cleanup: move MP_NOPTS_VALUE definition to mpcommon.h
[mplayer/glamo.git] / subreader.c
blobc9b3261c0ed01b656435390237b90de30537312c
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 "mpcommon.h"
36 #include "stream/stream.h"
37 #include "libavutil/common.h"
38 #include "libavutil/avstring.h"
40 #ifdef CONFIG_ENCA
41 #include <enca.h>
42 #endif
44 #define ERR ((void *) -1)
46 #ifdef CONFIG_ICONV
47 #include <iconv.h>
48 char *sub_cp=NULL;
49 #endif
50 #ifdef CONFIG_FRIBIDI
51 #include <fribidi/fribidi.h>
52 char *fribidi_charset = NULL; ///character set that will be passed to FriBiDi
53 int flip_hebrew = 1; ///flip subtitles using fribidi
54 int fribidi_flip_commas = 0; ///flip comma when fribidi is used
55 #endif
57 /* Maximal length of line of a subtitle */
58 #define LINE_LEN 1000
59 static float mpsub_position=0;
60 static float mpsub_multiplier=1.;
61 static int sub_slacktime = 20000; //20 sec
63 int sub_no_text_pp=0; // 1 => do not apply text post-processing
64 // like {\...} elimination in SSA format.
66 int sub_match_fuzziness=0; // level of sub name matching fuzziness
68 /* Use the SUB_* constant defined in the header file */
69 int sub_format=SUB_INVALID;
70 #ifdef CONFIG_SORTSUB
72 Some subtitling formats, namely AQT and Subrip09, define the end of a
73 subtitle as the beginning of the following. Since currently we read one
74 subtitle at time, for these format we keep two global *subtitle,
75 previous_aqt_sub and previous_subrip09_sub, pointing to previous subtitle,
76 so we can change its end when we read current subtitle starting time.
77 When CONFIG_SORTSUB is defined, we use a single global unsigned long,
78 previous_sub_end, for both (and even future) formats, to store the end of
79 the previous sub: it is initialized to 0 in sub_read_file and eventually
80 modified by sub_read_aqt_line or sub_read_subrip09_line.
82 unsigned long previous_sub_end;
83 #endif
85 static int eol(char p) {
86 return p=='\r' || p=='\n' || p=='\0';
89 /* Remove leading and trailing space */
90 static void trail_space(char *s) {
91 int i = 0;
92 while (isspace(s[i])) ++i;
93 if (i) strcpy(s, s + i);
94 i = strlen(s) - 1;
95 while (i > 0 && isspace(s[i])) s[i--] = '\0';
98 static char *stristr(const char *haystack, const char *needle) {
99 int len = 0;
100 const char *p = haystack;
102 if (!(haystack && needle)) return NULL;
104 len=strlen(needle);
105 while (*p != '\0') {
106 if (strncasecmp(p, needle, len) == 0) return (char*)p;
107 p++;
110 return NULL;
113 static void sami_add_line(subtitle *current, char *buffer, char **pos) {
114 char *p = *pos;
115 *p = 0;
116 trail_space(buffer);
117 if (*buffer && current->lines < SUB_MAX_TEXT)
118 current->text[current->lines++] = strdup(buffer);
119 *pos = buffer;
122 static subtitle *sub_read_line_sami(stream_t* st, subtitle *current, int utf16) {
123 static char line[LINE_LEN+1];
124 static char *s = NULL, *slacktime_s;
125 char text[LINE_LEN+1], *p=NULL, *q;
126 int state;
128 current->lines = current->start = current->end = 0;
129 current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
130 state = 0;
132 /* read the first line */
133 if (!s)
134 if (!(s = stream_read_line(st, line, LINE_LEN, utf16))) return 0;
136 do {
137 switch (state) {
139 case 0: /* find "START=" or "Slacktime:" */
140 slacktime_s = stristr (s, "Slacktime:");
141 if (slacktime_s)
142 sub_slacktime = strtol (slacktime_s+10, NULL, 0) / 10;
144 s = stristr (s, "Start=");
145 if (s) {
146 current->start = strtol (s + 6, &s, 0) / 10;
147 /* eat '>' */
148 for (; *s != '>' && *s != '\0'; s++);
149 s++;
150 state = 1; continue;
152 break;
154 case 1: /* find (optional) "<P", skip other TAGs */
155 for (; *s == ' ' || *s == '\t'; s++); /* strip blanks, if any */
156 if (*s == '\0') break;
157 if (*s != '<') { state = 3; p = text; continue; } /* not a TAG */
158 s++;
159 if (*s == 'P' || *s == 'p') { s++; state = 2; continue; } /* found '<P' */
160 for (; *s != '>' && *s != '\0'; s++); /* skip remains of non-<P> TAG */
161 if (s == '\0')
162 break;
163 s++;
164 continue;
166 case 2: /* find ">" */
167 if ((s = strchr (s, '>'))) { s++; state = 3; p = text; continue; }
168 break;
170 case 3: /* get all text until '<' appears */
171 if (*s == '\0') break;
172 else if (!strncasecmp (s, "<br>", 4)) {
173 sami_add_line(current, text, &p);
174 s += 4;
176 else if ((*s == '{') && !sub_no_text_pp) { state = 5; ++s; continue; }
177 else if (*s == '<') { state = 4; }
178 else if (!strncasecmp (s, "&nbsp;", 6)) { *p++ = ' '; s += 6; }
179 else if (*s == '\t') { *p++ = ' '; s++; }
180 else if (*s == '\r' || *s == '\n') { s++; }
181 else *p++ = *s++;
183 /* skip duplicated space */
184 if (p > text + 2) if (*(p-1) == ' ' && *(p-2) == ' ') p--;
186 continue;
188 case 4: /* get current->end or skip <TAG> */
189 q = stristr (s, "Start=");
190 if (q) {
191 current->end = strtol (q + 6, &q, 0) / 10 - 1;
192 *p = '\0'; trail_space (text);
193 if (text[0] != '\0')
194 current->text[current->lines++] = strdup (text);
195 if (current->lines > 0) { state = 99; break; }
196 state = 0; continue;
198 s = strchr (s, '>');
199 if (s) { s++; state = 3; continue; }
200 break;
201 case 5: /* get rid of {...} text, but read the alignment code */
202 if ((*s == '\\') && (*(s + 1) == 'a') && !sub_no_text_pp) {
203 if (stristr(s, "\\a1") != NULL) {
204 current->alignment = SUB_ALIGNMENT_BOTTOMLEFT;
205 s = s + 3;
207 if (stristr(s, "\\a2") != NULL) {
208 current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
209 s = s + 3;
210 } else if (stristr(s, "\\a3") != NULL) {
211 current->alignment = SUB_ALIGNMENT_BOTTOMRIGHT;
212 s = s + 3;
213 } else if ((stristr(s, "\\a4") != NULL) || (stristr(s, "\\a5") != NULL) || (stristr(s, "\\a8") != NULL)) {
214 current->alignment = SUB_ALIGNMENT_TOPLEFT;
215 s = s + 3;
216 } else if (stristr(s, "\\a6") != NULL) {
217 current->alignment = SUB_ALIGNMENT_TOPCENTER;
218 s = s + 3;
219 } else if (stristr(s, "\\a7") != NULL) {
220 current->alignment = SUB_ALIGNMENT_TOPRIGHT;
221 s = s + 3;
222 } else if (stristr(s, "\\a9") != NULL) {
223 current->alignment = SUB_ALIGNMENT_MIDDLELEFT;
224 s = s + 3;
225 } else if (stristr(s, "\\a10") != NULL) {
226 current->alignment = SUB_ALIGNMENT_MIDDLECENTER;
227 s = s + 4;
228 } else if (stristr(s, "\\a11") != NULL) {
229 current->alignment = SUB_ALIGNMENT_MIDDLERIGHT;
230 s = s + 4;
233 if (*s == '}') state = 3;
234 ++s;
235 continue;
238 /* read next line */
239 if (state != 99 && !(s = stream_read_line (st, line, LINE_LEN, utf16))) {
240 if (current->start > 0) {
241 break; // if it is the last subtitle
242 } else {
243 return 0;
247 } while (state != 99);
249 // For the last subtitle
250 if (current->end <= 0) {
251 current->end = current->start + sub_slacktime;
252 sami_add_line(current, text, &p);
255 return current;
259 static char *sub_readtext(char *source, char **dest) {
260 int len=0;
261 char *p=source;
263 // printf("src=%p dest=%p \n",source,dest);
265 while ( !eol(*p) && *p!= '|' ) {
266 p++,len++;
269 *dest= malloc (len+1);
270 if (!dest) {return ERR;}
272 strncpy(*dest, source, len);
273 (*dest)[len]=0;
275 while (*p=='\r' || *p=='\n' || *p=='|') p++;
277 if (*p) return p; // not-last text field
278 else return NULL; // last text field
281 static subtitle *sub_read_line_microdvd(stream_t *st,subtitle *current, int utf16) {
282 char line[LINE_LEN+1];
283 char line2[LINE_LEN+1];
284 char *p, *next;
285 int i;
287 do {
288 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
289 } while ((sscanf (line,
290 "{%ld}{}%[^\r\n]",
291 &(current->start), line2) < 2) &&
292 (sscanf (line,
293 "{%ld}{%ld}%[^\r\n]",
294 &(current->start), &(current->end), line2) < 3));
296 p=line2;
298 next=p, i=0;
299 while ((next =sub_readtext (next, &(current->text[i])))) {
300 if (current->text[i]==ERR) {return ERR;}
301 i++;
302 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
304 current->lines= ++i;
306 return current;
309 static subtitle *sub_read_line_mpl2(stream_t *st,subtitle *current, int utf16) {
310 char line[LINE_LEN+1];
311 char line2[LINE_LEN+1];
312 char *p, *next;
313 int i;
315 do {
316 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
317 } while ((sscanf (line,
318 "[%ld][%ld]%[^\r\n]",
319 &(current->start), &(current->end), line2) < 3));
320 current->start *= 10;
321 current->end *= 10;
322 p=line2;
324 next=p, i=0;
325 while ((next =sub_readtext (next, &(current->text[i])))) {
326 if (current->text[i]==ERR) {return ERR;}
327 i++;
328 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
330 current->lines= ++i;
332 return current;
335 static subtitle *sub_read_line_subrip(stream_t* st, subtitle *current, int utf16) {
336 char line[LINE_LEN+1];
337 int a1,a2,a3,a4,b1,b2,b3,b4;
338 char *p=NULL, *q=NULL;
339 int len;
341 while (1) {
342 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
343 if (sscanf (line, "%d:%d:%d.%d,%d:%d:%d.%d",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4) < 8) continue;
344 current->start = a1*360000+a2*6000+a3*100+a4;
