sub/OSD: move some related files to sub/

1 files changed, 2504 insertions, 0 deletions

diff --git a/sub/subreader.c b/sub/subreader.c
new file mode 100644
index 0000000000..1f9294ec47
--- /dev/null
+++ b/sub/subreader.c
@@ -0,0 +1,2504 @@
+/*
+ * Subtitle reader with format autodetection
+ *
+ * Copyright (c) 2001 laaz
+ * Some code cleanup & realloc() by A'rpi/ESP-team
+ *
+ * This file is part of MPlayer.
+ *
+ * MPlayer is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * MPlayer is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License along
+ * with MPlayer; if not, write to the Free Software Foundation, Inc.,
+ * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <ctype.h>
+
+#include <sys/types.h>
+#include <dirent.h>
+
+#include "config.h"
+#include "mp_msg.h"
+#include "subreader.h"
+#include "mpcommon.h"
+#include "subassconvert.h"
+#include "options.h"
+#include "stream/stream.h"
+#include "libavutil/common.h"
+#include "libavutil/avstring.h"
+
+#ifdef CONFIG_ENCA
+#include <enca.h>
+#endif
+
+#define ERR ((void *) -1)
+
+#ifdef CONFIG_ICONV
+#include <iconv.h>
+char *sub_cp=NULL;
+#endif
+#ifdef CONFIG_FRIBIDI
+#include <fribidi/fribidi.h>
+char *fribidi_charset = NULL;   ///character set that will be passed to FriBiDi
+int flip_hebrew = 1;            ///flip subtitles using fribidi
+int fribidi_flip_commas = 0;    ///flip comma when fribidi is used
+#endif
+
+// Parameter struct for the format-specific readline functions
+struct readline_args {
+    int utf16;
+    struct MPOpts *opts;
+};
+
+/* Maximal length of line of a subtitle */
+#define LINE_LEN 1000
+static float mpsub_position=0;
+static float mpsub_multiplier=1.;
+static int sub_slacktime = 20000; //20 sec
+
+int sub_no_text_pp=0;   // 1 => do not apply text post-processing
+                        // like {\...} elimination in SSA format.
+
+int sub_match_fuzziness=0; // level of sub name matching fuzziness
+
+/* Use the SUB_* constant defined in the header file */
+int sub_format=SUB_INVALID;
+#ifdef CONFIG_SORTSUB
+/*
+   Some subtitling formats, namely AQT and Subrip09, define the end of a
+   subtitle as the beginning of the following. Since currently we read one
+   subtitle at time, for these format we keep two global *subtitle,
+   previous_aqt_sub and previous_subrip09_sub, pointing to previous subtitle,
+   so we can change its end when we read current subtitle starting time.
+   When CONFIG_SORTSUB is defined, we use a single global unsigned long,
+   previous_sub_end, for both (and even future) formats, to store the end of
+   the previous sub: it is initialized to 0 in sub_read_file and eventually
+   modified by sub_read_aqt_line or sub_read_subrip09_line.
+ */
+unsigned long previous_sub_end;
+#endif
+
+static int eol(char p) {
+	return p=='\r' || p=='\n' || p=='\0';
+}
+
+/* Remove leading and trailing space */
+static void trail_space(char *s) {
+	int i = 0;
+	while (isspace(s[i])) ++i;
+	if (i) strcpy(s, s + i);
+	i = strlen(s) - 1;
+	while (i > 0 && isspace(s[i])) s[i--] = '\0';
+}
+
+static char *stristr(const char *haystack, const char *needle) {
+    int len = 0;
+    const char *p = haystack;
+
+    if (!(haystack && needle)) return NULL;
+
+    len=strlen(needle);
+    while (*p != '\0') {
+	if (strncasecmp(p, needle, len) == 0) return (char*)p;
+	p++;
+    }
+
+    return NULL;
+}
+
+static void sami_add_line(subtitle *current, char *buffer, char **pos) {
+    char *p = *pos;
+    *p = 0;
+    trail_space(buffer);
+    if (*buffer && current->lines < SUB_MAX_TEXT)
+        current->text[current->lines++] = strdup(buffer);
+    *pos = buffer;
+}
+
+static subtitle *sub_read_line_sami(stream_t* st, subtitle *current,
+                                    struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    static char line[LINE_LEN+1];
+    static char *s = NULL, *slacktime_s;
+    char text[LINE_LEN+1], *p=NULL, *q;
+    int state;
+
