2 * Copyright (C) 2008 by INdT
4 * This program is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public License
6 * as published by the Free Software Foundation; either version 2
7 * of the License, or (at your option) any later version.
9 * This program 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
12 * GNU General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public License
15 * along with this program; if not, write to the Free Software
16 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
18 * @author Andre Moreira Magalhaes <andre.magalhaes@openbossa.org>
31 #define _XOPEN_SOURCE 600
32 #include <lightmediascanner_plugin.h>
33 #include <lightmediascanner_db.h>
34 #include <sys/types.h>
43 #define ID3V2_HEADER_SIZE 10
44 #define ID3V2_FOOTER_SIZE 10
45 #define ID3V2_FRAME_HEADER_SIZE 10
47 #define ID3V1_NUM_GENRES 148
49 static const char *id3v1_genres[ID3V1_NUM_GENRES] = {
186 "Christian Gangsta Rap",
190 "Contemporary Christian",
200 struct id3v2_frame_header {
202 unsigned int frame_size;
204 int data_length_indicator;
214 } __attribute__((packed));
217 struct lms_plugin plugin;
218 lms_db_audio_t *audio_db;
221 static const char _name[] = "id3";
222 static const struct lms_string_size _exts[] = {
223 LMS_STATIC_STRING_SIZE(".mp3"),
224 LMS_STATIC_STRING_SIZE(".aac")
228 _to_uint(const char *data, int data_size)
230 unsigned int sum = 0;
231 int last = data_size > 4 ? 3 : data_size - 1;
234 for (i = 0; i <= last; i++)
235 sum |= (data[i] & 0x7f) << ((last - i) * 7);
241 _is_id3v2_second_synch_byte(unsigned char byte)
245 if ((byte & 0xE0) == 0xE0)
253 long buffer_offset = 0;
255 int buffer_size = sizeof(buffer);
256 const char pattern[] = "ID3";
259 /* These variables are used to keep track of a partial match that happens at
260 * the end of a buffer. */
261 int previous_partial_match = -1;
262 int previous_partial_synch_match = 0;
263 int first_synch_byte;
265 /* Start the search at the beginning of the file. */
266 lseek(fd, 0, SEEK_SET);
268 if ((nread = read(fd, &buffer, buffer_size)) != buffer_size)
271 /* check if pattern is in the beggining of the file */
272 if (memcmp(buffer, pattern, 3) == 0)
275 /* This loop is the crux of the find method. There are three cases that we
276 * want to account for:
277 * (1) The previously searched buffer contained a partial match of the search
278 * pattern and we want to see if the next one starts with the remainder of
281 * (2) The search pattern is wholly contained within the current buffer.
283 * (3) The current buffer ends with a partial match of the pattern. We will
284 * note this for use in the next iteration, where we will check for the rest
287 /* (1) previous partial match */
288 if (previous_partial_synch_match && _is_id3v2_second_synch_byte(buffer[0]))
291 if (previous_partial_match >= 0 && previous_partial_match < buffer_size) {
292 const int pattern_offset = buffer_size - previous_partial_match;
294 if (memcmp(buffer, pattern + pattern_offset, 3 - pattern_offset) == 0)
295 return buffer_offset - buffer_size + previous_partial_match;
298 /* (2) pattern contained in current buffer */
300 while ((p = memchr(p, 'I', buffer_size))) {
301 if (memcmp(p, pattern, 3) == 0)
302 return buffer_offset + (p - buffer);
306 p = memchr(buffer, 255, buffer_size);
308 first_synch_byte = p - buffer;
310 first_synch_byte = -1;
312 /* Here we have to loop because there could be several of the first
313 * (11111111) byte, and we want to check all such instances until we find
314 * a full match (11111111 111) or hit the end of the buffer. */
315 while (first_synch_byte >= 0) {
316 /* if this *is not* at the end of the buffer */
317 if (first_synch_byte < buffer_size - 1) {
318 if(_is_id3v2_second_synch_byte(buffer[first_synch_byte + 1]))
319 /* We've found the frame synch pattern. */
322 /* We found 11111111 at the end of the current buffer indicating a
323 * partial match of the synch pattern. The find() below should
324 * return -1 and break out of the loop. */
325 previous_partial_synch_match = 1;
328 /* Check in the rest of the buffer. */
329 p = memchr(p + 1, 255, buffer_size);
331 first_synch_byte = p - buffer;
