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 if (work->parse_size < frame_data_size) {
120 GST_WARNING ("ID3v2 frame %s has invalid size %d.", tag_name,
126 /* in v2.3 the frame sizes are not syncsafe, so the entire tag had to be
127 * unsynced. In v2.4 the frame sizes are syncsafe so it's just the frame
128 * data that needs un-unsyncing, but not the frame headers. */
129 if (ID3V2_VER_MAJOR (work->hdr.version) == 4) {
130 if ((work->hdr.flags & ID3V2_HDR_FLAG_UNSYNC) != 0 ||
131 ((work->frame_flags & ID3V2_FRAME_FORMAT_UNSYNCHRONISATION) != 0)) {
132 GST_DEBUG ("Un-unsyncing frame %s", work->frame_id);
133 uu_data = id3demux_ununsync_data (frame_data, &frame_data_size);
134 frame_data = uu_data;
135 GST_MEMDUMP ("ID3v2 frame (un-unsyced)", frame_data, frame_data_size);
139 work->parse_size = frame_data_size;
141 if (work->frame_flags & ID3V2_FRAME_FORMAT_COMPRESSION) {
143 uLongf destSize = work->parse_size;
146 uncompressed_data = g_malloc (work->parse_size);
148 dest = (Bytef *) uncompressed_data;
149 src = (Bytef *) frame_data;
151 if (uncompress (dest, &destSize, src, frame_data_size) != Z_OK) {
152 g_free (uncompressed_data);
156 if (destSize != work->parse_size) {
158 ("Decompressing ID3v2 frame %s did not produce expected size %d bytes (got %lu)",
159 tag_name, work->parse_size, destSize);
160 g_free (uncompressed_data);
164 work->parse_data = uncompressed_data;
166 GST_WARNING ("Compressed ID3v2 tag frame could not be decompressed"
167 " because gstid3demux was compiled without zlib support");
172 work->parse_data = frame_data;
175 if (work->frame_id[0] == 'T') {
176 if (strcmp (work->frame_id, "TDAT") == 0) {
177 parse_obsolete_tdat_frame (work);
179 } else if (strcmp (work->frame_id, "TXXX") == 0) {
180 /* Handle user text frame */
181 tag_str = parse_user_text_identification_frame (work, &tag_name);
183 /* Text identification frame */
184 tag_fields = parse_text_identification_frame (work);
186 } else if (work->frame_id[0] == 'W' && strcmp (work->frame_id, "WXXX") != 0) {
187 /* URL link frame: ISO-8859-1 encoded, one frame per tag */
188 tag_str = parse_url_link_frame (work, &tag_name);
189 } else if (!strcmp (work->frame_id, "COMM")) {
191 result = parse_comment_frame (work);
192 } else if (!strcmp (work->frame_id, "APIC")) {
193 /* Attached picture */
194 result = parse_picture_frame (work);
195 } else if (!strcmp (work->frame_id, "RVA2")) {
196 /* Relative volume */
197 result = parse_relative_volume_adjustment_two (work);
198 } else if (!strcmp (work->frame_id, "UFID")) {
199 /* Unique file identifier */
200 tag_str = parse_unique_file_identifier (work, &tag_name);
203 if (work->frame_flags & ID3V2_FRAME_FORMAT_COMPRESSION) {
204 g_free (uncompressed_data);
205 uncompressed_data = NULL;
206 work->parse_data = frame_data;
210 if (tag_str != NULL) {
211 /* g_print ("Tag %s value %s\n", tag_name, tag_str); */
212 result = id3v2_tag_to_taglist (work, tag_name, tag_str);
215 if (tag_fields != NULL) {
216 if (strcmp (work->frame_id, "TCON") == 0) {
217 /* Genre strings need special treatment */
218 result |= id3v2_genre_fields_to_taglist (work, tag_name, tag_fields);
222 for (t = 0; t < tag_fields->len; t++) {
223 tag_str = g_array_index (tag_fields, gchar *, t);
224 if (tag_str != NULL && tag_str[0] != '\0')
225 result |= id3v2_tag_to_taglist (work, tag_name, tag_str);
228 free_tag_strings (tag_fields);
237 parse_comment_frame (ID3TagsWorking * work)
242 GArray *fields = NULL;
243 gchar *description, *text;
245 if (work->parse_size < 6)
248 encoding = work->parse_data[0];