345 current->end = b1*360000+b2*6000+b3*100+b4;
347 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
349 p=q=line;
350 for (current->lines=1; current->lines < SUB_MAX_TEXT; current->lines++) {
351 for (q=p,len=0; *p && *p!='\r' && *p!='\n' && *p!='|' && strncmp(p,"[br]",4); p++,len++);
352 current->text[current->lines-1]=malloc (len+1);
353 if (!current->text[current->lines-1]) return ERR;
354 strncpy (current->text[current->lines-1], q, len);
355 current->text[current->lines-1][len]='\0';
356 if (!*p || *p=='\r' || *p=='\n') break;
357 if (*p=='|') p++;
358 else while (*p++!=']');
360 break;
362 return current;
365 static subtitle *sub_read_line_subviewer(stream_t *st,subtitle *current, int utf16) {
366 char line[LINE_LEN+1];
367 int a1,a2,a3,a4,b1,b2,b3,b4;
368 char *p=NULL;
369 int i,len;
371 while (!current->text[0]) {
372 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
373 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)
374 continue;
375 current->start = a1*360000+a2*6000+a3*100+a4/10;
376 current->end = b1*360000+b2*6000+b3*100+b4/10;
377 for (i=0; i<SUB_MAX_TEXT;) {
378 int blank = 1;
379 if (!stream_read_line (st, line, LINE_LEN, utf16)) break;
380 len=0;
381 for (p=line; *p!='\n' && *p!='\r' && *p; p++,len++)
382 if (*p != ' ' && *p != '\t')
383 blank = 0;
384 if (len && !blank) {
385 int j=0,skip=0;
386 char *curptr=current->text[i]=malloc (len+1);
387 if (!current->text[i]) return ERR;
388 //strncpy (current->text[i], line, len); current->text[i][len]='\0';
389 for(; j<len; j++) {
390 /* let's filter html tags ::atmos */
391 if(line[j]=='>') {
392 skip=0;
393 continue;
395 if(line[j]=='<') {
396 skip=1;
397 continue;
399 if(skip) {
400 continue;
402 *curptr=line[j];
403 curptr++;
405 *curptr='\0';
407 i++;
408 } else {
409 break;
412 current->lines=i;
414 return current;
417 static subtitle *sub_read_line_subviewer2(stream_t *st,subtitle *current, int utf16) {
418 char line[LINE_LEN+1];
419 int a1,a2,a3,a4;
420 char *p=NULL;
421 int i,len;
423 while (!current->text[0]) {
424 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
425 if (line[0]!='{')
426 continue;
427 if ((len=sscanf (line, "{T %d:%d:%d:%d",&a1,&a2,&a3,&a4)) < 4)
428 continue;
429 current->start = a1*360000+a2*6000+a3*100+a4/10;
430 for (i=0; i<SUB_MAX_TEXT;) {
431 if (!stream_read_line (st, line, LINE_LEN, utf16)) break;
432 if (line[0]=='}') break;
433 len=0;
434 for (p=line; *p!='\n' && *p!='\r' && *p; ++p,++len);
435 if (len) {
436 current->text[i]=malloc (len+1);
437 if (!current->text[i]) return ERR;
438 strncpy (current->text[i], line, len); current->text[i][len]='\0';
439 ++i;
440 } else {
441 break;
444 current->lines=i;
446 return current;
450 static subtitle *sub_read_line_vplayer(stream_t *st,subtitle *current, int utf16) {
451 char line[LINE_LEN+1];
452 int a1,a2,a3;
453 char *p=NULL, *next,separator;
454 int i,len,plen;
456 while (!current->text[0]) {
457 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
458 if ((len=sscanf (line, "%d:%d:%d%c%n",&a1,&a2,&a3,&separator,&plen)) < 4)
459 continue;
461 if (!(current->start = a1*360000+a2*6000+a3*100))
462 continue;
463 /* removed by wodzu
464 p=line;
465 // finds the body of the subtitle
466 for (i=0; i<3; i++){
467 p=strchr(p,':');
468 if (p==NULL) break;
469 ++p;
471 if (p==NULL) {
472 printf("SUB: Skipping incorrect subtitle line!\n");
473 continue;
476 // by wodzu: hey! this time we know what length it has! what is
477 // that magic for? it can't deal with space instead of third
478 // colon! look, what simple it can be:
479 p = &line[ plen ];
481 i=0;
482 if (*p!='|') {
484 next = p,i=0;
485 while ((next =sub_readtext (next, &(current->text[i])))) {
486 if (current->text[i]==ERR) {return ERR;}
487 i++;
488 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
490 current->lines=i+1;
493 return current;
496 static subtitle *sub_read_line_rt(stream_t *st,subtitle *current, int utf16) {
497 //TODO: This format uses quite rich (sub/super)set of xhtml
498 // I couldn't check it since DTD is not included.
499 // WARNING: full XML parses can be required for proper parsing
500 char line[LINE_LEN+1];
501 int a1,a2,a3,a4,b1,b2,b3,b4;
502 char *p=NULL,*next=NULL;
503 int i,len,plen;
505 while (!current->text[0]) {
506 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
507 //TODO: it seems that format of time is not easily determined, it may be 1:12, 1:12.0 or 0:1:12.0
508 //to describe the same moment in time. Maybe there are even more formats in use.
509 //if ((len=sscanf (line, "<Time Begin=\"%d:%d:%d.%d\" End=\"%d:%d:%d.%d\"",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4)) < 8)
510 plen=a1=a2=a3=a4=b1=b2=b3=b4=0;
511 if (
512 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&b3,&b4,&plen)) < 4) &&
513 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&b2,&b3,&b4,&plen)) < 5) &&
514 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&b2,&b3,&plen)) < 4) &&
515 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2,&a3,&b2,&b3,&b4,&plen)) < 5) &&
516 // ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&b2,&b3,&plen)) < 5) &&
517 ((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) &&
518 ((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) &&
519 //now try it without end time
520 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&plen)) < 2) &&
521 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&plen)) < 2) &&
522 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&plen)) < 3) &&
523 ((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d:%d.%d\"%*[^<]<clear/>%n",&a1,&a2,&a3,&a4,&plen)) < 4)
525 continue;
526 current->start = a1*360000+a2*6000+a3*100+a4/10;
527 current->end = b1*360000+b2*6000+b3*100+b4/10;
528 if (b1 == 0 && b2 == 0 && b3 == 0 && b4 == 0)
529 current->end = current->start+200;
530 p=line; p+=plen;i=0;
531 // TODO: I don't know what kind of convention is here for marking multiline subs, maybe <br/> like in xml?
532 next = strstr(line,"<clear/>");
533 if(next && strlen(next)>8){
534 next+=8;i=0;
535 while ((next =sub_readtext (next, &(current->text[i])))) {
536 if (current->text[i]==ERR) {return ERR;}
537 i++;
538 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
541 current->lines=i+1;
543 return current;
546 static subtitle *sub_read_line_ssa(stream_t *st,subtitle *current, int utf16) {
548 * Sub Station Alpha v4 (and v2?) scripts have 9 commas before subtitle
549 * other Sub Station Alpha scripts have only 8 commas before subtitle
550 * Reading the "ScriptType:" field is not reliable since many scripts appear
551 * w/o it
553 * http://www.scriptclub.org is a good place to find more examples
554 * http://www.eswat.demon.co.uk is where the SSA specs can be found
556 int comma;
557 static int max_comma = 32; /* let's use 32 for the case that the */
558 /* amount of commas increase with newer SSA versions */
560 int hour1, min1, sec1, hunsec1,
561 hour2, min2, sec2, hunsec2, nothing;
562 int num;
564 char line[LINE_LEN+1],
565 line3[LINE_LEN+1],
566 *line2;
567 char *tmp;
569 do {
570 if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
571 } while (sscanf (line, "Dialogue: Marked=%d,%d:%d:%d.%d,%d:%d:%d.%d"
572 "%[^\n\r]", &nothing,
573 &hour1, &min1, &sec1, &hunsec1,
574 &hour2, &min2, &sec2, &hunsec2,
575 line3) < 9
577 sscanf (line, "Dialogue: %d,%d:%d:%d.%d,%d:%d:%d.%d"
578 "%[^\n\r]", &nothing,
579 &hour1, &min1, &sec1, &hunsec1,
580 &hour2, &min2, &sec2, &hunsec2,
581 line3) < 9 );
583 line2=strchr(line3, ',');
584 if (!line2) return NULL;
586 for (comma = 4; comma < max_comma; comma ++)
588 tmp = line2;
589 if(!(tmp=strchr(++tmp, ','))) break;
590 if(*(++tmp) == ' ') break;
591 /* a space after a comma means we're already in a sentence */
592 line2 = tmp;
595 if(comma < max_comma)max_comma = comma;
596 /* eliminate the trailing comma */
597 if(*line2 == ',') line2++;
599 current->lines=0;num=0;
600 current->start = 360000*hour1 + 6000*min1 + 100*sec1 + hunsec1;
601 current->end = 360000*hour2 + 6000*min2 + 100*sec2 + hunsec2;
603 while (((tmp=strstr(line2, "\\n")) != NULL) || ((tmp=strstr(line2, "\\N")) != NULL) ){
604 current->text[num]=malloc(tmp-line2+1);
605 strncpy (current->text[num], line2, tmp-line2);
606 current->text[num][tmp-line2]='\0';
607 line2=tmp+2;
608 num++;
609 current->lines++;
610 if (current->lines >= SUB_MAX_TEXT) return current;
613 current->text[num]=strdup(line2);
614 current->lines++;
616 return current;
619 static void sub_pp_ssa(subtitle *sub) {
620 int l=sub->lines;
621 char *so,*de,*start;
623 while (l){
624 /* eliminate any text enclosed with {}, they are font and color settings */
625 so=de=sub->text[--l];
626 while (*so) {
627 if(*so == '{' && so[1]=='\\') {
628 for (start=so; *so && *so!='}'; so++);
629 if(*so) so++; else so=start;
631 if(*so) {
632 *de=*so;
633 so++; de++;
636 *de=*so;
641 * PJS subtitles reader.
642 * That's the "Phoenix Japanimation Society" format.
643 * I found some of them in http://www.scriptsclub.org/ (used for anime).
644 * The time is in tenths of second.