+    current->lines = current->start = current->end = 0;
+    current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
+    state = 0;
+
+    /* read the first line */
+    if (!s)
+	    if (!(s = stream_read_line(st, line, LINE_LEN, utf16))) return 0;
+
+    do {
+	switch (state) {
+
+	case 0: /* find "START=" or "Slacktime:" */
+	    slacktime_s = stristr (s, "Slacktime:");
+	    if (slacktime_s)
+                sub_slacktime = strtol (slacktime_s+10, NULL, 0) / 10;
+
+	    s = stristr (s, "Start=");
+	    if (s) {
+		current->start = strtol (s + 6, &s, 0) / 10;
+                /* eat '>' */
+                for (; *s != '>' && *s != '\0'; s++);
+                s++;
+		state = 1; continue;
+	    }
+	    break;
+
+	case 1: /* find (optional) "<P", skip other TAGs */
+	    for  (; *s == ' ' || *s == '\t'; s++); /* strip blanks, if any */
+	    if (*s == '\0') break;
+	    if (*s != '<') { state = 3; p = text; continue; } /* not a TAG */
+	    s++;
+	    if (*s == 'P' || *s == 'p') { s++; state = 2; continue; } /* found '<P' */
+	    for (; *s != '>' && *s != '\0'; s++); /* skip remains of non-<P> TAG */
+	    if (s == '\0')
+	      break;
+	    s++;
+	    continue;
+
+	case 2: /* find ">" */
+	    if ((s = strchr (s, '>'))) { s++; state = 3; p = text; continue; }
+	    break;
+
+	case 3: /* get all text until '<' appears */
+	    if (*s == '\0') break;
+	    else if (!strncasecmp (s, "<br>", 4)) {
+                sami_add_line(current, text, &p);
+		s += 4;
+	    }
+	    else if ((*s == '{') && !sub_no_text_pp) { state = 5; ++s; continue; }
+	    else if (*s == '<') { state = 4; }
+	    else if (!strncasecmp (s, "&nbsp;", 6)) { *p++ = ' '; s += 6; }
+	    else if (*s == '\t') { *p++ = ' '; s++; }
+	    else if (*s == '\r' || *s == '\n') { s++; }
+	    else *p++ = *s++;
+
+	    /* skip duplicated space */
+	    if (p > text + 2) if (*(p-1) == ' ' && *(p-2) == ' ') p--;
+
+	    continue;
+
+	case 4: /* get current->end or skip <TAG> */
+	    q = stristr (s, "Start=");
+	    if (q) {
+		current->end = strtol (q + 6, &q, 0) / 10 - 1;
+		*p = '\0'; trail_space (text);
+		if (text[0] != '\0')
+		    current->text[current->lines++] = strdup (text);
+		if (current->lines > 0) { state = 99; break; }
+		state = 0; continue;
+	    }
+	    s = strchr (s, '>');
+	    if (s) { s++; state = 3; continue; }
+	    break;
+       case 5: /* get rid of {...} text, but read the alignment code */
+	    if ((*s == '\\') && (*(s + 1) == 'a') && !sub_no_text_pp) {
+               if (stristr(s, "\\a1") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_BOTTOMLEFT;
+                   s = s + 3;
+               }
+               if (stristr(s, "\\a2") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
+                   s = s + 3;
+               } else if (stristr(s, "\\a3") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_BOTTOMRIGHT;
+                   s = s + 3;
+               } else if ((stristr(s, "\\a4") != NULL) || (stristr(s, "\\a5") != NULL) || (stristr(s, "\\a8") != NULL)) {
+                   current->alignment = SUB_ALIGNMENT_TOPLEFT;
+                   s = s + 3;
+               } else if (stristr(s, "\\a6") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_TOPCENTER;
+                   s = s + 3;
+               } else if (stristr(s, "\\a7") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_TOPRIGHT;
+                   s = s + 3;
+               } else if (stristr(s, "\\a9") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_MIDDLELEFT;
+                   s = s + 3;
+               } else if (stristr(s, "\\a10") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_MIDDLECENTER;
+                   s = s + 4;
+               } else if (stristr(s, "\\a11") != NULL) {
+                   current->alignment = SUB_ALIGNMENT_MIDDLERIGHT;
+                   s = s + 4;
+               }
+	    }
+	    if (*s == '}') state = 3;
+	    ++s;
+	    continue;
+	}
+
+	/* read next line */