333 first_synch_byte = -1;
336 /* (3) partial match */
337 if (buffer[nread - 1] == pattern[1])
338 previous_partial_match = nread - 1;
339 else if (memcmp(&buffer[nread - 2], pattern, 2) == 0)
340 previous_partial_match = nread - 2;
341 buffer_offset += buffer_size;
343 if ((nread = read(fd, &buffer, sizeof(buffer))) == -1)
351 _get_id3v2_frame_header_size(unsigned int version)
366 _parse_id3v2_frame_header(char *data, unsigned int version, struct id3v2_frame_header *fh)
372 memcpy(fh->frame_id, data, 3);
374 fh->frame_size = _to_uint(data + 3, 3);
376 fh->data_length_indicator = 0;
379 memcpy(fh->frame_id, data, 4);
380 fh->frame_size = _to_uint(data + 4, 4);
381 fh->compression = data[9] & 0x40;
382 fh->data_length_indicator = 0;
386 memcpy(fh->frame_id, data, 4);
387 fh->frame_size = _to_uint(data + 4, 4);
388 fh->compression = data[9] & 0x4;
389 fh->data_length_indicator = data[9] & 0x1;
395 _get_id3v2_frame_info(const char *frame_data, unsigned int frame_size, struct lms_string_size *s)
397 s->str = malloc(sizeof(char) * (frame_size + 1));
398 memcpy(s->str, frame_data, frame_size);
399 s->str[frame_size] = '\0';
404 _parse_id3v2_frame(struct id3v2_frame_header *fh, const char *frame_data, struct lms_audio_info *info)
406 unsigned int frame_size;
408 /* TODO proper handle text encoding
417 /* skip first byte - text encoding */
419 frame_size = fh->frame_size - 1;
421 if (memcmp(fh->frame_id, "TIT2", 4) == 0)
422 _get_id3v2_frame_info(frame_data, frame_size, &info->title);
423 else if (memcmp(fh->frame_id, "TPE1", 4) == 0)
424 _get_id3v2_frame_info(frame_data, frame_size, &info->artist);
425 else if (memcmp(fh->frame_id, "TALB", 4) == 0)
426 _get_id3v2_frame_info(frame_data, frame_size, &info->album);
427 else if (memcmp(fh->frame_id, "TCON", 4) == 0) {
428 /* TODO handle multiple genres */
431 _get_id3v2_frame_info(frame_data, frame_size, &info->genre);
434 for (i = 0; i < info->genre.len; ++i) {
435 if (!isdigit(info->genre.str[i]))
441 (info->genre.len > 0)) {
442 int genre = atoi(info->genre.str);
443 if (genre >= 0 && genre < ID3V1_NUM_GENRES) {
444 free(info->genre.str);
445 info->genre.str = strdup(id3v1_genres[(int) genre]);
446 info->genre.len = strlen(info->genre.str);
449 /* ignore other genres */
450 free(info->genre.str);
451 info->genre.str = NULL;
456 else if (memcmp(fh->frame_id, "TRCK", 4) == 0) {
457 struct lms_string_size trackno;
458 _get_id3v2_frame_info(frame_data, frame_size, &trackno);
459 info->trackno = atoi(trackno.str);
465 _parse_id3v2(int fd, long id3v2_offset, struct lms_audio_info *info)
467 char header_data[10], frame_header_data[10];
468 unsigned int tag_size, major_version, frame_data_pos, frame_data_length, frame_header_size;
469 int extended_header, footer_present;
470 struct id3v2_frame_header fh;
472 lseek(fd, id3v2_offset, SEEK_SET);
475 if (read(fd, header_data, ID3V2_HEADER_SIZE) != ID3V2_HEADER_SIZE)
478 tag_size = _to_uint(header_data + 6, 4);
483 major_version = header_data[3];
486 frame_data_length = tag_size;
488 /* check for extended header */
489 extended_header = header_data[5] & 0x20; /* bit 6 */
490 if (extended_header) {
491 /* skip extended header */
492 unsigned int extended_header_size;
493 char extended_header_data[4];
495 if (read(fd, extended_header_data, 4) != 4)
497 extended_header_size = _to_uint(extended_header_data, 4);
498 lseek(fd, extended_header_size - 4, SEEK_CUR);
499 frame_data_pos += extended_header_size;
500 frame_data_length -= extended_header_size;
503 footer_present = header_data[5] & 0x8; /* bit 4 */
504 if (footer_present && frame_data_length > ID3V2_FOOTER_SIZE)
505 frame_data_length -= ID3V2_FOOTER_SIZE;
507 frame_header_size = _get_id3v2_frame_header_size(major_version);
508 while (frame_data_pos < frame_data_length - frame_header_size) {
509 if (read(fd, frame_header_data, ID3V2_FRAME_HEADER_SIZE) != ID3V2_FRAME_HEADER_SIZE)
512 if (frame_header_data[0] == 0)
515 _parse_id3v2_frame_header(frame_header_data, major_version, &fh);
517 if (!fh.compression &&
518 fh.frame_id[0] == 'T' &&
519 memcmp(fh.frame_id, "TXXX", 4) != 0) {
522 if (fh.data_length_indicator)
523 lseek(fd, 4, SEEK_CUR);
