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.
28 #include <gst/tag/tag.h>
29 #include <gst/base/gsttypefindhelper.h>
37 GST_DEBUG_CATEGORY_EXTERN (id3demux_debug);
38 #define GST_CAT_DEFAULT (id3demux_debug)
40 static gboolean parse_comment_frame (ID3TagsWorking * work);
41 static gchar *parse_url_link_frame (ID3TagsWorking * work,
42 const gchar ** tag_name);
43 static GArray *parse_text_identification_frame (ID3TagsWorking * work);
44 static gchar *parse_user_text_identification_frame (ID3TagsWorking * work,
45 const gchar ** tag_name);
46 static gchar *parse_unique_file_identifier (ID3TagsWorking * work,
47 const gchar ** tag_name);
48 static gboolean parse_relative_volume_adjustment_two (ID3TagsWorking * work);
49 static void parse_obsolete_tdat_frame (ID3TagsWorking * work);
50 static gboolean id3v2_tag_to_taglist (ID3TagsWorking * work,
51 const gchar * tag_name, const gchar * tag_str);
52 /* Parse a single string into an array of gchar* */
53 static void parse_split_strings (guint8 encoding, gchar * data, gint data_size,
54 GArray ** out_fields);
55 static void free_tag_strings (GArray * fields);
57 id3v2_genre_fields_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
59 static gboolean parse_picture_frame (ID3TagsWorking * work);
61 #define ID3V2_ENCODING_ISO8859 0x00
62 #define ID3V2_ENCODING_UTF16 0x01
63 #define ID3V2_ENCODING_UTF16BE 0x02
64 #define ID3V2_ENCODING_UTF8 0x03
67 id3demux_id3v2_parse_frame (ID3TagsWorking * work)
69 const gchar *tag_name;
70 gboolean result = FALSE;
72 guint8 *frame_data = work->hdr.frame_data;
73 guint frame_data_size = work->cur_frame_size;
74 gchar *tag_str = NULL;
75 GArray *tag_fields = NULL;
76 guint8 *uu_data = NULL;
79 guint8 *uncompressed_data = NULL;
82 /* Check that the frame id is valid */
83 for (i = 0; i < 5 && work->frame_id[i] != '\0'; i++) {
84 if (!g_ascii_isalnum (work->frame_id[i])) {
85 GST_DEBUG ("Encountered invalid frame_id");
90 /* Can't handle encrypted frames right now (in case we ever do, we'll have
91 * to do the decryption after the un-unsynchronisation and decompression,
93 if (work->frame_flags & ID3V2_FRAME_FORMAT_ENCRYPTION) {
94 GST_WARNING ("Encrypted frames are not supported");
98 tag_name = gst_tag_from_id3_tag (work->frame_id);
99 if (tag_name == NULL &&
100 strncmp (work->frame_id, "RVA2", 4) != 0 &&
101 strncmp (work->frame_id, "TXXX", 4) != 0 &&
102 strncmp (work->frame_id, "TDAT", 4) != 0 &&
103 strncmp (work->frame_id, "UFID", 4) != 0) {
107 if (work->frame_flags & (ID3V2_FRAME_FORMAT_COMPRESSION |
108 ID3V2_FRAME_FORMAT_DATA_LENGTH_INDICATOR)) {
109 if (work->hdr.frame_data_size <= 4)
111 if (ID3V2_VER_MAJOR (work->hdr.version) == 3) {
112 work->parse_size = GST_READ_UINT32_BE (frame_data);
114 work->parse_size = read_synch_uint (frame_data, 4);
117 frame_data_size -= 4;
118 if (work->parse_size < frame_data_size) {
119 GST_WARNING ("ID3v2 frame %s has invalid size %d.", tag_name,
125 /* in v2.3 the frame sizes are not syncsafe, so the entire tag had to be
126 * unsynced. In v2.4 the frame sizes are syncsafe so it's just the frame
127 * data that needs un-unsyncing, but not the frame headers. */
128 if (ID3V2_VER_MAJOR (work->hdr.version) == 4) {
129 if ((work->hdr.flags & ID3V2_HDR_FLAG_UNSYNC) != 0 ||
130 ((work->frame_flags & ID3V2_FRAME_FORMAT_UNSYNCHRONISATION) != 0)) {
131 GST_DEBUG ("Un-unsyncing frame %s", work->frame_id);
132 uu_data = id3demux_ununsync_data (frame_data, &frame_data_size);
