font_load_ft.c: disable -fontconfig if compiled without libfontconfig
[mplayer.git] / subreader.c
blob3d422ddbdff71320209901072a6367fcc2e32010
1 /*
2 * Subtitle reader with format autodetection
4 * Copyright (c) 2001 laaz
5 * Some code cleanup & realloc() by A'rpi/ESP-team
7 * This file is part of MPlayer.
9 * MPlayer 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 * MPlayer 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 along
20 * with MPlayer; if not, write to the Free Software Foundation, Inc.,
21 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
24 #include <stdio.h>
25 #include <stdlib.h>
26 #include <string.h>
27 #include <ctype.h>
29 #include <sys/types.h>
30 #include <dirent.h>
32 #include "config.h"
33 #include "mp_msg.h"
34 #include "subreader.h"
35 #include "stream/stream.h"
36 #include "libavutil/common.h"
37 #include "libavutil/avstring.h"
39 #ifdef CONFIG_ENCA
40 #include <enca.h>
41 #endif
43 #define ERR ((void *) -1)
45 #ifdef CONFIG_ICONV
46 #include <iconv.h>
47 char *sub_cp=NULL;
48 #endif
49 #ifdef CONFIG_FRIBIDI
50 #include <fribidi/fribidi.h>
51 char *fribidi_charset = NULL; ///character set that will be passed to FriBiDi
52 int flip_hebrew = 1; ///flip subtitles using fribidi
53 int fribidi_flip_commas = 0; ///flip comma when fribidi is used
54 #endif
56 /* Maximal length of line of a subtitle */
57 #define LINE_LEN 1000
58 static float mpsub_position=0;
59 static float mpsub_multiplier=1.;
60 static int sub_slacktime = 20000; //20 sec
62 int sub_no_text_pp=0; // 1 => do not apply text post-processing
63 // like {\...} elimination in SSA format.
65 int sub_match_fuzziness=0; // level of sub name matching fuzziness
67 /* Use the SUB_* constant defined in the header file */
68 int sub_format=SUB_INVALID;
69 #ifdef CONFIG_SORTSUB
71 Some subtitling formats, namely AQT and Subrip09, define the end of a
72 subtitle as the beginning of the following. Since currently we read one
73 subtitle at time, for these format we keep two global *subtitle,
74 previous_aqt_sub and previous_subrip09_sub, pointing to previous subtitle,
75 so we can change its end when we read current subtitle starting time.
76 When CONFIG_SORTSUB is defined, we use a single global unsigned long,
77 previous_sub_end, for both (and even future) formats, to store the end of
78 the previous sub: it is initialized to 0 in sub_read_file and eventually
79 modified by sub_read_aqt_line or sub_read_subrip09_line.
81 unsigned long previous_sub_end;
82 #endif
84 static int eol(char p) {
85 return p=='\r' || p=='\n' || p=='\0';
88 /* Remove leading and trailing space */
89 static void trail_space(char *s) {
90 int i = 0;
91 while (isspace(s[i])) ++i;
92 if (i) strcpy(s, s + i);
93 i = strlen(s) - 1;
94 while (i > 0 && isspace(s[i])) s[i--] = '\0';
97 static char *stristr(const char *haystack, const char *needle) {
98 int len = 0;
99 const char *p = haystack;
101 if (!(haystack && needle)) return NULL;
103 len=strlen(needle);
104 while (*p != '\0') {
105 if (strncasecmp(p, needle, len) == 0) return (char*)p;
106 p++;
109 return NULL;
112 static void sami_add_line(subtitle *current, char *buffer, char **pos) {
113 char *p = *pos;
114 *p = 0;
115 trail_space(buffer);
116 if (*buffer && current->lines < SUB_MAX_TEXT)
117 current->text[current->lines++] = strdup(buffer);
118 *pos = buffer;
121 static subtitle *sub_read_line_sami(stream_t* st, subtitle *current, int utf16) {
122 static char line[LINE_LEN+1];
123 static char *s = NULL, *slacktime_s;
124 char text[LINE_LEN+1], *p=NULL, *q;
125 int state;
127 current->lines = current->start = current->end = 0;
128 current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
129 state = 0;
131 /* read the first line */
132 if (!s)
133 if (!(s = stream_read_line(st, line, LINE_LEN, utf16))) return 0;
135 do {
136 switch (state) {
138 case 0: /* find "START=" or "Slacktime:" */
139 slacktime_s = stristr (s, "Slacktime:");
140 if (slacktime_s)
141 sub_slacktime = strtol (slacktime_s+10, NULL, 0) / 10;
143 s = stristr (s, "Start=");
144 if (s) {
145 current->start = strtol (s + 6, &s, 0) / 10;
146 /* eat '>' */
147 for (; *s != '>' && *s != '\0'; s++);
148 s++;
149 state = 1; continue;
151 break;
153 case 1: /* find (optional) "<P", skip other TAGs */
154 for (; *s == ' ' || *s == '\t'; s++); /* strip blanks, if any */
155 if (*s == '\0') break;
156 if (*s != '<') { state = 3; p = text; continue; } /* not a TAG */
157 s++;
158 if (*s == 'P' || *s == 'p') { s++; state = 2; continue; } /* found '<P' */
159 for (; *s != '>' && *s != '\0'; s++); /* skip remains of non-<P> TAG */
160 if (s == '\0')
161 break;
162 s++;
163 continue;
165 case 2: /* find ">" */
166 if ((s = strchr (s, '>'))) { s++; state = 3; p = text; continue; }
167 break;
169 case 3: /* get all text until '<' appears */
170 if (*s == '\0') break;
171 else if (!strncasecmp (s, "<br>", 4)) {
172 sami_add_line(current, text, &p);
173 s += 4;
175 else if ((*s == '{') && !sub_no_text_pp) { state = 5; ++s; continue; }
176 else if (*s == '<') { state = 4; }
177 else if (!strncasecmp (s, "&nbsp;", 6)) { *p++ = ' '; s += 6; }
178 else if (*s == '\t') { *p++ = ' '; s++; }
179 else if (*s == '\r' || *s == '\n') { s++; }
180 else *p++ = *s++;
182 /* skip duplicated space */
183 if (p > text + 2) if (*(p-1) == ' ' && *(p-2) == ' ') p--;
185 continue;
187 case 4: /* get current->end or skip <TAG> */
188 q = stristr (s, "Start=");
189 if (q) {
190 current->end = strtol (q + 6, &q, 0) / 10 - 1;
191 *p = '\0'; trail_space (text);
192 if (text[0] != '\0')
193 current->text[current->lines++] = strdup (text);
194 if (current->lines > 0) { state = 99; break; }
195 state = 0; continue;
197 s = strchr (s, '>');
198 if (s) { s++; state = 3; continue; }
199 break;
200 case 5: /* get rid of {...} text, but read the alignment code */
201 if ((*s == '\\') && (*(s + 1) == 'a') && !sub_no_text_pp) {
202 if (stristr(s, "\\a1") != NULL) {
203 current->alignment = SUB_ALIGNMENT_BOTTOMLEFT;
204 s = s + 3;
206 if (stristr(s, "\\a2") != NULL) {
207 current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
208 s = s + 3;
209 } else if (stristr(s, "\\a3") != NULL) {
210 current->alignment = SUB_ALIGNMENT_BOTTOMRIGHT;
211 s = s + 3;
212 } else if ((stristr(s, "\\a4") != NULL) || (stristr(s, "\\a5") != NULL) || (stristr(s, "\\a8") != NULL)) {
213 current->alignment = SUB_ALIGNMENT_TOPLEFT;
214 s = s + 3;
215 } else if (stristr(s, "\\a6") != NULL) {
216 current->alignment = SUB_ALIGNMENT_TOPCENTER;
217 s = s + 3;
218 } else if (stristr(s, "\\a7") != NULL) {
219 current->alignment = SUB_ALIGNMENT_TOPRIGHT;
220 s = s + 3;
221 } else if (stristr(s, "\\a9") != NULL) {
222 current->alignment = SUB_ALIGNMENT_MIDDLELEFT;
223 s = s + 3;
224 } else if (stristr(s, "\\a10") != NULL) {
225 current->alignment = SUB_ALIGNMENT_MIDDLECENTER;
226 s = s + 4;
227 } else if (stristr(s, "\\a11") != NULL) {
228 current->alignment = SUB_ALIGNMENT_MIDDLERIGHT;
229 s = s + 4;
232 if (*s == '}') state = 3;
233 ++s;
234 continue;
237 /* read next line */
238 if (state != 99 && !(s = stream_read_line (st, line, LINE_LEN, utf16))) {
239 if (current->start > 0) {
240 break; // if it is the last subtitle
241 } else {
242 return 0;
246 } while (state != 99);
248 // For the last subtitle
249 if (current->end <= 0) {
250 current->end = current->start + sub_slacktime;
251 sami_add_line(current, text, &p);
254 return current;
258 static char *sub_readtext(char *source, char **dest) {
259 int len=0;
260 char *p=source;
262 // printf("src=%p dest=%p \n",source,dest);
264 while ( !eol(*p) && *p!= '|' ) {
265 p++,len++;
268 *dest= malloc (len+1);
269 if (!dest) {return ERR;}
271 strncpy(*dest, source, len);
272 (*dest)[len]=0;
274 while (*p=='\r' || *p=='\n' || *p=='|') p++;
276 if (*p) return p; // not-last text field
277 else return NULL; // last text field
280 static subtitle *sub_read_line_microdvd(stream_t *st,subtitle *current, int utf16) {
281 char line[LINE_LEN+1];
282 char line2[LINE_LEN+1];
283 char *p, *next;
284 int i;
286 do {
287 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
288 } while ((sscanf (line,
289 "{%ld}{}%[^\r\n]",
290 &(current->start), line2) < 2) &&
291 (sscanf (line,
292 "{%ld}{%ld}%[^\r\n]",
293 &(current->start), &(current->end), line2) < 3));
295 p=line2;
297 next=p, i=0;
298 while ((next =sub_readtext (next, &(current->text[i])))) {
299 if (current->text[i]==ERR) {return ERR;}
300 i++;
301 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
303 current->lines= ++i;
305 return current;
308 static subtitle *sub_read_line_mpl2(stream_t *st,subtitle *current, int utf16) {
309 char line[LINE_LEN+1];
310 char line2[LINE_LEN+1];
311 char *p, *next;
312 int i;
314 do {
315 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
316 } while ((sscanf (line,
317 "[%ld][%ld]%[^\r\n]",
318 &(current->start), &(current->end), line2) < 3));
319 current->start *= 10;
320 current->end *= 10;
321 p=line2;
323 next=p, i=0;
324 while ((next =sub_readtext (next, &(current->text[i])))) {
325 if (current->text[i]==ERR) {return ERR;}
326 i++;
327 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
329 current->lines= ++i;
331 return current;
334 static subtitle *sub_read_line_subrip(stream_t* st, subtitle *current, int utf16) {
335 char line[LINE_LEN+1];
336 int a1,a2,a3,a4,b1,b2,b3,b4;
337 char *p=NULL, *q=NULL;