249 language[0] = g_ascii_tolower (work->parse_data[1]);
250 language[1] = g_ascii_tolower (work->parse_data[2]);
251 language[2] = g_ascii_tolower (work->parse_data[3]);
254 parse_split_strings (encoding, (gchar *) work->parse_data + 4,
255 work->parse_size - 4, &fields);
257 if (fields == NULL || fields->len < 2) {
258 GST_WARNING ("Failed to decode comment frame");
261 description = g_array_index (fields, gchar *, 0);
262 text = g_array_index (fields, gchar *, 1);
264 if (!g_utf8_validate (text, -1, NULL)) {
265 GST_WARNING ("Converted string is not valid utf-8");
269 /* skip our own dummy descriptions (from id3v2mux) */
270 if (strlen (description) > 0 && g_utf8_validate (description, -1, NULL) &&
271 sscanf (description, "c%u", &dummy) != 1) {
274 /* must be either an ISO-639-1 or ISO-639-2 language code */
275 if (language[0] != '\0' &&
276 g_ascii_isalpha (language[0]) &&
277 g_ascii_isalpha (language[1]) &&
278 (g_ascii_isalpha (language[2]) || language[2] == '\0')) {
279 s = g_strdup_printf ("%s[%s]=%s", description, language, text);
281 s = g_strdup_printf ("%s=%s", description, text);
283 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
284 GST_TAG_EXTENDED_COMMENT, s, NULL);
286 } else if (text != NULL && *text != '\0') {
287 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
288 GST_TAG_COMMENT, text, NULL);
293 free_tag_strings (fields);
298 GST_WARNING ("failed to parse COMM frame");
299 free_tag_strings (fields);
305 parse_text_identification_frame (ID3TagsWorking * work)
308 GArray *fields = NULL;
310 if (work->parse_size < 2)
313 encoding = work->parse_data[0];
314 parse_split_strings (encoding, (gchar *) work->parse_data + 1,
315 work->parse_size - 1, &fields);
317 if (fields->len > 0) {
318 GST_LOG ("Read %d fields from Text ID frame of size %d with encoding %d"
319 ". First is '%s'", fields->len, work->parse_size - 1, encoding,
320 g_array_index (fields, gchar *, 0));
322 GST_LOG ("Read 0 fields from Text ID frame of size %d with encoding %d",
323 work->parse_size - 1, encoding);
331 link_is_known_license (const gchar * url)
333 return g_str_has_prefix (url, "http://creativecommons.org/licenses/");
337 parse_url_link_frame (ID3TagsWorking * work, const gchar ** tag_name)
340 gchar *nul, *data, *link;
344 if (work->parse_size == 0)
347 data = (gchar *) work->parse_data;
348 /* if there's more data then the string is long, we only want to parse the
349 * data up to the terminating zero to g_convert and ignore the rest, as
351 nul = memchr (data, '\0', work->parse_size);
353 len = (gsize) (nul - data);
355 len = work->parse_size;
358 link = g_convert (data, len, "UTF-8", "ISO-8859-1", NULL, NULL, NULL);
360 if (link == NULL || !gst_uri_is_valid (link)) {
361 GST_DEBUG ("Invalid URI in %s frame: %s", work->frame_id,
362 GST_STR_NULL (link));
367 /* we don't know if it's a link to a page that explains the copyright
368 * situation, or a link that points to/represents a license, the ID3 spec
369 * does not separate those two things; for now only put known license URIs
370 * into GST_TAG_LICENSE_URI and everything else into GST_TAG_COPYRIGHT_URI */
371 if (strcmp (work->frame_id, "WCOP") == 0) {
372 if (link_is_known_license (link))
373 *tag_name = GST_TAG_LICENSE_URI;
375 *tag_name = GST_TAG_COPYRIGHT_URI;
376 } else if (strcmp (work->frame_id, "WOAF") == 0) {
377 /* can't be bothered to create a CONTACT_URI tag for this, so let's just
378 * put into into GST_TAG_CONTACT, which is where it ends up when reading
379 * the info from vorbis comments as well */