646 * by set, based on code by szabi (dunnowhat sub format ;-)
648 static subtitle *sub_read_line_pjs(stream_t *st,subtitle *current, int utf16) {
649 char line[LINE_LEN+1];
650 char text[LINE_LEN+1], *s, *d;
652 if (!stream_read_line (st, line, LINE_LEN, utf16))
653 return NULL;
654 /* skip spaces */
655 for (s=line; *s && isspace(*s); s++);
656 /* allow empty lines at the end of the file */
657 if (*s==0)
658 return NULL;
659 /* get the time */
660 if (sscanf (s, "%ld,%ld,", &(current->start),
661 &(current->end)) <2) {
662 return ERR;
664 /* the files I have are in tenths of second */
665 current->start *= 10;
666 current->end *= 10;
667 /* walk to the beggining of the string */
668 for (; *s; s++) if (*s==',') break;
669 if (*s) {
670 for (s++; *s; s++) if (*s==',') break;
671 if (*s) s++;
673 if (*s!='"') {
674 return ERR;
676 /* copy the string to the text buffer */
677 for (s++, d=text; *s && *s!='"'; s++, d++)
678 *d=*s;
679 *d=0;
680 current->text[0] = strdup(text);
681 current->lines = 1;
683 return current;
686 static subtitle *sub_read_line_mpsub(stream_t *st, subtitle *current, int utf16) {
687 char line[LINE_LEN+1];
688 float a,b;
689 int num=0;
690 char *p, *q;
694 if (!stream_read_line(st, line, LINE_LEN, utf16)) return NULL;
695 } while (sscanf (line, "%f %f", &a, &b) !=2);
697 mpsub_position += a*mpsub_multiplier;
698 current->start=(int) mpsub_position;
699 mpsub_position += b*mpsub_multiplier;
700 current->end=(int) mpsub_position;
702 while (num < SUB_MAX_TEXT) {
703 if (!stream_read_line (st, line, LINE_LEN, utf16)) {
704 if (num == 0) return NULL;
705 else return current;
707 p=line;
708 while (isspace(*p)) p++;
709 if (eol(*p) && num > 0) return current;
710 if (eol(*p)) return NULL;
712 for (q=p; !eol(*q); q++);
713 *q='\0';
714 if (strlen(p)) {
715 current->text[num]=strdup(p);
716 // printf (">%s<\n",p);
717 current->lines = ++num;
718 } else {
719 if (num) return current;
720 else return NULL;
723 return NULL; // we should have returned before if it's OK
726 #ifndef CONFIG_SORTSUB
727 //we don't need this if we use previous_sub_end
728 subtitle *previous_aqt_sub = NULL;
729 #endif
731 static subtitle *sub_read_line_aqt(stream_t *st,subtitle *current, int utf16) {
732 char line[LINE_LEN+1];
733 char *next;
734 int i;
736 while (1) {
737 // try to locate next subtitle
738 if (!stream_read_line (st, line, LINE_LEN, utf16))
739 return NULL;
740 if (!(sscanf (line, "-->> %ld", &(current->start)) <1))
741 break;
744 #ifdef CONFIG_SORTSUB
745 previous_sub_end = (current->start) ? current->start - 1 : 0;
746 #else
747 if (previous_aqt_sub != NULL)
748 previous_aqt_sub->end = current->start-1;
750 previous_aqt_sub = current;
751 #endif
753 if (!stream_read_line (st, line, LINE_LEN, utf16))
754 return NULL;
756 sub_readtext((char *) &line,&current->text[0]);
757 current->lines = 1;
758 current->end = current->start; // will be corrected by next subtitle
760 if (!stream_read_line (st, line, LINE_LEN, utf16))
761 return current;
763 next = line,i=1;
764 while ((next =sub_readtext (next, &(current->text[i])))) {
765 if (current->text[i]==ERR) {return ERR;}
766 i++;
767 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
769 current->lines=i+1;
771 if (!strlen(current->text[0]) && !strlen(current->text[1])) {
772 #ifdef CONFIG_SORTSUB
773 previous_sub_end = 0;
774 #else
775 // void subtitle -> end of previous marked and exit
776 previous_aqt_sub = NULL;
777 #endif
778 return NULL;
781 return current;
784 #ifndef CONFIG_SORTSUB
785 subtitle *previous_subrip09_sub = NULL;
786 #endif
788 static subtitle *sub_read_line_subrip09(stream_t *st,subtitle *current, int utf16) {
789 char line[LINE_LEN+1];
790 int a1,a2,a3;
791 char * next=NULL;
792 int i,len;
794 while (1) {
795 // try to locate next subtitle
796 if (!stream_read_line (st, line, LINE_LEN, utf16))
797 return NULL;
798 if (!((len=sscanf (line, "[%d:%d:%d]",&a1,&a2,&a3)) < 3))
799 break;
802 current->start = a1*360000+a2*6000+a3*100;
804 #ifdef CONFIG_SORTSUB
805 previous_sub_end = (current->start) ? current->start - 1 : 0;
806 #else
807 if (previous_subrip09_sub != NULL)
808 previous_subrip09_sub->end = current->start-1;
810 previous_subrip09_sub = current;
811 #endif
813 if (!stream_read_line (st, line, LINE_LEN, utf16))
814 return NULL;
816 next = line,i=0;
818 current->text[0]=""; // just to be sure that string is clear
820 while ((next =sub_readtext (next, &(current->text[i])))) {
821 if (current->text[i]==ERR) {return ERR;}
822 i++;
823 if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
825 current->lines=i+1;
827 if (!strlen(current->text[0]) && (i==0)) {
828 #ifdef CONFIG_SORTSUB
829 previous_sub_end = 0;
830 #else
831 // void subtitle -> end of previous marked and exit
832 previous_subrip09_sub = NULL;
833 #endif
834 return NULL;
837 return current;
840 static subtitle *sub_read_line_jacosub(stream_t* st, subtitle * current, int utf16)
842 char line1[LINE_LEN], line2[LINE_LEN], directive[LINE_LEN], *p, *q;
843 unsigned a1, a2, a3, a4, b1, b2, b3, b4, comment = 0;
844 static unsigned jacoTimeres = 30;
845 static int jacoShift = 0;
847 memset(current, 0, sizeof(subtitle));
848 memset(line1, 0, LINE_LEN);
849 memset(line2, 0, LINE_LEN);
850 memset(directive, 0, LINE_LEN);
851 while (!current->text[0]) {
852 if (!stream_read_line(st, line1, LINE_LEN, utf16)) {
853 return NULL;
855 if (sscanf
856 (line1, "%u:%u:%u.%u %u:%u:%u.%u %[^\n\r]", &a1, &a2, &a3, &a4,
857 &b1, &b2, &b3, &b4, line2) < 9) {
858 if (sscanf(line1, "@%u @%u %[^\n\r]", &a4, &b4, line2) < 3) {
859 if (line1[0] == '#') {
860 int hours = 0, minutes = 0, seconds, delta, inverter =
862 unsigned units = jacoShift;
863 switch (toupper(line1[1])) {
864 case 'S':
865 if (isalpha(line1[2])) {
866 delta = 6;
867 } else {
868 delta = 2;
870 if (sscanf(&line1[delta], "%d", &hours)) {
871 if (hours < 0) {
872 hours *= -1;
873 inverter = -1;
875 if (sscanf(&line1[delta], "%*d:%d", &minutes)) {
876 if (sscanf
877 (&line1[delta], "%*d:%*d:%d",
878 &seconds)) {
879 sscanf(&line1[delta], "%*d:%*d:%*d.%d",
880 &units);
881 } else {
882 hours = 0;
883 sscanf(&line1[delta], "%d:%d.%d",
884 &minutes, &seconds, &units);
885 minutes *= inverter;
887 } else {
888 hours = minutes = 0;
889 sscanf(&line1[delta], "%d.%d", &seconds,
890 &units);
891 seconds *= inverter;
893 jacoShift =
894 ((hours * 3600 + minutes * 60 +
895 seconds) * jacoTimeres +
896 units) * inverter;
898 break;
899 case 'T':
900 if (isalpha(line1[2])) {
901 delta = 8;
902 } else {
903 delta = 2;
905 sscanf(&line1[delta], "%u", &jacoTimeres);
906 break;
909 continue;
910 } else {
911 current->start =
912 (unsigned long) ((a4 + jacoShift) * 100.0 /
913 jacoTimeres);
914 current->end =
915 (unsigned long) ((b4 + jacoShift) * 100.0 /
916 jacoTimeres);
918 } else {
919 current->start =
920 (unsigned
921 long) (((a1 * 3600 + a2 * 60 + a3) * jacoTimeres + a4 +
922 jacoShift) * 100.0 / jacoTimeres);
923 current->end =
924 (unsigned
925 long) (((b1 * 3600 + b2 * 60 + b3) * jacoTimeres + b4 +
926 jacoShift) * 100.0 / jacoTimeres);
928 current->lines = 0;
929 p = line2;
930 while ((*p == ' ') || (*p == '\t')) {
931 ++p;
933 if (isalpha(*p)||*p == '[') {
934 int cont, jLength;
936 if (sscanf(p, "%s %[^\n\r]", directive, line1) < 2)
937 return (subtitle *) ERR;
938 jLength = strlen(directive);
939 for (cont = 0; cont < jLength; ++cont) {
940 if (isalpha(*(directive + cont)))
941 *(directive + cont) = toupper(*(directive + cont));
943 if ((strstr(directive, "RDB") != NULL)
944 || (strstr(directive, "RDC") != NULL)
945 || (strstr(directive, "RLB") != NULL)
946 || (strstr(directive, "RLG") != NULL)) {
947 continue;
949 if (strstr(directive, "JL") != NULL) {
950 current->alignment = SUB_ALIGNMENT_BOTTOMLEFT;
951 } else if (strstr(directive, "JR") != NULL) {
952 current->alignment = SUB_ALIGNMENT_BOTTOMRIGHT;
953 } else {
954 current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
956 strcpy(line2, line1);
957 p = line2;
959 for (q = line1; (!eol(*p)) && (current->lines < SUB_MAX_TEXT); ++p) {
960 switch (*p) {
961 case '{':
962 comment++;
963 break;
964 case '}':
965 if (comment) {
966 --comment;
967 //the next line to get rid of a blank after the comment
968 if ((*(p + 1)) == ' ')
969 p++;
971 break;
972 case '~':
973 if (!comment) {
974 *q = ' ';
975 ++q;
977 break;
978 case ' ':
979 case '\t':
980 if ((*(p + 1) == ' ') || (*(p + 1) == '\t'))
981 break;
982 if (!comment) {
983 *q = ' ';
984 ++q;
986 break;
987 case '\\':
988 if (*(p + 1) == 'n') {
989 *q = '\0';
990 q = line1;
991 current->text[current->lines++] = strdup(line1);
992 ++p;
993 break;
995 if ((toupper(*(p + 1)) == 'C')
996 || (toupper(*(p + 1)) == 'F')) {