525 frame_data = malloc(sizeof(char) * fh.frame_size);
526 if (read(fd, frame_data, fh.frame_size) != fh.frame_size) {
531 _parse_id3v2_frame(&fh, frame_data, info);
535 if (fh.data_length_indicator)
536 lseek(fd, fh.frame_size + 4, SEEK_CUR);
538 lseek(fd, fh.frame_size, SEEK_CUR);
541 frame_data_pos += fh.frame_size + frame_header_size;
548 _parse_id3v1(int fd, struct lms_audio_info *info)
550 struct id3v1_tag tag;
551 if (read(fd, &tag, sizeof(struct id3v1_tag)) == -1)
554 info->title.str = strdup(tag.title);
555 info->title.len = strlen(info->title.str);
556 info->artist.str = strdup(tag.artist);
557 info->artist.len = strlen(info->artist.str);
558 info->album.str = strdup(tag.album);
559 info->album.len = strlen(info->album.str);
560 info->genre.str = strdup(id3v1_genres[(int) tag.genre]);
561 info->genre.len = strlen(info->genre.str);
562 if (tag.comments[28] == 0 && tag.comments[29] != 0)
563 info->trackno = (unsigned char) tag.comments[29];
569 _match(struct plugin *p, const char *path, int len, int base)
573 i = lms_which_extension(path, len, _exts, LMS_ARRAY_SIZE(_exts));
577 return (void*)(i + 1);
581 _parse(struct plugin *plugin, struct lms_context *ctxt, const struct lms_file_info *finfo, void *match)
583 struct lms_audio_info info = {0, {0}, {0}, {0}, {0}, 0, 0, 0};
587 fd = open(finfo->path, O_RDONLY);
593 id3v2_offset = _find_id3v2(fd);
594 if (id3v2_offset >= 0) {
596 fprintf(stderr, "id3v2 tag found in file %s with offset %ld\n",
597 finfo->path, id3v2_offset);
599 if (_parse_id3v2(fd, id3v2_offset, &info) != 0) {
607 fprintf(stderr, "id3v2 tag not found in file %s. trying id3v1\n", finfo->path);
609 /* check for id3v1 tag */
610 if (lseek(fd, -128, SEEK_END) == -1) {
615 if (read(fd, &tag, 3) == -1) {
620 if (memcmp(tag, "TAG", 3) == 0) {
622 fprintf(stderr, "id3v1 tag found in file %s\n", finfo->path);
624 if (_parse_id3v1(fd, &info) != 0) {
631 lms_string_size_strip_and_free(&info.title);
632 lms_string_size_strip_and_free(&info.artist);
633 lms_string_size_strip_and_free(&info.album);
634 lms_string_size_strip_and_free(&info.genre);
636 if (!info.title.str) {
638 ext_idx = ((int)match) - 1;
639 info.title.len = finfo->path_len - finfo->base - _exts[ext_idx].len;
640 info.title.str = malloc((info.title.len + 1) * sizeof(char));
641 memcpy(info.title.str, finfo->path + finfo->base, info.title.len);
642 info.title.str[info.title.len] = '\0';
644 lms_charset_conv(ctxt->cs_conv, &info.title.str, &info.title.len);
647 lms_charset_conv(ctxt->cs_conv, &info.artist.str, &info.artist.len);
649 lms_charset_conv(ctxt->cs_conv, &info.album.str, &info.album.len);
651 lms_charset_conv(ctxt->cs_conv, &info.genre.str, &info.genre.len);
654 fprintf(stderr, "file %s info\n", finfo->path);
655 fprintf(stderr, "\ttitle='%s'\n", info.title.str);
656 fprintf(stderr, "\tartist='%s'\n", info.artist.str);
657 fprintf(stderr, "\talbum='%s'\n", info.album.str);
658 fprintf(stderr, "\tgenre='%s'\n", info.genre.str);
659 fprintf(stderr, "\ttrack number='%d'\n", info.trackno);
663 r = lms_db_audio_add(plugin->audio_db, &info);
669 free(info.title.str);
671 free(info.artist.str);
673 free(info.album.str);
675 free(info.genre.str);
681 _setup(struct plugin *plugin, struct lms_context *ctxt)
683 plugin->audio_db = lms_db_audio_new(ctxt->db);
684 if (!plugin->audio_db)
690 _start(struct plugin *plugin, struct lms_context *ctxt)
692 return lms_db_audio_start(plugin->audio_db);
696 _finish(struct plugin *plugin, struct lms_context *ctxt)
698 if (plugin->audio_db)
699 lms_db_audio_free(plugin->audio_db);
704 _close(struct plugin *plugin)
710 API struct lms_plugin *
711 lms_plugin_open(void)
713 struct plugin *plugin;
715 plugin = (struct plugin *)malloc(sizeof(*plugin));
716 plugin->plugin.name = _name;
717 plugin->plugin.match = (lms_plugin_match_fn_t)_match;
718 plugin->plugin.parse = (lms_plugin_parse_fn_t)_parse;
719 plugin->plugin.close = (lms_plugin_close_fn_t)_close;
720 plugin->plugin.setup = (lms_plugin_setup_fn_t)_setup;
721 plugin->plugin.start = (lms_plugin_start_fn_t)_start;
722 plugin->plugin.finish = (lms_plugin_finish_fn_t)_finish;
724 return (struct lms_plugin *)plugin;