338 int len;
340 while (1) {
341 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
342 if (sscanf (line, "%d:%d:%d.%d,%d:%d:%d.%d",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4) < 8) continue;
343 current->start = a1*360000+a2*6000+a3*100+a4;
344 current->end = b1*360000+b2*6000+b3*100+b4;
346 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
348 p=q=line;
349 for (current->lines=1; current->lines < SUB_MAX_TEXT; current->lines++) {
350 for (q=p,len=0; *p && *p!='\r' && *p!='\n' && *p!='|' && strncmp(p,"[br]",4); p++,len++);
351 current->text[current->lines-1]=malloc (len+1);
352 if (!current->text[current->lines-1]) return ERR;
353 strncpy (current->text[current->lines-1], q, len);
354 current->text[current->lines-1][len]='\0';
355 if (!*p || *p=='\r' || *p=='\n') break;
356 if (*p=='|') p++;
357 else while (*p++!=']');
359 break;
361 return current;
364 static subtitle *sub_read_line_subviewer(stream_t *st,subtitle *current, int utf16) {
365 char line[LINE_LEN+1];
366 int a1,a2,a3,a4,b1,b2,b3,b4;
367 char *p=NULL;
368 int i,len;
370 while (!current->text[0]) {
371 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
372 if ((len=sscanf (line, "%d:%d:%d%[,.:]%d --> %d:%d:%d%[,.:]%d",&a1,&a2,&a3,(char *)&i,&a4,&b1,&b2,&b3,(char *)&i,&b4)) < 10)
373 continue;
374 current->start = a1*360000+a2*6000+a3*100+a4/10;
375 current->end = b1*360000+b2*6000+b3*100+b4/10;
376 for (i=0; i<SUB_MAX_TEXT;) {
377 int blank = 1;
378 if (!stream_read_line (st, line, LINE_LEN, utf16)) break;
379 len=0;
380 for (p=line; *p!='\n' && *p!='\r' && *p; p++,len++)
381 if (*p != ' ' && *p != '\t')
382 blank = 0;
383 if (len && !blank) {
384 int j=0,skip=0;
385 char *curptr=current->text[i]=malloc (len+1);
386 if (!current->text[i]) return ERR;
387 //strncpy (current->text[i], line, len); current->text[i][len]='\0';
388 for(; j<len; j++) {
389 /* let's filter html tags ::atmos */
390 if(line[j]=='>') {
391 skip=0;
392 continue;
394 if(line[j]=='<') {
395 skip=1;
396 continue;
398 if(skip) {
399 continue;
401 *curptr=line[j];
402 curptr++;
404 *curptr='\0';
406 i++;
407 } else {
408 break;
411 current->lines=i;
413 return current;
416 static subtitle *sub_read_line_subviewer2(stream_t *st,subtitle *current, int utf16) {
417 char line[LINE_LEN+1];
418 int a1,a2,a3,a4;
419 char *p=NULL;
420 int i,len;
422 while (!current->text[0]) {
423 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
424 if (line[0]!='{')
425 continue;
426 if ((len=sscanf (line, "{T %d:%d:%d:%d",&a1,&a2,&a3,&a4)) < 4)
427 continue;
428 current->start = a1*360000+a2*6000+a3*100+a4/10;
429 for (i=0; i<SUB_MAX_TEXT;) {
430 if (!stream_read_line (st, line, LINE_LEN, utf16)) break;
431 if (line[0]=='}') break;
432 len=0;
433 for (p=line; *p!='\n' && *p!='\r' && *p; ++p,++len);
434 if (len) {
435 current->text[i]=malloc (len+1);
436 if (!current->text[i]) return ERR;
437 strncpy (current->text[i], line, len); current->text[i][len]='\0';
438 ++i;
439 } else {
440 break;
443 current->lines=i;
445 return current;
449 static subtitle *sub_read_line_vplayer(stream_t *st,subtitle *current, int utf16) {
450 char line[LINE_LEN+1];
451 int a1,a2,a3;
452 char *p=NULL, *next,separator;
453 int i,len,plen;
455 while (!current->text[0]) {
456 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
457 if ((len=sscanf (line, "%d:%d:%d%c%n",&a1,&a2,&a3,&separator,&plen)) < 4)
458 continue;
460 if (!(current->start = a1*360000+a2*6000+a3*100))
461 continue;
462 /* removed by wodzu
463 p=line;
464 // finds the body of the subtitle
465 for (i=0; i<3; i++){
466 p=strchr(p,':');
467 if (p==NULL) break;
468 ++p;
470 if (p==NULL) {
471 printf("SUB: Skipping incorrect subtitle line!\n");
472 continue;
475 // by wodzu: hey! this time we know what length it has! what is
476 // that magic for? it can't deal with space instead of third
477 // colon! look, what simple it can be:
478 p = &line[ plen ];
480 i=0;
481 if (*p!='|') {
483 next = p,i=0;
484 while ((next =sub_readtext (next, &(current->text[i])))) {
485 if (current->text[i]==ERR) {return ERR;}
486 i++;
487 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
489 current->lines=i+1;
492 return current;
495 static subtitle *sub_read_line_rt(stream_t *st,subtitle *current, int utf16) {
496 //TODO: This format uses quite rich (sub/super)set of xhtml
497 // I couldn't check it since DTD is not included.
498 // WARNING: full XML parses can be required for proper parsing
499 char line[LINE_LEN+1];
500 int a1,a2,a3,a4,b1,b2,b3,b4;
501 char *p=NULL,*next=NULL;
502 int i,len,plen;
504 while (!current->text[0]) {
505 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
506 //TODO: it seems that format of time is not easily determined, it may be 1:12, 1:12.0 or 0:1:12.0
507 //to describe the same moment in time. Maybe there are even more formats in use.
508 //if ((len=sscanf (line, "<Time Begin=\"%d:%d:%d.%d\" End=\"%d:%d:%d.%d\"",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4)) < 8)
509 plen=a1=a2=a3=a4=b1=b2=b3=b4=0;
510 if (
511 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&b3,&b4,&plen)) < 4) &&
512 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&b2,&b3,&b4,&plen)) < 5) &&
513 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&b2,&b3,&plen)) < 4) &&
514 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2,&a3,&b2,&b3,&b4,&plen)) < 5) &&
515 // ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&b2,&b3,&plen)) < 5) &&
516 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&b2,&b3,&b4,&plen)) < 6) &&
517 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d:%d.%d\" %*[Ee]nd=\"%d:%d:%d.%d\"%*[^<]<clear/>%n",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4,&plen)) < 8) &&
518 //now try it without end time
519 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&plen)) < 2) &&
520 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&plen)) < 2) &&
521 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&plen)) < 3) &&
522 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d:%d.%d\"%*[^<]<clear/>%n",&a1,&a2,&a3,&a4,&plen)) < 4)
524 continue;
525 current->start = a1*360000+a2*6000+a3*100+a4/10;
526 current->end = b1*360000+b2*6000+b3*100+b4/10;
527 if (b1 == 0 && b2 == 0 && b3 == 0 && b4 == 0)
528 current->end = current->start+200;
529 p=line; p+=plen;i=0;
530 // TODO: I don't know what kind of convention is here for marking multiline subs, maybe <br/> like in xml?
531 next = strstr(line,"<clear/>");
532 if(next && strlen(next)>8){
533 next+=8;i=0;
534 while ((next =sub_readtext (next, &(current->text[i])))) {
535 if (current->text[i]==ERR) {return ERR;}
536 i++;
537 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
540 current->lines=i+1;
542 return current;
545 static subtitle *sub_read_line_ssa(stream_t *st,subtitle *current, int utf16) {
547 * Sub Station Alpha v4 (and v2?) scripts have 9 commas before subtitle
548 * other Sub Station Alpha scripts have only 8 commas before subtitle
549 * Reading the "ScriptType:" field is not reliable since many scripts appear
550 * w/o it
552 * http://www.scriptclub.org is a good place to find more examples
553 * http://www.eswat.demon.co.uk is where the SSA specs can be found
555 int comma;
556 static int max_comma = 32; /* let's use 32 for the case that the */
557 /* amount of commas increase with newer SSA versions */
559 int hour1, min1, sec1, hunsec1,
560 hour2, min2, sec2, hunsec2, nothing;
561 int num;
563 char line[LINE_LEN+1],
564 line3[LINE_LEN+1],
565 *line2;
566 char *tmp;
568 do {
569 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
570 } while (sscanf (line, "Dialogue: Marked=%d,%d:%d:%d.%d,%d:%d:%d.%d"
571 "%[^\n\r]", &nothing,
572 &hour1, &min1, &sec1, &hunsec1,
573 &hour2, &min2, &sec2, &hunsec2,
574 line3) < 9
576 sscanf (line, "Dialogue: %d,%d:%d:%d.%d,%d:%d:%d.%d"
577 "%[^\n\r]", &nothing,
578 &hour1, &min1, &sec1, &hunsec1,
579 &hour2, &min2, &sec2, &hunsec2,
580 line3) < 9 );
582 line2=strchr(line3, ',');
583 if (!line2) return NULL;
585 for (comma = 4; comma < max_comma; comma ++)
587 tmp = line2;
588 if(!(tmp=strchr(++tmp, ','))) break;
589 if(*(++tmp) == ' ') break;
590 /* a space after a comma means we're already in a sentence */
591 line2 = tmp;
594 if(comma < max_comma)max_comma = comma;
595 /* eliminate the trailing comma */
596 if(*line2 == ',') line2++;
598 current->lines=0;num=0;
599 current->start = 360000*hour1 + 6000*min1 + 100*sec1 + hunsec1;
600 current->end = 360000*hour2 + 6000*min2 + 100*sec2 + hunsec2;
602 while (((tmp=strstr(line2, "\\n")) != NULL) || ((tmp=strstr(line2, "\\N")) != NULL) ){
603 current->text[num]=malloc(tmp-line2+1);
604 strncpy (current->text[num], line2, tmp-line2);
605 current->text[num][tmp-line2]='\0';
606 line2=tmp+2;
607 num++;
608 current->lines++;
609 if (current->lines >= SUB_MAX_TEXT) return current;
612 current->text[num]=strdup(line2);
613 current->lines++;
615 return current;
618 static void sub_pp_ssa(subtitle *sub) {
619 int l=sub->lines;
620 char *so,*de,*start;
622 while (l){
623 /* eliminate any text enclosed with {}, they are font and color settings */
624 so=de=sub->text[--l];
625 while (*so) {
626 if(*so == '{' && so[1]=='\\') {
627 for (start=so; *so && *so!='}'; so++);
628 if(*so) so++; else so=start;
630 if(*so) {
631 *de=*so;
632 so++; de++;
635 *de=*so;
640 * PJS subtitles reader.
