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;
int len;
retry:
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;
if (!args->previous_sub_end)
args->previous_sub_end = (current->start) ? current->start - 1 : 0;
if (!stream_read_line (st, line, LINE_LEN, utf16))
return NULL;
current->text[0]=""; // just to be sure that string is clear
if (set_multiline_text(args, current, line, 0) == ERR)
return ERR;
if (!strlen(current->text[0]) && current->lines <= 1)
goto retry;
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_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_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];
int i;
/* Parse SubRip header */
if (!stream_read_line(st, line, LINE_LEN, utf16))
return NULL;
if (sscanf(line, "%d:%d:%d%*1[,.:]%d --> %d:%d:%d%*1[,.:]%d",
&a1, &a2, &a3, &a4, &b1, &b2, &b3, &b4) < 8)
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';
snprintf(&full_line[j], sizeof(full_line) - j, "%s", line);
full_line[LINE_LEN] = '\0';
j += len;
}
if (full_line[0]) {
current->text[0] = strdup(full_line);
current->lines = 1;
}
}
return current;
}
static char *pl_get_line0(struct pl_parser *p)
{
char *res = stream_read_line(p->s, p->buffer, sizeof(p->buffer), p->utf16);
if (res) {
int len = strlen(res);
if (len > 0 && res[len - 1] == '\n')
res[len - 1] = '\0';
}
return res;
}
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);
args->mpsub_position += a*args->mpsub_multiplier;
current->start=(int) args->mpsub_position;
args->mpsub_position += b*args->mpsub_multiplier;
current->end=(int) args->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
}
static int sub_autodetect (stream_t* st, int *uses_time, int utf16) {
char line[LINE_LEN+1];
int i,j=0;
while (j < 100) {
j++;
if (!stream_read_line (st, line, LINE_LEN, utf16))
return SUB_INVALID;
if (sscanf (line, "{%d}{%d}", &i, &i)==2)
{*uses_time=0;return SUB_MICRODVD;}
if (sscanf (line, "{%d}{}", &i)==1)
{*uses_time=0;return SUB_MICRODVD;}
if (sscanf (line, "[%d][%d]", &i, &i)==2)
{*uses_time=1;return SUB_MPL2;}
if (sscanf (line, "%d:%d:%d.%d,%d:%d:%d.%d", &i, &i, &i, &i, &i, &i, &i, &i)==8)
{*uses_time=1;return SUB_SUBRIP;}
if (sscanf (line, "%d:%d:%d%*1[,.:]%d --> %d:%d:%d%*1[,.:]%d", &i, &i, &i, &i, &i, &i, &i, &i) == 8)
{*uses_time=1;return SUB_SUBVIEWER;}
if (sscanf (line, "{T %d:%d:%d:%d",&i, &i, &i, &i)==4)
{*uses_time=1;return SUB_SUBVIEWER2;}
if (strstr (line, "<SAMI>"))
{*uses_time=1; return SUB_SAMI;}
if (sscanf(line, "%d:%d:%d.%d %d:%d:%d.%d", &i, &i, &i, &i, &i, &i, &i, &i) == 8)
{*uses_time = 1; return SUB_JACOSUB;}
if (sscanf(line, "@%d @%d", &i, &i) == 2)
{*uses_time = 1; return SUB_JACOSUB;}
if (sscanf (line, "%d:%d:%d:", &i, &i, &i )==3)
{*uses_time=1;return SUB_VPLAYER;}
if (sscanf (line, "%d:%d:%d ", &i, &i, &i )==3)
{*uses_time=1;return SUB_VPLAYER;}
if (!strncasecmp(line, "<window", 7))
{*uses_time=1;return SUB_RT;}
if (!memcmp(line, "Dialogue: Marked", 16))
{*uses_time=1; return SUB_SSA;}
if (!memcmp(line, "Dialogue: ", 10))
{*uses_time=1; return SUB_SSA;}
if (sscanf (line, "%d,%d,\"%c", &i, &i, (char *) &i) == 3)
{*uses_time=1;return SUB_PJS;}
if (sscanf (line, "FORMAT=%d", &i) == 1)
{*uses_time=0; return SUB_MPSUB;}
if (!memcmp(line, "FORMAT=TIME", 11))
{*uses_time=1; return SUB_MPSUB;}
if (strstr (line, "-->>"))
{*uses_time=0; return SUB_AQTITLE;}
if (sscanf (line, "[%d:%d:%d]", &i, &i, &i)==3)
{*uses_time=1;return SUB_SUBRIP09;}
}
return SUB_INVALID; // too many bad lines
}
static subtitle *sub_read_line_aqt(stream_t *st,subtitle *current,
struct readline_args *args)
{
int utf16 = args->utf16;
char line[LINE_LEN+1];
retry:
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;
}
if (!args->previous_sub_end)
args->previous_sub_end = (current->start) ? current->start - 1 : 0;
if (!stream_read_line (st, line, LINE_LEN, utf16))