133 frame_data = uu_data;
134 GST_MEMDUMP ("ID3v2 frame (un-unsyced)", frame_data, frame_data_size);
138 work->parse_size = frame_data_size;
140 if (work->frame_flags & ID3V2_FRAME_FORMAT_COMPRESSION) {
142 uLongf destSize = work->parse_size;
145 uncompressed_data = g_malloc (work->parse_size);
147 dest = (Bytef *) uncompressed_data;
148 src = (Bytef *) frame_data;
150 if (uncompress (dest, &destSize, src, frame_data_size) != Z_OK) {
151 g_free (uncompressed_data);
155 if (destSize != work->parse_size) {
157 ("Decompressing ID3v2 frame %s did not produce expected size %d bytes (got %lu)",
158 tag_name, work->parse_size, destSize);
159 g_free (uncompressed_data);
163 work->parse_data = uncompressed_data;
165 GST_WARNING ("Compressed ID3v2 tag frame could not be decompressed"
166 " because gstid3demux was compiled without zlib support");
171 work->parse_data = frame_data;
174 if (work->frame_id[0] == 'T') {
175 if (strcmp (work->frame_id, "TDAT") == 0) {
176 parse_obsolete_tdat_frame (work);
178 } else if (strcmp (work->frame_id, "TXXX") == 0) {
179 /* Handle user text frame */
180 tag_str = parse_user_text_identification_frame (work, &tag_name);
182 /* Text identification frame */
183 tag_fields = parse_text_identification_frame (work);
185 } else if (work->frame_id[0] == 'W' && strcmp (work->frame_id, "WXXX") != 0) {
186 /* URL link frame: ISO-8859-1 encoded, one frame per tag */
187 tag_str = parse_url_link_frame (work, &tag_name);
188 } else if (!strcmp (work->frame_id, "COMM")) {
190 result = parse_comment_frame (work);
191 } else if (!strcmp (work->frame_id, "APIC")) {
192 /* Attached picture */
193 result = parse_picture_frame (work);
194 } else if (!strcmp (work->frame_id, "RVA2")) {
195 /* Relative volume */
196 result = parse_relative_volume_adjustment_two (work);
197 } else if (!strcmp (work->frame_id, "UFID")) {
198 /* Unique file identifier */
199 tag_str = parse_unique_file_identifier (work, &tag_name);
202 if (work->frame_flags & ID3V2_FRAME_FORMAT_COMPRESSION) {
203 g_free (uncompressed_data);
204 uncompressed_data = NULL;
205 work->parse_data = frame_data;
209 if (tag_str != NULL) {
210 /* g_print ("Tag %s value %s\n", tag_name, tag_str); */
211 result = id3v2_tag_to_taglist (work, tag_name, tag_str);
214 if (tag_fields != NULL) {
215 if (strcmp (work->frame_id, "TCON") == 0) {
216 /* Genre strings need special treatment */
217 result |= id3v2_genre_fields_to_taglist (work, tag_name, tag_fields);
221 for (t = 0; t < tag_fields->len; t++) {
222 tag_str = g_array_index (tag_fields, gchar *, t);
223 if (tag_str != NULL && tag_str[0] != '\0')
224 result |= id3v2_tag_to_taglist (work, tag_name, tag_str);
227 free_tag_strings (tag_fields);
236 parse_comment_frame (ID3TagsWorking * work)
241 GArray *fields = NULL;
242 gchar *description, *text;
244 if (work->parse_size < 6)
247 encoding = work->parse_data[0];
248 language[0] = g_ascii_tolower (work->parse_data[1]);
249 language[1] = g_ascii_tolower (work->parse_data[2]);
250 language[2] = g_ascii_tolower (work->parse_data[3]);
253 parse_split_strings (encoding, (gchar *) work->parse_data + 4,
254 work->parse_size - 4, &fields);
256 if (fields == NULL || fields->len < 2) {
257 GST_WARNING ("Failed to decode comment frame");
260 description = g_array_index (fields, gchar *, 0);
261 text = g_array_index (fields, gchar *, 1);
263 if (!g_utf8_validate (text, -1, NULL)) {