641 * That's the "Phoenix Japanimation Society" format.
642 * I found some of them in http://www.scriptsclub.org/ (used for anime).
643 * The time is in tenths of second.
645 * by set, based on code by szabi (dunnowhat sub format ;-)
647 static subtitle *sub_read_line_pjs(stream_t *st,subtitle *current, int utf16) {
648 char line[LINE_LEN+1];
649 char text[LINE_LEN+1], *s, *d;
651 if (!stream_read_line (st, line, LINE_LEN, utf16))
652 return NULL;
653 /* skip spaces */
654 for (s=line; *s && isspace(*s); s++);
655 /* allow empty lines at the end of the file */
656 if (*s==0)
657 return NULL;
658 /* get the time */
659 if (sscanf (s, "%ld,%ld,", &(current->start),
660 &(current->end)) <2) {
661 return ERR;
663 /* the files I have are in tenths of second */
664 current->start *= 10;
665 current->end *= 10;
666 /* walk to the beggining of the string */
667 for (; *s; s++) if (*s==',') break;
668 if (*s) {
669 for (s++; *s; s++) if (*s==',') break;
670 if (*s) s++;
672 if (*s!='"') {
673 return ERR;
675 /* copy the string to the text buffer */
676 for (s++, d=text; *s && *s!='"'; s++, d++)
677 *d=*s;
678 *d=0;
679 current->text[0] = strdup(text);
680 current->lines = 1;
682 return current;
685 static subtitle *sub_read_line_mpsub(stream_t *st, subtitle *current, int utf16) {
686 char line[LINE_LEN+1];
687 float a,b;
688 int num=0;
689 char *p, *q;
693 if (!stream_read_line(st, line, LINE_LEN, utf16)) return NULL;
694 } while (sscanf (line, "%f %f", &a, &b) !=2);
696 mpsub_position += a*mpsub_multiplier;
697 current->start=(int) mpsub_position;
698 mpsub_position += b*mpsub_multiplier;
699 current->end=(int) mpsub_position;
701 while (num < SUB_MAX_TEXT) {
702 if (!stream_read_line (st, line, LINE_LEN, utf16)) {
703 if (num == 0) return NULL;
704 else return current;
706 p=line;
707 while (isspace(*p)) p++;
708 if (eol(*p) && num > 0) return current;
709 if (eol(*p)) return NULL;
711 for (q=p; !eol(*q); q++);
712 *q='\0';
713 if (strlen(p)) {
714 current->text[num]=strdup(p);
715 // printf (">%s<\n",p);
716 current->lines = ++num;
717 } else {
718 if (num) return current;
719 else return NULL;
722 return NULL; // we should have returned before if it's OK
725 #ifndef CONFIG_SORTSUB
726 //we don't need this if we use previous_sub_end
727 subtitle *previous_aqt_sub = NULL;
728 #endif
730 static subtitle *sub_read_line_aqt(stream_t *st,subtitle *current, int utf16) {
731 char line[LINE_LEN+1];
732 char *next;
733 int i;
735 while (1) {
736 // try to locate next subtitle
737 if (!stream_read_line (st, line, LINE_LEN, utf16))
738 return NULL;
739 if (!(sscanf (line, "-->> %ld", &(current->start)) <1))
740 break;
743 #ifdef CONFIG_SORTSUB
744 previous_sub_end = (current->start) ? current->start - 1 : 0;
745 #else
746 if (previous_aqt_sub != NULL)
747 previous_aqt_sub->end = current->start-1;
749 previous_aqt_sub = current;
750 #endif
752 if (!stream_read_line (st, line, LINE_LEN, utf16))
753 return NULL;
755 sub_readtext((char *) &line,&current->text[0]);
756 current->lines = 1;
757 current->end = current->start; // will be corrected by next subtitle
759 if (!stream_read_line (st, line, LINE_LEN, utf16))
760 return current;
762 next = line,i=1;
763 while ((next =sub_readtext (next, &(current->text[i])))) {
764 if (current->text[i]==ERR) {return ERR;}
765 i++;
766 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
768 current->lines=i+1;
770 if (!strlen(current->text[0]) && !strlen(current->text[1])) {
771 #ifdef CONFIG_SORTSUB
772 previous_sub_end = 0;
773 #else
774 // void subtitle -> end of previous marked and exit
775 previous_aqt_sub = NULL;
776 #endif
777 return NULL;
780 return current;
783 #ifndef CONFIG_SORTSUB
784 subtitle *previous_subrip09_sub = NULL;
785 #endif
787 static subtitle *sub_read_line_subrip09(stream_t *st,subtitle *current, int utf16) {
788 char line[LINE_LEN+1];
789 int a1,a2,a3;
790 char * next=NULL;
791 int i,len;
793 while (1) {
794 // try to locate next subtitle
795 if (!stream_read_line (st, line, LINE_LEN, utf16))
796 return NULL;
797 if (!((len=sscanf (line, "[%d:%d:%d]",&a1,&a2,&a3)) < 3))
798 break;
801 current->start = a1*360000+a2*6000+a3*100;
803 #ifdef CONFIG_SORTSUB
804 previous_sub_end = (current->start) ? current->start - 1 : 0;
805 #else
806 if (previous_subrip09_sub != NULL)
807 previous_subrip09_sub->end = current->start-1;
809 previous_subrip09_sub = current;
810 #endif
812 if (!stream_read_line (st, line, LINE_LEN, utf16))
813 return NULL;
815 next = line,i=0;
817 current->text[0]=""; // just to be sure that string is clear
819 while ((next =sub_readtext (next, &(current->text[i])))) {
820 if (current->text[i]==ERR) {return ERR;}
821 i++;
822 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
824 current->lines=i+1;
826 if (!strlen(current->text[0]) && (i==0)) {
827 #ifdef CONFIG_SORTSUB
828 previous_sub_end = 0;
829 #else
830 // void subtitle -> end of previous marked and exit
831 previous_subrip09_sub = NULL;
832 #endif
833 return NULL;
836 return current;
839 static subtitle *sub_read_line_jacosub(stream_t* st, subtitle * current, int utf16)
841 char line1[LINE_LEN], line2[LINE_LEN], directive[LINE_LEN], *p, *q;
842 unsigned a1, a2, a3, a4, b1, b2, b3, b4, comment = 0;
843 static unsigned jacoTimeres = 30;
844 static int jacoShift = 0;
846 memset(current, 0, sizeof(subtitle));
847 memset(line1, 0, LINE_LEN);
848 memset(line2, 0, LINE_LEN);
849 memset(directive, 0, LINE_LEN);
850 while (!current->text[0]) {
851 if (!stream_read_line(st, line1, LINE_LEN, utf16)) {
852 return NULL;
854 if (sscanf
855 (line1, "%u:%u:%u.%u %u:%u:%u.%u %[^\n\r]", &a1, &a2, &a3, &a4,
856 &b1, &b2, &b3, &b4, line2) < 9) {
857 if (sscanf(line1, "@%u @%u %[^\n\r]", &a4, &b4, line2) < 3) {
858 if (line1[0] == '#') {
859 int hours = 0, minutes = 0, seconds, delta, inverter =
861 unsigned units = jacoShift;
862 switch (toupper(line1[1])) {
863 case 'S':
864 if (isalpha(line1[2])) {
865 delta = 6;
866 } else {
867 delta = 2;
869 if (sscanf(&line1[delta], "%d", &hours)) {
870 if (hours < 0) {
871 hours *= -1;
872 inverter = -1;
874 if (sscanf(&line1[delta], "%*d:%d", &minutes)) {
875 if (sscanf
876 (&line1[delta], "%*d:%*d:%d",
877 &seconds)) {
878 sscanf(&line1[delta], "%*d:%*d:%*d.%d",
879 &units);
880 } else {
881 hours = 0;
882 sscanf(&line1[delta], "%d:%d.%d",
883 &minutes, &seconds, &units);
884 minutes *= inverter;
886 } else {
887 hours = minutes = 0;
888 sscanf(&line1[delta], "%d.%d", &seconds,
889 &units);
890 seconds *= inverter;
892 jacoShift =
893 ((hours * 3600 + minutes * 60 +
894 seconds) * jacoTimeres +
895 units) * inverter;
897 break;
898 case 'T':
899 if (isalpha(line1[2])) {
900 delta = 8;
901 } else {
902 delta = 2;
904 sscanf(&line1[delta], "%u", &jacoTimeres);
905 break;
908 continue;
909 } else {
910 current->start =
911 (unsigned long) ((a4 + jacoShift) * 100.0 /
912 jacoTimeres);
913 current->end =
914 (unsigned long) ((b4 + jacoShift) * 100.0 /
915 jacoTimeres);
917 } else {
918 current->start =
919 (unsigned
920 long) (((a1 * 3600 + a2 * 60 + a3) * jacoTimeres + a4 +
921 jacoShift) * 100.0 / jacoTimeres);
922 current->end =
923 (unsigned
924 long) (((b1 * 3600 + b2 * 60 + b3) * jacoTimeres + b4 +
925 jacoShift) * 100.0 / jacoTimeres);
927 current->lines = 0;
928 p = line2;
929 while ((*p == ' ') || (*p == '\t')) {
930 ++p;
932 if (isalpha(*p)||*p == '[') {
933 int cont, jLength;
935 if (sscanf(p, "%s %[^\n\r]", directive, line1) < 2)
936 return (subtitle *) ERR;
937 jLength = strlen(directive);
938 for (cont = 0; cont < jLength; ++cont) {
939 if (isalpha(*(directive + cont)))
940 *(directive + cont) = toupper(*(directive + cont));
942 if ((strstr(directive, "RDB") != NULL)
943 || (strstr(directive, "RDC") != NULL)
944 || (strstr(directive, "RLB") != NULL)
945 || (strstr(directive, "RLG") != NULL)) {
946 continue;
948 if (strstr(directive, "JL") != NULL) {
949 current->alignment = SUB_ALIGNMENT_BOTTOMLEFT;
950 } else if (strstr(directive, "JR") != NULL) {
951 current->alignment = SUB_ALIGNMENT_BOTTOMRIGHT;
952 } else {
953 current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
955 strcpy(line2, line1);
956 p = line2;
958 for (q = line1; (!eol(*p)) && (current->lines < SUB_MAX_TEXT); ++p) {
959 switch (*p) {
960 case '{':
961 comment++;
962 break;
963 case '}':
964 if (comment) {
965 --comment;
966 //the next line to get rid of a blank after the comment
967 if ((*(p + 1)) == ' ')
968 p++;
970 break;
971 case '~':
972 if (!comment) {
973 *q = ' ';
974 ++q;
976 break;
977 case ' ':
978 case '\t':
979 if ((*(p + 1) == ' ') || (*(p + 1) == '\t'))
980 break;
981 if (!comment) {
982 *q = ' ';
983 ++q;
985 break;
986 case '\\':
987 if (*(p + 1) == 'n') {
988 *q = '\0';
989 q = line1;
990 current->text[current->lines++] = strdup(line1);