997 ++p,++p;
998 break;
1000 if ((*(p + 1) == 'B') || (*(p + 1) == 'b') || (*(p + 1) == 'D') || //actually this means "insert current date here"
1001 (*(p + 1) == 'I') || (*(p + 1) == 'i') || (*(p + 1) == 'N') || (*(p + 1) == 'T') || //actually this means "insert current time here"
1002 (*(p + 1) == 'U') || (*(p + 1) == 'u')) {
1003 ++p;
1004 break;
1006 if ((*(p + 1) == '\\') ||
1007 (*(p + 1) == '~') || (*(p + 1) == '{')) {
1008 ++p;
1009 } else if (eol(*(p + 1))) {
1010 if (!stream_read_line(st, directive, LINE_LEN, utf16))
1011 return NULL;
1012 trail_space(directive);
1013 av_strlcat(line2, directive, LINE_LEN);
1014 break;
1016 default:
1017 if (!comment) {
1018 *q = *p;
1019 ++q;
1021 } //-- switch
1022 } //-- for
1023 *q = '\0';
1024 current->text[current->lines] = strdup(line1);
1025 } //-- while
1026 current->lines++;
1027 return current;
1030 static int sub_autodetect (stream_t* st, int *uses_time, int utf16) {
1031 char line[LINE_LEN+1];
1032 int i,j=0;
1034 while (j < 100) {
1035 j++;
1036 if (!stream_read_line (st, line, LINE_LEN, utf16))
1037 return SUB_INVALID;
1039 if (sscanf (line, "{%d}{%d}", &i, &i)==2)
1040 {*uses_time=0;return SUB_MICRODVD;}
1041 if (sscanf (line, "{%d}{}", &i)==1)
1042 {*uses_time=0;return SUB_MICRODVD;}
1043 if (sscanf (line, "[%d][%d]", &i, &i)==2)
1044 {*uses_time=1;return SUB_MPL2;}
1045 if (sscanf (line, "%d:%d:%d.%d,%d:%d:%d.%d", &i, &i, &i, &i, &i, &i, &i, &i)==8)
1046 {*uses_time=1;return SUB_SUBRIP;}
1047 if (sscanf (line, "%d:%d:%d%[,.:]%d --> %d:%d:%d%[,.:]%d", &i, &i, &i, (char *)&i, &i, &i, &i, &i, (char *)&i, &i)==10)
1048 {*uses_time=1;return SUB_SUBVIEWER;}
1049 if (sscanf (line, "{T %d:%d:%d:%d",&i, &i, &i, &i)==4)
1050 {*uses_time=1;return SUB_SUBVIEWER2;}
1051 if (strstr (line, "<SAMI>"))
1052 {*uses_time=1; return SUB_SAMI;}
1053 if (sscanf(line, "%d:%d:%d.%d %d:%d:%d.%d", &i, &i, &i, &i, &i, &i, &i, &i) == 8)
1054 {*uses_time = 1; return SUB_JACOSUB;}
1055 if (sscanf(line, "@%d @%d", &i, &i) == 2)
1056 {*uses_time = 1; return SUB_JACOSUB;}
1057 if (sscanf (line, "%d:%d:%d:", &i, &i, &i )==3)
1058 {*uses_time=1;return SUB_VPLAYER;}
1059 if (sscanf (line, "%d:%d:%d ", &i, &i, &i )==3)
1060 {*uses_time=1;return SUB_VPLAYER;}
1061 if (!strncasecmp(line, "<window", 7))
1062 {*uses_time=1;return SUB_RT;}
1063 if (!memcmp(line, "Dialogue: Marked", 16))
1064 {*uses_time=1; return SUB_SSA;}
1065 if (!memcmp(line, "Dialogue: ", 10))
1066 {*uses_time=1; return SUB_SSA;}
1067 if (sscanf (line, "%d,%d,\"%c", &i, &i, (char *) &i) == 3)
1068 {*uses_time=1;return SUB_PJS;}
1069 if (sscanf (line, "FORMAT=%d", &i) == 1)
1070 {*uses_time=0; return SUB_MPSUB;}
1071 if (!memcmp(line, "FORMAT=TIME", 11))
1072 {*uses_time=1; return SUB_MPSUB;}
1073 if (strstr (line, "-->>"))
1074 {*uses_time=0; return SUB_AQTITLE;}
1075 if (sscanf (line, "[%d:%d:%d]", &i, &i, &i)==3)
1076 {*uses_time=1;return SUB_SUBRIP09;}
1079 return SUB_INVALID; // too many bad lines
1082 extern int sub_utf8;
1083 int sub_utf8_prev=0;
1085 extern float sub_delay;
1086 extern float sub_fps;
1088 #ifdef CONFIG_ICONV
1089 static iconv_t icdsc = (iconv_t)(-1);
1091 void subcp_open (stream_t *st)
1093 char *tocp = "UTF-8";
1095 if (sub_cp){
1096 const char *cp_tmp = sub_cp;
1097 #ifdef CONFIG_ENCA
1098 char enca_lang[3], enca_fallback[100];
1099 if (sscanf(sub_cp, "enca:%2s:%99s", enca_lang, enca_fallback) == 2
1100 || sscanf(sub_cp, "ENCA:%2s:%99s", enca_lang, enca_fallback) == 2) {
1101 if (st && st->flags & MP_STREAM_SEEK ) {
1102 cp_tmp = guess_cp(st, enca_lang, enca_fallback);
1103 } else {
1104 cp_tmp = enca_fallback;
1105 if (st)
1106 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: enca failed, stream must be seekable.\n");
1109 #endif
1110 if ((icdsc = iconv_open (tocp, cp_tmp)) != (iconv_t)(-1)){
1111 mp_msg(MSGT_SUBREADER,MSGL_V,"SUB: opened iconv descriptor.\n");
1112 sub_utf8 = 2;
1113 } else
1114 mp_msg(MSGT_SUBREADER,MSGL_ERR,"SUB: error opening iconv descriptor.\n");
1118 void subcp_close (void)
1120 if (icdsc != (iconv_t)(-1)){
1121 (void) iconv_close (icdsc);
1122 icdsc = (iconv_t)(-1);
1123 mp_msg(MSGT_SUBREADER,MSGL_V,"SUB: closed iconv descriptor.\n");
1127 subtitle* subcp_recode (subtitle *sub)
1129 int l=sub->lines;
1130 size_t ileft, oleft;
1131 char *op, *ip, *ot;
1132 if(icdsc == (iconv_t)(-1)) return sub;
1134 while (l){
1135 ip = sub->text[--l];
1136 ileft = strlen(ip);
1137 oleft = 4 * ileft;
1139 if (!(ot = malloc(oleft + 1))){
1140 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error allocating mem.\n");
1141 continue;
1143 op = ot;
1144 if (iconv(icdsc, &ip, &ileft,
1145 &op, &oleft) == (size_t)(-1)) {
1146 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error recoding line.\n");
1147 free(ot);
1148 continue;
1150 // In some stateful encodings, we must clear the state to handle the last character
1151 if (iconv(icdsc, NULL, NULL,
1152 &op, &oleft) == (size_t)(-1)) {
1153 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error recoding line, can't clear encoding state.\n");
1155 *op='\0' ;
1156 free (sub->text[l]);
1157 sub->text[l] = ot;
1159 return sub;
1161 #endif
1163 #ifdef CONFIG_FRIBIDI
1165 * Do conversion necessary for right-to-left language support via fribidi.
1166 * @param sub subtitle to convert
1167 * @param sub_utf8 whether the subtitle is encoded in UTF-8
1168 * @param from first new subtitle, all lines before this are assumed to be already converted
1170 static subtitle* sub_fribidi (subtitle *sub, int sub_utf8, int from)
1172 FriBidiChar logical[LINE_LEN+1], visual[LINE_LEN+1]; // Hopefully these two won't smash the stack
1173 char *ip = NULL, *op = NULL;
1174 size_t len,orig_len;
1175 int l=sub->lines;
1176 int char_set_num;
1177 fribidi_boolean log2vis;
1178 if (!flip_hebrew)
1179 return sub;
1180 fribidi_set_mirroring(1);
1181 fribidi_set_reorder_nsm(0);
1183 if( sub_utf8 == 0 ) {
1184 char_set_num = fribidi_parse_charset (fribidi_charset?fribidi_charset:"ISO8859-8");
1185 }else {
1186 char_set_num = fribidi_parse_charset ("UTF-8");
1188 while (l > from) {
1189 ip = sub->text[--l];
1190 orig_len = len = strlen( ip ); // We assume that we don't use full unicode, only UTF-8 or ISO8859-x
1191 if(len > LINE_LEN) {
1192 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: sub->text is longer than LINE_LEN.\n");
1193 l++;
1194 break;
1196 len = fribidi_charset_to_unicode (char_set_num, ip, len, logical);
1197 #if FRIBIDI_INTERFACE_VERSION < 3
1198 FriBidiCharType base = fribidi_flip_commas?FRIBIDI_TYPE_ON:FRIBIDI_TYPE_L;
1199 #else
1200 FriBidiParType base = fribidi_flip_commas?FRIBIDI_TYPE_ON:FRIBIDI_TYPE_L;
1201 #endif
1202 log2vis = fribidi_log2vis (logical, len, &base,
1203 /* output */
1204 visual, NULL, NULL, NULL);
1205 if(log2vis) {
1206 len = fribidi_remove_bidi_marks (visual, len, NULL, NULL,
1207 NULL);
1208 if((op = malloc((FFMAX(2*orig_len,2*len) + 1))) == NULL) {
1209 mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: error allocating mem.\n");
1210 l++;
1211 break;
1213 fribidi_unicode_to_charset ( char_set_num, visual, len,op);
1214 free (ip);
1215 sub->text[l] = op;
1218 if (!from && l){
1219 for (l = sub->lines; l;)
1220 free (sub->text[--l]);
1221 return ERR;
1223 return sub;
1226 #endif
1228 static void adjust_subs_time(subtitle* sub, float subtime, float fps, int block,
1229 int sub_num, int sub_uses_time) {
1230 int n,m;
1231 subtitle* nextsub;
1232 int i = sub_num;
1233 unsigned long subfms = (sub_uses_time ? 100 : fps) * subtime;
1234 unsigned long overlap = (sub_uses_time ? 100 : fps) / 5; // 0.2s
1236 n=m=0;
1237 if (i) for (;;){
1238 if (sub->end <= sub->start){
1239 sub->end = sub->start + subfms;
1240 m++;
1241 n++;
1243 if (!--i) break;
1244 nextsub = sub + 1;
1245 if(block){
1246 if ((sub->end > nextsub->start) && (sub->end <= nextsub->start + overlap)) {
1247 // these subtitles overlap for less than 0.2 seconds
1248 // and would result in very short overlapping subtitle
1249 // so let's fix the problem here, before overlapping code
1250 // get its hands on them
1251 unsigned delta = sub->end - nextsub->start, half = delta / 2;
1252 sub->end -= half + 1;
1253 nextsub->start += delta - half;
1255 if (sub->end >= nextsub->start){
1256 sub->end = nextsub->start - 1;
1257 if (sub->end - sub->start > subfms)
1258 sub->end = sub->start + subfms;
1259 if (!m)
1260 n++;
1264 /* Theory:
1265 * Movies are often converted from FILM (24 fps)
1266 * to PAL (25) by simply speeding it up, so we
1267 * to multiply the original timestmaps by
1268 * (Movie's FPS / Subtitle's (guessed) FPS)
1269 * so eg. for 23.98 fps movie and PAL time based
1270 * subtitles we say -subfps 25 and we're fine!