264 GST_WARNING ("Converted string is not valid utf-8");
268 /* skip our own dummy descriptions (from id3v2mux) */
269 if (strlen (description) > 0 && g_utf8_validate (description, -1, NULL) &&
270 sscanf (description, "c%u", &dummy) != 1) {
273 /* must be either an ISO-639-1 or ISO-639-2 language code */
274 if (language[0] != '\0' &&
275 g_ascii_isalpha (language[0]) &&
276 g_ascii_isalpha (language[1]) &&
277 (g_ascii_isalpha (language[2]) || language[2] == '\0')) {
278 s = g_strdup_printf ("%s[%s]=%s", description, language, text);
280 s = g_strdup_printf ("%s=%s", description, text);
282 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
283 GST_TAG_EXTENDED_COMMENT, s, NULL);
285 } else if (text != NULL && *text != '\0') {
286 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
287 GST_TAG_COMMENT, text, NULL);
292 free_tag_strings (fields);
297 GST_WARNING ("failed to parse COMM frame");
298 free_tag_strings (fields);
304 parse_text_identification_frame (ID3TagsWorking * work)
307 GArray *fields = NULL;
309 if (work->parse_size < 2)
312 encoding = work->parse_data[0];
313 parse_split_strings (encoding, (gchar *) work->parse_data + 1,
314 work->parse_size - 1, &fields);
316 if (fields->len > 0) {
317 GST_LOG ("Read %d fields from Text ID frame of size %d with encoding %d"
318 ". First is '%s'", fields->len, work->parse_size - 1, encoding,
319 g_array_index (fields, gchar *, 0));
321 GST_LOG ("Read 0 fields from Text ID frame of size %d with encoding %d",
322 work->parse_size - 1, encoding);
330 link_is_known_license (const gchar * url)
332 return g_str_has_prefix (url, "http://creativecommons.org/licenses/");
336 parse_url_link_frame (ID3TagsWorking * work, const gchar ** tag_name)
339 gchar *nul, *data, *link;
343 if (work->parse_size == 0)
346 data = (gchar *) work->parse_data;
347 /* if there's more data then the string is long, we only want to parse the
348 * data up to the terminating zero to g_convert and ignore the rest, as
350 nul = memchr (data, '\0', work->parse_size);
352 len = (gsize) (nul - data);
354 len = work->parse_size;
357 link = g_convert (data, len, "UTF-8", "ISO-8859-1", NULL, NULL, NULL);
359 if (link == NULL || !gst_uri_is_valid (link)) {
360 GST_DEBUG ("Invalid URI in %s frame: %s", work->frame_id,
361 GST_STR_NULL (link));
366 /* we don't know if it's a link to a page that explains the copyright
367 * situation, or a link that points to/represents a license, the ID3 spec
368 * does not separate those two things; for now only put known license URIs
369 * into GST_TAG_LICENSE_URI and everything else into GST_TAG_COPYRIGHT_URI */
370 if (strcmp (work->frame_id, "WCOP") == 0) {
371 if (link_is_known_license (link))
372 *tag_name = GST_TAG_LICENSE_URI;
374 *tag_name = GST_TAG_COPYRIGHT_URI;
375 } else if (strcmp (work->frame_id, "WOAF") == 0) {
376 /* can't be bothered to create a CONTACT_URI tag for this, so let's just
377 * put into into GST_TAG_CONTACT, which is where it ends up when reading
378 * the info from vorbis comments as well */
379 *tag_name = GST_TAG_CONTACT;
387 parse_user_text_identification_frame (ID3TagsWorking * work,
388 const gchar ** tag_name)
392 GArray *fields = NULL;
396 if (work->parse_size < 2)
399 encoding = work->parse_data[0];
401 parse_split_strings (encoding, (gchar *) work->parse_data + 1,
402 work->parse_size - 1, &fields);
407 if (fields->len != 2) {
408 GST_WARNING ("Expected 2 fields in TXXX frame, but got %d", fields->len);