380 *tag_name = GST_TAG_CONTACT;
388 parse_user_text_identification_frame (ID3TagsWorking * work,
389 const gchar ** tag_name)
393 GArray *fields = NULL;
397 if (work->parse_size < 2)
400 encoding = work->parse_data[0];
402 parse_split_strings (encoding, (gchar *) work->parse_data + 1,
403 work->parse_size - 1, &fields);
408 if (fields->len != 2) {
409 GST_WARNING ("Expected 2 fields in TXXX frame, but got %d", fields->len);
410 free_tag_strings (fields);
415 gst_tag_from_id3_user_tag ("TXXX", g_array_index (fields, gchar *, 0));
417 GST_LOG ("TXXX frame of size %d. Mapped descriptor '%s' to GStreamer tag %s",
418 work->parse_size - 1, g_array_index (fields, gchar *, 0),
419 GST_STR_NULL (*tag_name));
422 ret = g_strdup (g_array_index (fields, gchar *, 1));
423 /* GST_LOG ("%s = %s", *tag_name, GST_STR_NULL (ret)); */
428 free_tag_strings (fields);
433 parse_id_string (ID3TagsWorking * work, gchar ** p_str, gint * p_len,
438 if (work->parse_size < 2)
441 for (len = 0; len < work->parse_size - 1; ++len) {
442 if (work->parse_data[len] == '\0')
446 datalen = work->parse_size - (len + 1);
447 if (len == 0 || datalen <= 0)
450 *p_str = g_strndup ((gchar *) work->parse_data, len);
452 *p_datalen = datalen;
458 parse_unique_file_identifier (ID3TagsWorking * work, const gchar ** tag_name)
461 gchar *owner_id, *data, *ret = NULL;
463 GST_LOG ("parsing UFID frame of size %d", work->parse_size);
465 if (!parse_id_string (work, &owner_id, &len, &datalen))
468 data = (gchar *) work->parse_data + len + 1;
469 GST_LOG ("UFID owner ID: %s (+ %d bytes of data)", owner_id, datalen);
471 if (strcmp (owner_id, "http://musicbrainz.org") == 0 &&
472 g_utf8_validate (data, datalen, NULL)) {
473 *tag_name = GST_TAG_MUSICBRAINZ_TRACKID;
474 ret = g_strndup (data, datalen);
476 GST_INFO ("Unknown UFID owner ID: %s", owner_id);
483 /* parse data and return length of the next string in the given encoding,
484 * including the NUL terminator */
486 scan_encoded_string (guint8 encoding, gchar * data, gint data_size)
491 case ID3V2_ENCODING_ISO8859:
492 case ID3V2_ENCODING_UTF8:
493 for (i = 0; i < data_size; ++i) {
498 case ID3V2_ENCODING_UTF16:
499 case ID3V2_ENCODING_UTF16BE:
500 /* we don't care about BOMs here and treat them as part of the string */
501 /* Find '\0\0' terminator */
502 for (i = 0; i < data_size - 1; i += 2) {
503 if (data[i] == '\0' && data[i + 1] == '\0')
515 parse_picture_frame (ID3TagsWorking * work)
517 guint8 txt_encoding, pic_type;
518 gchar *mime_str = NULL;
521 GST_LOG ("APIC frame (ID3v2.%u)", ID3V2_VER_MAJOR (work->hdr.version));
523 if (work->parse_size < 1 + 1 + 1 + 1 + 1)
524 goto not_enough_data;
526 txt_encoding = work->parse_data[0];
530 /* Read image format; in early ID3v2 versions this is a fixed-length
531 * 3-character string without terminator; in later versions (>= 2.3.0)
532 * this is a NUL-terminated string of variable length */
533 if (ID3V2_VER_MAJOR (work->hdr.version) < 3) {
534 if (work->parse_size < 3)
535 goto not_enough_data;
537 mime_str = g_strndup ((gchar *) work->parse_data, 3);
540 if (!parse_id_string (work, &mime_str, &len, &datalen))
542 ++len; /* for string terminator */
545 if (work->parse_size < len + 1 + 1 + 1)
546 goto not_enough_data;
548 work->parse_data += len;
549 work->parse_size -= len;
551 /* Read image type */
552 pic_type = work->parse_data[0];
556 GST_LOG ("APIC frame mime type : %s", GST_STR_NULL (mime_str));
557 GST_LOG ("APIC frame picture type : 0x%02x", (guint) pic_type);
559 if (work->parse_size < 1 + 1)
