1 /*****************************************************************************
3 *****************************************************************************
4 * Copyright (C) 2003-2006 the VideoLAN team
7 * Authors: Derk-Jan Hartman <hartman at videolan.org>
8 * This is adapted code from the GPL'ed MPlayer (http://mplayerhq.hu)
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston MA 02110-1301, USA.
23 *****************************************************************************/
27 * This file contains functions to dectect subtitle files.
32 #include <vlc/input.h>
50 * What's between a directory and a filename?
53 #define DIRECTORY_SEPARATOR '\\'
55 #define DIRECTORY_SEPARATOR '/'
59 * We are not going to autodetect more subtitle files than this.
61 #define MAX_SUBTITLE_FILES 128
65 * The possible extensions for subtitle files we support
67 static const char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "txt", "ssa", "idx", NULL };
68 /* extensions from unsupported types */
69 /* rt, aqt, jss, js, ass */
71 static void strcpy_trim( char *d, char *s )
73 /* skip leading whitespace */
74 while( *s && !isalnum(*s) )
81 while( *s && isalnum(*s) )
87 /* trim excess whitespace */
88 while( *s && !isalnum(*s) )
98 static void strcpy_strip_ext( char *d, char *s )
100 char *tmp = strrchr(s, '.');
107 strlcpy(d, s, tmp - s);
115 static void strcpy_get_ext( char *d, char *s )
117 char *tmp = strrchr(s, '.');
122 } else strcpy( d, tmp + 1 );
125 static int whiteonly( char *s )
129 if( isalnum( *s ) ) return 0;
135 typedef struct _subfn
142 static int compare_sub_priority( const void *a, const void *b )
144 if (((subfn*)a)->priority > ((subfn*)b)->priority)
149 if (((subfn*)a)->priority < ((subfn*)b)->priority)
155 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
157 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
161 /* Utility function for scandir */
162 static int Filter( const char *psz_dir_content )
164 /* does it end with a subtitle extension? */
165 const char *tmp = strrchr( psz_dir_content, '.');
173 for( i = 0; sub_exts[i]; i++ )
174 if( strcmp( sub_exts[i], tmp ) == 0 )
182 * Convert a list of paths separated by ',' to a char**
184 static char **paths_to_list( char *psz_dir, char *psz_path )
186 unsigned int i, k, i_nb_subdirs;
187 char **subdirs; /* list of subdirectories to look in */
189 if( !psz_dir ) return NULL;
190 if( !psz_path ) return NULL;
193 for( k = 0; k < strlen( psz_path ); k++ )
195 if( psz_path[k] == ',' )
201 if( i_nb_subdirs > 0 )
203 char *psz_parser = NULL, *psz_temp = NULL;
205 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
206 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
208 psz_parser = psz_path;
209 while( psz_parser && *psz_parser )
212 psz_subdir = psz_parser;
213 psz_parser = strchr( psz_subdir, ',' );
218 while( *psz_parser == ' ' )
223 if( strlen( psz_subdir ) > 0 )
225 psz_temp = (char *)malloc( strlen(psz_dir)
226 + strlen(psz_subdir) + 2 );
229 sprintf( psz_temp, "%s%s%c",
230 psz_subdir[0] == '.' ? psz_dir : "",
232 psz_subdir[strlen(psz_subdir) - 1] ==
233 DIRECTORY_SEPARATOR ? '\0' : DIRECTORY_SEPARATOR );
234 subdirs[i] = psz_temp;
250 * Detect subtitle files.
252 * When called this function will split up the psz_name string into a
253 * directory, filename and extension. It then opens the directory
254 * in which the file resides and tries to find possible matches of
258 * \param p_this the calling \ref input_thread_t
259 * \param psz_path a list of subdirectories (separated by a ',') to look in.
260 * \param psz_name the complete filename to base the search on.
261 * \return a NULL terminated array of filenames with detected possible subtitles.