409 free_tag_strings (fields);
414 gst_tag_from_id3_user_tag ("TXXX", g_array_index (fields, gchar *, 0));
416 GST_LOG ("TXXX frame of size %d. Mapped descriptor '%s' to GStreamer tag %s",
417 work->parse_size - 1, g_array_index (fields, gchar *, 0),
418 GST_STR_NULL (*tag_name));
421 ret = g_strdup (g_array_index (fields, gchar *, 1));
422 /* GST_LOG ("%s = %s", *tag_name, GST_STR_NULL (ret)); */
427 free_tag_strings (fields);
432 parse_id_string (ID3TagsWorking * work, gchar ** p_str, gint * p_len,
437 if (work->parse_size < 2)
440 for (len = 0; len < work->parse_size - 1; ++len) {
441 if (work->parse_data[len] == '\0')
445 datalen = work->parse_size - (len + 1);
446 if (len == 0 || datalen <= 0)
449 *p_str = g_strndup ((gchar *) work->parse_data, len);
451 *p_datalen = datalen;
457 parse_unique_file_identifier (ID3TagsWorking * work, const gchar ** tag_name)
460 gchar *owner_id, *data, *ret = NULL;
462 GST_LOG ("parsing UFID frame of size %d", work->parse_size);
464 if (!parse_id_string (work, &owner_id, &len, &datalen))
467 data = (gchar *) work->parse_data + len + 1;
468 GST_LOG ("UFID owner ID: %s (+ %d bytes of data)", owner_id, datalen);
470 if (strcmp (owner_id, "http://musicbrainz.org") == 0 &&
471 g_utf8_validate (data, datalen, NULL)) {
472 *tag_name = GST_TAG_MUSICBRAINZ_TRACKID;
473 ret = g_strndup (data, datalen);
475 GST_INFO ("Unknown UFID owner ID: %s", owner_id);
482 /* parse data and return length of the next string in the given encoding,
483 * including the NUL terminator */
485 scan_encoded_string (guint8 encoding, gchar * data, gint data_size)
490 case ID3V2_ENCODING_ISO8859:
491 case ID3V2_ENCODING_UTF8:
492 for (i = 0; i < data_size; ++i) {
497 case ID3V2_ENCODING_UTF16:
498 case ID3V2_ENCODING_UTF16BE:
499 /* we don't care about BOMs here and treat them as part of the string */
500 /* Find '\0\0' terminator */
501 for (i = 0; i < data_size - 1; i += 2) {
502 if (data[i] == '\0' && data[i + 1] == '\0')
514 parse_picture_frame (ID3TagsWorking * work)
516 guint8 txt_encoding, pic_type;
517 gchar *mime_str = NULL;
520 GST_LOG ("APIC frame (ID3v2.%u)", ID3V2_VER_MAJOR (work->hdr.version));
522 if (work->parse_size < 1 + 1 + 1 + 1 + 1)
523 goto not_enough_data;
525 txt_encoding = work->parse_data[0];
529 /* Read image format; in early ID3v2 versions this is a fixed-length
530 * 3-character string without terminator; in later versions (>= 2.3.0)
531 * this is a NUL-terminated string of variable length */
532 if (ID3V2_VER_MAJOR (work->hdr.version) < 3) {
533 if (work->parse_size < 3)
534 goto not_enough_data;
536 mime_str = g_strndup ((gchar *) work->parse_data, 3);
539 if (!parse_id_string (work, &mime_str, &len, &datalen))
541 ++len; /* for string terminator */
544 if (work->parse_size < len + 1 + 1 + 1)
545 goto not_enough_data;
547 work->parse_data += len;
548 work->parse_size -= len;
550 /* Read image type */
551 pic_type = work->parse_data[0];
555 GST_LOG ("APIC frame mime type : %s", GST_STR_NULL (mime_str));
556 GST_LOG ("APIC frame picture type : 0x%02x", (guint) pic_type);
558 if (work->parse_size < 1 + 1)
559 goto not_enough_data;
561 len = scan_encoded_string (txt_encoding, (gchar *) work->parse_data,
567 /* just skip the description string ... */
568 GST_LOG ("Skipping description string (%d bytes in original coding)", len);
570 if (work->parse_size < len + 1)
571 goto not_enough_data;
573 work->parse_data += len;
574 work->parse_size -= len;