1273 /* timed sub fps correction ::atmos */
1274 /* the frame-based case is handled in mpcommon.c
1275 * where find_sub is called */
1276 if(sub_uses_time && sub_fps) {
1277 sub->start *= sub_fps/fps;
1278 sub->end *= sub_fps/fps;
1281 sub = nextsub;
1282 m = 0;
1284 if (n) mp_msg(MSGT_SUBREADER,MSGL_V,"SUB: Adjusted %d subtitle(s).\n", n);
1287 struct subreader {
1288 subtitle * (*read)(stream_t *st,subtitle *dest,int utf16);
1289 void (*post)(subtitle *dest);
1290 const char *name;
1293 #ifdef CONFIG_ENCA
1294 const char* guess_buffer_cp(unsigned char* buffer, int buflen, const char *preferred_language, const char *fallback)
1296 const char **languages;
1297 size_t langcnt;
1298 EncaAnalyser analyser;
1299 EncaEncoding encoding;
1300 const char *detected_sub_cp = NULL;
1301 int i;
1303 languages = enca_get_languages(&langcnt);
1304 mp_msg(MSGT_SUBREADER, MSGL_V, "ENCA supported languages: ");
1305 for (i = 0; i < langcnt; i++) {
1306 mp_msg(MSGT_SUBREADER, MSGL_V, "%s ", languages[i]);
1308 mp_msg(MSGT_SUBREADER, MSGL_V, "\n");
1310 for (i = 0; i < langcnt; i++) {
1311 if (strcasecmp(languages[i], preferred_language) != 0) continue;
1312 analyser = enca_analyser_alloc(languages[i]);
1313 encoding = enca_analyse_const(analyser, buffer, buflen);
1314 enca_analyser_free(analyser);
1315 if (encoding.charset != ENCA_CS_UNKNOWN) {
1316 detected_sub_cp = enca_charset_name(encoding.charset, ENCA_NAME_STYLE_ICONV);
1317 break;
1321 free(languages);
1323 if (!detected_sub_cp) {
1324 detected_sub_cp = fallback;
1325 mp_msg(MSGT_SUBREADER, MSGL_INFO, "ENCA detection failed: fallback to %s\n", fallback);
1326 }else{
1327 mp_msg(MSGT_SUBREADER, MSGL_INFO, "ENCA detected charset: %s\n", detected_sub_cp);
1330 return detected_sub_cp;
1333 #define MAX_GUESS_BUFFER_SIZE (256*1024)
1334 const char* guess_cp(stream_t *st, const char *preferred_language, const char *fallback)
1336 size_t buflen;
1337 unsigned char *buffer;
1338 const char *detected_sub_cp = NULL;
1340 buffer = malloc(MAX_GUESS_BUFFER_SIZE);
1341 buflen = stream_read(st,buffer, MAX_GUESS_BUFFER_SIZE);
1343 detected_sub_cp = guess_buffer_cp(buffer, buflen, preferred_language, fallback);
1345 free(buffer);
1346 stream_reset(st);
1347 stream_seek(st,0);
1349 return detected_sub_cp;
1351 #undef MAX_GUESS_BUFFER_SIZE
1352 #endif
1354 sub_data* sub_read_file (char *filename, float fps) {
1355 int utf16;
1356 stream_t* fd;
1357 int n_max, n_first, i, j, sub_first, sub_orig;
1358 subtitle *first, *second, *sub, *return_sub, *alloced_sub = NULL;
1359 sub_data *subt_data;
1360 int uses_time = 0, sub_num = 0, sub_errs = 0;
1361 static const struct subreader sr[]=
1363 { sub_read_line_microdvd, NULL, "microdvd" },
1364 { sub_read_line_subrip, NULL, "subrip" },
1365 { sub_read_line_subviewer, NULL, "subviewer" },
1366 { sub_read_line_sami, NULL, "sami" },
1367 { sub_read_line_vplayer, NULL, "vplayer" },
1368 { sub_read_line_rt, NULL, "rt" },
1369 { sub_read_line_ssa, sub_pp_ssa, "ssa" },
1370 { sub_read_line_pjs, NULL, "pjs" },
1371 { sub_read_line_mpsub, NULL, "mpsub" },
1372 { sub_read_line_aqt, NULL, "aqt" },
1373 { sub_read_line_subviewer2, NULL, "subviewer 2.0" },
1374 { sub_read_line_subrip09, NULL, "subrip 0.9" },
1375 { sub_read_line_jacosub, NULL, "jacosub" },
1376 { sub_read_line_mpl2, NULL, "mpl2" }
1378 const struct subreader *srp;
1380 if(filename==NULL) return NULL; //qnx segfault
1381 fd=open_stream (filename, NULL, NULL); if (!fd) return NULL;
1383 sub_format = SUB_INVALID;
1384 for (utf16 = 0; sub_format == SUB_INVALID && utf16 < 3; utf16++) {
1385 sub_format=sub_autodetect (fd, &uses_time, utf16);
1386 stream_reset(fd);
1387 stream_seek(fd,0);
1389 utf16--;
1391 mpsub_multiplier = (uses_time ? 100.0 : 1.0);
1392 if (sub_format==SUB_INVALID) {mp_msg(MSGT_SUBREADER,MSGL_WARN,"SUB: Could not determine file format\n");return NULL;}
1393 srp=sr+sub_format;
1394 mp_msg(MSGT_SUBREADER, MSGL_V, "SUB: Detected subtitle file format: %s\n", srp->name);
1396 #ifdef CONFIG_ICONV
1397 sub_utf8_prev=sub_utf8;
1399 int l,k;
1400 k = -1;
1401 if ((l=strlen(filename))>4){
1402 char *exts[] = {".utf", ".utf8", ".utf-8" };
1403 for (k=3;--k>=0;)
1404 if (l >= strlen(exts[k]) && !strcasecmp(filename+(l - strlen(exts[k])), exts[k])){
1405 sub_utf8 = 1;
1406 break;
1409 if (k<0) subcp_open(fd);
1411 #endif
1413 sub_num=0;n_max=32;
1414 first=malloc(n_max*sizeof(subtitle));
1415 if(!first){
1416 #ifdef CONFIG_ICONV
1417 subcp_close();
1418 sub_utf8=sub_utf8_prev;
1419 #endif
1420 return NULL;
1423 #ifdef CONFIG_SORTSUB
1424 alloced_sub =
1425 sub = malloc(sizeof(subtitle));
1426 //This is to deal with those formats (AQT & Subrip) which define the end of a subtitle
1427 //as the beginning of the following
1428 previous_sub_end = 0;
1429 #endif
1430 while(1){
1431 if(sub_num>=n_max){
1432 n_max+=16;
1433 first=realloc(first,n_max*sizeof(subtitle));
1435 #ifndef CONFIG_SORTSUB
1436 sub = &first[sub_num];
1437 #endif
1438 memset(sub, '\0', sizeof(subtitle));
1439 sub=srp->read(fd,sub,utf16);
1440 if(!sub) break; // EOF
1441 #ifdef CONFIG_ICONV
1442 if ((sub!=ERR) && sub_utf8 == 2) sub=subcp_recode(sub);
1443 #endif
1444 #ifdef CONFIG_FRIBIDI
1445 if (sub!=ERR) sub=sub_fribidi(sub,sub_utf8,0);
1446 #endif
1447 if ( sub == ERR )
1449 #ifdef CONFIG_ICONV
1450 subcp_close();
1451 #endif
1452 free(first);
1453 free(alloced_sub);
1454 return NULL;
1456 // Apply any post processing that needs recoding first
1457 if ((sub!=ERR) && !sub_no_text_pp && srp->post) srp->post(sub);
1458 #ifdef CONFIG_SORTSUB
1459 if(!sub_num || (first[sub_num - 1].start <= sub->start)){
1460 first[sub_num].start = sub->start;
1461 first[sub_num].end = sub->end;
1462 first[sub_num].lines = sub->lines;
1463 first[sub_num].alignment = sub->alignment;
1464 for(i = 0; i < sub->lines; ++i){
1465 first[sub_num].text[i] = sub->text[i];
1467 if (previous_sub_end){
1468 first[sub_num - 1].end = previous_sub_end;
1469 previous_sub_end = 0;
1471 } else {
1472 for(j = sub_num - 1; j >= 0; --j){
1473 first[j + 1].start = first[j].start;
1474 first[j + 1].end = first[j].end;
1475 first[j + 1].lines = first[j].lines;
1476 first[j + 1].alignment = first[j].alignment;
1477 for(i = 0; i < first[j].lines; ++i){
1478 first[j + 1].text[i] = first[j].text[i];
1480 if(!j || (first[j - 1].start <= sub->start)){
1481 first[j].start = sub->start;
1482 first[j].end = sub->end;
1483 first[j].lines = sub->lines;
1484 first[j].alignment = sub->alignment;
1485 for(i = 0; i < SUB_MAX_TEXT; ++i){
1486 first[j].text[i] = sub->text[i];
1488 if (previous_sub_end){
1489 first[j].end = first[j - 1].end;
1490 first[j - 1].end = previous_sub_end;
1491 previous_sub_end = 0;
1493 break;
1497 #endif
1498 if(sub==ERR) ++sub_errs; else ++sub_num; // Error vs. Valid
1501 free_stream(fd);
1503 #ifdef CONFIG_ICONV
1504 subcp_close();
1505 #endif
1506 free(alloced_sub);
1508 // printf ("SUB: Subtitle format %s time.\n", uses_time?"uses":"doesn't use");
1509 mp_msg(MSGT_SUBREADER, MSGL_V,"SUB: Read %i subtitles, %i bad line(s).\n",
1510 sub_num, sub_errs);
1512 if(sub_num<=0){
1513 free(first);
1514 return NULL;
1517 // we do overlap if the user forced it (suboverlap_enable == 2) or
1518 // the user didn't forced no-overlapsub and the format is Jacosub or Ssa.
