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>
42 #define ID3V2_GET_FRAME_INFO(frame_data, frame_size, str, len) { \
43 str = malloc(sizeof(char) * (frame_size + 1)); \
44 memcpy(str, frame_data, frame_size); \
45 str[frame_size] = '\0'; \
49 #define ID3V2_HEADER_SIZE 10
50 #define ID3V2_FOOTER_SIZE 10
51 #define ID3V2_FRAME_HEADER_SIZE 10
53 #define ID3V1_NUM_GENRES 148
55 static const char *id3v1_genres[ID3V1_NUM_GENRES] = {
192 "Christian Gangsta Rap",
196 "Contemporary Christian",
206 struct id3v2_frame_header {
208 unsigned int frame_size;
210 int data_length_indicator;
220 } __attribute__((packed));
223 struct lms_plugin plugin;
224 lms_db_audio_t *audio_db;
227 static const char _name[] = "id3";
228 static const struct lms_string_size _exts[] = {
229 LMS_STATIC_STRING_SIZE(".mp3"),
230 LMS_STATIC_STRING_SIZE(".aac")
234 _is_id3v2_second_synch_byte(unsigned char byte)
238 if ((byte & 0xE0) == 0xE0)
246 long buffer_offset = 0;
249 int buffer_size = sizeof(buffer);
250 const char pattern[] = "ID3";
253 /* These variables are used to keep track of a partial match that happens at
254 * the end of a buffer. */
255 int previous_partial_match = -1;
256 int previous_partial_synch_match = 0;
257 int first_synch_byte;
259 /* Start the search at the beginning of the file. */
260 lseek(fd, 0, SEEK_SET);
262 /* This loop is the crux of the find method. There are three cases that we
263 * want to account for:
264 * (1) The previously searched buffer contained a partial match of the search
265 * pattern and we want to see if the next one starts with the remainder of
268 * (2) The search pattern is wholly contained within the current buffer.
270 * (3) The current buffer ends with a partial match of the pattern. We will
271 * note this for use in the next iteration, where we will check for the rest
274 if ((nread = read(fd, &buffer, sizeof(buffer))) <= 0)
277 /* (1) previous partial match */
278 if (previous_partial_synch_match && _is_id3v2_second_synch_byte(buffer[0]))
281 if (previous_partial_match >= 0 && previous_partial_match < buffer_size) {
282 const int pattern_offset = buffer_size - previous_partial_match;
284 if (memcmp(buffer, pattern + pattern_offset, 3 - pattern_offset) == 0)
285 return buffer_offset - buffer_size + previous_partial_match;
288 /* (2) pattern contained in current buffer */
290 while ((p = memchr(p, 'I', buffer_size))) {
291 if (memcmp(p, pattern, 3) == 0)
292 return buffer_offset + (p - buffer);
296 p = memchr(buffer, 255, buffer_size);
298 first_synch_byte = p - buffer;
300 first_synch_byte = -1;
302 /* Here we have to loop because there could be several of the first
303 * (11111111) byte, and we want to check all such instances until we find
304 * a full match (11111111 111) or hit the end of the buffer. */
305 while (first_synch_byte >= 0) {
306 /* if this *is not* at the end of the buffer */
307 if (first_synch_byte < buffer_size - 1) {
308 if(_is_id3v2_second_synch_byte(buffer[first_synch_byte + 1]))
309 /* We've found the frame synch pattern. */
312 /* We found 11111111 at the end of the current buffer indicating a
313 * partial match of the synch pattern. The find() below should
314 * return -1 and break out of the loop. */
315 previous_partial_synch_match = 1;
318 /* Check in the rest of the buffer. */
319 p = memchr(p + 1, 255, buffer_size);
321 first_synch_byte = p - buffer;
323 first_synch_byte = -1;
326 /* (3) partial match */
327 if (buffer[nread - 1] == pattern[1])
328 previous_partial_match = nread - 1;
329 else if (memcmp(&buffer[nread - 2], pattern, 2) == 0)
330 previous_partial_match = nread - 2;
331 buffer_offset += buffer_size;
333 if ((nread = read(fd, &buffer, sizeof(buffer))) <= 0)
341 _to_uint(const char *data, int data_size)
343 unsigned int sum = 0;
344 int last = data_size > 4 ? 3 : data_size - 1;
347 for (i = 0; i <= last; i++)
348 sum |= (data[i] & 0x7f) << ((last - i) * 7);
