1 /*****************************************************************************
2 * asx.c : ASX playlist format import
3 *****************************************************************************
4 * Copyright (C) 2005-2006 the VideoLAN team
7 * Authors: Derk-Jan Hartman <hartman at videolan dot org>
9 * This program is free software; you can redistribute it and/or modify
10 * it under the terms of the GNU General Public License as published by
11 * the Free Software Foundation; either version 2 of the License, or
12 * (at your option) any later version.
14 * This program is distributed in the hope that it will be useful,
15 * but WITHOUT ANY WARRANTY; without even the implied warranty of
16 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
17 * GNU General Public License for more details.
19 * You should have received a copy of the GNU General Public License
20 * along with this program; if not, write to the Free Software
21 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
22 *****************************************************************************/
24 /* See also: http://msdn.microsoft.com/library/en-us/wmplay10/mmp_sdk/windowsmediametafilereference.asp
27 /*****************************************************************************
29 *****************************************************************************/
34 #include <vlc_common.h>
35 #include <vlc_demux.h>
38 #include <vlc_charset.h>
51 /*****************************************************************************
53 *****************************************************************************/
54 static int Demux( demux_t *p_demux);
55 static int Control( demux_t *p_demux, int i_query, va_list args );
57 static int StoreString( demux_t *p_demux, char **ppsz_string,
58 const char *psz_source_start,
59 const char *psz_source_end )
61 demux_sys_t *p_sys = p_demux->p_sys;
62 unsigned len = psz_source_end - psz_source_start;
66 char *buf = *ppsz_string = malloc ((len * (1 + !p_sys->b_utf8)) + 1);
72 memcpy (buf, psz_source_start, len);
73 (*ppsz_string)[len] = '\0';
74 EnsureUTF8 (*ppsz_string);
78 /* Latin-1 -> UTF-8 */
79 for (unsigned i = 0; i < len; i++)
81 unsigned char c = psz_source_start[i];
84 *buf++ = 0xc0 | (c >> 6);
85 *buf++ = 0x80 | (c & 0x3f);
92 buf = realloc (*ppsz_string, buf - *ppsz_string);
99 static char *SkipBlanks(char *s, size_t i_strlen )
101 while( i_strlen > 0 ) {
118 static int ParseTime(char *s, size_t i_strlen)
120 // need to parse hour:minutes:sec.fraction string
123 const char *end = s + i_strlen;
124 // skip leading spaces if any
125 s = SkipBlanks(s, i_strlen);
128 while( (s < end) && isdigit(*s) )
130 int newval = val*10 + (*s - '0');
141 s = SkipBlanks(s, end-s);
145 s = SkipBlanks(s, end-s);
146 result = result * 60;
148 while( (s < end) && isdigit(*s) )
150 int newval = val*10 + (*s - '0');
161 s = SkipBlanks(s, end-s);
165 s = SkipBlanks(s, end-s);
166 result = result * 60;
168 while( (s < end) && isdigit(*s) )
170 int newval = val*10 + (*s - '0');
181 // TODO: one day, we may need to parse fraction for sub-second resolution
187 /*****************************************************************************
188 * Import_ASX: main import function
189 *****************************************************************************/
190 int Import_ASX( vlc_object_t *p_this )
192 demux_t *p_demux = (demux_t *)p_this;
193 const uint8_t *p_peek;
194 CHECK_PEEK( p_peek, 10 );
196 // skip over possible leading empty lines and empty spaces
197 p_peek = (uint8_t *)SkipBlanks((char *)p_peek, 6);
199 if( POKE( p_peek, "<asx", 4 ) || demux_IsPathExtension( p_demux, ".asx" ) ||
200 demux_IsPathExtension( p_demux, ".wax" ) || demux_IsPathExtension( p_demux, ".wvx" ) ||
201 demux_IsForced( p_demux, "asx-open" ) )
208 STANDARD_DEMUX_INIT_MSG( "found valid ASX playlist" );
209 p_demux->p_sys->psz_prefix = FindPrefix( p_demux );
210 p_demux->p_sys->psz_data = NULL;
211 p_demux->p_sys->i_data_len = -1;
212 p_demux->p_sys->b_utf8 = false;
