subreader.c

自己移植的linux下的流媒体播放器原代码,支持mms协议,支持ftp和http协议.

/*
 * 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 "subreader.h"

#ifdef HAVE_ENCA
#include <enca.h>
#endif

#define ERR ((void *) -1)

#ifdef USE_ICONV
#include <iconv.h>
char *sub_cp=NULL;
#endif
#ifdef USE_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

extern char* dvdsub_lang;

/* 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 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;
#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;
}

subtitle *sub_read_line_sami(FILE *fd, subtitle *current) {
    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 = fgets(line, LINE_LEN, fd))) 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 (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 (!strncasecmp (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 (!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 = fgets (line, LINE_LEN, fd))) {
	    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;
	*p = '\0'; trail_space (text);
	if (text[0] != '\0')
	    current->text[current->lines++] = strdup (text);
    }
    
    return current;
}


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= (char *)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
}

subtitle *sub_read_line_microdvd(FILE *fd,subtitle *current) {
    char line[LINE_LEN+1];
    char line2[LINE_LEN+1];
    char *p, *next;
    int i;

    do {
	if (!fgets (line, LINE_LEN, fd)) 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) { mp_msg(MSGT_SUBREADER,MSGL_WARN,"Too many lines in a subtitle\n");current->lines=i;return current;}
    }
    current->lines= ++i;

    return current;
}

subtitle *sub_read_line_mpl2(FILE *fd,subtitle *current) {
    char line[LINE_LEN+1];
    char line2[LINE_LEN+1];
    char *p, *next;
    int i;

    do {
	if (!fgets (line, LINE_LEN, fd)) 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;
}

subtitle *sub_read_line_subrip(FILE *fd, subtitle *current) {
    char line[LINE_LEN+1];
    int a1,a2,a3,a4,b1,b2,b3,b4;
    char *p=NULL, *q=NULL;
    int len;
    
    while (1) {
	if (!fgets (line, LINE_LEN, fd)) 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 (!fgets (line, LINE_LEN, fd)) 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]=(char *)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;
}

subtitle *sub_read_line_subviewer(FILE *fd,subtitle *current) {
    char line[LINE_LEN+1];
    int a1,a2,a3,a4,b1,b2,b3,b4;
    char *p=NULL;
    int i,len;
    
    while (!current->text[0]) {
	if (!fgets (line, LINE_LEN, fd)) 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;) {
	    if (!fgets (line, LINE_LEN, fd)) break;
	    len=0;
	    for (p=line; *p!='\n' && *p!='\r' && *p; p++,len++);
	    if (len) {
                int j=0,skip=0;
		char *curptr=current->text[i]=(char *)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;
}

subtitle *sub_read_line_subviewer2(FILE *fd,subtitle *current) {
    char line[LINE_LEN+1];
    int a1,a2,a3,a4;
    char *p=NULL;
    int i,len;
   
    while (!current->text[0]) {
        if (!fgets (line, LINE_LEN, fd)) 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 (!fgets (line, LINE_LEN, fd)) break;
            if (line[0]=='}') break;
            len=0;
            for (p=line; *p!='\n' && *p!='\r' && *p; ++p,++len);
            if (len) {
                current->text[i]=(char *)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;
}


subtitle *sub_read_line_vplayer(FILE *fd,subtitle *current) {
	char line[LINE_LEN+1];
	int a1,a2,a3;
	char *p=NULL, *next,separator;
	int i,len,plen;

	while (!current->text[0]) {
		if (!fgets (line, LINE_LEN, fd)) 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!='|') {