560 goto not_enough_data;
562 len = scan_encoded_string (txt_encoding, (gchar *) work->parse_data,
568 /* just skip the description string ... */
569 GST_LOG ("Skipping description string (%d bytes in original coding)", len);
571 if (work->parse_size < len + 1)
572 goto not_enough_data;
574 work->parse_data += len;
575 work->parse_size -= len;
577 GST_DEBUG ("image data is %u bytes", work->parse_size);
579 if (work->parse_size <= 0)
580 goto not_enough_data;
582 if (!gst_tag_list_add_id3_image (work->tags, (guint8 *) work->parse_data,
583 work->parse_size, pic_type)) {
592 GST_DEBUG ("not enough data, skipping APIC frame");
593 /* fall through to error */
597 GST_DEBUG ("problem parsing APIC frame, skipping");
603 #define ID3V2_RVA2_CHANNEL_MASTER 1
606 parse_relative_volume_adjustment_two (ID3TagsWorking * work)
608 const gchar *gain_tag_name = NULL;
609 const gchar *peak_tag_name = NULL;
610 gdouble gain_dB, peak_val;
612 guint8 *data, chan, peak_bits;
614 gint len, datalen, i;
616 if (!parse_id_string (work, &id, &len, &datalen))
619 if (datalen < (1 + 2 + 1)) {
620 GST_WARNING ("broken RVA2 frame, data size only %d bytes", datalen);
625 data = work->parse_data + len + 1;
626 chan = GST_READ_UINT8 (data);
627 gain_dB = (gdouble) ((gint16) GST_READ_UINT16_BE (data + 1)) / 512.0;
628 /* The meaning of the peak value is not defined in the ID3v2 spec. However,
629 * the first/only implementation of this seems to have been in XMMS, and
630 * other libs (like mutagen) seem to follow that implementation as well:
631 * see http://bugs.xmms.org/attachment.cgi?id=113&action=view */
632 peak_bits = GST_READ_UINT8 (data + 1 + 2);
633 if (peak_bits > 64) {
634 GST_WARNING ("silly peak precision of %d bits, ignoring", (gint) peak_bits);
638 datalen -= 1 + 2 + 1;
639 if (peak_bits == 16) {
640 peak = GST_READ_UINT16_BE (data);
643 for (i = 0; i < (GST_ROUND_UP_8 (peak_bits) / 8) && datalen > 0; ++i) {
645 peak |= GST_READ_UINT8 (data);
651 peak = peak << (64 - GST_ROUND_UP_8 (peak_bits));
653 gst_guint64_to_gdouble (peak) / gst_util_guint64_to_gdouble (G_MAXINT64);
654 GST_LOG ("RVA2 frame: id=%s, chan=%u, adj=%.2fdB, peak_bits=%u, peak=%.2f",
655 id, chan, gain_dB, (guint) peak_bits, peak_val);
657 if (chan == ID3V2_RVA2_CHANNEL_MASTER && strcmp (id, "track") == 0) {
658 gain_tag_name = GST_TAG_TRACK_GAIN;
659 peak_tag_name = GST_TAG_TRACK_PEAK;
660 } else if (chan == ID3V2_RVA2_CHANNEL_MASTER && strcmp (id, "album") == 0) {
661 gain_tag_name = GST_TAG_ALBUM_GAIN;
662 peak_tag_name = GST_TAG_ALBUM_PEAK;
664 GST_INFO ("Unhandled RVA2 frame id '%s' for channel %d", id, chan);
668 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
669 gain_tag_name, gain_dB, NULL);
671 if (peak_tag_name && peak_bits > 0) {
672 gst_tag_list_add (work->tags, GST_TAG_MERGE_APPEND,
673 peak_tag_name, peak_val, NULL);
678 return (gain_tag_name != NULL || peak_tag_name != NULL);
682 parse_obsolete_tdat_frame (ID3TagsWorking * work)
684 if (work->parse_size >= 5 &&
685 work->parse_data[0] == ID3V2_ENCODING_ISO8859 &&
686 g_ascii_isdigit (work->parse_data[1]) &&
687 g_ascii_isdigit (work->parse_data[2]) &&
688 g_ascii_isdigit (work->parse_data[3]) &&
689 g_ascii_isdigit (work->parse_data[4])) {
690 work->pending_day = (10 * g_ascii_digit_value (work->parse_data[1])) +
691 g_ascii_digit_value (work->parse_data[2]);
692 work->pending_month = (10 * g_ascii_digit_value (work->parse_data[3])) +
693 g_ascii_digit_value (work->parse_data[4]);
694 GST_LOG ("date (dd/mm) %02u/%02u", work->pending_day, work->pending_month);