213 p_demux->p_sys->b_skip_ads = config_GetInt( p_demux, "playlist-skip-ads" );
218 /*****************************************************************************
219 * Deactivate: frees unused data
220 *****************************************************************************/
221 void Close_ASX( vlc_object_t *p_this )
223 demux_t *p_demux = (demux_t *)p_this;
224 demux_sys_t *p_sys = p_demux->p_sys;
226 free( p_sys->psz_prefix );
227 free( p_sys->psz_data );
231 static int Demux( demux_t *p_demux )
233 demux_sys_t *p_sys = p_demux->p_sys;
234 char *psz_parse = NULL;
235 char *psz_backup = NULL;
236 bool b_entry = false;
240 if( p_sys->i_data_len < 0 )
243 p_sys->i_data_len = stream_Size( p_demux->s ) +1; /* This is a cheat to prevent unnecessary realloc */
244 if( p_sys->i_data_len <= 0 && p_sys->i_data_len < 16384 ) p_sys->i_data_len = 1024;
245 p_sys->psz_data = malloc( p_sys->i_data_len +1);
247 /* load the complete file */
250 int i_read = stream_Read( p_demux->s, &p_sys->psz_data[i_pos], p_sys->i_data_len - i_pos );
251 p_sys->psz_data[i_pos + i_read] = '\0';
253 if( i_read < p_sys->i_data_len - i_pos ) break; /* Done */
255 /* XXX this looks fishy and inefficient */
257 p_sys->i_data_len += 1024;
258 p_sys->psz_data = realloc( p_sys->psz_data, p_sys->i_data_len * sizeof( char * ) +1 );
260 if( p_sys->i_data_len <= 0 ) return -1;
263 psz_parse = p_sys->psz_data;
264 /* Find first element */
265 if( ( psz_parse = strcasestr( psz_parse, "<ASX" ) ) )
268 char *psz_string = NULL;
271 char *psz_base_asx = NULL;
272 char *psz_title_asx = NULL;
273 char *psz_artist_asx = NULL;
274 char *psz_copyright_asx = NULL;
275 char *psz_moreinfo_asx = NULL;
276 char *psz_abstract_asx = NULL;
278 char *psz_base_entry = NULL;
279 char *psz_title_entry = NULL;
280 char *psz_artist_entry = NULL;
281 char *psz_copyright_entry = NULL;
282 char *psz_moreinfo_entry = NULL;
283 char *psz_abstract_entry = NULL;
284 int i_entry_count = 0;
285 bool b_skip_entry = false;
287 char *psz_href = NULL;
291 psz_parse = strcasestr( psz_parse, ">" );
293 while( psz_parse && ( psz_parse = strcasestr( psz_parse, "<" ) ) )
295 if( !strncasecmp( psz_parse, "<!--", 4 ) )
297 /* this is a comment */
298 if( ( psz_parse = strcasestr( psz_parse, "-->" ) ) )
302 else if( !strncasecmp( psz_parse, "<PARAM ", 7 ) )
304 bool b_encoding_flag = false;
305 psz_parse = SkipBlanks(psz_parse+7, (unsigned)-1);
306 if( !strncasecmp( psz_parse, "name", 4 ) )
308 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
310 psz_backup = ++psz_parse;
311 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
313 i_strlen = psz_parse-psz_backup;
314 if( i_strlen < 1 ) continue;
315 msg_Dbg( p_demux, "param name strlen: %d", i_strlen);
316 psz_string = malloc( i_strlen + 1);
317 memcpy( psz_string, psz_backup, i_strlen );
318 psz_string[i_strlen] = '\0';
319 msg_Dbg( p_demux, "param name: %s", psz_string);
320 b_encoding_flag = !strcasecmp( psz_string, "encoding" );
328 if( !strncasecmp( psz_parse, "value", 5 ) )
330 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
332 psz_backup = ++psz_parse;
333 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
335 i_strlen = psz_parse-psz_backup;
336 if( i_strlen < 1 ) continue;
337 msg_Dbg( p_demux, "param value strlen: %d", i_strlen);
338 psz_string = malloc( i_strlen +1);
339 memcpy( psz_string, psz_backup, i_strlen );
340 psz_string[i_strlen] = '\0';
341 msg_Dbg( p_demux, "param value: %s", psz_string);
342 if( b_encoding_flag && !strcasecmp( psz_string, "utf-8" ) ) p_sys->b_utf8 = true;
349 if( ( psz_parse = strcasestr( psz_parse, "/>" ) ) )
353 else if( !strncasecmp( psz_parse, "<BANNER", 7 ) )
355 /* We skip this element */
356 if( ( psz_parse = strcasestr( psz_parse, "</BANNER>" ) ) )
360 else if( !strncasecmp( psz_parse, "<PREVIEWDURATION", 16 ) ||
361 !strncasecmp( psz_parse, "<LOGURL", 7 ) ||
362 !strncasecmp( psz_parse, "<Skin", 5 ) )
364 /* We skip this element */
365 if( ( psz_parse = strcasestr( psz_parse, "/>" ) ) )