354 _get_id3v2_frame_header_size(unsigned int version)
369 _parse_id3v2_frame_header(char *data, unsigned int version, struct id3v2_frame_header *fh)
375 memcpy(fh->frame_id, data, 3);
377 fh->frame_size = _to_uint(data + 3, 3);
379 fh->data_length_indicator = 0;
382 memcpy(fh->frame_id, data, 4);
383 fh->frame_size = _to_uint(data + 4, 4);
384 fh->compression = data[9] & 0x40;
385 fh->data_length_indicator = 0;
389 memcpy(fh->frame_id, data, 4);
390 fh->frame_size = _to_uint(data + 4, 4);
391 fh->compression = data[9] & 0x4;
392 fh->data_length_indicator = data[9] & 0x1;
398 _parse_id3v2_frame(struct id3v2_frame_header *fh, const char *frame_data, struct lms_audio_info *info)
400 unsigned int frame_size;
402 /* TODO proper handle text encoding
411 /* skip first byte - text encoding */
413 frame_size = fh->frame_size - 1;
415 if (memcmp(fh->frame_id, "TIT2", 4) == 0)
416 ID3V2_GET_FRAME_INFO(frame_data, frame_size, info->title.str, info->title.len)
417 else if (memcmp(fh->frame_id, "TPE1", 4) == 0)
418 ID3V2_GET_FRAME_INFO(frame_data, frame_size, info->artist.str, info->artist.len)
419 else if (memcmp(fh->frame_id, "TALB", 4) == 0)
420 ID3V2_GET_FRAME_INFO(frame_data, frame_size, info->album.str, info->album.len)
421 else if (memcmp(fh->frame_id, "TCON", 4) == 0) {
422 /* TODO handle multiple genres */
425 ID3V2_GET_FRAME_INFO(frame_data, frame_size, info->genre.str, info->genre.len)
428 for (i = 0; i < info->genre.len; ++i) {
429 if (!isdigit(info->genre.str[i]))
435 (info->genre.len > 0)) {
436 int genre = atoi(info->genre.str);
437 if (genre >= 0 && genre < ID3V1_NUM_GENRES) {
438 free(info->genre.str);
439 info->genre.str = strdup(id3v1_genres[(int) genre]);
440 info->genre.len = strlen(info->genre.str);
443 /* ignore other genres */
444 free(info->genre.str);
445 info->genre.str = NULL;
450 else if (memcmp(fh->frame_id, "TRCK", 4) == 0) {
452 unsigned int trackno_len;
453 ID3V2_GET_FRAME_INFO(frame_data, frame_size, trackno, trackno_len);
454 info->trackno = atoi(trackno);
459 _parse_id3v2(int fd, long id3v2_offset, struct lms_audio_info *info)
461 char header_data[10], frame_header_data[10];
462 unsigned int tag_size, major_version, frame_data_pos, frame_data_length, frame_header_size;
463 int extended_header, footer_present;
464 struct id3v2_frame_header fh;
466 lseek(fd, id3v2_offset, SEEK_SET);
469 if (read(fd, header_data, ID3V2_HEADER_SIZE) != ID3V2_HEADER_SIZE)
472 tag_size = _to_uint(header_data + 6, 4);
477 major_version = header_data[3];
480 frame_data_length = tag_size;
482 /* check for extended header */
483 extended_header = header_data[5] & 0x20; /* bit 6 */
484 if (extended_header) {
485 /* skip extended header */
486 unsigned int extended_header_size;
487 char extended_header_data[4];
489 if (read(fd, extended_header_data, 4) != 4)
491 extended_header_size = _to_uint(extended_header_data, 4);
492 lseek(fd, extended_header_size - 4, SEEK_CUR);
493 frame_data_pos += extended_header_size;
494 frame_data_length -= extended_header_size;
497 footer_present = header_data[5] & 0x8; /* bit 4 */
498 if (footer_present && frame_data_length > ID3V2_FOOTER_SIZE)
499 frame_data_length -= ID3V2_FOOTER_SIZE;
501 frame_header_size = _get_id3v2_frame_header_size(major_version);
502 while (frame_data_pos < frame_data_length - frame_header_size) {
503 if (read(fd, frame_header_data, ID3V2_FRAME_HEADER_SIZE) != ID3V2_FRAME_HEADER_SIZE)
506 if (frame_header_data[0] == 0)
509 _parse_id3v2_frame_header(frame_header_data, major_version, &fh);
511 if (!fh.compression &&
512 fh.frame_id[0] == 'T' &&
513 memcmp(fh.frame_id, "TXXX", 4) != 0) {
516 if (fh.data_length_indicator)
517 lseek(fd, 4, SEEK_CUR);
519 frame_data = malloc(sizeof(char) * fh.frame_size);
520 if (read(fd, frame_data, fh.frame_size) != fh.frame_size) {
525 _parse_id3v2_frame(&fh, frame_data, info);
529 if (fh.data_length_indicator)
530 lseek(fd, fh.frame_size + 4, SEEK_CUR);