1519 // this is because usually overlapping subtitles are found in these formats,
1520 // while in others they are probably result of bad timing
1521 if ((suboverlap_enabled == 2) ||
1522 ((suboverlap_enabled) && ((sub_format == SUB_JACOSUB) || (sub_format == SUB_SSA)))) {
1523 adjust_subs_time(first, 6.0, fps, 0, sub_num, uses_time);/*~6 secs AST*/
1524 // here we manage overlapping subtitles
1525 sub_orig = sub_num;
1526 n_first = sub_num;
1527 sub_num = 0;
1528 second = NULL;
1529 // for each subtitle in first[] we deal with its 'block' of
1530 // bonded subtitles
1531 for (sub_first = 0; sub_first < n_first; ++sub_first) {
1532 unsigned long global_start = first[sub_first].start,
1533 global_end = first[sub_first].end, local_start, local_end;
1534 int lines_to_add = first[sub_first].lines, sub_to_add = 0,
1535 **placeholder = NULL, higher_line = 0, counter, start_block_sub = sub_num;
1536 char real_block = 1;
1538 // here we find the number of subtitles inside the 'block'
1539 // and its span interval. this works well only with sorted
1540 // subtitles
1541 while ((sub_first + sub_to_add + 1 < n_first) && (first[sub_first + sub_to_add + 1].start < global_end)) {
1542 ++sub_to_add;
1543 lines_to_add += first[sub_first + sub_to_add].lines;
1544 if (first[sub_first + sub_to_add].start < global_start) {
1545 global_start = first[sub_first + sub_to_add].start;
1547 if (first[sub_first + sub_to_add].end > global_end) {
1548 global_end = first[sub_first + sub_to_add].end;
1552 /* Avoid n^2 memory use for the "placeholder" data structure
1553 * below with subtitles that have a huge number of
1554 * consecutive overlapping lines. */
1555 lines_to_add = FFMIN(lines_to_add, SUB_MAX_TEXT);
1557 // we need a structure to keep trace of the screen lines
1558 // used by the subs, a 'placeholder'
1559 counter = 2 * sub_to_add + 1; // the maximum number of subs derived
1560 // from a block of sub_to_add+1 subs
1561 placeholder = malloc(sizeof(int *) * counter);
1562 for (i = 0; i < counter; ++i) {
1563 placeholder[i] = malloc(sizeof(int) * lines_to_add);
1564 for (j = 0; j < lines_to_add; ++j) {
1565 placeholder[i][j] = -1;
1569 counter = 0;
1570 local_end = global_start - 1;
1571 do {
1572 int ls;
1574 // here we find the beginning and the end of a new
1575 // subtitle in the block
1576 local_start = local_end + 1;
1577 local_end = global_end;
1578 for (j = 0; j <= sub_to_add; ++j) {
1579 if ((first[sub_first + j].start - 1 > local_start) && (first[sub_first + j].start - 1 < local_end)) {
1580 local_end = first[sub_first + j].start - 1;
1581 } else if ((first[sub_first + j].end > local_start) && (first[sub_first + j].end < local_end)) {
1582 local_end = first[sub_first + j].end;
1585 // here we allocate the screen lines to subs we must
1586 // display in current local_start-local_end interval.
1587 // if the subs were yet presents in the previous interval
1588 // they keep the same lines, otherside they get unused lines
1589 for (j = 0; j <= sub_to_add; ++j) {
1590 if ((first[sub_first + j].start <= local_end) && (first[sub_first + j].end > local_start)) {
1591 unsigned long sub_lines = first[sub_first + j].lines, fragment_length = lines_to_add + 1,
1592 tmp = 0;
1593 char boolean = 0;
1594 int fragment_position = -1;
1596 // if this is not the first new sub of the block
1597 // we find if this sub was present in the previous
1598 // new sub
1599 if (counter)
1600 for (i = 0; i < lines_to_add; ++i) {
1601 if (placeholder[counter - 1][i] == sub_first + j) {
1602 placeholder[counter][i] = sub_first + j;
1603 boolean = 1;
1606 if (boolean)
1607 continue;
1609 // we are looking for the shortest among all groups of
1610 // sequential blank lines whose length is greater than or
1611 // equal to sub_lines. we store in fragment_position the
1612 // position of the shortest group, in fragment_length its
1613 // length, and in tmp the length of the group currently
1614 // examinated
1615 for (i = 0; i < lines_to_add; ++i) {
1616 if (placeholder[counter][i] == -1) {
1617 // placeholder[counter][i] is part of the current group
1618 // of blank lines
1619 ++tmp;
1620 } else {
1621 if (tmp == sub_lines) {
1622 // current group's size fits exactly the one we
1623 // need, so we stop looking
1624 fragment_position = i - tmp;
1625 tmp = 0;
1626 break;
1628 if ((tmp) && (tmp > sub_lines) && (tmp < fragment_length)) {
1629 // current group is the best we found till here,
1630 // but is still bigger than the one we are looking
1631 // for, so we keep on looking
1632 fragment_length = tmp;
1633 fragment_position = i - tmp;
1634 tmp = 0;
1635 } else {
1636 // current group doesn't fit at all, so we forget it
1637 tmp = 0;
1641 if (tmp) {
1642 // last screen line is blank, a group ends with it
1643 if ((tmp >= sub_lines) && (tmp < fragment_length)) {
1644 fragment_position = i - tmp;
1647 if (fragment_position == -1) {
1648 // it was not possible to find free screen line(s) for a subtitle,
1649 // usually this means a bug in the code; however we do not overlap
1650 mp_msg(MSGT_SUBREADER, MSGL_WARN, "SUB: we could not find a suitable position for an overlapping subtitle\n");
1651 higher_line = SUB_MAX_TEXT + 1;
1652 break;
1653 } else {
1654 for (tmp = 0; tmp < sub_lines; ++tmp) {
1655 placeholder[counter][fragment_position + tmp] = sub_first + j;
1660 for (j = higher_line + 1; j < lines_to_add; ++j) {
1661 if (placeholder[counter][j] != -1)
1662 higher_line = j;
1663 else
1664 break;
1666 if (higher_line >= SUB_MAX_TEXT) {
1667 // the 'block' has too much lines, so we don't overlap the
1668 // subtitles
1669 second = realloc(second, (sub_num + sub_to_add + 1) * sizeof(subtitle));
1670 for (j = 0; j <= sub_to_add; ++j) {
1671 int ls;
1672 memset(&second[sub_num + j], '\0', sizeof(subtitle));
1673 second[sub_num + j].start = first[sub_first + j].start;
1674 second[sub_num + j].end = first[sub_first + j].end;
1675 second[sub_num + j].lines = first[sub_first + j].lines;
1676 second[sub_num + j].alignment = first[sub_first + j].alignment;
1677 for (ls = 0; ls < second[sub_num + j].lines; ls++) {
1678 second[sub_num + j].text[ls] = strdup(first[sub_first + j].text[ls]);
1681 sub_num += sub_to_add + 1;
1682 sub_first += sub_to_add;
1683 real_block = 0;
1684 break;
1687 // we read the placeholder structure and create the new
1688 // subs.