369 else if( !strncasecmp( psz_parse, "<BASE ", 6 ) )
371 psz_parse = SkipBlanks(psz_parse+6, (unsigned)-1);
372 if( !strncasecmp( psz_parse, "HREF", 4 ) )
374 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
376 psz_backup = ++psz_parse;
377 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
379 StoreString( p_demux, (b_entry ? &psz_base_entry : &psz_base_asx), psz_backup, psz_parse );
385 if( ( psz_parse = strcasestr( psz_parse, "/>" ) ) )
389 else if( !strncasecmp( psz_parse, "<TITLE>", 7 ) )
391 psz_backup = psz_parse+=7;
392 if( ( psz_parse = strcasestr( psz_parse, "</TITLE>" ) ) )
394 StoreString( p_demux, (b_entry ? &psz_title_entry : &psz_title_asx), psz_backup, psz_parse );
399 else if( !strncasecmp( psz_parse, "<Author>", 8 ) )
401 psz_backup = psz_parse+=8;
402 if( ( psz_parse = strcasestr( psz_parse, "</Author>" ) ) )
404 StoreString( p_demux, (b_entry ? &psz_artist_entry : &psz_artist_asx), psz_backup, psz_parse );
409 else if( !strncasecmp( psz_parse, "<Copyright", 10 ) )
411 psz_backup = psz_parse+=11;
412 if( ( psz_parse = strcasestr( psz_parse, "</Copyright>" ) ) )
414 StoreString( p_demux, (b_entry ? &psz_copyright_entry : &psz_copyright_asx), psz_backup, psz_parse );
419 else if( !strncasecmp( psz_parse, "<MoreInfo ", 10 ) )
421 psz_parse = SkipBlanks(psz_parse+10, (unsigned)-1);
422 if( !strncasecmp( psz_parse, "HREF", 4 ) )
424 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
426 psz_backup = ++psz_parse;
427 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
429 StoreString( p_demux, (b_entry ? &psz_moreinfo_entry : &psz_moreinfo_asx), psz_backup, psz_parse );
435 if( ( psz_parse = strcasestr( psz_parse, "/>" ) ) )
439 else if( !strncasecmp( psz_parse, "<ABSTRACT>", 10 ) )
441 psz_backup = psz_parse+=10;
442 if( ( psz_parse = strcasestr( psz_parse, "</ABSTRACT>" ) ) )
444 StoreString( p_demux, (b_entry ? &psz_abstract_entry : &psz_abstract_asx), psz_backup, psz_parse );
449 else if( !strncasecmp( psz_parse, "<EntryRef ", 10 ) )
451 psz_parse = SkipBlanks(psz_parse+10, (unsigned)-1);
452 if( !strncasecmp( psz_parse, "HREF", 4 ) )
454 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
456 psz_backup = ++psz_parse;
457 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
459 i_strlen = psz_parse-psz_backup;
460 if( i_strlen < 1 ) continue;
461 psz_string = malloc( i_strlen +1);
462 memcpy( psz_string, psz_backup, i_strlen );
463 psz_string[i_strlen] = '\0';
464 input_item_t *p_input;
465 p_input = input_item_New( p_demux, psz_string, psz_title_asx );
466 input_item_CopyOptions( p_current_input, p_input );
467 input_item_AddSubItem( p_current_input, p_input );
468 vlc_gc_decref( p_input );
475 if( ( psz_parse = strcasestr( psz_parse, "/>" ) ) )
479 else if( !strncasecmp( psz_parse, "</Entry>", 8 ) )
481 input_item_t *p_entry = NULL;
482 char *psz_name = NULL;
484 char * ppsz_options[2];
487 /* add a new entry */
491 msg_Err( p_demux, "end of entry without start?" );
497 msg_Err( p_demux, "entry without href?" );
501 if( p_sys->b_skip_ads && b_skip_entry )
503 char *psz_current_input_name = input_item_GetName( p_current_input );
505 msg_Dbg( p_demux, "skipped entry %d %s (%s)",
507 ( psz_title_entry ? psz_title_entry : psz_current_input_name ), psz_href );
508 free( psz_current_input_name );
512 if( i_starttime || i_duration )
516 if( asprintf(ppsz_options+i_options, ":start-time=%d", i_starttime) == -1 )
517 *(ppsz_options+i_options) = NULL;
523 if( asprintf(ppsz_options+i_options, ":stop-time=%d", i_starttime + i_duration) == -1 )
524 *(ppsz_options+i_options) = NULL;
530 /* create the new entry */
531 char *psz_current_input_name = input_item_GetName( p_current_input );
532 if( asprintf( &psz_name, "%d %s", i_entry_count, ( psz_title_entry ? psz_title_entry : psz_current_input_name ) ) != -1 )
534 p_entry = input_item_NewExt( p_demux, psz_href, psz_name,
535 i_options, (const char * const *)ppsz_options, VLC_INPUT_OPTION_TRUSTED, -1 );
537 input_item_CopyOptions( p_current_input, p_entry );
