA2 frame id '%s' for channel %d", id, chan);
675 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
676 gain_tag_name, gain_dB, NULL);
678 if (peak_tag_name && peak_bits > 0) {
679 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
680 peak_tag_name, peak_val, NULL);
685 return (gain_tag_name != NULL || peak_tag_name != NULL);
689 parse_obsolete_tdat_frame (ID3TagsWorking * work)
691 if (work->parse_size >= 5 &&
692 work->parse_data[0] == ID3V2_ENCODING_ISO8859 &&
693 g_ascii_isdigit (work->parse_data[1]) &&
694 g_ascii_isdigit (work->parse_data[2]) &&
695 g_ascii_isdigit (work->parse_data[3]) &&
696 g_ascii_isdigit (work->parse_data[4])) {
697 work->pending_day = (10 * g_ascii_digit_value (work->parse_data[1])) +
698 g_ascii_digit_value (work->parse_data[2]);
699 work->pending_month = (10 * g_ascii_digit_value (work->parse_data[3])) +
700 g_ascii_digit_value (work->parse_data[4]);
701 GST_LOG ("date (dd/mm) %02u/%02u", work->pending_day, work->pending_month);
706 id3v2_tag_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
707 const gchar * tag_str)
709 GType tag_type = gst_tag_get_type (tag_name);
710 GstTagList *tag_list = work->tags;
720 if (sscanf (tag_str, "%d/%d", ¤t, &total) == 2) {
722 GST_WARNING ("Ignoring invalid value for total %d in tag %s",
725 if (strcmp (tag_name, GST_TAG_TRACK_NUMBER) == 0) {
726 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
727 GST_TAG_TRACK_COUNT, total, NULL);
728 } else if (strcmp (tag_name, GST_TAG_ALBUM_VOLUME_NUMBER) == 0) {
729 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
730 GST_TAG_ALBUM_VOLUME_COUNT, total, NULL);
733 } else if (sscanf (tag_str, "%d", ¤t) != 1) {
734 /* Not an integer in the string */
735 GST_WARNING ("Tag string for tag %s does not contain an integer - "
736 "ignoring", tag_name);
741 GST_WARNING ("Ignoring invalid value %d in tag %s", current, tag_name);
743 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND, tag_name, current,
752 g_assert (strcmp (tag_name, GST_TAG_DURATION) == 0);
753 tmp = strtoul (tag_str, NULL, 10);
757 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
758 GST_TAG_DURATION, tmp * 1000 * 1000, NULL);
765 /* make sure we add each unique string only once per tag, we don't want
766 * to have the same genre in the genre list multiple times, for example,
767 * or the same DiscID in there twice just because it's contained in the
768 * tag multiple times under different TXXX user tags */
769 num = gst_tag_list_get_tag_size (tag_list, tag_name);
770 for (i = 0; i < num; ++i) {
771 val = gst_tag_list_get_value_index (tag_list, tag_name, i);
772 if (val != NULL && strcmp (g_value_get_string (val), tag_str) == 0)
776 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
777 tag_name, tag_str, NULL);
785 GValue dest = { 0, };
787 /* Ensure that any date string is complete */
788 if (tag_type == GST_TYPE_DATE) {
789 guint year = 1901, month = 1, day = 1;
791 /* Dates can be yyyy-MM-dd, yyyy-MM or yyyy, but we need
793 if (sscanf (tag_str, "%04u-%02u-%02u", &year, &month, &day) == 0)
796 tmp = g_strdup_printf ("%04u-%02u-%02u", year, month, day);
800 /* handles anything else */
801 g_value_init (&src, G_TYPE_STRING);
802 g_value_set_string (&src, (const gchar *) tag_str);
803 g_value_init (&dest, tag_type);
805 if (g_value_transform (&src, &dest)) {
806 gst_tag_list_add_values (tag_list, GST_TAG_MERGE_APPEND,
807 tag_name, &dest, NULL);
808 } else if (tag_type == G_TYPE_DOUBLE) {
809 /* replaygain tags in TXXX frames ... */
810 g_value_set_double (&dest, g_strtod (tag_str, NULL));
811 gst_tag_list_add_values (tag_list, GST_TAG_MERGE_KEEP,
812 tag_name, &dest, NULL);
813 GST_LOG ("Converted string '%s' to double %f", tag_str,
814 g_value_get_double (&dest));
816 GST_WARNING ("Failed to transform tag from string to type '%s'",
817 g_type_name (tag_type));
820 g_value_unset (&src);
821 g_value_unset (&dest);
830 /* Check that an array of characters contains only digits */
832 id3v2_are_digits (const gchar * chars, gint size)
836 for (i = 0; i < size; i++) {