262 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
264 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
268 int j, i_result2, i_sub_count = 0, i_fname_len = 0;
269 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
272 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
274 subfn *result = NULL; /* unsorted results */
275 char **result2; /* sorted results */
277 char *psz_fname_original = strdup( psz_name );
278 char *psz_fname = psz_fname_original;
280 if( psz_fname == NULL ) return NULL;
282 if( !strncmp( psz_fname, "file://", 7 ) )
287 /* extract filename & dirname from psz_fname */
288 tmp = strrchr( psz_fname, DIRECTORY_SEPARATOR );
293 f_fname = malloc( strlen(tmp) );
295 strcpy( f_fname, tmp+1 ); // we skip the separator, so it will still fit in the allocated space
296 dirlen = strlen(psz_fname) - strlen(tmp) + 2; // add the separator
297 f_dir = malloc( dirlen + 1 );
299 strlcpy( f_dir, psz_fname, dirlen );
303 /* Get the current working directory */
306 f_dir = getcwd( NULL, 0 );
310 if( psz_fname_original ) free( psz_fname_original );
313 dirlen = strlen( f_dir );
314 f_dir = (char *)realloc(f_dir, dirlen +2 );
315 f_dir[dirlen] = DIRECTORY_SEPARATOR;
316 f_dir[dirlen+1] = '\0';
317 f_fname = FromLocaleDup( psz_fname );
320 i_fname_len = strlen( f_fname );
321 f_fname_noext = malloc(i_fname_len + 1);
322 f_fname_trim = malloc(i_fname_len + 1 );
324 strcpy_strip_ext( f_fname_noext, f_fname );
325 strcpy_trim( f_fname_trim, f_fname_noext );
327 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
329 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
331 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
333 tmp_subdirs = paths_to_list( f_dir, psz_path );
334 subdirs = tmp_subdirs;
336 for( j = -1; (j == -1) || ( (j >= 0) && (subdirs != NULL) &&
337 (*subdirs != NULL) ); j++)
339 const char *psz_dir = j < 0 ? f_dir : *subdirs;
340 char **ppsz_dir_content;
343 if( psz_dir == NULL )
346 /* parse psz_src dir */
347 i_dir_content = utf8_scandir( psz_dir, &ppsz_dir_content, Filter,
350 if( i_dir_content != -1 )
354 msg_Dbg( p_this, "looking for a subtitle file in %s", psz_dir );
355 for( a = 0; a < i_dir_content; a++ )
357 char *psz_name = vlc_fix_readdir_charset( p_this,
358 ppsz_dir_content[a] );
359 char tmp_fname_noext[strlen( psz_name ) + 1];
360 char tmp_fname_trim[strlen( psz_name ) + 1];
361 char tmp_fname_ext[strlen( psz_name ) + 1];
365 if( psz_name == NULL )
368 /* retrieve various parts of the filename */
369 strcpy_strip_ext( tmp_fname_noext, psz_name );
370 strcpy_get_ext( tmp_fname_ext, psz_name );
371 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
373 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
375 /* matches the movie name exactly */
379 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
381 /* contains the movie name */
382 tmp += strlen( f_fname_trim );
383 if( whiteonly( tmp ) )
385 /* chars in front of the movie name */
390 /* chars after (and possibly in front of)
397 /* doesn't contain the movie name */
398 if( j == 0 ) i_prio = 1;
400 if( i_prio >= fuzzy.i_int )
403 char psz_path[strlen( psz_dir ) + strlen( psz_name ) + 1];
405 sprintf( psz_path, "%s%s", psz_dir, psz_name );
407 "autodetected subtitle: %s with priority %d",
409 /* FIXME: a portable wrapper for stat() or access() would be more suited */
410 if( ( f = utf8_fopen( psz_path, "rt" ) ) )
414 "autodetected subtitle: %s with priority %d",
416 result[i_sub_count].priority = i_prio;
417 result[i_sub_count].psz_fname = strdup( psz_path );
418 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
423 msg_Dbg( p_this, "fopen failed" );
426 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
429 for( a = 0; a < i_dir_content; a++ )
430 free( ppsz_dir_content[a] );
431 if( ppsz_dir_content ) free( ppsz_dir_content );
433 if( j >= 0 ) if( *subdirs ) free( *subdirs++ );
436 if( tmp_subdirs ) free( tmp_subdirs );
437 if( f_fname_trim ) free( f_fname_trim );
438 if( f_fname_noext ) free( f_fname_noext );
439 if( f_fname ) free( f_fname );
440 if( f_dir ) free( f_dir );
442 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
444 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
446 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
449 for( j = 0; j < i_sub_count; j++ )
451 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
454 for( i = 0; i < i_sub_count; i++ )
456 if( result[i].psz_fname && result[j].psz_fname &&
457 !strncasecmp( result[j].psz_fname, result[i].psz_fname,
458 sizeof( result[j].psz_fname) - 4 ) &&
459 !strcasecmp( result[i].psz_ext, "idx" ) )
462 if( i >= i_sub_count )
464 result2[i_result2] = result[j].psz_fname;
470 result2[i_result2] = result[j].psz_fname;
475 if( psz_fname_original ) free( psz_fname_original );
476 if( result ) free( result );