+	if (state != 99 && !(s = stream_read_line (st, line, LINE_LEN, utf16))) {
+	    if (current->start > 0) {
+		break; // if it is the last subtitle
+	    } else {
+		return 0;
+	    }
+	}
+
+    } while (state != 99);
+
+    // For the last subtitle
+    if (current->end <= 0) {
+        current->end = current->start + sub_slacktime;
+        sami_add_line(current, text, &p);
+    }
+
+    return current;
+}
+
+
+static char *sub_readtext(char *source, char **dest) {
+    int len=0;
+    char *p=source;
+
+//    printf("src=%p  dest=%p  \n",source,dest);
+
+    while ( !eol(*p) && *p!= '|' ) {
+	p++,len++;
+    }
+
+    *dest= malloc (len+1);
+    if (!dest) {return ERR;}
+
+    strncpy(*dest, source, len);
+    (*dest)[len]=0;
+
+    while (*p=='\r' || *p=='\n' || *p=='|') p++;
+
+    if (*p) return p;  // not-last text field
+    else return NULL;  // last text field
+}
+
+static subtitle *sub_read_line_microdvd(stream_t *st,subtitle *current,
+                                        struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    char line2[LINE_LEN+1];
+    char *p, *next;
+    int i;
+
+    do {
+	if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+    } while ((sscanf (line,
+		      "{%ld}{}%[^\r\n]",
+		      &(current->start), line2) < 2) &&
+	     (sscanf (line,
+		      "{%ld}{%ld}%[^\r\n]",
+		      &(current->start), &(current->end), line2) < 3));
+
+    if (args->opts->ass_enabled) {
+        subassconvert_microdvd(line2, line, LINE_LEN + 1);
+        p = line;
+    } else
+        p = line2;
+
+    next=p, i=0;
+    while ((next =sub_readtext (next, &(current->text[i])))) {
+        if (current->text[i]==ERR) {return ERR;}
+	i++;
+	if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
+    }
+    current->lines= ++i;
+
+    return current;
+}
+
+static subtitle *sub_read_line_mpl2(stream_t *st,subtitle *current,
+                                    struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    char line2[LINE_LEN+1];
+    char *p, *next;
+    int i;
+
+    do {
+	if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+    } while ((sscanf (line,
+		      "[%ld][%ld]%[^\r\n]",
+		      &(current->start), &(current->end), line2) < 3));
+    current->start *= 10;
+    current->end *= 10;
+    p=line2;
+
+    next=p, i=0;
+    while ((next =sub_readtext (next, &(current->text[i])))) {
+        if (current->text[i]==ERR) {return ERR;}
+	i++;
+	if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
+    }
+    current->lines= ++i;
+
+    return current;
+}
+
+static subtitle *sub_read_line_subrip(stream_t* st, subtitle *current,
+                                    struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    int a1,a2,a3,a4,b1,b2,b3,b4;
+    char *p=NULL, *q=NULL;
+    int len;
+
+    while (1) {
+	if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+	if (sscanf (line, "%d:%d:%d.%d,%d:%d:%d.%d",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4) < 8) continue;
+	current->start = a1*360000+a2*6000+a3*100+a4;
+	current->end   = b1*360000+b2*6000+b3*100+b4;
+
+	if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+
+	p=q=line;
+	for (current->lines=1; current->lines < SUB_MAX_TEXT; current->lines++) {
+	    for (q=p,len=0; *p && *p!='\r' && *p!='\n' && *p!='|' && strncmp(p,"[br]",4); p++,len++);
+	    current->text[current->lines-1]=malloc (len+1);
+	    if (!current->text[current->lines-1]) return ERR;
+	    strncpy (current->text[current->lines-1], q, len);
+	    current->text[current->lines-1][len]='\0';
+	    if (!*p || *p=='\r' || *p=='\n') break;
+	    if (*p=='|') p++;
+	    else while (*p++!=']');
+	}
+	break;
+    }
+    return current;
+}
+
+static subtitle *sub_ass_read_line_subviewer(stream_t *st, subtitle *current,
+                                             struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    int a1, a2, a3, a4, b1, b2, b3, b4, j = 0;
+