576 GST_DEBUG ("image data is %u bytes", work->parse_size);
578 if (work->parse_size <= 0)
579 goto not_enough_data;
581 if (!gst_tag_list_add_id3_image (work->tags, (guint8 *) work->parse_data,
582 work->parse_size, pic_type)) {
591 GST_DEBUG ("not enough data, skipping APIC frame");
592 /* fall through to error */
596 GST_DEBUG ("problem parsing APIC frame, skipping");
602 #define ID3V2_RVA2_CHANNEL_MASTER 1
605 parse_relative_volume_adjustment_two (ID3TagsWorking * work)
607 const gchar *gain_tag_name = NULL;
608 const gchar *peak_tag_name = NULL;
609 gdouble gain_dB, peak_val;
611 guint8 *data, chan, peak_bits;
613 gint len, datalen, i;
615 if (!parse_id_string (work, &id, &len, &datalen))
618 if (datalen < (1 + 2 + 1)) {
619 GST_WARNING ("broken RVA2 frame, data size only %d bytes", datalen);
624 data = work->parse_data + len + 1;
625 chan = GST_READ_UINT8 (data);
626 gain_dB = (gdouble) ((gint16) GST_READ_UINT16_BE (data + 1)) / 512.0;
627 /* The meaning of the peak value is not defined in the ID3v2 spec. However,
628 * the first/only implementation of this seems to have been in XMMS, and
629 * other libs (like mutagen) seem to follow that implementation as well:
630 * see http://bugs.xmms.org/attachment.cgi?id=113&action=view */
631 peak_bits = GST_READ_UINT8 (data + 1 + 2);
632 if (peak_bits > 64) {
633 GST_WARNING ("silly peak precision of %d bits, ignoring", (gint) peak_bits);
637 datalen -= 1 + 2 + 1;
638 if (peak_bits == 16) {
639 peak = GST_READ_UINT16_BE (data);
642 for (i = 0; i < (GST_ROUND_UP_8 (peak_bits) / 8) && datalen > 0; ++i) {
644 peak |= GST_READ_UINT8 (data);
650 peak = peak << (64 - GST_ROUND_UP_8 (peak_bits));
652 gst_guint64_to_gdouble (peak) / gst_util_guint64_to_gdouble (G_MAXINT64);
653 GST_LOG ("RVA2 frame: id=%s, chan=%u, adj=%.2fdB, peak_bits=%u, peak=%.2f",
654 id, chan, gain_dB, (guint) peak_bits, peak_val);
656 if (chan == ID3V2_RVA2_CHANNEL_MASTER && strcmp (id, "track") == 0) {
657 gain_tag_name = GST_TAG_TRACK_GAIN;
658 peak_tag_name = GST_TAG_TRACK_PEAK;
659 } else if (chan == ID3V2_RVA2_CHANNEL_MASTER && strcmp (id, "album") == 0) {
660 gain_tag_name = GST_TAG_ALBUM_GAIN;
661 peak_tag_name = GST_TAG_ALBUM_PEAK;
663 GST_INFO ("Unhandled RVA2 frame id '%s' for channel %d", id, chan);
667 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
668 gain_tag_name, gain_dB, NULL);
670 if (peak_tag_name && peak_bits > 0) {
671 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
672 peak_tag_name, peak_val, NULL);
677 return (gain_tag_name != NULL || peak_tag_name != NULL);
681 parse_obsolete_tdat_frame (ID3TagsWorking * work)
683 if (work->parse_size >= 5 &&
684 work->parse_data[0] == ID3V2_ENCODING_ISO8859 &&
685 g_ascii_isdigit (work->parse_data[1]) &&
686 g_ascii_isdigit (work->parse_data[2]) &&
687 g_ascii_isdigit (work->parse_data[3]) &&
688 g_ascii_isdigit (work->parse_data[4])) {
689 work->pending_day = (10 * g_ascii_digit_value (work->parse_data[1])) +
690 g_ascii_digit_value (work->parse_data[2]);
691 work->pending_month = (10 * g_ascii_digit_value (work->parse_data[3])) +
692 g_ascii_digit_value (work->parse_data[4]);
693 GST_LOG ("date (dd/mm) %02u/%02u", work->pending_day, work->pending_month);
698 id3v2_tag_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
699 const gchar * tag_str)
701 GType tag_type = gst_tag_get_type (tag_name);
702 GstTagList *tag_list = work->tags;
712 if (sscanf (tag_str, "%d/%d", ¤t, &total) == 2) {