return NULL;
sub_readtext((char *) &line,¤t->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;
if (set_multiline_text(args, current, line, 1) == ERR)
return ERR;
if (!strlen(current->text[0]) && !strlen(current->text[1]))
goto retry;
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 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;
// 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;
return set_multiline_text(args, current, next, 0);
}
}
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];
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;
return set_multiline_text(args, current, line2, 0);
}
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, separator;
int 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("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 ];
if (*p!='|') {
//
return set_multiline_text(args, current, p, 0);
}
}
return current;
}
/*
* 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_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
long) (((a1 * 3600 + a2 * 60 + a3) * jacoTimeres + a4 +
jacoShift) * 100.0 / jacoTimeres);
current->end =
(unsigned
long) (((b1 * 3600 + b2 * 60 + b3) * jacoTimeres + b4 +
jacoShift) * 100.0 / jacoTimeres);
}
current->lines = 0;
p = line2;
while ((*p == ' ') || (*p == '\t')) {
++p;
}
if (isalpha(*p)||*p == '[') {
int cont, jLength;
if (sscanf(p, "%s %[^\n\r]", directive, line1) < 2)
return (subtitle *) ERR;
jLength = strlen(directive);
for (cont = 0; cont < jLength; ++cont) {
if (isalpha(*(directive + cont)))
*(directive + cont) = toupper(*(directive + cont));
}
if ((strstr(directive, "RDB") != NULL)
|| (strstr(directive, "RDC") != NULL)
|| (strstr(directive, "RLB") != NULL)
|| (strstr(directive, "RLG") != NULL)) {
continue;
}
if (strstr(directive, "JL") != NULL) {
current->alignment = SUB_ALIGNMENT_BOTTOMLEFT;
} else if (strstr(directive, "JR") != NULL) {
current->alignment = SUB_ALIGNMENT_BOTTOMRIGHT;
} else {
current->alignment = SUB_ALIGNMENT_BOTTOMCENTER;
}
snprintf(line2, sizeof(line2), "%s", line1);
p = line2;
}
for (q = line1; (!eol(*p)) && (current->lines < SUB_MAX_TEXT); ++p) {
switch (*p) {
case '{':
comment++;
break;
case '}':
if (comment) {
--comment;
//the next line to get rid of a blank after the comment
if ((*(p + 1)) == ' ')
p++;
}
break;
case '~':
if (!comment) {
*q = ' ';
++q;
}
break;
case ' ':
case '\t':
if ((*(p + 1) == ' ') || (*(p + 1) == '\t'))
break;
if (!comment) {
*q = ' ';
++q;
}
break;
case '\\':
if (*(p + 1) == 'n') {
*q = '\0';
q = line1;
current->text[current->lines++] = strdup(line1);
++p;
break;
}
if ((toupper(*(p + 1)) == 'C')
|| (toupper(*(p + 1)) == 'F')) {
++p,++p;
break;
}
if ((*(p + 1) == 'B') || (*(p + 1) == 'b') || (*(p + 1) == 'D') || //actually this means "insert current date here"
(*(p + 1) == 'I') || (*(p + 1) == 'i') || (*(p + 1) == 'N') || (*(p + 1) == 'T') || //actually this means "insert current time here"
(*(p + 1) == 'U') || (*(p + 1) == 'u')) {
++p;
break;
}
if ((*(p + 1) == '\\') ||
(*(p + 1) == '~') || (*(p + 1) == '{')) {
++p;
} else if (eol(*(p + 1))) {
if (!stream_read_line(st, directive, LINE_LEN, utf16))
return NULL;
trail_space(directive);
av_strlcat(line2, directive, LINE_LEN);
break;
}
default:
if (!comment) {
*q = *p;
++q;
}
} //-- switch
} //-- for
*q = '\0';
if (current->lines < SUB_MAX_TEXT)
current->text[current->lines] = strdup(line1);
} //-- while
if (current->lines < SUB_MAX_TEXT)
current->lines++;
return current;
}
static subtitle *sub_read_line_ssa(stream_t *st,subtitle *current,
struct readline_args *args)
{
/* Instead of hardcoding the expected fields and their order on
* each dialogue line, this code should parse the "Format: " line
* which lists the fields used in the script. As is, this may not
* work correctly with all scripts.