+    while (!current->text[0]) {
+        char line[LINE_LEN + 1], full_line[LINE_LEN + 1], sep;
+        int i;
+
+        /* Parse SubRip header */
+        if (!stream_read_line(st, line, LINE_LEN, utf16))
+            return NULL;
+        if (sscanf(line, "%d:%d:%d%[,.:]%d --> %d:%d:%d%[,.:]%d",
+                     &a1, &a2, &a3, &sep, &a4, &b1, &b2, &b3, &sep, &b4) < 10)
+            continue;
+
+        current->start = a1 * 360000 + a2 * 6000 + a3 * 100 + a4 / 10;
+        current->end   = b1 * 360000 + b2 * 6000 + b3 * 100 + b4 / 10;
+
+        /* Concat lines */
+        full_line[0] = 0;
+        for (i = 0; i < SUB_MAX_TEXT; i++) {
+            int blank = 1, len = 0;
+            char *p;
+
+            if (!stream_read_line(st, line, LINE_LEN, utf16))
+                break;
+
+            for (p = line; *p != '\n' && *p != '\r' && *p; p++, len++)
+                if (*p != ' ' && *p != '\t')
+                    blank = 0;
+
+            if (blank)
+                break;
+
+            *p = 0;
+
+            if (!(j + 1 + len < sizeof(full_line) - 1))
+                break;
+
+            if (j != 0)
+                full_line[j++] = '\n';
+            strcpy(&full_line[j], line);
+            j += len;
+        }
+
+        /* Use the ASS/SSA converter to transform the whole lines */
+        if (full_line[0]) {
+            char converted_line[LINE_LEN + 1];
+            subassconvert_subrip(full_line, converted_line, LINE_LEN + 1);
+            current->text[0] = strdup(converted_line);
+            current->lines = 1;
+        }
+    }
+    return current;
+}
+
+static subtitle *sub_read_line_subviewer(stream_t *st,subtitle *current,
+                                         struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    int a1,a2,a3,a4,b1,b2,b3,b4;
+    char *p=NULL;
+    int i,len;
+
+    if (args->opts->ass_enabled)
+        return sub_ass_read_line_subviewer(st, current, args);
+    while (!current->text[0]) {
+	if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+	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)
+	    continue;
+	current->start = a1*360000+a2*6000+a3*100+a4/10;
+	current->end   = b1*360000+b2*6000+b3*100+b4/10;
+	for (i=0; i<SUB_MAX_TEXT;) {
+	    int blank = 1;
+	    if (!stream_read_line (st, line, LINE_LEN, utf16)) break;
+	    len=0;
+	    for (p=line; *p!='\n' && *p!='\r' && *p; p++,len++)
+		if (*p != ' ' && *p != '\t')
+		    blank = 0;
+	    if (len && !blank) {
+                int j=0,skip=0;
+		char *curptr=current->text[i]=malloc (len+1);
+		if (!current->text[i]) return ERR;
+		//strncpy (current->text[i], line, len); current->text[i][len]='\0';
+                for(; j<len; j++) {
+		    /* let's filter html tags ::atmos */
+		    if(line[j]=='>') {
+			skip=0;
+			continue;
+		    }
+		    if(line[j]=='<') {
+			skip=1;
+			continue;
+		    }
+		    if(skip) {
+			continue;
+		    }
+		    *curptr=line[j];
+		    curptr++;
+		}
+		*curptr='\0';
+
+		i++;
+	    } else {
+		break;
+	    }
+	}
+	current->lines=i;
+    }
+    return current;
+}
+
+static subtitle *sub_read_line_subviewer2(stream_t *st,subtitle *current,
+                                          struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    int a1,a2,a3,a4;
+    char *p=NULL;
+    int i,len;
+
+    while (!current->text[0]) {
+        if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+	if (line[0]!='{')
+	    continue;
+        if ((len=sscanf (line, "{T %d:%d:%d:%d",&a1,&a2,&a3,&a4)) < 4)
+            continue;
+        current->start = a1*360000+a2*6000+a3*100+a4/10;
+        for (i=0; i<SUB_MAX_TEXT;) {
+            if (!stream_read_line (st, line, LINE_LEN, utf16)) break;
+            if (line[0]=='}') break;
+            len=0;
+            for (p=line; *p!='\n' && *p!='\r' && *p; ++p,++len);
+            if (len) {
+                current->text[i]=malloc (len+1);
+                if (!current->text[i]) return ERR;