714 GST_WARNING ("Ignoring invalid value for total %d in tag %s",
717 if (strcmp (tag_name, GST_TAG_TRACK_NUMBER) == 0) {
718 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
719 GST_TAG_TRACK_COUNT, total, NULL);
720 } else if (strcmp (tag_name, GST_TAG_ALBUM_VOLUME_NUMBER) == 0) {
721 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
722 GST_TAG_ALBUM_VOLUME_COUNT, total, NULL);
725 } else if (sscanf (tag_str, "%d", ¤t) != 1) {
726 /* Not an integer in the string */
727 GST_WARNING ("Tag string for tag %s does not contain an integer - "
728 "ignoring", tag_name);
733 GST_WARNING ("Ignoring invalid value %d in tag %s", current, tag_name);
735 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND, tag_name, current,
744 g_assert (strcmp (tag_name, GST_TAG_DURATION) == 0);
745 tmp = strtoul (tag_str, NULL, 10);
749 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
750 GST_TAG_DURATION, tmp * 1000 * 1000, NULL);
757 /* make sure we add each unique string only once per tag, we don't want
758 * to have the same genre in the genre list multiple times, for example,
759 * or the same DiscID in there twice just because it's contained in the
760 * tag multiple times under different TXXX user tags */
761 num = gst_tag_list_get_tag_size (tag_list, tag_name);
762 for (i = 0; i < num; ++i) {
763 val = gst_tag_list_get_value_index (tag_list, tag_name, i);
764 if (val != NULL && strcmp (g_value_get_string (val), tag_str) == 0)
768 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
769 tag_name, tag_str, NULL);
777 GValue dest = { 0, };
779 /* Ensure that any date string is complete */
780 if (tag_type == GST_TYPE_DATE) {
781 guint year = 1901, month = 1, day = 1;
783 /* Dates can be yyyy-MM-dd, yyyy-MM or yyyy, but we need
785 if (sscanf (tag_str, "%04u-%02u-%02u", &year, &month, &day) == 0)
788 tmp = g_strdup_printf ("%04u-%02u-%02u", year, month, day);
792 /* handles anything else */
793 g_value_init (&src, G_TYPE_STRING);
794 g_value_set_string (&src, (const gchar *) tag_str);
795 g_value_init (&dest, tag_type);
797 if (g_value_transform (&src, &dest)) {
798 gst_tag_list_add_values (tag_list, GST_TAG_MERGE_APPEND,
799 tag_name, &dest, NULL);
800 } else if (tag_type == G_TYPE_DOUBLE) {
801 /* replaygain tags in TXXX frames ... */
802 g_value_set_double (&dest, g_strtod (tag_str, NULL));
803 gst_tag_list_add_values (tag_list, GST_TAG_MERGE_KEEP,
804 tag_name, &dest, NULL);
805 GST_LOG ("Converted string '%s' to double %f", tag_str,
806 g_value_get_double (&dest));
808 GST_WARNING ("Failed to transform tag from string to type '%s'",
809 g_type_name (tag_type));
812 g_value_unset (&src);
813 g_value_unset (&dest);
822 /* Check that an array of characters contains only digits */
824 id3v2_are_digits (const gchar * chars, gint size)
828 for (i = 0; i < size; i++) {
829 if (!g_ascii_isdigit (chars[i]))
836 id3v2_genre_string_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
837 const gchar * tag_str, gint len)
839 g_return_val_if_fail (tag_str != NULL, FALSE);
841 /* If it's a number, it might be a defined genre */
842 if (id3v2_are_digits (tag_str, len)) {
843 tag_str = gst_tag_id3_genre_get (strtol (tag_str, NULL, 10));
844 return id3v2_tag_to_taglist (work, tag_name, tag_str);
846 /* Otherwise it might be "RX" or "CR" */
848 if (g_ascii_strncasecmp ("rx", tag_str, len) == 0)
849 return id3v2_tag_to_taglist (work, tag_name, "Remix");