540 psz_name = ppsz_options[--i_options];
545 if( psz_title_entry ) input_item_SetTitle( p_entry, psz_title_entry );
546 if( psz_artist_entry ) input_item_SetArtist( p_entry, psz_artist_entry );
547 if( psz_copyright_entry ) input_item_SetCopyright( p_entry, psz_copyright_entry );
548 if( psz_moreinfo_entry ) input_item_SetURL( p_entry, psz_moreinfo_entry );
549 if( psz_abstract_entry ) input_item_SetDescription( p_entry, psz_abstract_entry );
550 input_item_AddSubItem( p_current_input, p_entry );
551 vlc_gc_decref( p_entry );
553 free( psz_current_input_name );
557 FREENULL( psz_href );
558 FREENULL( psz_title_entry );
559 FREENULL( psz_base_entry );
560 FREENULL( psz_artist_entry );
561 FREENULL( psz_copyright_entry );
562 FREENULL( psz_moreinfo_entry );
563 FREENULL( psz_abstract_entry );
566 else if( !strncasecmp( psz_parse, "<Entry", 6 ) )
568 char *psz_clientskip;
572 msg_Err( p_demux, "We already are in an entry section" );
577 psz_clientskip = strcasestr( psz_parse, "clientskip=\"no\"" );
578 psz_parse = strcasestr( psz_parse, ">" );
580 /* If clientskip was enabled ... this is an ad */
581 b_skip_entry = (NULL != psz_clientskip) && (psz_clientskip < psz_parse);
583 // init entry details
588 else if( !strncasecmp( psz_parse, "<Ref ", 5 ) )
590 psz_parse = SkipBlanks(psz_parse+5, (unsigned)-1);
593 msg_Err( p_demux, "A ref outside an entry section" );
597 if( !strncasecmp( psz_parse, "HREF", 4 ) )
599 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
601 psz_backup = ++psz_parse;
602 psz_backup = SkipBlanks(psz_backup, (unsigned)-1);
603 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
606 i_strlen = psz_parse-psz_backup;
607 if( i_strlen < 1 ) continue;
610 psz_href = malloc( i_strlen +1);
611 memcpy( psz_href, psz_backup, i_strlen );
612 psz_href[i_strlen] = '\0';
613 psz_tmp = psz_href + (i_strlen-1);
614 while( psz_tmp >= psz_href &&
615 ( *psz_tmp == '\r' || *psz_tmp == '\n' ) )
625 if( ( psz_parse = strcasestr( psz_parse, ">" ) ) )
629 else if( !strncasecmp( psz_parse, "<starttime ", 11 ) )
631 psz_parse = SkipBlanks(psz_parse+11, (unsigned)-1);
634 msg_Err( p_demux, "starttime outside an entry section" );
638 if( !strncasecmp( psz_parse, "value", 5 ) )
640 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
642 psz_backup = ++psz_parse;
643 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
645 i_strlen = psz_parse-psz_backup;
646 if( i_strlen < 1 ) continue;
648 i_starttime = ParseTime(psz_backup, i_strlen);
654 if( ( psz_parse = strcasestr( psz_parse, ">" ) ) )
658 else if( !strncasecmp( psz_parse, "<duration ", 11 ) )
660 psz_parse = SkipBlanks(psz_parse+5, (unsigned)-1);
663 msg_Err( p_demux, "duration outside an entry section" );
667 if( !strncasecmp( psz_parse, "value", 5 ) )
669 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
671 psz_backup = ++psz_parse;
672 if( ( psz_parse = strcasestr( psz_parse, "\"" ) ) )
674 i_strlen = psz_parse-psz_backup;
675 if( i_strlen < 1 ) continue;
677 i_duration = ParseTime(psz_backup, i_strlen);
683 if( ( psz_parse = strcasestr( psz_parse, ">" ) ) )
687 else if( !strncasecmp( psz_parse, "</ASX", 5 ) )
689 if( psz_title_asx ) input_item_SetTitle( p_current_input, psz_title_asx );
690 if( psz_artist_asx ) input_item_SetArtist( p_current_input, psz_artist_asx );
691 if( psz_copyright_asx ) input_item_SetCopyright( p_current_input, psz_copyright_asx );
692 if( psz_moreinfo_asx ) input_item_SetURL( p_current_input, psz_moreinfo_asx );
693 if( psz_abstract_asx ) input_item_SetDescription( p_current_input, psz_abstract_asx );
694 FREENULL( psz_base_asx );
695 FREENULL( psz_title_asx );
696 FREENULL( psz_artist_asx );
697 FREENULL( psz_copyright_asx );
698 FREENULL( psz_moreinfo_asx );
699 FREENULL( psz_abstract_asx );
705 /* FIXME Unsupported elements */
713 HANDLE_PLAY_AND_RELEASE;
714 return 0; /* Needed for correct operation of go back */
717 static int Control( demux_t *p_demux, int i_query, va_list args )
719 VLC_UNUSED(p_demux); VLC_UNUSED(i_query); VLC_UNUSED(args);