+                strncpy (current->text[i], line, len); current->text[i][len]='\0';
+                ++i;
+            } else {
+                break;
+            }
+        }
+        current->lines=i;
+    }
+    return current;
+}
+
+
+static subtitle *sub_read_line_vplayer(stream_t *st,subtitle *current,
+                                       struct readline_args *args)
+{
+        int utf16 = args->utf16;
+	char line[LINE_LEN+1];
+	int a1,a2,a3;
+	char *p=NULL, *next,separator;
+	int i,len,plen;
+
+	while (!current->text[0]) {
+		if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+		if ((len=sscanf (line, "%d:%d:%d%c%n",&a1,&a2,&a3,&separator,&plen)) < 4)
+			continue;
+
+		if (!(current->start = a1*360000+a2*6000+a3*100))
+			continue;
+                /* removed by wodzu
+		p=line;
+ 		// finds the body of the subtitle
+ 		for (i=0; i<3; i++){
+		   p=strchr(p,':');
+		   if (p==NULL) break;
+		   ++p;
+		}
+		if (p==NULL) {
+		    printf("SUB: Skipping incorrect subtitle line!\n");
+		    continue;
+		}
+                */
+                // by wodzu: hey! this time we know what length it has! what is
+                // that magic for? it can't deal with space instead of third
+                // colon! look, what simple it can be:
+                p = &line[ plen ];
+
+ 		i=0;
+		if (*p!='|') {
+			//
+			next = p,i=0;
+			while ((next =sub_readtext (next, &(current->text[i])))) {
+				if (current->text[i]==ERR) {return ERR;}
+				i++;
+				if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
+			}
+			current->lines=i+1;
+		}
+	}
+	return current;
+}
+
+static subtitle *sub_read_line_rt(stream_t *st,subtitle *current,
+                                    struct readline_args *args)
+{
+    int utf16 = args->utf16;
+
+	//TODO: This format uses quite rich (sub/super)set of xhtml
+	// I couldn't check it since DTD is not included.
+	// WARNING: full XML parses can be required for proper parsing
+    char line[LINE_LEN+1];
+    int a1,a2,a3,a4,b1,b2,b3,b4;
+    char *p=NULL,*next=NULL;
+    int i,len,plen;
+
+    while (!current->text[0]) {
+	if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+	//TODO: it seems that format of time is not easily determined, it may be 1:12, 1:12.0 or 0:1:12.0
+	//to describe the same moment in time. Maybe there are even more formats in use.
+	//if ((len=sscanf (line, "<Time Begin=\"%d:%d:%d.%d\" End=\"%d:%d:%d.%d\"",&a1,&a2,&a3,&a4,&b1,&b2,&b3,&b4)) < 8)
+	plen=a1=a2=a3=a4=b1=b2=b3=b4=0;
+	if (
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&b3,&b4,&plen)) < 4) &&
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&b2,&b3,&b4,&plen)) < 5) &&
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&b2,&b3,&plen)) < 4) &&
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\" %*[Ee]nd=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2,&a3,&b2,&b3,&b4,&plen)) < 5) &&
+//	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\" %*[Ee]nd=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&b2,&b3,&plen)) < 5) &&
+	((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) &&
+	((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) &&
+	//now try it without end time
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d.%d\"%*[^<]<clear/>%n",&a3,&a4,&plen)) < 2) &&
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d\"%*[^<]<clear/>%n",&a2,&a3,&plen)) < 2) &&
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d.%d\"%*[^<]<clear/>%n",&a2,&a3,&a4,&plen)) < 3) &&
+	((len=sscanf (line, "<%*[tT]ime %*[bB]egin=\"%d:%d:%d.%d\"%*[^<]<clear/>%n",&a1,&a2,&a3,&a4,&plen)) < 4)
+	)
+	    continue;
+	current->start = a1*360000+a2*6000+a3*100+a4/10;
+	current->end   = b1*360000+b2*6000+b3*100+b4/10;
+	if (b1 == 0 && b2 == 0 && b3 == 0 && b4 == 0)
+	  current->end = current->start+200;
+	p=line;	p+=plen;i=0;
+	// TODO: I don't know what kind of convention is here for marking multiline subs, maybe <br/> like in xml?