851 if (g_ascii_strncasecmp ("cr", tag_str, len) == 0)
852 return id3v2_tag_to_taglist (work, tag_name, "Cover");
855 /* Otherwise it's a string */
856 return id3v2_tag_to_taglist (work, tag_name, tag_str);
860 id3v2_genre_fields_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
863 gchar *tag_str = NULL;
864 gboolean result = FALSE;
867 for (i = 0; i < tag_fields->len; i++) {
870 tag_str = g_array_index (tag_fields, gchar *, 0);
874 len = strlen (tag_str);
875 /* Only supposed to see '(n)' type numeric genre strings in ID3 <= 2.3.0
876 * but apparently we see them in 2.4.0 sometimes too */
877 if (TRUE || work->hdr.version <= 0x300) { /* <= 2.3.0 */
878 /* Check for genre numbers wrapped in parentheses, possibly
879 * followed by a string */
882 gboolean found = FALSE;
884 /* Double parenthesis ends the numeric genres, but we need
885 * to swallow the first one so we actually output '(' */
886 if (tag_str[0] == '(' && tag_str[1] == '(') {
892 /* If the first char is not a parenthesis, then stop
893 * looking for parenthesised genre strings */
894 if (tag_str[0] != '(')
897 for (pos = 1; pos < len; pos++) {
898 if (tag_str[pos] == ')') {
901 tmp_str = g_strndup (tag_str + 1, pos - 1);
903 id3v2_genre_string_to_taglist (work, tag_name, tmp_str,
912 /* If we encounter a non-digit while searching for a closing
913 * parenthesis, we should not try and interpret this as a
914 * numeric genre string */
915 if (!g_ascii_isdigit (tag_str[pos]))
919 break; /* There was no closing parenthesis */
923 if (len > 0 && tag_str != NULL)
924 result |= id3v2_genre_string_to_taglist (work, tag_name, tag_str, len);
929 static const gchar utf16enc[] = "UTF-16";
930 static const gchar utf16leenc[] = "UTF-16LE";
931 static const gchar utf16beenc[] = "UTF-16BE";
934 find_utf16_bom (gchar * data, const gchar ** p_in_encoding)
936 guint16 marker = (GST_READ_UINT8 (data) << 8) | GST_READ_UINT8 (data + 1);
940 *p_in_encoding = utf16leenc;
943 *p_in_encoding = utf16beenc;
952 string_utf8_dup (const gchar * start, const guint size)
958 /* Should we try the charsets specified
959 * via environment variables FIRST ? */
960 if (g_utf8_validate (start, size, NULL)) {
961 utf8 = g_strndup (start, size);
965 env = g_getenv ("GST_ID3V1_TAG_ENCODING");
966 if (!env || *env == '\0')
967 env = g_getenv ("GST_ID3_TAG_ENCODING");
968 if (!env || *env == '\0')
969 env = g_getenv ("GST_TAG_ENCODING");
971 /* Try charsets specified via the environment */
972 if (env && *env != '\0') {
975 csets = g_strsplit (env, G_SEARCHPATH_SEPARATOR_S, -1);
977 for (c = csets; c && *c; ++c) {
979 g_convert (start, size, "UTF-8", *c, &bytes_read, NULL, NULL))) {
980 if (bytes_read == size) {
981 GST_DEBUG ("Using charset %s to interperate id3 tags\n", c);
990 /* Try current locale (if not UTF-8) */
991 if (!g_get_charset (&env)) {
992 if ((utf8 = g_locale_to_utf8 (start, size, &bytes_read, NULL, NULL))) {
993 if (bytes_read == size) {
1001 /* Try ISO-8859-1 */
1003 g_convert (start, size, "UTF-8", "ISO-8859-1", &bytes_read, NULL, NULL);
1004 if (utf8 != NULL && bytes_read == size) {
1019 parse_insert_string_field (guint8 encoding, gchar * data, gint data_size,
1022 gchar *field = NULL;
1025 case ID3V2_ENCODING_UTF16:
1026 case ID3V2_ENCODING_UTF16BE:
1028 const gchar *in_encode;
1030 if (encoding == ID3V2_ENCODING_UTF16)
1031 in_encode = utf16enc;
1033 in_encode = utf16beenc;
1035 /* Sometimes we see strings with multiple BOM markers at the start.