532 lseek(fd, fh.frame_size, SEEK_CUR);
535 frame_data_pos += fh.frame_size + frame_header_size;
542 _parse_id3v1(int fd, struct lms_audio_info *info)
544 struct id3v1_tag tag;
545 if (read(fd, &tag, sizeof(struct id3v1_tag)) == -1)
548 info->title.str = strdup(tag.title);
549 info->title.len = strlen(info->title.str);
550 info->artist.str = strdup(tag.artist);
551 info->artist.len = strlen(info->artist.str);
552 info->album.str = strdup(tag.album);
553 info->album.len = strlen(info->album.str);
554 info->genre.str = strdup(id3v1_genres[(int) tag.genre]);
555 info->genre.len = strlen(info->genre.str);
556 if (tag.comments[28] == 0 && tag.comments[29] != 0)
557 info->trackno = (unsigned char) tag.comments[29];
563 _match(struct plugin *p, const char *path, int len, int base)
567 i = lms_which_extension(path, len, _exts, LMS_ARRAY_SIZE(_exts));
571 return (void*)(i + 1);
575 _parse(struct plugin *plugin, struct lms_context *ctxt, const struct lms_file_info *finfo, void *match)
577 struct lms_audio_info info = {0, {0}, {0}, {0}, {0}, 0, 0, 0};
581 fd = open(finfo->path, O_RDONLY);
587 id3v2_offset = _find_id3v2(fd);
588 if (id3v2_offset >= 0) {
590 fprintf(stderr, "id3v2 tag found in file %s with offset %ld\n",
591 finfo->path, id3v2_offset);
593 if (_parse_id3v2(fd, id3v2_offset, &info) != 0) {
601 fprintf(stderr, "id3v2 tag not found in file %s. trying id3v1\n", finfo->path);
603 /* check for id3v1 tag */
604 if (lseek(fd, -128, SEEK_END) == -1) {
609 if (read(fd, &tag, 3) == -1) {
614 if (memcmp(tag, "TAG", 3) == 0) {
616 fprintf(stderr, "id3v1 tag found in file %s\n", finfo->path);
618 if (_parse_id3v1(fd, &info) != 0) {
625 lms_string_size_strip_and_free(&info.title);
626 lms_string_size_strip_and_free(&info.artist);
627 lms_string_size_strip_and_free(&info.album);
628 lms_string_size_strip_and_free(&info.genre);
630 if (!info.title.str) {
632 ext_idx = ((int)match) - 1;
633 info.title.len = finfo->path_len - finfo->base - _exts[ext_idx].len;
634 info.title.str = malloc((info.title.len + 1) * sizeof(char));
635 memcpy(info.title.str, finfo->path + finfo->base, info.title.len);
636 info.title.str[info.title.len] = '\0';
638 lms_charset_conv(ctxt->cs_conv, &info.title.str, &info.title.len);
641 lms_charset_conv(ctxt->cs_conv, &info.artist.str, &info.artist.len);
643 lms_charset_conv(ctxt->cs_conv, &info.album.str, &info.album.len);
645 lms_charset_conv(ctxt->cs_conv, &info.genre.str, &info.genre.len);
648 fprintf(stderr, "file %s info\n", finfo->path);
649 fprintf(stderr, "\ttitle='%s'\n", info.title.str);
650 fprintf(stderr, "\tartist='%s'\n", info.artist.str);
651 fprintf(stderr, "\talbum='%s'\n", info.album.str);
652 fprintf(stderr, "\tgenre='%s'\n", info.genre.str);
653 fprintf(stderr, "\ttrack number='%d'\n", info.trackno);
657 r = lms_db_audio_add(plugin->audio_db, &info);
663 free(info.title.str);
665 free(info.artist.str);
667 free(info.album.str);
669 free(info.genre.str);
675 _setup(struct plugin *plugin, struct lms_context *ctxt)
677 plugin->audio_db = lms_db_audio_new(ctxt->db);
678 if (!plugin->audio_db)
684 _start(struct plugin *plugin, struct lms_context *ctxt)
686 return lms_db_audio_start(plugin->audio_db);
690 _finish(struct plugin *plugin, struct lms_context *ctxt)
692 if (plugin->audio_db)
693 lms_db_audio_free(plugin->audio_db);
698 _close(struct plugin *plugin)
704 API struct lms_plugin *
705 lms_plugin_open(void)
707 struct plugin *plugin;
709 plugin = (struct plugin *)malloc(sizeof(*plugin));
710 plugin->plugin.name = _name;
711 plugin->plugin.match = (lms_plugin_match_fn_t)_match;
712 plugin->plugin.parse = (lms_plugin_parse_fn_t)_parse;
713 plugin->plugin.close = (lms_plugin_close_fn_t)_close;
714 plugin->plugin.setup = (lms_plugin_setup_fn_t)_setup;
715 plugin->plugin.start = (lms_plugin_start_fn_t)_start;
716 plugin->plugin.finish = (lms_plugin_finish_fn_t)_finish;
718 return (struct lms_plugin *)plugin;