+	next = strstr(line,"<clear/>");
+	if(next && strlen(next)>8){
+	  next+=8;i=0;
+	  while ((next =sub_readtext (next, &(current->text[i])))) {
+		if (current->text[i]==ERR) {return ERR;}
+		i++;
+		if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
+	  }
+	}
+			current->lines=i+1;
+    }
+    return current;
+}
+
+static subtitle *sub_read_line_ssa(stream_t *st,subtitle *current,
+                                    struct readline_args *args)
+{
+/*
+ * Sub Station Alpha v4 (and v2?) scripts have 9 commas before subtitle
+ * other Sub Station Alpha scripts have only 8 commas before subtitle
+ * Reading the "ScriptType:" field is not reliable since many scripts appear
+ * w/o it
+ *
+ * http://www.scriptclub.org is a good place to find more examples
+ * http://www.eswat.demon.co.uk is where the SSA specs can be found
+ */
+        int utf16 = args->utf16;
+        int comma;
+        static int max_comma = 32; /* let's use 32 for the case that the */
+                    /*  amount of commas increase with newer SSA versions */
+
+	int hour1, min1, sec1, hunsec1,
+	    hour2, min2, sec2, hunsec2, nothing;
+	int num;
+
+	char line[LINE_LEN+1],
+	     line3[LINE_LEN+1],
+	     *line2;
+	char *tmp;
+
+	do {
+		if (!stream_read_line (st, line, LINE_LEN, utf16)) return NULL;
+	} while (sscanf (line, "Dialogue: Marked=%d,%d:%d:%d.%d,%d:%d:%d.%d"
+			"%[^\n\r]", &nothing,
+			&hour1, &min1, &sec1, &hunsec1,
+			&hour2, &min2, &sec2, &hunsec2,
+			line3) < 9
+		 &&
+		 sscanf (line, "Dialogue: %d,%d:%d:%d.%d,%d:%d:%d.%d"
+			 "%[^\n\r]", &nothing,
+			 &hour1, &min1, &sec1, &hunsec1,
+			 &hour2, &min2, &sec2, &hunsec2,
+			 line3) < 9	    );
+
+        line2=strchr(line3, ',');
+        if (!line2) return NULL;
+
+        for (comma = 4; comma < max_comma; comma ++)
+          {
+            tmp = line2;
+            if(!(tmp=strchr(++tmp, ','))) break;
+            if(*(++tmp) == ' ') break;
+                  /* a space after a comma means we're already in a sentence */
+            line2 = tmp;
+          }
+
+        if(comma < max_comma)max_comma = comma;
+	/* eliminate the trailing comma */
+	if(*line2 == ',') line2++;
+
+	current->lines=0;num=0;
+	current->start = 360000*hour1 + 6000*min1 + 100*sec1 + hunsec1;
+	current->end   = 360000*hour2 + 6000*min2 + 100*sec2 + hunsec2;
+
+        while (((tmp=strstr(line2, "\\n")) != NULL) || ((tmp=strstr(line2, "\\N")) != NULL) ){
+		current->text[num]=malloc(tmp-line2+1);
+		strncpy (current->text[num], line2, tmp-line2);
+		current->text[num][tmp-line2]='\0';
+		line2=tmp+2;
+		num++;
+		current->lines++;
+		if (current->lines >=  SUB_MAX_TEXT) return current;
+	}
+
+	current->text[num]=strdup(line2);
+	current->lines++;
+
+	return current;
+}
+
+static void sub_pp_ssa(subtitle *sub) {
+	int l=sub->lines;
+	char *so,*de,*start;
+
+	while (l){
+            	/* eliminate any text enclosed with {}, they are font and color settings */
+            	so=de=sub->text[--l];
+            	while (*so) {
+            		if(*so == '{' && so[1]=='\\') {
+            			for (start=so; *so && *so!='}'; so++);
+            			if(*so) so++; else so=start;
+            		}
+            		if(*so) {
+            			*de=*so;
+            			so++; de++;
+            		}
+            	}
+            	*de=*so;
+        }
+}
+
+/*
+ * PJS subtitles reader.
+ * That's the "Phoenix Japanimation Society" format.
+ * I found some of them in http://www.scriptsclub.org/ (used for anime).
+ * The time is in tenths of second.