991 ++p;
992 break;
994 if ((toupper(*(p + 1)) == 'C')
995 || (toupper(*(p + 1)) == 'F')) {
996 ++p,++p;
997 break;
999 if ((*(p + 1) == 'B') || (*(p + 1) == 'b') || (*(p + 1) == 'D') || //actually this means "insert current date here"
1000 (*(p + 1) == 'I') || (*(p + 1) == 'i') || (*(p + 1) == 'N') || (*(p + 1) == 'T') || //actually this means "insert current time here"
1001 (*(p + 1) == 'U') || (*(p + 1) == 'u')) {
1002 ++p;
1003 break;
1005 if ((*(p + 1) == '\\') ||
1006 (*(p + 1) == '~') || (*(p + 1) == '{')) {
1007 ++p;
1008 } else if (eol(*(p + 1))) {
1009 if (!stream_read_line(st, directive, LINE_LEN, utf16))
1010 return NULL;
1011 trail_space(directive);
1012 av_strlcat(line2, directive, LINE_LEN);
1013 break;
1015 default:
1016 if (!comment) {
1017 *q = *p;
1018 ++q;
1020 } //-- switch
1021 } //-- for
1022 *q = '\0';
1023 current->text[current->lines] = strdup(line1);
1024 } //-- while
1025 current->lines++;
1026 return current;
1029 static int sub_autodetect (stream_t* st, int *uses_time, int utf16) {
1030 char line[LINE_LEN+1];
1031 int i,j=0;
1033 while (j < 100) {
1034 j++;
1035 if (!stream_read_line (st, line, LINE_LEN, utf16))
1036 return SUB_INVALID;
1038 if (sscanf (line, "{%d}{%d}", &i, &i)==2)
1039 {*uses_time=0;return SUB_MICRODVD;}
1040 if (sscanf (line, "{%d}{}", &i)==1)
1041 {*uses_time=0;return SUB_MICRODVD;}
1042 if (sscanf (line, "[%d][%d]", &i, &i)==2)
1043 {*uses_time=1;return SUB_MPL2;}
1044 if (sscanf (line, "%d:%d:%d.%d,%d:%d:%d.%d", &i, &i, &i, &i, &i, &i, &i, &i)==8)
1045 {*uses_time=1;return SUB_SUBRIP;}
1046 if (sscanf (line, "%d:%d:%d%[,.:]%d --> %d:%d:%d%[,.:]%d", &i, &i, &i, (char *)&i, &i, &i, &i, &i, (char *)&i, &i)==10)
1047 {*uses_time=1;return SUB_SUBVIEWER;}
1048 if (sscanf (line, "{T %d:%d:%d:%d",&i, &i, &i, &i)==4)
1049 {*uses_time=1;return SUB_SUBVIEWER2;}
1050 if (strstr (line, "<SAMI>"))
1051 {*uses_time=1; return SUB_SAMI;}
1052 if (sscanf(line, "%d:%d:%d.%d %d:%d:%d.%d", &i, &i, &i, &i, &i, &i, &i, &i) == 8)
1053 {*uses_time = 1; return SUB_JACOSUB;}
1054 if (sscanf(line, "@%d @%d", &i, &i) == 2)
1055 {*uses_time = 1; return SUB_JACOSUB;}
1056 if (sscanf (line, "%d:%d:%d:", &i, &i, &i )==3)
1057 {*uses_time=1;return SUB_VPLAYER;}
1058 if (sscanf (line, "%d:%d:%d ", &i, &i, &i )==3)
1059 {*uses_time=1;return SUB_VPLAYER;}
1060 if (!strncasecmp(line, "<window", 7))
1061 {*uses_time=1;return SUB_RT;}
1062 if (!memcmp(line, "Dialogue: Marked", 16))
1063 {*uses_time=1; return SUB_SSA;}
1064 if (!memcmp(line, "Dialogue: ", 10))
1065 {*uses_time=1; return SUB_SSA;}
1066 if (sscanf (line, "%d,%d,\"%c", &i, &i, (char *) &i) == 3)
1067 {*uses_time=1;return SUB_PJS;}
1068 if (sscanf (line, "FORMAT=%d", &i) == 1)
1069 {*uses_time=0; return SUB_MPSUB;}
1070 if (!memcmp(line, "FORMAT=TIME", 11))
1071 {*uses_time=1; return SUB_MPSUB;}
1072 if (strstr (line, "-->>"))
1073 {*uses_time=0; return SUB_AQTITLE;}
1074 if (sscanf (line, "[%d:%d:%d]", &i, &i, &i)==3)
1075 {*uses_time=1;return SUB_SUBRIP09;}
1078 return SUB_INVALID; // too many bad lines
1081 extern int sub_utf8;
1082 int sub_utf8_prev=0;
1084 extern float sub_delay;
1085 extern float sub_fps;
1087 #ifdef CONFIG_ICONV
1088 static iconv_t icdsc = (iconv_t)(-1);
1090 void subcp_open (stream_t *st)
1092 char *tocp = "UTF-8";
1094 if (sub_cp){
1095 const char *cp_tmp = sub_cp;
1096 #ifdef CONFIG_ENCA
1097 char enca_lang[3], enca_fallback[100];
1098 if (sscanf(sub_cp, "enca:%2s:%99s", enca_lang, enca_fallback) == 2
1099 || sscanf(sub_cp, "ENCA:%2s:%99s", enca_lang, enca_fallback) == 2) {
1100 if (st && st->flags & MP_STREAM_SEEK ) {
1101 cp_tmp = guess_cp(st, enca_lang, enca_fallback);
1102 } else {
1103 cp_tmp = enca_fallback;
1104 if (st)
1105 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: enca failed, stream must be seekable.\n");
1108 #endif
1109 if ((icdsc = iconv_open (tocp, cp_tmp)) != (iconv_t)(-1)){
1110 mp_msg(MSGT_SUBREADER,MSGL_V,"SUB: opened iconv descriptor.\n");
1111 sub_utf8 = 2;
1112 } else
1113 mp_msg(MSGT_SUBREADER,MSGL_ERR,"SUB: error opening iconv descriptor.\n");
1117 void subcp_close (void)
1119 if (icdsc != (iconv_t)(-1)){
1120 (void) iconv_close (icdsc);
1121 icdsc = (iconv_t)(-1);
1122 mp_msg(MSGT_SUBREADER,MSGL_V,"SUB: closed iconv descriptor.\n");
1126 subtitle* subcp_recode (subtitle *sub)
1128 int l=sub->lines;
1129 size_t ileft, oleft;
1130 char *op, *ip, *ot;
1131 if(icdsc == (iconv_t)(-1)) return sub;
1133 while (l){
1134 ip = sub->text[--l];
1135 ileft = strlen(ip);
1136 oleft = 4 * ileft;
1138 if (!(ot = malloc(oleft + 1))){
1139 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error allocating mem.\n");
1140 continue;
1142 op = ot;
1143 if (iconv(icdsc, &ip, &ileft,
1144 &op, &oleft) == (size_t)(-1)) {
1145 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error recoding line.\n");
1146 free(ot);
1147 continue;
1149 // In some stateful encodings, we must clear the state to handle the last character
1150 if (iconv(icdsc, NULL, NULL,
1151 &op, &oleft) == (size_t)(-1)) {
1152 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error recoding line, can't clear encoding state.\n");
1154 *op='\0' ;
1155 free (sub->text[l]);
1156 sub->text[l] = ot;
1158 return sub;
1160 #endif
1162 #ifdef CONFIG_FRIBIDI
1164 * Do conversion necessary for right-to-left language support via fribidi.
1165 * @param sub subtitle to convert
1166 * @param sub_utf8 whether the subtitle is encoded in UTF-8
1167 * @param from first new subtitle, all lines before this are assumed to be already converted
1169 static subtitle* sub_fribidi (subtitle *sub, int sub_utf8, int from)
1171 FriBidiChar logical[LINE_LEN+1], visual[LINE_LEN+1]; // Hopefully these two won't smash the stack
1172 char *ip = NULL, *op = NULL;
1173 size_t len,orig_len;
1174 int l=sub->lines;
1175 int char_set_num;
1176 fribidi_boolean log2vis;
1177 if (!flip_hebrew)
1178 return sub;
1179 fribidi_set_mirroring(1);
1180 fribidi_set_reorder_nsm(0);
1182 if( sub_utf8 == 0 ) {
1183 char_set_num = fribidi_parse_charset (fribidi_charset?fribidi_charset:"ISO8859-8");
1184 }else {
1185 char_set_num = fribidi_parse_charset ("UTF-8");
1187 while (l > from) {
1188 ip = sub->text[--l];
1189 orig_len = len = strlen( ip ); // We assume that we don't use full unicode, only UTF-8 or ISO8859-x
1190 if(len > LINE_LEN) {
1191 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: sub->text is longer than LINE_LEN.\n");
1192 l++;
1193 break;
1195 len = fribidi_charset_to_unicode (char_set_num, ip, len, logical);
1196 #if FRIBIDI_INTERFACE_VERSION < 3
1197 FriBidiCharType base = fribidi_flip_commas?FRIBIDI_TYPE_ON:FRIBIDI_TYPE_L;
1198 #else
1199 FriBidiParType base = fribidi_flip_commas?FRIBIDI_TYPE_ON:FRIBIDI_TYPE_L;
1200 #endif
1201 log2vis = fribidi_log2vis (logical, len, &base,
1202 /* output */
1203 visual, NULL, NULL, NULL);
1204 if(log2vis) {
1205 len = fribidi_remove_bidi_marks (visual, len, NULL, NULL,
1206 NULL);
1207 if((op = malloc((FFMAX(2*orig_len,2*len) + 1))) == NULL) {
1208 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error allocating mem.\n");
1209 l++;
1210 break;
1212 fribidi_unicode_to_charset ( char_set_num, visual, len,op);
1213 free (ip);
1214 sub->text[l] = op;
1217 if (!from && l){
1218 for (l = sub->lines; l;)
1219 free (sub->text[--l]);
1220 return ERR;
1222 return sub;
1225 #endif
1227 static void adjust_subs_time(subtitle* sub, float subtime, float fps, int block,
1228 int sub_num, int sub_uses_time) {
1229 int n,m;
1230 subtitle* nextsub;
1231 int i = sub_num;
1232 unsigned long subfms = (sub_uses_time ? 100 : fps) * subtime;
1233 unsigned long overlap = (sub_uses_time ? 100 : fps) / 5; // 0.2s
1235 n=m=0;
1236 if (i) for (;;){
1237 if (sub->end <= sub->start){
1238 sub->end = sub->start + subfms;
1239 m++;
1240 n++;
1242 if (!--i) break;
1243 nextsub = sub + 1;
1244 if(block){
1245 if ((sub->end > nextsub->start) && (sub->end <= nextsub->start + overlap)) {
1246 // these subtitles overlap for less than 0.2 seconds
1247 // and would result in very short overlapping subtitle
1248 // so let's fix the problem here, before overlapping code
1249 // get its hands on them
1250 unsigned delta = sub->end - nextsub->start, half = delta / 2;
1251 sub->end -= half + 1;
1252 nextsub->start += delta - half;
1254 if (sub->end >= nextsub->start){
1255 sub->end = nextsub->start - 1;
1256 if (sub->end - sub->start > subfms)
1257 sub->end = sub->start + subfms;
1258 if (!m)
1259 n++;
1263 /* Theory:
1264 * Movies are often converted from FILM (24 fps)
1265 * to PAL (25) by simply speeding it up, so we
1266 * to multiply the original timestmaps by
1267 * (Movie's FPS / Subtitle's (guessed) FPS)
1268 * so eg. for 23.98 fps movie and PAL time based
1269 * subtitles we say -subfps 25 and we're fine!
