1 /*****************************************************************************
3 *****************************************************************************
4 * Copyright (C) 2003-2004 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 extentions 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, '.');
108 strncpy(d, s, tmp - s);
118 static void strcpy_get_ext( char *d, char *s )
120 char *tmp = strrchr(s, '.');
125 } else strcpy( d, tmp + 1 );
128 static int whiteonly( char *s )
132 if( isalnum( *s ) ) return 0;
138 typedef struct _subfn
145 static int compare_sub_priority( const void *a, const void *b )
147 if (((subfn*)a)->priority > ((subfn*)b)->priority)
152 if (((subfn*)a)->priority < ((subfn*)b)->priority)
158 return strcoll(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
160 return strcmp(((subfn*)a)->psz_fname, ((subfn*)b)->psz_fname);
164 /* Utility function for scandir */
165 static int Filter( const struct dirent *p_dir_content )
170 if( p_dir_content == NULL || p_dir_content->d_name == NULL ) return VLC_FALSE;
171 /* does it end with a subtitle extension? */
172 tmp = strrchr( p_dir_content->d_name, '.');
179 for( i = 0; sub_exts[i]; i++ )
181 if( strcmp( sub_exts[i], tmp+1 ) == 0 )
192 * Convert a list of paths separated by ',' to a char**
194 static char **paths_to_list( char *psz_dir, char *psz_path )
196 unsigned int i, k, i_nb_subdirs;
197 char **subdirs; /* list of subdirectories to look in */
199 if( !psz_dir ) return NULL;
200 if( !psz_path ) return NULL;
203 for( k = 0; k < strlen( psz_path ); k++ )
205 if( psz_path[k] == ',' )
211 if( i_nb_subdirs > 0 )
213 char *psz_parser = NULL, *psz_temp = NULL;
215 subdirs = (char**)malloc( sizeof(char*) * ( i_nb_subdirs + 1 ) );
216 memset( subdirs, 0, sizeof(char*) * ( i_nb_subdirs + 1 ) );
218 psz_parser = psz_path;
219 while( psz_parser && *psz_parser )
222 psz_subdir = psz_parser;
223 psz_parser = strchr( psz_subdir, ',' );
228 while( *psz_parser == ' ' )
233 if( strlen( psz_subdir ) > 0 )
235 psz_temp = (char *)malloc( strlen(psz_dir)
236 + strlen(psz_subdir) + 2 );
239 sprintf( psz_temp, "%s%s%c",
240 psz_subdir[0] == '.' ? psz_dir : "",
242 psz_subdir[strlen(psz_subdir) - 1] ==
243 DIRECTORY_SEPARATOR ? '\0' : DIRECTORY_SEPARATOR );
244 subdirs[i] = psz_temp;
260 * Detect subtitle files.
262 * When called this function will split up the psz_name string into a
263 * directory, filename and extension. It then opens the directory
264 * in which the file resides and tries to find possible matches of
268 * \param p_this the calling \ref input_thread_t
269 * \param psz_path a list of subdirectories (separated by a ',') to look in.
270 * \param psz_name the complete filename to base the search on.
271 * \return a NULL terminated array of filenames with detected possible subtitles.
272 * The array contains max MAX_SUBTITLE_FILES items and you need to free it after use.
274 char **subtitles_Detect( input_thread_t *p_this, char *psz_path,
278 int j, i_result2, i_dir_content, i_sub_count = 0, i_fname_len = 0;
279 char *f_dir = NULL, *f_fname = NULL, *f_fname_noext = NULL, *f_fname_trim = NULL;
282 char tmp_fname_noext[PATH_MAX];
283 char tmp_fname_trim[PATH_MAX];
284 char tmp_fname_ext[PATH_MAX];
286 struct dirent **pp_dir_content;
287 char **tmp_subdirs, **subdirs; /* list of subdirectories to look in */
289 subfn *result = NULL; /* unsorted results */
290 char **result2; /* sorted results */
292 char *psz_fname_original = strdup( psz_name );
293 char *psz_fname = psz_fname_original;
295 if( psz_fname == NULL ) return NULL;
297 if( !strncmp( psz_fname, "file://", 7 ) )
302 /* extract filename & dirname from psz_fname */
303 tmp = strrchr( psz_fname, DIRECTORY_SEPARATOR );
308 f_fname = malloc( strlen(tmp) );
310 strcpy( f_fname, tmp+1 ); // we skip the seperator, so it will still fit in the allocated space
311 dirlen = strlen(psz_fname) - strlen(tmp) + 1; // add the seperator
312 f_dir = malloc( dirlen + 1 );
315 strncpy( f_dir, psz_fname, dirlen );
321 /* Get the current working directory */
324 f_dir = getcwd( NULL, 0 );