+ *
+ * by set, based on code by szabi (dunnowhat sub format ;-)
+ */
+static subtitle *sub_read_line_pjs(stream_t *st,subtitle *current,
+                                   struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    char text[LINE_LEN+1], *s, *d;
+
+    if (!stream_read_line (st, line, LINE_LEN, utf16))
+	return NULL;
+    /* skip spaces */
+    for (s=line; *s && isspace(*s); s++);
+    /* allow empty lines at the end of the file */
+    if (*s==0)
+	return NULL;
+    /* get the time */
+    if (sscanf (s, "%ld,%ld,", &(current->start),
+		&(current->end)) <2) {
+	return ERR;
+    }
+    /* the files I have are in tenths of second */
+    current->start *= 10;
+    current->end *= 10;
+    /* walk to the beggining of the string */
+    for (; *s; s++) if (*s==',') break;
+    if (*s) {
+	for (s++; *s; s++) if (*s==',') break;
+	if (*s) s++;
+    }
+    if (*s!='"') {
+	return ERR;
+    }
+    /* copy the string to the text buffer */
+    for (s++, d=text; *s && *s!='"'; s++, d++)
+	*d=*s;
+    *d=0;
+    current->text[0] = strdup(text);
+    current->lines = 1;
+
+    return current;
+}
+
+static subtitle *sub_read_line_mpsub(stream_t *st, subtitle *current,
+                                     struct readline_args *args)
+{
+        int utf16 = args->utf16;
+	char line[LINE_LEN+1];
+	float a,b;
+	int num=0;
+	char *p, *q;
+
+	do
+	{
+		if (!stream_read_line(st, line, LINE_LEN, utf16)) return NULL;
+	} while (sscanf (line, "%f %f", &a, &b) !=2);
+
+	mpsub_position += a*mpsub_multiplier;
+	current->start=(int) mpsub_position;
+	mpsub_position += b*mpsub_multiplier;
+	current->end=(int) mpsub_position;
+
+	while (num < SUB_MAX_TEXT) {
+		if (!stream_read_line (st, line, LINE_LEN, utf16)) {
+			if (num == 0) return NULL;
+			else return current;
+		}
+		p=line;
+		while (isspace(*p)) p++;
+		if (eol(*p) && num > 0) return current;
+		if (eol(*p)) return NULL;
+
+		for (q=p; !eol(*q); q++);
+		*q='\0';
+		if (strlen(p)) {
+			current->text[num]=strdup(p);
+//			printf (">%s<\n",p);
+			current->lines = ++num;
+		} else {
+			if (num) return current;
+			else return NULL;
+		}
+	}
+	return NULL; // we should have returned before if it's OK
+}
+
+#ifndef CONFIG_SORTSUB
+//we don't need this if we use previous_sub_end
+subtitle *previous_aqt_sub = NULL;
+#endif
+
+static subtitle *sub_read_line_aqt(stream_t *st,subtitle *current,
+                                   struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    char *next;
+    int i;
+
+    while (1) {
+    // try to locate next subtitle
+        if (!stream_read_line (st, line, LINE_LEN, utf16))
+		return NULL;
+        if (!(sscanf (line, "-->> %ld", &(current->start)) <1))
+		break;
+    }
+
+#ifdef CONFIG_SORTSUB
+    previous_sub_end = (current->start) ? current->start - 1 : 0;
+#else
+    if (previous_aqt_sub != NULL)
+	previous_aqt_sub->end = current->start-1;
+
+    previous_aqt_sub = current;
+#endif
+
+    if (!stream_read_line (st, line, LINE_LEN, utf16))
+	return NULL;
+
+    sub_readtext((char *) &line,&current->text[0]);
+    current->lines = 1;
+    current->end = current->start; // will be corrected by next subtitle
+
+    if (!stream_read_line (st, line, LINE_LEN, utf16))
+	return current;
+
+    next = line,i=1;
+    while ((next =sub_readtext (next, &(current->text[i])))) {
+	if (current->text[i]==ERR) {return ERR;}
+	i++;
+	if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
+	}
+    current->lines=i+1;
+
+    if (!strlen(current->text[0]) && !strlen(current->text[1])) {
+#ifdef CONFIG_SORTSUB
+	previous_sub_end = 0;
+#else
+	// void subtitle -> end of previous marked and exit
+	previous_aqt_sub = NULL;
+#endif
+	return NULL;
+	}
+
+    return current;
+}
+