1272 /* timed sub fps correction ::atmos */
1273 /* the frame-based case is handled in mpcommon.c
1274 * where find_sub is called */
1275 if(sub_uses_time && sub_fps) {
1276 sub->start *= sub_fps/fps;
1277 sub->end *= sub_fps/fps;
1280 sub = nextsub;
1281 m = 0;
1283 if (n) mp_msg(MSGT_SUBREADER,MSGL_V,"SUB: Adjusted %d subtitle(s).\n", n);
1286 struct subreader {
1287 subtitle * (*read)(stream_t *st,subtitle *dest,int utf16);
1288 void (*post)(subtitle *dest);
1289 const char *name;
1292 #ifdef CONFIG_ENCA
1293 const char* guess_buffer_cp(unsigned char* buffer, int buflen, const char *preferred_language, const char *fallback)
1295 const char **languages;
1296 size_t langcnt;
1297 EncaAnalyser analyser;
1298 EncaEncoding encoding;
1299 const char *detected_sub_cp = NULL;
1300 int i;
1302 languages = enca_get_languages(&langcnt);
1303 mp_msg(MSGT_SUBREADER, MSGL_V, "ENCA supported languages: ");
1304 for (i = 0; i < langcnt; i++) {
1305 mp_msg(MSGT_SUBREADER, MSGL_V, "%s ", languages[i]);
1307 mp_msg(MSGT_SUBREADER, MSGL_V, "\n");
1309 for (i = 0; i < langcnt; i++) {
1310 if (strcasecmp(languages[i], preferred_language) != 0) continue;
1311 analyser = enca_analyser_alloc(languages[i]);
1312 encoding = enca_analyse_const(analyser, buffer, buflen);
1313 enca_analyser_free(analyser);
1314 if (encoding.charset != ENCA_CS_UNKNOWN) {
1315 detected_sub_cp = enca_charset_name(encoding.charset, ENCA_NAME_STYLE_ICONV);
1316 break;
1320 free(languages);
1322 if (!detected_sub_cp) {
1323 detected_sub_cp = fallback;
1324 mp_msg(MSGT_SUBREADER, MSGL_INFO, "ENCA detection failed: fallback to %s\n", fallback);
1325 }else{
1326 mp_msg(MSGT_SUBREADER, MSGL_INFO, "ENCA detected charset: %s\n", detected_sub_cp);
1329 return detected_sub_cp;
1332 #define MAX_GUESS_BUFFER_SIZE (256*1024)
1333 const char* guess_cp(stream_t *st, const char *preferred_language, const char *fallback)
1335 size_t buflen;
1336 unsigned char *buffer;
1337 const char *detected_sub_cp = NULL;
1339 buffer = malloc(MAX_GUESS_BUFFER_SIZE);
1340 buflen = stream_read(st,buffer, MAX_GUESS_BUFFER_SIZE);
1342 detected_sub_cp = guess_buffer_cp(buffer, buflen, preferred_language, fallback);
1344 free(buffer);
1345 stream_reset(st);
1346 stream_seek(st,0);
1348 return detected_sub_cp;
1350 #undef MAX_GUESS_BUFFER_SIZE
1351 #endif
1353 sub_data* sub_read_file (char *filename, float fps) {
1354 int utf16;
1355 stream_t* fd;
1356 int n_max, n_first, i, j, sub_first, sub_orig;
1357 subtitle *first, *second, *sub, *return_sub, *alloced_sub = NULL;
1358 sub_data *subt_data;
1359 int uses_time = 0, sub_num = 0, sub_errs = 0;
1360 static const struct subreader sr[]=
1362 { sub_read_line_microdvd, NULL, "microdvd" },
1363 { sub_read_line_subrip, NULL, "subrip" },
1364 { sub_read_line_subviewer, NULL, "subviewer" },
1365 { sub_read_line_sami, NULL, "sami" },
1366 { sub_read_line_vplayer, NULL, "vplayer" },
1367 { sub_read_line_rt, NULL, "rt" },
1368 { sub_read_line_ssa, sub_pp_ssa, "ssa" },
1369 { sub_read_line_pjs, NULL, "pjs" },
1370 { sub_read_line_mpsub, NULL, "mpsub" },
1371 { sub_read_line_aqt, NULL, "aqt" },
1372 { sub_read_line_subviewer2, NULL, "subviewer 2.0" },
1373 { sub_read_line_subrip09, NULL, "subrip 0.9" },
1374 { sub_read_line_jacosub, NULL, "jacosub" },
1375 { sub_read_line_mpl2, NULL, "mpl2" }
1377 const struct subreader *srp;
1379 if(filename==NULL) return NULL; //qnx segfault
1380 fd=open_stream (filename, NULL, NULL); if (!fd) return NULL;
1382 sub_format = SUB_INVALID;
1383 for (utf16 = 0; sub_format == SUB_INVALID && utf16 < 3; utf16++) {
1384 sub_format=sub_autodetect (fd, &uses_time, utf16);
1385 stream_reset(fd);
1386 stream_seek(fd,0);
1388 utf16--;
1390 mpsub_multiplier = (uses_time ? 100.0 : 1.0);
1391 if (sub_format==SUB_INVALID) {mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: Could not determine file format\n");return NULL;}
1392 srp=sr+sub_format;
1393 mp_msg(MSGT_SUBREADER, MSGL_V, "SUB: Detected subtitle file format: %s\n", srp->name);
1395 #ifdef CONFIG_ICONV
1396 sub_utf8_prev=sub_utf8;
1398 int l,k;
1399 k = -1;
1400 if ((l=strlen(filename))>4){
1401 char *exts[] = {".utf", ".utf8", ".utf-8" };
1402 for (k=3;--k>=0;)
1403 if (l >= strlen(exts[k]) && !strcasecmp(filename+(l - strlen(exts[k])), exts[k])){
1404 sub_utf8 = 1;
1405 break;
1408 if (k<0) subcp_open(fd);
1410 #endif
1412 sub_num=0;n_max=32;
1413 first=malloc(n_max*sizeof(subtitle));
1414 if(!first){
1415 #ifdef CONFIG_ICONV
1416 subcp_close();
1417 sub_utf8=sub_utf8_prev;
1418 #endif
1419 return NULL;
1422 #ifdef CONFIG_SORTSUB
1423 alloced_sub =
1424 sub = malloc(sizeof(subtitle));
1425 //This is to deal with those formats (AQT & Subrip) which define the end of a subtitle
1426 //as the beginning of the following
1427 previous_sub_end = 0;
1428 #endif
1429 while(1){
1430 if(sub_num>=n_max){
1431 n_max+=16;
1432 first=realloc(first,n_max*sizeof(subtitle));
1434 #ifndef CONFIG_SORTSUB
1435 sub = &first[sub_num];
1436 #endif
1437 memset(sub, '\0', sizeof(subtitle));
1438 sub=srp->read(fd,sub,utf16);
1439 if(!sub) break; // EOF
1440 #ifdef CONFIG_ICONV
1441 if ((sub!=ERR) && sub_utf8 == 2) sub=subcp_recode(sub);
1442 #endif
1443 #ifdef CONFIG_FRIBIDI
1444 if (sub!=ERR) sub=sub_fribidi(sub,sub_utf8,0);
1445 #endif
1446 if ( sub == ERR )
1448 #ifdef CONFIG_ICONV
1449 subcp_close();
1450 #endif
1451 free(first);
1452 free(alloced_sub);
1453 return NULL;
1455 // Apply any post processing that needs recoding first
1456 if ((sub!=ERR) && !sub_no_text_pp && srp->post) srp->post(sub);
1457 #ifdef CONFIG_SORTSUB
1458 if(!sub_num || (first[sub_num - 1].start <= sub->start)){
1459 first[sub_num].start = sub->start;
1460 first[sub_num].end = sub->end;
1461 first[sub_num].lines = sub->lines;
1462 first[sub_num].alignment = sub->alignment;
1463 for(i = 0; i < sub->lines; ++i){
1464 first[sub_num].text[i] = sub->text[i];
1466 if (previous_sub_end){
1467 first[sub_num - 1].end = previous_sub_end;
1468 previous_sub_end = 0;
1470 } else {
1471 for(j = sub_num - 1; j >= 0; --j){
1472 first[j + 1].start = first[j].start;
1473 first[j + 1].end = first[j].end;
1474 first[j + 1].lines = first[j].lines;
1475 first[j + 1].alignment = first[j].alignment;
1476 for(i = 0; i < first[j].lines; ++i){
1477 first[j + 1].text[i] = first[j].text[i];
1479 if(!j || (first[j - 1].start <= sub->start)){
1480 first[j].start = sub->start;
1481 first[j].end = sub->end;
1482 first[j].lines = sub->lines;
1483 first[j].alignment = sub->alignment;
1484 for(i = 0; i < SUB_MAX_TEXT; ++i){
1485 first[j].text[i] = sub->text[i];
1487 if (previous_sub_end){
1488 first[j].end = first[j - 1].end;
1489 first[j - 1].end = previous_sub_end;
1490 previous_sub_end = 0;
1492 break;
1496 #endif
1497 if(sub==ERR) ++sub_errs; else ++sub_num; // Error vs. Valid
1500 free_stream(fd);
1502 #ifdef CONFIG_ICONV
1503 subcp_close();
1504 #endif
1505 free(alloced_sub);
1507 // printf ("SUB: Subtitle format %s time.\n", uses_time?"uses":"doesn't use");
1508 mp_msg(MSGT_SUBREADER, MSGL_V,"SUB: Read %i subtitles, %i bad line(s).\n",
1509 sub_num, sub_errs);
1511 if(sub_num<=0){
1512 free(first);
1513 return NULL;
1516 // we do overlap if the user forced it (suboverlap_enable == 2) or
1517 // the user didn't forced no-overlapsub and the format is Jacosub or Ssa.