699 id3v2_tag_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
700 const gchar * tag_str)
702 GType tag_type = gst_tag_get_type (tag_name);
703 GstTagList *tag_list = work->tags;
713 if (sscanf (tag_str, "%d/%d", ¤t, &total) == 2) {
715 GST_WARNING ("Ignoring invalid value for total %d in tag %s",
718 if (strcmp (tag_name, GST_TAG_TRACK_NUMBER) == 0) {
719 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
720 GST_TAG_TRACK_COUNT, total, NULL);
721 } else if (strcmp (tag_name, GST_TAG_ALBUM_VOLUME_NUMBER) == 0) {
722 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
723 GST_TAG_ALBUM_VOLUME_COUNT, total, NULL);
726 } else if (sscanf (tag_str, "%d", ¤t) != 1) {
727 /* Not an integer in the string */
728 GST_WARNING ("Tag string for tag %s does not contain an integer - "
729 "ignoring", tag_name);
734 GST_WARNING ("Ignoring invalid value %d in tag %s", current, tag_name);
736 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND, tag_name, current,
745 g_assert (strcmp (tag_name, GST_TAG_DURATION) == 0);
746 tmp = strtoul (tag_str, NULL, 10);
750 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
751 GST_TAG_DURATION, tmp * 1000 * 1000, NULL);
758 /* make sure we add each unique string only once per tag, we don't want
759 * to have the same genre in the genre list multiple times, for example,
760 * or the same DiscID in there twice just because it's contained in the
761 * tag multiple times under different TXXX user tags */
762 num = gst_tag_list_get_tag_size (tag_list, tag_name);
763 for (i = 0; i < num; ++i) {
764 val = gst_tag_list_get_value_index (tag_list, tag_name, i);
765 if (val != NULL && strcmp (g_value_get_string (val), tag_str) == 0)
769 gst_tag_list_add (tag_list, GST_TAG_MERGE_APPEND,
770 tag_name, tag_str, NULL);
778 GValue dest = { 0, };
780 /* Ensure that any date string is complete */
781 if (tag_type == GST_TYPE_DATE) {
782 guint year = 1901, month = 1, day = 1;
784 /* Dates can be yyyy-MM-dd, yyyy-MM or yyyy, but we need
786 if (sscanf (tag_str, "%04u-%02u-%02u", &year, &month, &day) == 0)
789 tmp = g_strdup_printf ("%04u-%02u-%02u", year, month, day);
793 /* handles anything else */
794 g_value_init (&src, G_TYPE_STRING);
795 g_value_set_string (&src, (const gchar *) tag_str);
796 g_value_init (&dest, tag_type);
798 if (g_value_transform (&src, &dest)) {
799 gst_tag_list_add_values (tag_list, GST_TAG_MERGE_APPEND,
800 tag_name, &dest, NULL);
801 } else if (tag_type == G_TYPE_DOUBLE) {
802 /* replaygain tags in TXXX frames ... */
803 g_value_set_double (&dest, g_strtod (tag_str, NULL));
804 gst_tag_list_add_values (tag_list, GST_TAG_MERGE_KEEP,
805 tag_name, &dest, NULL);
806 GST_LOG ("Converted string '%s' to double %f", tag_str,
807 g_value_get_double (&dest));
809 GST_WARNING ("Failed to transform tag from string to type '%s'",
810 g_type_name (tag_type));
813 g_value_unset (&src);
814 g_value_unset (&dest);
823 /* Check that an array of characters contains only digits */
825 id3v2_are_digits (const gchar * chars, gint size)
829 for (i = 0; i < size; i++) {
830 if (!g_ascii_isdigit (chars[i]))
837 id3v2_genre_string_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
838 const gchar * tag_str, gint len)
840 g_return_val_if_fail (tag_str != NULL, FALSE);
842 /* If it's a number, it might be a defined genre */
843 if (id3v2_are_digits (tag_str, len)) {
844 tag_str = gst_tag_id3_genre_get (strtol (tag_str, NULL, 10));
845 return id3v2_tag_to_taglist (work, tag_name, tag_str);
847 /* Otherwise it might be "RX" or "CR" */
849 if (g_ascii_strncasecmp ("rx", tag_str, len) == 0)