328 if( psz_fname_original ) free( psz_fname_original );
331 dirlen = strlen( f_dir );
332 f_dir = (char *)realloc(f_dir, dirlen +2 );
333 f_dir[dirlen] = DIRECTORY_SEPARATOR;
334 f_dir[dirlen+1] = '\0';
335 f_fname = strdup( psz_fname );
338 i_fname_len = strlen( f_fname );
339 f_fname_noext = malloc(i_fname_len + 1);
340 f_fname_trim = malloc(i_fname_len + 1 );
342 strcpy_strip_ext( f_fname_noext, f_fname );
343 strcpy_trim( f_fname_trim, f_fname_noext );
345 result = (subfn*)malloc( sizeof(subfn) * MAX_SUBTITLE_FILES );
347 memset( result, 0, sizeof(subfn) * MAX_SUBTITLE_FILES );
349 var_Get( p_this, "sub-autodetect-fuzzy", &fuzzy );
351 tmp_subdirs = paths_to_list( f_dir, psz_path );
352 subdirs = tmp_subdirs;
354 for( j = -1; (j == -1) || ( (j >= 0) && (subdirs != NULL) && (*subdirs != NULL) );
357 char *psz_locale_dir;
359 pp_dir_content = NULL;
362 if( j < 0 && f_dir == NULL )
365 /* parse psz_src dir */
366 psz_locale_dir = ToLocale( j < 0 ? f_dir : *subdirs );
367 i_dir_content = scandir( psz_locale_dir, &pp_dir_content, Filter, NULL );
368 LocaleFree( psz_locale_dir );
370 if( i_dir_content != -1 )
374 msg_Dbg( p_this, "looking for a subtitle file in %s", j < 0 ? f_dir : *subdirs );
375 for( a = 0; a < i_dir_content; a++ )
378 struct dirent *p_dir_content = pp_dir_content[a];
379 char *psz_inUTF8 = FromLocale( p_dir_content->d_name );
380 char *p_fixed_name = vlc_fix_readdir_charset( p_this, psz_inUTF8 );
382 LocaleFree( psz_inUTF8 );
384 /* retrieve various parts of the filename */
385 strcpy_strip_ext( tmp_fname_noext, p_fixed_name );
386 strcpy_get_ext( tmp_fname_ext, p_fixed_name );
387 strcpy_trim( tmp_fname_trim, tmp_fname_noext );
389 if( !i_prio && !strcmp( tmp_fname_trim, f_fname_trim ) )
391 /* matches the movie name exactly */
395 ( tmp = strstr( tmp_fname_trim, f_fname_trim ) ) )
397 /* contains the movie name */
398 tmp += strlen( f_fname_trim );
399 if( whiteonly( tmp ) )
401 /* chars in front of the movie name */
406 /* chars after (and possibly in front of)
413 /* doesn't contain the movie name */
414 if( j == 0 ) i_prio = 1;
416 if( i_prio >= fuzzy.i_int )
420 char *psz_locale_path;
422 asprintf( &psz_UTF8_path, "%s%s", j < 0 ? f_dir : *subdirs, p_fixed_name );
423 psz_locale_path = ToLocale( psz_UTF8_path );
424 msg_Dbg( p_this, "autodetected subtitle: %s with priority %d", p_fixed_name, i_prio );
425 if( ( f = fopen( psz_locale_path, "rt" ) ) )
428 LocaleFree( psz_locale_path );
429 msg_Dbg( p_this, "autodetected subtitle: %s with priority %d", p_fixed_name, i_prio );
430 result[i_sub_count].priority = i_prio;
431 result[i_sub_count].psz_fname = psz_UTF8_path;
432 result[i_sub_count].psz_ext = strdup(tmp_fname_ext);
437 msg_Dbg( p_this, "fopen failed" );
438 if( psz_UTF8_path ) free( psz_UTF8_path );
439 LocaleFree( psz_locale_path );
442 if( i_sub_count >= MAX_SUBTITLE_FILES ) break;
443 if( p_fixed_name ) free( p_fixed_name );
445 for( a = 0; a < i_dir_content; a++ )
446 if( pp_dir_content[a] ) free( pp_dir_content[a] );
447 if( pp_dir_content ) free( pp_dir_content );
449 if( j >= 0 ) if( *subdirs ) free( *subdirs++ );
452 if( tmp_subdirs ) free( tmp_subdirs );
453 if( f_fname_trim ) free( f_fname_trim );
454 if( f_fname_noext ) free( f_fname_noext );
455 if( f_fname ) free( f_fname );
456 if( f_dir ) free( f_dir );
458 qsort( result, i_sub_count, sizeof( subfn ), compare_sub_priority );
460 result2 = (char**)malloc( sizeof(char*) * ( i_sub_count + 1 ) );
462 memset( result2, 0, sizeof(char*) * ( i_sub_count + 1 ) );
465 for( j = 0; j < i_sub_count; j++ )
467 if( result[j].psz_ext && !strcasecmp( result[j].psz_ext, "sub" ) )
470 for( i = 0; i < i_sub_count; i++ )
472 if( result[i].psz_fname && result[j].psz_fname &&
473 !strncasecmp( result[j].psz_fname, result[i].psz_fname, sizeof( result[j].psz_fname) - 4 ) &&
474 !strcasecmp( result[i].psz_ext, "idx" ) )
477 if( i >= i_sub_count )
479 result2[i_result2] = result[j].psz_fname;
485 result2[i_result2] = result[j].psz_fname;
490 if( psz_fname_original ) free( psz_fname_original );
491 if( result ) free( result );