1518 // this is because usually overlapping subtitles are found in these formats,
1519 // while in others they are probably result of bad timing
1520 if ((suboverlap_enabled == 2) ||
1521 ((suboverlap_enabled) && ((sub_format == SUB_JACOSUB) || (sub_format == SUB_SSA)))) {
1522 adjust_subs_time(first, 6.0, fps, 0, sub_num, uses_time);/*~6 secs AST*/
1523 // here we manage overlapping subtitles
1524 sub_orig = sub_num;
1525 n_first = sub_num;
1526 sub_num = 0;
1527 second = NULL;
1528 // for each subtitle in first[] we deal with its 'block' of
1529 // bonded subtitles
1530 for (sub_first = 0; sub_first < n_first; ++sub_first) {
1531 unsigned long global_start = first[sub_first].start,
1532 global_end = first[sub_first].end, local_start, local_end;
1533 int lines_to_add = first[sub_first].lines, sub_to_add = 0,
1534 **placeholder = NULL, higher_line = 0, counter, start_block_sub = sub_num;
1535 char real_block = 1;
1537 // here we find the number of subtitles inside the 'block'
1538 // and its span interval. this works well only with sorted
1539 // subtitles
1540 while ((sub_first + sub_to_add + 1 < n_first) && (first[sub_first + sub_to_add + 1].start < global_end)) {
1541 ++sub_to_add;
1542 lines_to_add += first[sub_first + sub_to_add].lines;
1543 if (first[sub_first + sub_to_add].start < global_start) {
1544 global_start = first[sub_first + sub_to_add].start;
1546 if (first[sub_first + sub_to_add].end > global_end) {
1547 global_end = first[sub_first + sub_to_add].end;
1551 /* Avoid n^2 memory use for the "placeholder" data structure
1552 * below with subtitles that have a huge number of
1553 * consecutive overlapping lines. */
1554 lines_to_add = FFMIN(lines_to_add, SUB_MAX_TEXT);
1556 // we need a structure to keep trace of the screen lines
1557 // used by the subs, a 'placeholder'
1558 counter = 2 * sub_to_add + 1; // the maximum number of subs derived
1559 // from a block of sub_to_add+1 subs
1560 placeholder = malloc(sizeof(int *) * counter);
1561 for (i = 0; i < counter; ++i) {
1562 placeholder[i] = malloc(sizeof(int) * lines_to_add);
1563 for (j = 0; j < lines_to_add; ++j) {
1564 placeholder[i][j] = -1;
1568 counter = 0;
1569 local_end = global_start - 1;
1570 do {
1571 int ls;
1573 // here we find the beginning and the end of a new
1574 // subtitle in the block
1575 local_start = local_end + 1;
1576 local_end = global_end;
1577 for (j = 0; j <= sub_to_add; ++j) {
1578 if ((first[sub_first + j].start - 1 > local_start) && (first[sub_first + j].start - 1 < local_end)) {
1579 local_end = first[sub_first + j].start - 1;
1580 } else if ((first[sub_first + j].end > local_start) && (first[sub_first + j].end < local_end)) {
1581 local_end = first[sub_first + j].end;
1584 // here we allocate the screen lines to subs we must
1585 // display in current local_start-local_end interval.
1586 // if the subs were yet presents in the previous interval
1587 // they keep the same lines, otherside they get unused lines
1588 for (j = 0; j <= sub_to_add; ++j) {
1589 if ((first[sub_first + j].start <= local_end) && (first[sub_first + j].end > local_start)) {
1590 unsigned long sub_lines = first[sub_first + j].lines, fragment_length = lines_to_add + 1,
1591 tmp = 0;
1592 char boolean = 0;
1593 int fragment_position = -1;
1595 // if this is not the first new sub of the block
1596 // we find if this sub was present in the previous
1597 // new sub
1598 if (counter)
1599 for (i = 0; i < lines_to_add; ++i) {
1600 if (placeholder[counter - 1][i] == sub_first + j) {
1601 placeholder[counter][i] = sub_first + j;
1602 boolean = 1;
1605 if (boolean)
1606 continue;
1608 // we are looking for the shortest among all groups of
1609 // sequential blank lines whose length is greater than or
1610 // equal to sub_lines. we store in fragment_position the
1611 // position of the shortest group, in fragment_length its
1612 // length, and in tmp the length of the group currently
1613 // examinated
1614 for (i = 0; i < lines_to_add; ++i) {
1615 if (placeholder[counter][i] == -1) {
1616 // placeholder[counter][i] is part of the current group
1617 // of blank lines
1618 ++tmp;
1619 } else {
1620 if (tmp == sub_lines) {
1621 // current group's size fits exactly the one we
1622 // need, so we stop looking
1623 fragment_position = i - tmp;
1624 tmp = 0;
1625 break;
1627 if ((tmp) && (tmp > sub_lines) && (tmp < fragment_length)) {
1628 // current group is the best we found till here,
1629 // but is still bigger than the one we are looking
1630 // for, so we keep on looking
1631 fragment_length = tmp;
1632 fragment_position = i - tmp;
1633 tmp = 0;
1634 } else {
1635 // current group doesn't fit at all, so we forget it
1636 tmp = 0;
1640 if (tmp) {
1641 // last screen line is blank, a group ends with it
1642 if ((tmp >= sub_lines) && (tmp < fragment_length)) {
1643 fragment_position = i - tmp;
1646 if (fragment_position == -1) {
1647 // it was not possible to find free screen line(s) for a subtitle,
1648 // usually this means a bug in the code; however we do not overlap
1649 mp_msg(MSGT_SUBREADER, MSGL_WARN, "SUB: we could not find a suitable position for an overlapping subtitle\n");
1650 higher_line = SUB_MAX_TEXT + 1;
1651 break;
1652 } else {
1653 for (tmp = 0; tmp < sub_lines; ++tmp) {
1654 placeholder[counter][fragment_position + tmp] = sub_first + j;
1659 for (j = higher_line + 1; j < lines_to_add; ++j) {
1660 if (placeholder[counter][j] != -1)
1661 higher_line = j;
1662 else
1663 break;
1665 if (higher_line >= SUB_MAX_TEXT) {
1666 // the 'block' has too much lines, so we don't overlap the
1667 // subtitles
1668 second = realloc(second, (sub_num + sub_to_add + 1) * sizeof(subtitle));
1669 for (j = 0; j <= sub_to_add; ++j) {
1670 int ls;
1671 memset(&second[sub_num + j], '\0', sizeof(subtitle));
1672 second[sub_num + j].start = first[sub_first + j].start;
1673 second[sub_num + j].end = first[sub_first + j].end;
1674 second[sub_num + j].lines = first[sub_first + j].lines;
1675 second[sub_num + j].alignment = first[sub_first + j].alignment;
1676 for (ls = 0; ls < second[sub_num + j].lines; ls++) {
1677 second[sub_num + j].text[ls] = strdup(first[sub_first + j].text[ls]);
1680 sub_num += sub_to_add + 1;
1681 sub_first += sub_to_add;
1682 real_block = 0;
1683 break;
1686 // we read the placeholder structure and create the new
1687 // subs.
1688 second = realloc(second, (sub_num + 1) * sizeof(subtitle));
1689 memset(&second[sub_num], '\0', sizeof(subtitle));
1690 second[sub_num].start = local_start;
1691 second[sub_num].end = local_end;
1692 second[sub_num].alignment = first[sub_first].alignment;
1693 n_max = (lines_to_add < SUB_MAX_TEXT) ? lines_to_add : SUB_MAX_TEXT;
1694 for (i = 0, j = 0; j < n_max; ++j) {
1695 if (placeholder[counter][j] != -1) {
1696 int lines = first[placeholder[counter][j]].lines;
1697 for (ls = 0; ls < lines; ++ls) {
1698 second[sub_num].text[i++] = strdup(first[placeholder[counter][j]].text[ls]);
1700 j += lines - 1;
1701 } else {
1702 second[sub_num].text[i++] = strdup(" ");
1705 ++sub_num;
1706 ++counter;
1707 } while (local_end < global_end);
1708 if (real_block)
1709 for (i = 0; i < counter; ++i)
1710 second[start_block_sub + i].lines = higher_line + 1;
1712 counter = 2 * sub_to_add + 1;
1713 for (i = 0; i < counter; ++i) {
1714 free(placeholder[i]);
1716 free(placeholder);
1717 sub_first += sub_to_add;
1720 for (j = sub_orig - 1; j >= 0; --j) {
1721 for (i = first[j].lines - 1; i >= 0; --i) {
1722 free(first[j].text[i]);
1725 free(first);
1727 return_sub = second;
1728 } else { //if(suboverlap_enabled)
1729 adjust_subs_time(first, 6.0, fps, 1, sub_num, uses_time);/*~6 secs AST*/
1730 return_sub = first;
1732 if (return_sub == NULL) return NULL;
1733 subt_data = malloc(sizeof(sub_data));
1734 subt_data->filename = strdup(filename);
1735 subt_data->sub_uses_time = uses_time;
1736 subt_data->sub_num = sub_num;
1737 subt_data->sub_errs = sub_errs;
1738 subt_data->subtitles = return_sub;
1739 return subt_data;
1742 #if 0
1743 char * strreplace( char * in,char * what,char * whereof )
1745 int i;
1746 char * tmp;
1748 if ( ( in == NULL )||( what == NULL )||( whereof == NULL )||( ( tmp=strstr( in,what ) ) == NULL ) ) return NULL;