1036 * In that case, we assume the innermost one is correct. If that fails
1037 * to produce valid UTF-8, we try the other endianness anyway */
1038 while (data_size > 2 && find_utf16_bom (data, &in_encode)) {
1039 data += 2; /* skip BOM */
1043 field = g_convert (data, data_size, "UTF-8", in_encode, NULL, NULL, NULL);
1045 if (field == NULL || g_utf8_validate (field, -1, NULL) == FALSE) {
1046 /* As a fallback, try interpreting UTF-16 in the other endianness */
1047 if (in_encode == utf16beenc)
1048 field = g_convert (data, data_size, "UTF-8", utf16leenc,
1054 case ID3V2_ENCODING_ISO8859:
1055 if (g_utf8_validate (data, data_size, NULL))
1056 field = g_strndup (data, data_size);
1058 /* field = g_convert (data, data_size, "UTF-8", "ISO-8859-1",
1059 NULL, NULL, NULL); */
1060 field = string_utf8_dup (data, data_size);
1063 field = g_strndup (data, data_size);
1068 if (g_utf8_validate (field, -1, NULL)) {
1069 g_array_append_val (fields, field);
1073 GST_DEBUG ("%s was bad UTF-8 after conversion from encoding %d. Ignoring",
1080 parse_split_strings (guint8 encoding, gchar * data, gint data_size,
1081 GArray ** out_fields)
1083 GArray *fields = g_array_new (FALSE, TRUE, sizeof (gchar *));
1087 g_return_if_fail (out_fields != NULL);
1090 case ID3V2_ENCODING_ISO8859:
1091 for (text_pos = 0; text_pos < data_size; text_pos++) {
1092 if (data[text_pos] == 0) {
1093 parse_insert_string_field (encoding, data + prev,
1094 text_pos - prev + 1, fields);
1095 prev = text_pos + 1;
1098 if (data_size - prev > 0 && data[prev] != 0x00) {
1099 parse_insert_string_field (encoding, data + prev,
1100 data_size - prev, fields);
1104 case ID3V2_ENCODING_UTF8:
1105 for (prev = 0, text_pos = 0; text_pos < data_size; text_pos++) {
1106 if (data[text_pos] == '\0') {
1107 parse_insert_string_field (encoding, data + prev,
1108 text_pos - prev + 1, fields);
1109 prev = text_pos + 1;
1112 if (data_size - prev > 0 && data[prev] != 0x00) {
1113 parse_insert_string_field (encoding, data + prev,
1114 data_size - prev, fields);
1117 case ID3V2_ENCODING_UTF16:
1118 case ID3V2_ENCODING_UTF16BE:
1120 /* Find '\0\0' terminator */
1121 for (text_pos = 0; text_pos < data_size - 1; text_pos += 2) {
1122 if (data[text_pos] == '\0' && data[text_pos + 1] == '\0') {
1123 /* found a delimiter */
1124 parse_insert_string_field (encoding, data + prev,
1125 text_pos - prev + 2, fields);
1126 text_pos++; /* Advance to the 2nd NULL terminator */
1127 prev = text_pos + 1;
1131 if (data_size - prev > 1 &&
1132 (data[prev] != 0x00 || data[prev + 1] != 0x00)) {
1133 /* There were 2 or more non-null chars left, convert those too */
1134 parse_insert_string_field (encoding, data + prev,
1135 data_size - prev, fields);
1140 if (fields->len > 0)
1141 *out_fields = fields;
1143 g_array_free (fields, TRUE);
1147 free_tag_strings (GArray * fields)
1153 for (i = 0; i < fields->len; i++) {
1154 c = g_array_index (fields, gchar *, i);
1157 g_array_free (fields, TRUE);