837 if (!g_ascii_isdigit (chars[i]))
844 id3v2_genre_string_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
845 const gchar * tag_str, gint len)
847 g_return_val_if_fail (tag_str != NULL, FALSE);
849 /* If it's a number, it might be a defined genre */
850 if (id3v2_are_digits (tag_str, len)) {
851 tag_str = gst_tag_id3_genre_get (strtol (tag_str, NULL, 10));
852 return id3v2_tag_to_taglist (work, tag_name, tag_str);
854 /* Otherwise it might be "RX" or "CR" */
856 if (g_ascii_strncasecmp ("rx", tag_str, len) == 0)
857 return id3v2_tag_to_taglist (work, tag_name, "Remix");
859 if (g_ascii_strncasecmp ("cr", tag_str, len) == 0)
860 return id3v2_tag_to_taglist (work, tag_name, "Cover");
863 /* Otherwise it's a string */
864 return id3v2_tag_to_taglist (work, tag_name, tag_str);
868 id3v2_genre_fields_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
871 gchar *tag_str = NULL;
872 gboolean result = FALSE;
875 for (i = 0; i < tag_fields->len; i++) {
878 tag_str = g_array_index (tag_fields, gchar *, i);
882 len = strlen (tag_str);
883 /* Only supposed to see '(n)' type numeric genre strings in ID3 <= 2.3.0
884 * but apparently we see them in 2.4.0 sometimes too */
885 if (TRUE || work->hdr.version <= 0x300) { /* <= 2.3.0 */
886 /* Check for genre numbers wrapped in parentheses, possibly
887 * followed by a string */
890 gboolean found = FALSE;
892 /* Double parenthesis ends the numeric genres, but we need
893 * to swallow the first one so we actually output '(' */
894 if (tag_str[0] == '(' && tag_str[1] == '(') {
900 /* If the first char is not a parenthesis, then stop
901 * looking for parenthesised genre strings */
902 if (tag_str[0] != '(')
905 for (pos = 1; pos < len; pos++) {
906 if (tag_str[pos] == ')') {
909 tmp_str = g_strndup (tag_str + 1, pos - 1);
911 id3v2_genre_string_to_taglist (work, tag_name, tmp_str,
920 /* If we encounter a non-digit while searching for a closing
921 * parenthesis, we should not try and interpret this as a
922 * numeric genre string */
923 if (!g_ascii_isdigit (tag_str[pos]))
927 break; /* There was no closing parenthesis */
931 if (len > 0 && tag_str != NULL)
932 result |= id3v2_genre_string_to_taglist (work, tag_name, tag_str, len);
937 static const gchar utf16enc[] = "UTF-16";
938 static const gchar utf16leenc[] = "UTF-16LE";
939 static const gchar utf16beenc[] = "UTF-16BE";
942 find_utf16_bom (gchar * data, const gchar ** p_in_encoding)
944 guint16 marker = (GST_READ_UINT8 (data) << 8) | GST_READ_UINT8 (data + 1);
948 *p_in_encoding = utf16leenc;
951 *p_in_encoding = utf16beenc;
960 string_utf8_dup (const gchar * start, const guint size)
966 /* Should we try the charsets specified
967 * via environment variables FIRST ? */
968 if (g_utf8_validate (start, size, NULL)) {
969 utf8 = g_strndup (start, size);
973 env = g_getenv ("GST_ID3V1_TAG_ENCODING");
974 if (!env || *env == '\0')
975 env = g_getenv ("GST_ID3_TAG_ENCODING");
976 if (!env || *env == '\0')
977 env = g_getenv ("GST_TAG_ENCODING");
979 /* Try charsets specified via the environment */
980 if (env && *env != '\0') {
983 csets = g_strsplit (env, G_SEARCHPATH_SEPARATOR_S, -1);
985 for (c = csets; c && *c; ++c) {
987 g_convert (start, size, "UTF-8", *c, &bytes_read, NULL, NULL))) {
988 if (bytes_read == size) {
989 GST_DEBUG ("Using charset %s to interperate id3 tags\n", *c);
998 /* Try current locale (if not UTF-8) */
999 if (!g_get_charset (&env)) {
1000 if ((utf8 = g_locale_to_utf8 (start, size, &bytes_read, NULL, NULL))) {
1001 if (bytes_read == size) {
1009 /* Try ISO-8859-1 */
1011 g_convert (start, size, "UTF-8", "ISO-8859-1", &bytes_read, NULL, NULL);
1012 if (utf8 != NULL && bytes_read == size) {
1027 parse_insert_string_field (guint8 encoding, gchar * data, gint data_size,
1030 gchar *field = NULL;
1033 case ID3V2_ENCODING_UTF16:
1034 case ID3V2_ENCODING_UTF16BE:
1036 const gchar *in_encode;
1038 if (encoding == ID3V2_ENCODING_UTF16)
1039 in_encode = utf16enc;
1041 in_encode = utf16beenc;
1043 /* Sometimes we see strings with multiple BOM markers at the start.