1749 for( i=0;i<strlen( whereof );i++ ) tmp[i]=whereof[i];
1750 if ( strlen( what ) > strlen( whereof ) ) tmp[i]=0;
1751 return in;
1753 #endif
1756 static void strcpy_trim(char *d, char *s)
1758 // skip leading whitespace
1759 while (*s && isspace(*s)) {
1760 s++;
1762 for (;;) {
1763 // copy word
1764 while (*s && !isspace(*s)) {
1765 *d = tolower(*s);
1766 s++; d++;
1768 if (*s == 0) break;
1769 // trim excess whitespace
1770 while (*s && isspace(*s)) {
1771 s++;
1773 if (*s == 0) break;
1774 *d++ = ' ';
1776 *d = 0;
1779 static void strcpy_strip_ext(char *d, char *s)
1781 char *tmp = strrchr(s,'.');
1782 if (!tmp) {
1783 strcpy(d, s);
1784 return;
1785 } else {
1786 strncpy(d, s, tmp-s);
1787 d[tmp-s] = 0;
1789 while (*d) {
1790 *d = tolower(*d);
1791 d++;
1795 static void strcpy_get_ext(char *d, char *s)
1797 char *tmp = strrchr(s,'.');
1798 if (!tmp) {
1799 strcpy(d, "");
1800 return;
1801 } else {
1802 strcpy(d, tmp+1);
1806 static int whiteonly(char *s)
1808 while (*s) {
1809 if (!isspace(*s)) return 0;
1810 s++;
1812 return 1;
1815 typedef struct subfn
1817 int priority;
1818 char *fname;
1819 } subfn;
1821 static int compare_sub_priority(const void *a, const void *b)
1823 if (((const subfn*)a)->priority > ((const subfn*)b)->priority) {
1824 return -1;
1825 } else if (((const subfn*)a)->priority < ((const subfn*)b)->priority) {
1826 return 1;
1827 } else {
1828 return strcoll(((const subfn*)a)->fname, ((const subfn*)b)->fname);
1832 char** sub_filenames(const char* path, char *fname)
1834 char *f_dir, *f_fname, *f_fname_noext, *f_fname_trim, *tmp, *tmp_sub_id;
1835 char *tmp_fname_noext, *tmp_fname_trim, *tmp_fname_ext, *tmpresult;
1837 int len, pos, found, i, j;
1838 char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "rt", "txt", "ssa", "aqt", "jss", "js", "ass", NULL};
1839 subfn *result;
1840 char **result2;
1842 int subcnt;
1844 FILE *f;
1846 DIR *d;
1847 struct dirent *de;
1849 len = (strlen(fname) > 256 ? strlen(fname) : 256)
1850 +(strlen(path) > 256 ? strlen(path) : 256)+2;
1852 f_dir = malloc(len);
1853 f_fname = malloc(len);
1854 f_fname_noext = malloc(len);
1855 f_fname_trim = malloc(len);
1857 tmp_fname_noext = malloc(len);
1858 tmp_fname_trim = malloc(len);
1859 tmp_fname_ext = malloc(len);
1861 tmpresult = malloc(len);
1863 result = calloc(MAX_SUBTITLE_FILES, sizeof(*result));
1865 subcnt = 0;
1867 tmp = strrchr(fname,'/');
1868 #if HAVE_DOS_PATHS
1869 if(!tmp)tmp = strrchr(fname,'\\');
1870 if(!tmp)tmp = strrchr(fname,':');
1871 #endif
1873 // extract filename & dirname from fname
1874 if (tmp) {
1875 strcpy(f_fname, tmp+1);
1876 pos = tmp - fname;
1877 strncpy(f_dir, fname, pos+1);
1878 f_dir[pos+1] = 0;
1879 } else {
1880 strcpy(f_fname, fname);
1881 strcpy(f_dir, "./");
1884 strcpy_strip_ext(f_fname_noext, f_fname);
1885 strcpy_trim(f_fname_trim, f_fname_noext);
1887 /* The code using sub language here is broken - it assumes strict
1888 * "videoname languagename" syntax for the subtitle file, which is
1889 * very unlikely to match especially if language name uses "en,de"
1890 * syntax... */
1891 tmp_sub_id = NULL;
1892 #if 0
1893 if (dvdsub_lang && !whiteonly(dvdsub_lang)) {
1894 tmp_sub_id = malloc(strlen(dvdsub_lang)+1);
1895 strcpy_trim(tmp_sub_id, dvdsub_lang);
1897 #endif
1899 // 0 = nothing
1900 // 1 = any subtitle file
1901 // 2 = any sub file containing movie name
1902 // 3 = sub file containing movie name and the lang extension
1903 for (j = 0; j <= 1; j++) {
1904 d = opendir(j == 0 ? f_dir : path);
1905 if (d) {
1906 while ((de = readdir(d))) {
1907 // retrieve various parts of the filename
1908 strcpy_strip_ext(tmp_fname_noext, de->d_name);
1909 strcpy_get_ext(tmp_fname_ext, de->d_name);
1910 strcpy_trim(tmp_fname_trim, tmp_fname_noext);
1912 // does it end with a subtitle extension?
1913 found = 0;
1914 #ifdef CONFIG_ICONV
1915 #ifdef CONFIG_ENCA
1916 for (i = ((sub_cp && strncasecmp(sub_cp, "enca", 4) != 0) ? 3 : 0); sub_exts[i]; i++) {
1917 #else
1918 for (i = (sub_cp ? 3 : 0); sub_exts[i]; i++) {
1919 #endif
1920 #else
1921 for (i = 0; sub_exts[i]; i++) {
1922 #endif
1923 if (strcasecmp(sub_exts[i], tmp_fname_ext) == 0) {
1924 found = 1;
1925 break;
1929 // we have a (likely) subtitle file
1930 if (found) {
1931 int prio = 0;
1932 if (!prio && tmp_sub_id)
1934 sprintf(tmpresult, "%s %s", f_fname_trim, tmp_sub_id);
1935 if (strcmp(tmp_fname_trim, tmpresult) == 0 && sub_match_fuzziness >= 1) {
1936 // matches the movie name + lang extension
1937 prio = 5;
1940 if (!prio && strcmp(tmp_fname_trim, f_fname_trim) == 0) {
1941 // matches the movie name
1942 prio = 4;
1944 if (!prio && (tmp = strstr(tmp_fname_trim, f_fname_trim)) && (sub_match_fuzziness >= 1)) {
1945 // contains the movie name
1946 tmp += strlen(f_fname_trim);
1947 if (tmp_sub_id && strstr(tmp, tmp_sub_id)) {
1948 // with sub_id specified prefer localized subtitles
1949 prio = 3;
1950 } else if ((tmp_sub_id == NULL) && whiteonly(tmp)) {
1951 // without sub_id prefer "plain" name
1952 prio = 3;
1953 } else {
1954 // with no localized subs found, try any else instead
1955 prio = 2;
1958 if (!prio) {
1959 // doesn't contain the movie name
1960 // don't try in the mplayer subtitle directory
1961 if ((j == 0) && (sub_match_fuzziness >= 2)) {
1962 prio = 1;
1966 mp_msg(MSGT_SUBREADER, MSGL_DBG2, "Potential sub file: "
1967 "\"%s\" Priority: %d\n", de->d_name, prio);
1968 if (prio) {
1969 prio += prio;
1970 #ifdef CONFIG_ICONV
1971 if (i<3){ // prefer UTF-8 coded
1972 prio++;
1974 #endif
1975 sprintf(tmpresult, "%s%s", j == 0 ? f_dir : path, de->d_name);
1976 // fprintf(stderr, "%s priority %d\n", tmpresult, prio);
1977 if ((f = fopen(tmpresult, "rt"))) {
1978 fclose(f);
1979 result[subcnt].priority = prio;
1980 result[subcnt].fname = strdup(tmpresult);
1981 subcnt++;
1986 if (subcnt >= MAX_SUBTITLE_FILES) break;
1988 closedir(d);
1993 free(tmp_sub_id);
1995 free(f_dir);
1996 free(f_fname);
1997 free(f_fname_noext);
1998 free(f_fname_trim);
2000 free(tmp_fname_noext);
2001 free(tmp_fname_trim);
2002 free(tmp_fname_ext);
2004 free(tmpresult);
2006 qsort(result, subcnt, sizeof(subfn), compare_sub_priority);
2008 result2 = calloc(subcnt + 1, sizeof(*result2));
2010 for (i = 0; i < subcnt; i++) {
2011 result2[i] = result[i].fname;
2013 result2[subcnt] = NULL;
2015 free(result);
2017 return result2;
2020 void list_sub_file(sub_data* subd){
2021 int i,j;
2022 subtitle *subs = subd->subtitles;
2024 for(j=0; j < subd->sub_num; j++){
2025 subtitle* egysub=&subs[j];
2026 mp_msg(MSGT_SUBREADER,MSGL_INFO,"%i line%c (%li-%li)\n",
2027 egysub->lines,
2028 (1==egysub->lines)?' ':'s',
2029 egysub->start,
2030 egysub->end);
2031 for (i=0; i<egysub->lines; i++) {
2032 mp_msg(MSGT_SUBREADER,MSGL_INFO,"\t\t%d: %s%s", i,egysub->text[i], i==egysub->lines-1?"":" \n ");
2034 mp_msg(MSGT_SUBREADER,MSGL_INFO,"\n");
2037 mp_msg(MSGT_SUBREADER,MSGL_INFO,"Subtitle format %s time.\n",
2038 subd->sub_uses_time ? "uses":"doesn't use");
2039 mp_msg(MSGT_SUBREADER,MSGL_INFO,"Read %i subtitles, %i errors.\n", subd->sub_num, subd->sub_errs);
2042 void dump_srt(sub_data* subd, float fps){
2043 int i,j;
2044 int h,m,s,ms;
2045 FILE * fd;
2046 subtitle * onesub;
2047 unsigned long temp;
2048 subtitle *subs = subd->subtitles;
2050 if (!subd->sub_uses_time && sub_fps == 0)
2051 sub_fps = fps;
2052 fd=fopen("dumpsub.srt","w");
2053 if(!fd)
2055 perror("dump_srt: fopen");
2056 return;
2058 for(i=0; i < subd->sub_num; i++)
2060 onesub=subs+i; //=&subs[i];
2061 fprintf(fd,"%d\n",i+1);//line number
2063 temp=onesub->start;
2064 if (!subd->sub_uses_time)
2065 temp = temp * 100 / sub_fps;
2066 temp -= sub_delay * 100;
2067 h=temp/360000;temp%=360000; //h =1*100*60*60
2068 m=temp/6000; temp%=6000; //m =1*100*60
2069 s=temp/100; temp%=100; //s =1*100
2070 ms=temp*10; //ms=1*10
2071 fprintf(fd,"%02d:%02d:%02d,%03d --> ",h,m,s,ms);
2073 temp=onesub->end;
2074 if (!subd->sub_uses_time)
2075 temp = temp * 100 / sub_fps;
2076 temp -= sub_delay * 100;
2077 h=temp/360000;temp%=360000;
2078 m=temp/6000; temp%=6000;
2079 s=temp/100; temp%=100;
2080 ms=temp*10;
2081 fprintf(fd,"%02d:%02d:%02d,%03d\n",h,m,s,ms);