850 return id3v2_tag_to_taglist (work, tag_name, "Remix");
852 if (g_ascii_strncasecmp ("cr", tag_str, len) == 0)
853 return id3v2_tag_to_taglist (work, tag_name, "Cover");
856 /* Otherwise it's a string */
857 return id3v2_tag_to_taglist (work, tag_name, tag_str);
861 id3v2_genre_fields_to_taglist (ID3TagsWorking * work, const gchar * tag_name,
864 gchar *tag_str = NULL;
865 gboolean result = FALSE;
868 for (i = 0; i < tag_fields->len; i++) {
871 tag_str = g_array_index (tag_fields, gchar *, 0);
875 len = strlen (tag_str);
876 /* Only supposed to see '(n)' type numeric genre strings in ID3 <= 2.3.0
877 * but apparently we see them in 2.4.0 sometimes too */
878 if (TRUE || work->hdr.version <= 0x300) { /* <= 2.3.0 */
879 /* Check for genre numbers wrapped in parentheses, possibly
880 * followed by a string */
883 gboolean found = FALSE;
885 /* Double parenthesis ends the numeric genres, but we need
886 * to swallow the first one so we actually output '(' */
887 if (tag_str[0] == '(' && tag_str[1] == '(') {
893 /* If the first char is not a parenthesis, then stop
894 * looking for parenthesised genre strings */
895 if (tag_str[0] != '(')
898 for (pos = 1; pos < len; pos++) {
899 if (tag_str[pos] == ')') {
902 tmp_str = g_strndup (tag_str + 1, pos - 1);
904 id3v2_genre_string_to_taglist (work, tag_name, tmp_str,
913 /* If we encounter a non-digit while searching for a closing
914 * parenthesis, we should not try and interpret this as a
915 * numeric genre string */
916 if (!g_ascii_isdigit (tag_str[pos]))
920 break; /* There was no closing parenthesis */
924 if (len > 0 && tag_str != NULL)
925 result |= id3v2_genre_string_to_taglist (work, tag_name, tag_str, len);
930 static const gchar utf16enc[] = "UTF-16";
931 static const gchar utf16leenc[] = "UTF-16LE";
932 static const gchar utf16beenc[] = "UTF-16BE";
935 find_utf16_bom (gchar * data, const gchar ** p_in_encoding)
937 guint16 marker = (GST_READ_UINT8 (data) << 8) | GST_READ_UINT8 (data + 1);
941 *p_in_encoding = utf16leenc;
944 *p_in_encoding = utf16beenc;
953 string_utf8_dup (const gchar * start, const guint size)
959 /* Should we try the charsets specified
960 * via environment variables FIRST ? */
961 if (g_utf8_validate (start, size, NULL)) {
962 utf8 = g_strndup (start, size);
966 env = g_getenv ("GST_ID3V1_TAG_ENCODING");
967 if (!env || *env == '\0')
968 env = g_getenv ("GST_ID3_TAG_ENCODING");
969 if (!env || *env == '\0')
970 env = g_getenv ("GST_TAG_ENCODING");
972 /* Try charsets specified via the environment */
973 if (env && *env != '\0') {
976 csets = g_strsplit (env, G_SEARCHPATH_SEPARATOR_S, -1);
978 for (c = csets; c && *c; ++c) {
980 g_convert (start, size, "UTF-8", *c, &bytes_read, NULL, NULL))) {
981 if (bytes_read == size) {
982 GST_DEBUG ("Using charset %s to interperate id3 tags\n", *c);
991 /* Try current locale (if not UTF-8) */
992 if (!g_get_charset (&env)) {
993 if ((utf8 = g_locale_to_utf8 (start, size, &bytes_read, NULL, NULL))) {
994 if (bytes_read == size) {
1002 /* Try ISO-8859-1 */
1004 g_convert (start, size, "UTF-8", "ISO-8859-1", &bytes_read, NULL, NULL);
1005 if (utf8 != NULL && bytes_read == size) {
1020 parse_insert_string_field (guint8 encoding, gchar * data, gint data_size,
1023 gchar *field = NULL;
1026 case ID3V2_ENCODING_UTF16:
1027 case ID3V2_ENCODING_UTF16BE:
1029 const gchar *in_encode;
1031 if (encoding == ID3V2_ENCODING_UTF16)
1032 in_encode = utf16enc;
1034 in_encode = utf16beenc;
1036 /* Sometimes we see strings with multiple BOM markers at the start.
