/* * Subtitle reader with format autodetection * * Written by laaz * Some code cleanup & realloc() by A'rpi/ESP-team * */#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 "../include/subreader.h"#include "../include/stream.h"#ifdef HAVE_ENCA#include <enca.h>#endif#define ERR ((void *) -1)#ifdef USE_ICONV#include <iconv.h>RMascii *sub_cp=NULL;#endif#ifdef USE_FRIBIDI#include <fribidi/fribidi.h>RMascii *fribidi_charset = NULL;   ///character set that will be passed to FriBiDiint flip_hebrew = 1;            ///flip subtitles using fribidiint fribidi_flip_commas = 0;    ///flip comma when fribidi is used#endifRMascii* dvdsub_lang = "de";//int suboverlap_enabled = 2;/* Maximal length of line of a subtitle */#define LINE_LEN 1000//static RMreal mpsub_position=0;static RMreal mpsub_multiplier=1.0;static int sub_slacktime = 20000; //20 secint sub_no_text_pp=0;   // 1 => do not apply text post-processing                        // like {\...} elimination in SSA format.int sub_match_fuzziness=1; // level of sub name matching fuzzinesssub_context subs;/* Use the SUB_* constant defined in the header file */int sub_format=SUB_INVALID;#ifdef USE_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 USE_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;#endifstatic int eol(RMascii p) {	return (p=='\r' || p=='\n' || p=='\0');}/* Remove leading and trailing space */static void trail_space(RMascii *s) {	int i = 0;	while (isspace(s[i])) ++i;	if (i) RMCopyAscii(s, s + i);	i = RMasciiLength(s) - 1;	while (i > 0 && isspace(s[i])) s[i--] = '\0';}static RMascii *stristr(const RMascii *haystack, const RMascii *needle) {    int len = 0;    const RMascii *p = haystack;    if (!(haystack && needle)) return NULL;    len=RMasciiLength(needle);    while (*p != '\0') {	if (RMNCompareAsciiCaseInsensitively(p, needle, len) == 0) return (RMascii*)p;	p++;    }    return NULL;}static subtitle *sub_read_line_sami(stream_t* st, subtitle *current) {    static RMascii line[LINE_LEN+1];    static RMascii *s = NULL, *slacktime_s;    RMascii text[LINE_LEN+1], *p=NULL, *q;    int state; //   int i;    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))) return 0;  //  printf("line %s\n", line);    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 (optionnal) "<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 (!RMNCompareAsciiCaseInsensitively (s, "<br>", 4)) {		*p = '\0'; p = text; trail_space (text);		if (text[0] != '\0')		    current->text[current->lines++] = strdup (text);		s += 4;	    }	    else if ((*s == '{') && !sub_no_text_pp) { state = 5; ++s; continue; }	    else if (*s == '<') { state = 4; }	    else if (!RMNCompareAsciiCaseInsensitively (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))) {	    if (current->start > 0) {		break; // if it is the last subtitle	    } else {		return 0;	    }	}//	for(i = 0; i < current->lines; i ++)//		printf("currect: %s\n" , current->text[i]);    } while (state != 99);    // For the last subtitle    if (current->end <= 0) {        current->end = current->start + sub_slacktime;	*p = '\0'; trail_space (text);	if (text[0] != '\0')	    current->text[current->lines++] = strdup (text);    }    return current;}static RMascii *sub_readtext(RMascii *source, RMascii **dest) {    int len=0;    RMascii *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) {    RMascii line[LINE_LEN+1];    RMascii line2[LINE_LEN+1];    RMascii *p, *next;    int i;    do {	if (!stream_read_line (st, line, LINE_LEN)) return NULL;    } while ((sscanf (line,		      "{%ld}{}%[^\r\n]",		      &(current->start), line2) < 2) &&	     (sscanf (line,		      "{%ld}{%ld}%[^\r\n]",		      &(current->start), &(current->end), line2) < 3));    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) { current->lines=i;return current;}    }    current->lines= ++i;    return current;}static subtitle *sub_read_line_mpl2(stream_t *st,subtitle *current) {    RMascii line[LINE_LEN+1];    RMascii line2[LINE_LEN+1];    RMascii *p, *next;    int i;    do {	if (!stream_read_line (st, line, LINE_LEN)) 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) { current->lines=i;return current;}    }    current->lines= ++i;    return current;}static subtitle *sub_read_line_subrip(stream_t* st, subtitle *current) {    RMascii line[LINE_LEN+1];    int a1,a2,a3,a4,b1,b2,b3,b4;    RMascii *p=NULL, *q=NULL;    int len;    while (1) {	if (!stream_read_line (st, line, LINE_LEN)) 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)) 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_read_line_subviewer(stream_t *st,subtitle *current) {    RMascii line[LINE_LEN+1];    int a1,a2,a3,a4,b1,b2,b3,b4;    RMascii *p=NULL;    int i,len;    while (!current->text[0]) {	if (!stream_read_line (st, line, LINE_LEN)) return NULL;//	printf("sub_read_line_subviewer: %s\n", line);	if ((len=sscanf (line, "%d:%d:%d%[,.:]%d --> %d:%d:%d%[,.:]%d",&a1,&a2,&a3,(RMascii *)&i,&a4,&b1,&b2,&b3,(RMascii *)&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;) {	    if (!stream_read_line (st, line, LINE_LEN)) break;	    len=0;	    for (p=line; *p!='\n' && *p!='\r' && *p; p++,len++);	    if (len) {                int j=0,skip=0;		RMascii *curptr=MALLOC (len+1);		current->text[i]=curptr;		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) {    RMascii line[LINE_LEN+1];    int a1,a2,a3,a4;    RMascii *p=NULL;    int i,len;    while (!current->text[0]) {        if (!stream_read_line (st, line, LINE_LEN)) 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)) 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;                RMNCopyAscii (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) {	RMascii line[LINE_LEN+1];	int a1,a2,a3;