2083 for(j=0;j<onesub->lines;j++)
2084 fprintf(fd,"%s\n",onesub->text[j]);
2086 fprintf(fd,"\n");
2088 fclose(fd);
2089 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.srt\'.\n");
2092 void dump_mpsub(sub_data* subd, float fps){
2093 int i,j;
2094 FILE *fd;
2095 float a,b;
2096 subtitle *subs = subd->subtitles;
2098 mpsub_position = subd->sub_uses_time? (sub_delay*100) : (sub_delay*fps);
2099 if (sub_fps==0) sub_fps=fps;
2101 fd=fopen ("dump.mpsub", "w");
2102 if (!fd) {
2103 perror ("dump_mpsub: fopen");
2104 return;
2108 if (subd->sub_uses_time) fprintf (fd,"FORMAT=TIME\n\n");
2109 else fprintf (fd, "FORMAT=%5.2f\n\n", fps);
2111 for(j=0; j < subd->sub_num; j++){
2112 subtitle* egysub=&subs[j];
2113 if (subd->sub_uses_time) {
2114 a=((egysub->start-mpsub_position)/100.0);
2115 b=((egysub->end-egysub->start)/100.0);
2116 if ( (float)((int)a) == a)
2117 fprintf (fd, "%.0f",a);
2118 else
2119 fprintf (fd, "%.2f",a);
2121 if ( (float)((int)b) == b)
2122 fprintf (fd, " %.0f\n",b);
2123 else
2124 fprintf (fd, " %.2f\n",b);
2125 } else {
2126 fprintf (fd, "%ld %ld\n", (long)((egysub->start*(fps/sub_fps))-((mpsub_position*(fps/sub_fps)))),
2127 (long)(((egysub->end)-(egysub->start))*(fps/sub_fps)));
2130 mpsub_position = egysub->end;
2131 for (i=0; i<egysub->lines; i++) {
2132 fprintf (fd, "%s\n",egysub->text[i]);
2134 fprintf (fd, "\n");
2136 fclose (fd);
2137 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dump.mpsub\'.\n");
2140 void dump_microdvd(sub_data* subd, float fps) {
2141 int i, delay;
2142 FILE *fd;
2143 subtitle *subs = subd->subtitles;
2144 if (sub_fps == 0)
2145 sub_fps = fps;
2146 fd = fopen("dumpsub.sub", "w");
2147 if (!fd) {
2148 perror("dumpsub.sub: fopen");
2149 return;
2151 delay = sub_delay * sub_fps;
2152 for (i = 0; i < subd->sub_num; ++i) {
2153 int j, start, end;
2154 start = subs[i].start;
2155 end = subs[i].end;
2156 if (subd->sub_uses_time) {
2157 start = start * sub_fps / 100 ;
2158 end = end * sub_fps / 100;
2160 else {
2161 start = start * sub_fps / fps;
2162 end = end * sub_fps / fps;
2164 start -= delay;
2165 end -= delay;
2166 fprintf(fd, "{%d}{%d}", start, end);
2167 for (j = 0; j < subs[i].lines; ++j)
2168 fprintf(fd, "%s%s", j ? "|" : "", subs[i].text[j]);
2169 fprintf(fd, "\n");
2171 fclose(fd);
2172 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.sub\'.\n");
2175 void dump_jacosub(sub_data* subd, float fps) {
2176 int i,j;
2177 int h,m,s,cs;
2178 FILE * fd;
2179 subtitle * onesub;
2180 unsigned long temp;
2181 subtitle *subs = subd->subtitles;
2183 if (!subd->sub_uses_time && sub_fps == 0)
2184 sub_fps = fps;
2185 fd=fopen("dumpsub.jss","w");
2186 if(!fd)
2188 perror("dump_jacosub: fopen");
2189 return;
2191 fprintf(fd, "#TIMERES %d\n", (subd->sub_uses_time) ? 100 : (int)sub_fps);
2192 for(i=0; i < subd->sub_num; i++)
2194 onesub=subs+i; //=&subs[i];
2196 temp=onesub->start;
2197 if (!subd->sub_uses_time)
2198 temp = temp * 100 / sub_fps;
2199 temp -= sub_delay * 100;
2200 h=temp/360000;temp%=360000; //h =1*100*60*60
2201 m=temp/6000; temp%=6000; //m =1*100*60
2202 s=temp/100; temp%=100; //s =1*100
2203 cs=temp; //cs=1*10
2204 fprintf(fd,"%02d:%02d:%02d.%02d ",h,m,s,cs);
2206 temp=onesub->end;
2207 if (!subd->sub_uses_time)
2208 temp = temp * 100 / sub_fps;
2209 temp -= sub_delay * 100;
2210 h=temp/360000;temp%=360000;
2211 m=temp/6000; temp%=6000;
2212 s=temp/100; temp%=100;
2213 cs=temp;
2214 fprintf(fd,"%02d:%02d:%02d.%02d {~} ",h,m,s,cs);
2216 for(j=0;j<onesub->lines;j++)
2217 fprintf(fd,"%s%s",j ? "\\n" : "", onesub->text[j]);
2219 fprintf(fd,"\n");
2221 fclose(fd);
2222 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.js\'.\n");
2225 void dump_sami(sub_data* subd, float fps) {
2226 int i,j;
2227 FILE * fd;
2228 subtitle * onesub;
2229 unsigned long temp;
2230 subtitle *subs = subd->subtitles;
2232 if (!subd->sub_uses_time && sub_fps == 0)
2233 sub_fps = fps;
2234 fd=fopen("dumpsub.smi","w");
2235 if(!fd)
2237 perror("dump_jacosub: fopen");
2238 return;
2240 fprintf(fd, "<SAMI>\n"
2241 "<HEAD>\n"
2242 " <STYLE TYPE=\"Text/css\">\n"
2243 " <!--\n"
2244 " P {margin-left: 29pt; margin-right: 29pt; font-size: 24pt; text-align: center; font-family: Tahoma; font-weight: bold; color: #FCDD03; background-color: #000000;}\n"
2245 " .SUBTTL {Name: 'Subtitles'; Lang: en-US; SAMIType: CC;}\n"
2246 " -->\n"
2247 " </STYLE>\n"
2248 "</HEAD>\n"
2249 "<BODY>\n");
2250 for(i=0; i < subd->sub_num; i++)
2252 onesub=subs+i; //=&subs[i];
2254 temp=onesub->start;
2255 if (!subd->sub_uses_time)
2256 temp = temp * 100 / sub_fps;
2257 temp -= sub_delay * 100;
2258 fprintf(fd,"\t<SYNC Start=%lu>\n"
2259 "\t <P>", temp * 10);
2261 for(j=0;j<onesub->lines;j++)
2262 fprintf(fd,"%s%s",j ? "<br>" : "", onesub->text[j]);
2264 fprintf(fd,"\n");
2266 temp=onesub->end;
2267 if (!subd->sub_uses_time)
2268 temp = temp * 100 / sub_fps;
2269 temp -= sub_delay * 100;
2270 fprintf(fd,"\t<SYNC Start=%lu>\n"
2271 "\t <P>&nbsp;\n", temp * 10);
2273 fprintf(fd, "</BODY>\n"
2274 "</SAMI>\n");
2275 fclose(fd);
2276 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.smi\'.\n");
2279 void sub_free( sub_data * subd )
2281 int i, j;
2283 if ( !subd ) return;
2285 for (i = 0; i < subd->sub_num; i++)
2286 for (j = 0; j < subd->subtitles[i].lines; j++)
2287 free( subd->subtitles[i].text[j] );
2288 free( subd->subtitles );
2289 free( subd->filename );
2290 free( subd );
2293 #define MAX_SUBLINE 512
2295 * \brief parse text and append it to subtitle in sub
2296 * \param sub subtitle struct to add text to
2297 * \param txt text to parse
2298 * \param len length of text in txt
2299 * \param endpts pts at which this subtitle text should be removed again
2301 * <> and {} are interpreted as comment delimiters, "\n", "\N", '\n', '\r'
2302 * and '\0' are interpreted as newlines, duplicate, leading and trailing
2303 * newlines are ignored.
2305 void sub_add_text(subtitle *sub, const char *txt, int len, double endpts) {
2306 int comment = 0;
2307 int double_newline = 1; // ignore newlines at the beginning
2308 int i, pos;
2309 char *buf;
2310 #ifdef CONFIG_FRIBIDI
2311 int orig_lines = sub->lines;
2312 #endif
2313 if (sub->lines >= SUB_MAX_TEXT) return;
2314 pos = 0;
2315 buf = malloc(MAX_SUBLINE + 1);
2316 sub->text[sub->lines] = buf;
2317 sub->endpts[sub->lines] = endpts;
2318 for (i = 0; i < len && pos < MAX_SUBLINE; i++) {
2319 char c = txt[i];
2320 if (c == '<') comment |= 1;
2321 if (c == '{') comment |= 2;
2322 if (comment) {
2323 if (c == '}') comment &= ~2;
2324 if (c == '>') comment &= ~1;
2325 continue;
2327 if (pos == MAX_SUBLINE - 1) {
2328 i--;
2329 c = 0;
2331 if (c == '\\' && i + 1 < len) {
2332 c = txt[++i];
2333 if (c == 'n' || c == 'N') c = 0;
2335 if (c == '\n' || c == '\r') c = 0;
2336 if (c) {
2337 double_newline = 0;
2338 buf[pos++] = c;
2339 } else if (!double_newline) {
2340 if (sub->lines >= SUB_MAX_TEXT - 1) {
2341 mp_msg(MSGT_VO, MSGL_WARN, "Too many subtitle lines\n");
2342 break;
2344 double_newline = 1;
2345 buf[pos] = 0;
2346 sub->lines++;
2347 pos = 0;
2348 buf = malloc(MAX_SUBLINE + 1);
2349 sub->text[sub->lines] = buf;
2350 sub->endpts[sub->lines] = endpts;
2353 buf[pos] = 0;
2354 if (sub->lines < SUB_MAX_TEXT &&
2355 strlen(sub->text[sub->lines]))
2356 sub->lines++;
2357 #ifdef CONFIG_FRIBIDI
2358 sub = sub_fribidi(sub, sub_utf8, orig_lines);
2359 #endif
2362 #define MP_NOPTS_VALUE (-1LL<<63)
2364 * \brief remove outdated subtitle lines.
2365 * \param sub subtitle struct to modify
2366 * \param pts current pts. All lines with endpts <= this will be removed.
2367 * Use MP_NOPTS_VALUE to remove all lines
2368 * \return 1 if sub was modified, 0 otherwise.
2370 int sub_clear_text(subtitle *sub, double pts) {
2371 int i = 0;
2372 int changed = 0;
2373 while (i < sub->lines) {
2374 double endpts = sub->endpts[i];
2375 if (pts == MP_NOPTS_VALUE || (endpts != MP_NOPTS_VALUE && pts >= endpts)) {
2376 int j;
2377 free(sub->text[i]);
2378 for (j = i + 1; j < sub->lines; j++) {
2379 sub->text[j - 1] = sub->text[j];
2380 sub->endpts[j - 1] = sub->endpts[j];
2382 sub->lines--;
2383 changed = 1;
2384 } else
2385 i++;
2387 return changed;