1689 second = realloc(second, (sub_num + 1) * sizeof(subtitle));
1690 memset(&second[sub_num], '\0', sizeof(subtitle));
1691 second[sub_num].start = local_start;
1692 second[sub_num].end = local_end;
1693 second[sub_num].alignment = first[sub_first].alignment;
1694 n_max = (lines_to_add < SUB_MAX_TEXT) ? lines_to_add : SUB_MAX_TEXT;
1695 for (i = 0, j = 0; j < n_max; ++j) {
1696 if (placeholder[counter][j] != -1) {
1697 int lines = first[placeholder[counter][j]].lines;
1698 for (ls = 0; ls < lines; ++ls) {
1699 second[sub_num].text[i++] = strdup(first[placeholder[counter][j]].text[ls]);
1701 j += lines - 1;
1702 } else {
1703 second[sub_num].text[i++] = strdup(" ");
1706 ++sub_num;
1707 ++counter;
1708 } while (local_end < global_end);
1709 if (real_block)
1710 for (i = 0; i < counter; ++i)
1711 second[start_block_sub + i].lines = higher_line + 1;
1713 counter = 2 * sub_to_add + 1;
1714 for (i = 0; i < counter; ++i) {
1715 free(placeholder[i]);
1717 free(placeholder);
1718 sub_first += sub_to_add;
1721 for (j = sub_orig - 1; j >= 0; --j) {
1722 for (i = first[j].lines - 1; i >= 0; --i) {
1723 free(first[j].text[i]);
1726 free(first);
1728 return_sub = second;
1729 } else { //if(suboverlap_enabled)
1730 adjust_subs_time(first, 6.0, fps, 1, sub_num, uses_time);/*~6 secs AST*/
1731 return_sub = first;
1733 if (return_sub == NULL) return NULL;
1734 subt_data = malloc(sizeof(sub_data));
1735 subt_data->filename = strdup(filename);
1736 subt_data->sub_uses_time = uses_time;
1737 subt_data->sub_num = sub_num;
1738 subt_data->sub_errs = sub_errs;
1739 subt_data->subtitles = return_sub;
1740 return subt_data;
1743 #if 0
1744 char * strreplace( char * in,char * what,char * whereof )
1746 int i;
1747 char * tmp;
1749 if ( ( in == NULL )||( what == NULL )||( whereof == NULL )||( ( tmp=strstr( in,what ) ) == NULL ) ) return NULL;
1750 for( i=0;i<strlen( whereof );i++ ) tmp[i]=whereof[i];
1751 if ( strlen( what ) > strlen( whereof ) ) tmp[i]=0;
1752 return in;
1754 #endif
1757 static void strcpy_trim(char *d, char *s)
1759 // skip leading whitespace
1760 while (*s && isspace(*s)) {
1761 s++;
1763 for (;;) {
1764 // copy word
1765 while (*s && !isspace(*s)) {
1766 *d = tolower(*s);
1767 s++; d++;
1769 if (*s == 0) break;
1770 // trim excess whitespace
1771 while (*s && isspace(*s)) {
1772 s++;
1774 if (*s == 0) break;
1775 *d++ = ' ';
1777 *d = 0;
1780 static void strcpy_strip_ext(char *d, char *s)
1782 char *tmp = strrchr(s,'.');
1783 if (!tmp) {
1784 strcpy(d, s);
1785 return;
1786 } else {
1787 strncpy(d, s, tmp-s);
1788 d[tmp-s] = 0;
1790 while (*d) {
1791 *d = tolower(*d);
1792 d++;
1796 static void strcpy_get_ext(char *d, char *s)
1798 char *tmp = strrchr(s,'.');
1799 if (!tmp) {
1800 strcpy(d, "");
1801 return;
1802 } else {
1803 strcpy(d, tmp+1);
1807 static int whiteonly(char *s)
1809 while (*s) {
1810 if (!isspace(*s)) return 0;
1811 s++;
1813 return 1;
1816 typedef struct subfn
1818 int priority;
1819 char *fname;
1820 } subfn;
1822 static int compare_sub_priority(const void *a, const void *b)
1824 if (((const subfn*)a)->priority > ((const subfn*)b)->priority) {
1825 return -1;
1826 } else if (((const subfn*)a)->priority < ((const subfn*)b)->priority) {
1827 return 1;
1828 } else {
1829 return strcoll(((const subfn*)a)->fname, ((const subfn*)b)->fname);
1833 char** sub_filenames(const char* path, char *fname)
1835 char *f_dir, *f_fname, *f_fname_noext, *f_fname_trim, *tmp, *tmp_sub_id;
1836 char *tmp_fname_noext, *tmp_fname_trim, *tmp_fname_ext, *tmpresult;
1838 int len, pos, found, i, j;
1839 char * sub_exts[] = { "utf", "utf8", "utf-8", "sub", "srt", "smi", "rt", "txt", "ssa", "aqt", "jss", "js", "ass", NULL};
1840 subfn *result;
1841 char **result2;
1843 int subcnt;
1845 FILE *f;
1847 DIR *d;
1848 struct dirent *de;
1850 len = (strlen(fname) > 256 ? strlen(fname) : 256)
1851 +(strlen(path) > 256 ? strlen(path) : 256)+2;
1853 f_dir = malloc(len);
1854 f_fname = malloc(len);
1855 f_fname_noext = malloc(len);
1856 f_fname_trim = malloc(len);
1858 tmp_fname_noext = malloc(len);
1859 tmp_fname_trim = malloc(len);
1860 tmp_fname_ext = malloc(len);
1862 tmpresult = malloc(len);
1864 result = calloc(MAX_SUBTITLE_FILES, sizeof(*result));
1866 subcnt = 0;
1868 tmp = strrchr(fname,'/');
1869 #if HAVE_DOS_PATHS
1870 if(!tmp)tmp = strrchr(fname,'\\');
1871 if(!tmp)tmp = strrchr(fname,':');
1872 #endif
1874 // extract filename & dirname from fname
1875 if (tmp) {
1876 strcpy(f_fname, tmp+1);
1877 pos = tmp - fname;
1878 strncpy(f_dir, fname, pos+1);
1879 f_dir[pos+1] = 0;
1880 } else {
1881 strcpy(f_fname, fname);
1882 strcpy(f_dir, "./");
1885 strcpy_strip_ext(f_fname_noext, f_fname);
1886 strcpy_trim(f_fname_trim, f_fname_noext);
1888 /* The code using sub language here is broken - it assumes strict
1889 * "videoname languagename" syntax for the subtitle file, which is
1890 * very unlikely to match especially if language name uses "en,de"
1891 * syntax... */
1892 tmp_sub_id = NULL;
1893 #if 0
1894 if (dvdsub_lang && !whiteonly(dvdsub_lang)) {
1895 tmp_sub_id = malloc(strlen(dvdsub_lang)+1);
1896 strcpy_trim(tmp_sub_id, dvdsub_lang);
1898 #endif
1900 // 0 = nothing
1901 // 1 = any subtitle file
1902 // 2 = any sub file containing movie name
1903 // 3 = sub file containing movie name and the lang extension
1904 for (j = 0; j <= 1; j++) {
1905 d = opendir(j == 0 ? f_dir : path);
1906 if (d) {
1907 while ((de = readdir(d))) {
1908 // retrieve various parts of the filename
1909 strcpy_strip_ext(tmp_fname_noext, de->d_name);
1910 strcpy_get_ext(tmp_fname_ext, de->d_name);
1911 strcpy_trim(tmp_fname_trim, tmp_fname_noext);
1913 // does it end with a subtitle extension?
1914 found = 0;
1915 #ifdef CONFIG_ICONV
1916 #ifdef CONFIG_ENCA
1917 for (i = ((sub_cp && strncasecmp(sub_cp, "enca", 4) != 0) ? 3 : 0); sub_exts[i]; i++) {
1918 #else
1919 for (i = (sub_cp ? 3 : 0); sub_exts[i]; i++) {
1920 #endif
1921 #else
1922 for (i = 0; sub_exts[i]; i++) {
1923 #endif
1924 if (strcasecmp(sub_exts[i], tmp_fname_ext) == 0) {
1925 found = 1;
1926 break;
1930 // we have a (likely) subtitle file
1931 if (found) {
1932 int prio = 0;
1933 if (!prio && tmp_sub_id)
1935 sprintf(tmpresult, "%s %s", f_fname_trim, tmp_sub_id);
1936 if (strcmp(tmp_fname_trim, tmpresult) == 0 && sub_match_fuzziness >= 1) {
1937 // matches the movie name + lang extension
1938 prio = 5;
1941 if (!prio && strcmp(tmp_fname_trim, f_fname_trim) == 0) {
1942 // matches the movie name
1943 prio = 4;
1945 if (!prio && (tmp = strstr(tmp_fname_trim, f_fname_trim)) && (sub_match_fuzziness >= 1)) {
1946 // contains the movie name
1947 tmp += strlen(f_fname_trim);
1948 if (tmp_sub_id && strstr(tmp, tmp_sub_id)) {
1949 // with sub_id specified prefer localized subtitles
1950 prio = 3;
1951 } else if ((tmp_sub_id == NULL) && whiteonly(tmp)) {
1952 // without sub_id prefer "plain" name
1953 prio = 3;
1954 } else {
1955 // with no localized subs found, try any else instead
1956 prio = 2;
1959 if (!prio) {
1960 // doesn't contain the movie name
1961 // don't try in the mplayer subtitle directory
1962 if ((j == 0) && (sub_match_fuzziness >= 2)) {
1963 prio = 1;
1967 mp_msg(MSGT_SUBREADER, MSGL_DBG2, "Potential sub file: "
1968 "\"%s\" Priority: %d\n", de->d_name, prio);
1969 if (prio) {
1970 prio += prio;
1971 #ifdef CONFIG_ICONV
1972 if (i<3){ // prefer UTF-8 coded
1973 prio++;
1975 #endif
1976 sprintf(tmpresult, "%s%s", j == 0 ? f_dir : path, de->d_name);
1977 // fprintf(stderr, "%s priority %d\n", tmpresult, prio);
1978 if ((f = fopen(tmpresult, "rt"))) {
1979 fclose(f);
1980 result[subcnt].priority = prio;
1981 result[subcnt].fname = strdup(tmpresult);
1982 subcnt++;
1987 if (subcnt >= MAX_SUBTITLE_FILES) break;
1989 closedir(d);
1994 free(tmp_sub_id);
1996 free(f_dir);
1997 free(f_fname);
1998 free(f_fname_noext);
1999 free(f_fname_trim);
2001 free(tmp_fname_noext);
2002 free(tmp_fname_trim);
2003 free(tmp_fname_ext);
2005 free(tmpresult);
2007 qsort(result, subcnt, sizeof(subfn), compare_sub_priority);
2009 result2 = calloc(subcnt + 1, sizeof(*result2));
2011 for (i = 0; i < subcnt; i++) {
2012 result2[i] = result[i].fname;
2014 result2[subcnt] = NULL;
2016 free(result);
2018 return result2;
2021 void list_sub_file(sub_data* subd){
2022 int i,j;
2023 subtitle *subs = subd->subtitles;
2025 for(j=0; j < subd->sub_num; j++){
2026 subtitle* egysub=&subs[j];
2027 mp_msg(MSGT_SUBREADER,MSGL_INFO,"%i line%c (%li-%li)\n",
2028 egysub->lines,
2029 (1==egysub->lines)?' ':'s',
2030 egysub->start,
2031 egysub->end);
2032 for (i=0; i<egysub->lines; i++) {
2033 mp_msg(MSGT_SUBREADER,MSGL_INFO,"\t\t%d: %s%s", i,egysub->text[i], i==egysub->lines-1?"":" \n ");