1037 * In that case, we assume the innermost one is correct. If that fails
1038 * to produce valid UTF-8, we try the other endianness anyway */
1039 while (data_size > 2 && find_utf16_bom (data, &in_encode)) {
1040 data += 2; /* skip BOM */
1044 field = g_convert (data, data_size, "UTF-8", in_encode, NULL, NULL, NULL);
1046 if (field == NULL || g_utf8_validate (field, -1, NULL) == FALSE) {
1047 /* As a fallback, try interpreting UTF-16 in the other endianness */
1048 if (in_encode == utf16beenc)
1049 field = g_convert (data, data_size, "UTF-8", utf16leenc,
1055 case ID3V2_ENCODING_ISO8859:
1056 if (g_utf8_validate (data, data_size, NULL))
1057 field = g_strndup (data, data_size);
1059 /* field = g_convert (data, data_size, "UTF-8", "ISO-8859-1",
1060 NULL, NULL, NULL); */
1061 field = string_utf8_dup (data, data_size);
1064 field = g_strndup (data, data_size);
1069 if (g_utf8_validate (field, -1, NULL)) {
1070 g_array_append_val (fields, field);
1074 GST_DEBUG ("%s was bad UTF-8 after conversion from encoding %d. Ignoring",
1081 parse_split_strings (guint8 encoding, gchar * data, gint data_size,
1082 GArray ** out_fields)
1084 GArray *fields = g_array_new (FALSE, TRUE, sizeof (gchar *));
1088 g_return_if_fail (out_fields != NULL);
1091 case ID3V2_ENCODING_ISO8859:
1092 for (text_pos = 0; text_pos < data_size; text_pos++) {
1093 if (data[text_pos] == 0) {
1094 parse_insert_string_field (encoding, data + prev,
1095 text_pos - prev + 1, fields);
1096 prev = text_pos + 1;
1099 if (data_size - prev > 0 && data[prev] != 0x00) {
1100 parse_insert_string_field (encoding, data + prev,
1101 data_size - prev, fields);
1105 case ID3V2_ENCODING_UTF8:
1106 for (prev = 0, text_pos = 0; text_pos < data_size; text_pos++) {
1107 if (data[text_pos] == '\0') {
1108 parse_insert_string_field (encoding, data + prev,
1109 text_pos - prev + 1, fields);
1110 prev = text_pos + 1;
1113 if (data_size - prev > 0 && data[prev] != 0x00) {
1114 parse_insert_string_field (encoding, data + prev,
1115 data_size - prev, fields);
1118 case ID3V2_ENCODING_UTF16:
1119 case ID3V2_ENCODING_UTF16BE:
1121 /* Find '\0\0' terminator */
1122 for (text_pos = 0; text_pos < data_size - 1; text_pos += 2) {
1123 if (data[text_pos] == '\0' && data[text_pos + 1] == '\0') {
1124 /* found a delimiter */
1125 parse_insert_string_field (encoding, data + prev,
1126 text_pos - prev + 2, fields);
1127 text_pos++; /* Advance to the 2nd NULL terminator */
1128 prev = text_pos + 1;
1132 if (data_size - prev > 1 &&
1133 (data[prev] != 0x00 || data[prev + 1] != 0x00)) {
1134 /* There were 2 or more non-null chars left, convert those too */
1135 parse_insert_string_field (encoding, data + prev,
1136 data_size - prev, fields);
1141 if (fields->len > 0)
1142 *out_fields = fields;
1144 g_array_free (fields, TRUE);
1148 free_tag_strings (GArray * fields)
1154 for (i = 0; i < fields->len; i++) {
1155 c = g_array_index (fields, gchar *, i);
1158 g_array_free (fields, TRUE);