+#ifndef CONFIG_SORTSUB
+subtitle *previous_subrip09_sub = NULL;
+#endif
+
+static subtitle *sub_read_line_subrip09(stream_t *st,subtitle *current,
+                                    struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line[LINE_LEN+1];
+    int a1,a2,a3;
+    char * next=NULL;
+    int i,len;
+
+    while (1) {
+    // try to locate next subtitle
+        if (!stream_read_line (st, line, LINE_LEN, utf16))
+		return NULL;
+        if (!((len=sscanf (line, "[%d:%d:%d]",&a1,&a2,&a3)) < 3))
+		break;
+    }
+
+    current->start = a1*360000+a2*6000+a3*100;
+
+#ifdef CONFIG_SORTSUB
+    previous_sub_end = (current->start) ? current->start - 1 : 0;
+#else
+    if (previous_subrip09_sub != NULL)
+	previous_subrip09_sub->end = current->start-1;
+
+    previous_subrip09_sub = current;
+#endif
+
+    if (!stream_read_line (st, line, LINE_LEN, utf16))
+	return NULL;
+
+    next = line,i=0;
+
+    current->text[0]=""; // just to be sure that string is clear
+
+    while ((next =sub_readtext (next, &(current->text[i])))) {
+	if (current->text[i]==ERR) {return ERR;}
+	i++;
+	if (i>=SUB_MAX_TEXT) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
+	}
+    current->lines=i+1;
+
+    if (!strlen(current->text[0]) && (i==0)) {
+#ifdef CONFIG_SORTSUB
+	previous_sub_end = 0;
+#else
+	// void subtitle -> end of previous marked and exit
+	previous_subrip09_sub = NULL;
+#endif
+	return NULL;
+	}
+
+    return current;
+}
+
+static subtitle *sub_read_line_jacosub(stream_t* st, subtitle * current,
+                                       struct readline_args *args)
+{
+    int utf16 = args->utf16;
+    char line1[LINE_LEN], line2[LINE_LEN], directive[LINE_LEN], *p, *q;
+    unsigned a1, a2, a3, a4, b1, b2, b3, b4, comment = 0;
+    static unsigned jacoTimeres = 30;
+    static int jacoShift = 0;
+
+    memset(current, 0, sizeof(subtitle));
+    memset(line1, 0, LINE_LEN);
+    memset(line2, 0, LINE_LEN);
+    memset(directive, 0, LINE_LEN);
+    while (!current->text[0]) {
+	if (!stream_read_line(st, line1, LINE_LEN, utf16)) {
+	    return NULL;
+	}
+	if (sscanf
+	    (line1, "%u:%u:%u.%u %u:%u:%u.%u %[^\n\r]", &a1, &a2, &a3, &a4,
+	     &b1, &b2, &b3, &b4, line2) < 9) {
+	    if (sscanf(line1, "@%u @%u %[^\n\r]", &a4, &b4, line2) < 3) {
+		if (line1[0] == '#') {
+		    int hours = 0, minutes = 0, seconds, delta, inverter =
+			1;
+		    unsigned units = jacoShift;
+		    switch (toupper(line1[1])) {
+		    case 'S':
+			if (isalpha(line1[2])) {
+			    delta = 6;
+			} else {
+			    delta = 2;
+			}
+			if (sscanf(&line1[delta], "%d", &hours)) {
+			    if (hours < 0) {
+				hours *= -1;
+				inverter = -1;
+			    }
+			    if (sscanf(&line1[delta], "%*d:%d", &minutes)) {
+				if (sscanf
+				    (&line1[delta], "%*d:%*d:%d",
+				     &seconds)) {
+				    sscanf(&line1[delta], "%*d:%*d:%*d.%d",
+					   &units);
+				} else {
+				    hours = 0;
+				    sscanf(&line1[delta], "%d:%d.%d",
+					   &minutes, &seconds, &units);
+				    minutes *= inverter;
+				}
+			    } else {
+				hours = minutes = 0;
+				sscanf(&line1[delta], "%d.%d", &seconds,
+				       &units);
+				seconds *= inverter;
+			    }
+			    jacoShift =
+				((hours * 3600 + minutes * 60 +
+				  seconds) * jacoTimeres +
+				 units) * inverter;
+			}
+			break;
+		    case 'T':
+			if (isalpha(line1[2])) {
+			    delta = 8;
+			} else {
+			    delta = 2;
+			}
+			sscanf(&line1[delta], "%u", &jacoTimeres);
+			break;
+		    }
+		}
+		continue;
+	    } else {
+		current->start =
+		    (unsigned long) ((a4 + jacoShift) * 100.0 /
+				     jacoTimeres);
+		current->end =
+		    (unsigned long) ((b4 + jacoShift) * 100.0 /
+				     jacoTimeres);
+	    }
+	} else {
+	    current->start =
+		(unsigned