2035 mp_msg(MSGT_SUBREADER,MSGL_INFO,"\n");
2038 mp_msg(MSGT_SUBREADER,MSGL_INFO,"Subtitle format %s time.\n",
2039 subd->sub_uses_time ? "uses":"doesn't use");
2040 mp_msg(MSGT_SUBREADER,MSGL_INFO,"Read %i subtitles, %i errors.\n", subd->sub_num, subd->sub_errs);
2043 void dump_srt(sub_data* subd, float fps){
2044 int i,j;
2045 int h,m,s,ms;
2046 FILE * fd;
2047 subtitle * onesub;
2048 unsigned long temp;
2049 subtitle *subs = subd->subtitles;
2051 if (!subd->sub_uses_time && sub_fps == 0)
2052 sub_fps = fps;
2053 fd=fopen("dumpsub.srt","w");
2054 if(!fd)
2056 perror("dump_srt: fopen");
2057 return;
2059 for(i=0; i < subd->sub_num; i++)
2061 onesub=subs+i; //=&subs[i];
2062 fprintf(fd,"%d\n",i+1);//line number
2064 temp=onesub->start;
2065 if (!subd->sub_uses_time)
2066 temp = temp * 100 / sub_fps;
2067 temp -= sub_delay * 100;
2068 h=temp/360000;temp%=360000; //h =1*100*60*60
2069 m=temp/6000; temp%=6000; //m =1*100*60
2070 s=temp/100; temp%=100; //s =1*100
2071 ms=temp*10; //ms=1*10
2072 fprintf(fd,"%02d:%02d:%02d,%03d --> ",h,m,s,ms);
2074 temp=onesub->end;
2075 if (!subd->sub_uses_time)
2076 temp = temp * 100 / sub_fps;
2077 temp -= sub_delay * 100;
2078 h=temp/360000;temp%=360000;
2079 m=temp/6000; temp%=6000;
2080 s=temp/100; temp%=100;
2081 ms=temp*10;
2082 fprintf(fd,"%02d:%02d:%02d,%03d\n",h,m,s,ms);
2084 for(j=0;j<onesub->lines;j++)
2085 fprintf(fd,"%s\n",onesub->text[j]);
2087 fprintf(fd,"\n");
2089 fclose(fd);
2090 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.srt\'.\n");
2093 void dump_mpsub(sub_data* subd, float fps){
2094 int i,j;
2095 FILE *fd;
2096 float a,b;
2097 subtitle *subs = subd->subtitles;
2099 mpsub_position = subd->sub_uses_time? (sub_delay*100) : (sub_delay*fps);
2100 if (sub_fps==0) sub_fps=fps;
2102 fd=fopen ("dump.mpsub", "w");
2103 if (!fd) {
2104 perror ("dump_mpsub: fopen");
2105 return;
2109 if (subd->sub_uses_time) fprintf (fd,"FORMAT=TIME\n\n");
2110 else fprintf (fd, "FORMAT=%5.2f\n\n", fps);
2112 for(j=0; j < subd->sub_num; j++){
2113 subtitle* egysub=&subs[j];
2114 if (subd->sub_uses_time) {
2115 a=((egysub->start-mpsub_position)/100.0);
2116 b=((egysub->end-egysub->start)/100.0);
2117 if ( (float)((int)a) == a)
2118 fprintf (fd, "%.0f",a);
2119 else
2120 fprintf (fd, "%.2f",a);
2122 if ( (float)((int)b) == b)
2123 fprintf (fd, " %.0f\n",b);
2124 else
2125 fprintf (fd, " %.2f\n",b);
2126 } else {
2127 fprintf (fd, "%ld %ld\n", (long)((egysub->start*(fps/sub_fps))-((mpsub_position*(fps/sub_fps)))),
2128 (long)(((egysub->end)-(egysub->start))*(fps/sub_fps)));
2131 mpsub_position = egysub->end;
2132 for (i=0; i<egysub->lines; i++) {
2133 fprintf (fd, "%s\n",egysub->text[i]);
2135 fprintf (fd, "\n");
2137 fclose (fd);
2138 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dump.mpsub\'.\n");
2141 void dump_microdvd(sub_data* subd, float fps) {
2142 int i, delay;
2143 FILE *fd;
2144 subtitle *subs = subd->subtitles;
2145 if (sub_fps == 0)
2146 sub_fps = fps;
2147 fd = fopen("dumpsub.sub", "w");
2148 if (!fd) {
2149 perror("dumpsub.sub: fopen");
2150 return;
2152 delay = sub_delay * sub_fps;
2153 for (i = 0; i < subd->sub_num; ++i) {
2154 int j, start, end;
2155 start = subs[i].start;
2156 end = subs[i].end;
2157 if (subd->sub_uses_time) {
2158 start = start * sub_fps / 100 ;
2159 end = end * sub_fps / 100;
2161 else {
2162 start = start * sub_fps / fps;
2163 end = end * sub_fps / fps;
2165 start -= delay;
2166 end -= delay;
2167 fprintf(fd, "{%d}{%d}", start, end);
2168 for (j = 0; j < subs[i].lines; ++j)
2169 fprintf(fd, "%s%s", j ? "|" : "", subs[i].text[j]);
2170 fprintf(fd, "\n");
2172 fclose(fd);
2173 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.sub\'.\n");
2176 void dump_jacosub(sub_data* subd, float fps) {
2177 int i,j;
2178 int h,m,s,cs;
2179 FILE * fd;
2180 subtitle * onesub;
2181 unsigned long temp;
2182 subtitle *subs = subd->subtitles;
2184 if (!subd->sub_uses_time && sub_fps == 0)
2185 sub_fps = fps;
2186 fd=fopen("dumpsub.jss","w");
2187 if(!fd)
2189 perror("dump_jacosub: fopen");
2190 return;
2192 fprintf(fd, "#TIMERES %d\n", (subd->sub_uses_time) ? 100 : (int)sub_fps);
2193 for(i=0; i < subd->sub_num; i++)
2195 onesub=subs+i; //=&subs[i];
2197 temp=onesub->start;
2198 if (!subd->sub_uses_time)
2199 temp = temp * 100 / sub_fps;
2200 temp -= sub_delay * 100;
2201 h=temp/360000;temp%=360000; //h =1*100*60*60
2202 m=temp/6000; temp%=6000; //m =1*100*60
2203 s=temp/100; temp%=100; //s =1*100
2204 cs=temp; //cs=1*10
2205 fprintf(fd,"%02d:%02d:%02d.%02d ",h,m,s,cs);
2207 temp=onesub->end;
2208 if (!subd->sub_uses_time)
2209 temp = temp * 100 / sub_fps;
2210 temp -= sub_delay * 100;
2211 h=temp/360000;temp%=360000;
2212 m=temp/6000; temp%=6000;
2213 s=temp/100; temp%=100;
2214 cs=temp;
2215 fprintf(fd,"%02d:%02d:%02d.%02d {~} ",h,m,s,cs);
2217 for(j=0;j<onesub->lines;j++)
2218 fprintf(fd,"%s%s",j ? "\\n" : "", onesub->text[j]);
2220 fprintf(fd,"\n");
2222 fclose(fd);
2223 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.js\'.\n");
2226 void dump_sami(sub_data* subd, float fps) {
2227 int i,j;
2228 FILE * fd;
2229 subtitle * onesub;
2230 unsigned long temp;
2231 subtitle *subs = subd->subtitles;
2233 if (!subd->sub_uses_time && sub_fps == 0)
2234 sub_fps = fps;
2235 fd=fopen("dumpsub.smi","w");
2236 if(!fd)
2238 perror("dump_jacosub: fopen");
2239 return;
2241 fprintf(fd, "<SAMI>\n"
2242 "<HEAD>\n"
2243 " <STYLE TYPE=\"Text/css\">\n"
2244 " <!--\n"
2245 " 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"
2246 " .SUBTTL {Name: 'Subtitles'; Lang: en-US; SAMIType: CC;}\n"
2247 " -->\n"
2248 " </STYLE>\n"
2249 "</HEAD>\n"
2250 "<BODY>\n");
2251 for(i=0; i < subd->sub_num; i++)
2253 onesub=subs+i; //=&subs[i];
2255 temp=onesub->start;
2256 if (!subd->sub_uses_time)
2257 temp = temp * 100 / sub_fps;
2258 temp -= sub_delay * 100;
2259 fprintf(fd,"\t<SYNC Start=%lu>\n"
2260 "\t <P>", temp * 10);
2262 for(j=0;j<onesub->lines;j++)
2263 fprintf(fd,"%s%s",j ? "<br>" : "", onesub->text[j]);
2265 fprintf(fd,"\n");
2267 temp=onesub->end;
2268 if (!subd->sub_uses_time)
2269 temp = temp * 100 / sub_fps;
2270 temp -= sub_delay * 100;
2271 fprintf(fd,"\t<SYNC Start=%lu>\n"
2272 "\t <P>&nbsp;\n", temp * 10);
2274 fprintf(fd, "</BODY>\n"
2275 "</SAMI>\n");
2276 fclose(fd);
2277 mp_msg(MSGT_SUBREADER,MSGL_INFO,"SUB: Subtitles dumped in \'dumpsub.smi\'.\n");
2280 void sub_free( sub_data * subd )
2282 int i, j;
2284 if ( !subd ) return;
2286 for (i = 0; i < subd->sub_num; i++)
2287 for (j = 0; j < subd->subtitles[i].lines; j++)
2288 free( subd->subtitles[i].text[j] );
2289 free( subd->subtitles );
2290 free( subd->filename );
2291 free( subd );
2294 #define MAX_SUBLINE 512
2296 * \brief parse text and append it to subtitle in sub
2297 * \param sub subtitle struct to add text to
2298 * \param txt text to parse
2299 * \param len length of text in txt
2300 * \param endpts pts at which this subtitle text should be removed again
2302 * <> and {} are interpreted as comment delimiters, "\n", "\N", '\n', '\r'
2303 * and '\0' are interpreted as newlines, duplicate, leading and trailing
2304 * newlines are ignored.
2306 void sub_add_text(subtitle *sub, const char *txt, int len, double endpts) {
2307 int comment = 0;
2308 int double_newline = 1; // ignore newlines at the beginning
2309 int i, pos;
2310 char *buf;
2311 #ifdef CONFIG_FRIBIDI
2312 int orig_lines = sub->lines;
2313 #endif
2314 if (sub->lines >= SUB_MAX_TEXT) return;
2315 pos = 0;
2316 buf = malloc(MAX_SUBLINE + 1);
2317 sub->text[sub->lines] = buf;
2318 sub->endpts[sub->lines] = endpts;
2319 for (i = 0; i < len && pos < MAX_SUBLINE; i++) {
2320 char c = txt[i];
2321 if (c == '<') comment |= 1;
2322 if (c == '{') comment |= 2;
2323 if (comment) {
2324 if (c == '}') comment &= ~2;
2325 if (c == '>') comment &= ~1;
2326 continue;
2328 if (pos == MAX_SUBLINE - 1) {
2329 i--;
2330 c = 0;
2332 if (c == '\\' && i + 1 < len) {
2333 c = txt[++i];
2334 if (c == 'n' || c == 'N') c = 0;
2336 if (c == '\n' || c == '\r') c = 0;
2337 if (c) {
2338 double_newline = 0;
2339 buf[pos++] = c;
2340 } else if (!double_newline) {
2341 if (sub->lines >= SUB_MAX_TEXT - 1) {
2342 mp_msg(MSGT_VO, MSGL_WARN, "Too many subtitle lines\n");
2343 break;
2345 double_newline = 1;
2346 buf[pos] = 0;
2347 sub->lines++;
2348 pos = 0;
2349 buf = malloc(MAX_SUBLINE + 1);
2350 sub->text[sub->lines] = buf;
2351 sub->endpts[sub->lines] = endpts;
2354 buf[pos] = 0;
2355 if (sub->lines < SUB_MAX_TEXT &&
2356 strlen(sub->text[sub->lines]))
2357 sub->lines++;
2358 #ifdef CONFIG_FRIBIDI
2359 sub = sub_fribidi(sub, sub_utf8, orig_lines);
2360 #endif
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;