1044 * In that case, we assume the innermost one is correct. If that fails
1045 * to produce valid UTF-8, we try the other endianness anyway */
1046 while (data_size > 2 && find_utf16_bom (data, &in_encode)) {
1047 data += 2; /* skip BOM */
1051 field = g_convert (data, data_size, "UTF-8", in_encode, NULL, NULL, NULL);
1053 if (field == NULL || g_utf8_validate (field, -1, NULL) == FALSE) {
1054 /* As a fallback, try interpreting UTF-16 in the other endianness */
1055 if (in_encode == utf16beenc)
1056 field = g_convert (data, data_size, "UTF-8", utf16leenc,
1062 case ID3V2_ENCODING_ISO8859:
1063 if (g_utf8_validate (data, data_size, NULL))
1064 field = g_strndup (data, data_size);
1066 /* field = g_convert (data, data_size, "UTF-8", "ISO-8859-1",
1067 NULL, NULL, NULL); */
1068 field = string_utf8_dup (data, data_size);
1071 field = g_strndup (data, data_size);
1076 if (g_utf8_validate (field, -1, NULL)) {
1077 g_array_append_val (fields, field);
1081 GST_DEBUG ("%s was bad UTF-8 after conversion from encoding %d. Ignoring",
1088 parse_split_strings (guint8 encoding, gchar * data, gint data_size,
1089 GArray ** out_fields)
1091 GArray *fields = g_array_new (FALSE, TRUE, sizeof (gchar *));
1095 g_return_if_fail (out_fields != NULL);
1098 case ID3V2_ENCODING_ISO8859:
1099 for (text_pos = 0; text_pos < data_size; text_pos++) {
1100 if (data[text_pos] == 0) {
1101 parse_insert_string_field (encoding, data + prev,
1102 text_pos - prev + 1, fields);
1103 prev = text_pos + 1;
1106 if (data_size - prev > 0 && data[prev] != 0x00) {
1107 parse_insert_string_field (encoding, data + prev,
1108 data_size - prev, fields);
1112 case ID3V2_ENCODING_UTF8:
1113 for (prev = 0, text_pos = 0; text_pos < data_size; text_pos++) {
1114 if (data[text_pos] == '\0') {
1115 parse_insert_string_field (encoding, data + prev,
1116 text_pos - prev + 1, fields);
1117 prev = text_pos + 1;
1120 if (data_size - prev > 0 && data[prev] != 0x00) {
1121 parse_insert_string_field (encoding, data + prev,
1122 data_size - prev, fields);
1125 case ID3V2_ENCODING_UTF16:
1126 case ID3V2_ENCODING_UTF16BE:
1128 /* Find '\0\0' terminator */
1129 for (text_pos = 0; text_pos < data_size - 1; text_pos += 2) {
1130 if (data[text_pos] == '\0' && data[text_pos + 1] == '\0') {
1131 /* found a delimiter */
1132 parse_insert_string_field (encoding, data + prev,
1133 text_pos - prev + 2, fields);
1134 text_pos++; /* Advance to the 2nd NULL terminator */
1135 prev = text_pos + 1;
1139 if (data_size - prev > 1 &&
1140 (data[prev] != 0x00 || data[prev + 1] != 0x00)) {
1141 /* There were 2 or more non-null chars left, convert those too */
1142 parse_insert_string_field (encoding, data + prev,
1143 data_size - prev, fields);
1148 if (fields->len > 0)
1149 *out_fields = fields;
1151 g_array_free (fields, TRUE);
1155 free_tag_strings (GArray * fields)
1161 for (i = 0; i < fields->len; i++) {
1162 c = g_array_index (fields, gchar *, i);
1165 g_array_free (fields, TRUE);