*/
int utf16 = args->utf16;
int comma;
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]", ¬hing,
&hour1, &min1, &sec1, &hunsec1,
&hour2, &min2, &sec2, &hunsec2,
line3) < 9
&&
sscanf (line, "Dialogue: %d,%d:%d:%d.%d,%d:%d:%d.%d"
"%[^\n\r]", ¬hing,
&hour1, &min1, &sec1, &hunsec1,
&hour2, &min2, &sec2, &hunsec2,
line3) < 9 );
line2=strchr(line3, ',');
if (!line2) return NULL;
for (comma = 3; comma < 9; comma ++)
if (!(line2 = strchr(++line2, ',')))
return NULL;
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 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)
args->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 (p - text >= LINE_LEN)
sami_add_line(current, text, &p);
if (*s == '\0') break;
else if (!strncasecmp (s, "<br>", 4)) {
sami_add_line(current, text, &p);
s += 4;
}
else if (*s == '{') { state = 5; ++s; continue; }
else if (*s == '<') { state = 4; }
else if (!strncasecmp (s, " ", 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')) {
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 + args->sub_slacktime;
sami_add_line(current, text, &p);
}
return current;
}
unsigned int stacktrace_capture( void** trace, unsigned int max_depth, unsigned int skip_frames )
{
unsigned int num_frames = 0;
if( !trace )
return 0;
if( !max_depth )
max_depth = BUILD_SIZE_STACKTRACE_DEPTH;
if( max_depth > BUILD_SIZE_STACKTRACE_DEPTH )
max_depth = BUILD_SIZE_STACKTRACE_DEPTH;
if( !_stackwalk_initialized )
{
if( !_initialize_stackwalker() )
{
memset( trace, 0, sizeof( void* ) * max_depth );
return num_frames;
}
}
#if FOUNDATION_PLATFORM_WINDOWS && ( FOUNDATION_COMPILER_MSVC || FOUNDATION_COMPILER_INTEL )
// Add 1 skip frame for this function call
++skip_frames;
# if USE_CAPTURESTACKBACKTRACE
if( CallRtlCaptureStackBackTrace )
{
void* local_trace[BUILD_SIZE_STACKTRACE_DEPTH];
if( max_depth + skip_frames > BUILD_SIZE_STACKTRACE_DEPTH )
max_depth = BUILD_SIZE_STACKTRACE_DEPTH - skip_frames;
num_frames = (unsigned int)CallRtlCaptureStackBackTrace( skip_frames, max_depth, local_trace, 0 );
if( num_frames > max_depth )
num_frames = max_depth;
memcpy( trace, local_trace, sizeof( void* ) * num_frames );
memset( trace + num_frames, 0, sizeof( void* ) * ( max_depth - num_frames ) );
}
else
{
# else
{
# endif
# if FOUNDATION_PLATFORM_ARCH_X86_64
// Raise an exception so helper has access to context record.
__try
{
RaiseException( 0, // Application-defined exception code.
0, // Zero indicates continuable exception.
0, // Number of arguments in args array (ignored if args is null)
0 ); // Array of arguments
}
__except( _capture_stack_trace_helper( trace, max_depth, skip_frames, (GetExceptionInformation())->ContextRecord ) )
{
}
# else
// Use a bit of inline assembly to capture the information relevant to stack walking which is
// basically EIP and EBP.
CONTEXT context;
memset( &context, 0, sizeof( CONTEXT ) );
context.ContextFlags = CONTEXT_FULL;
log_warnf( WARNING_DEPRECATED, "********** REIMPLEMENT FALLBACK STACKTRACE **********" );
/* Use a fake function call to pop the return address and retrieve EIP.*/
__asm
{
call FakeStackTraceCall
FakeStackTraceCall:
pop eax
mov context.Eip, eax
mov context.Ebp, ebp
mov context.Esp, esp
}
// Capture the back trace.
_capture_stack_trace_helper( trace, max_depth, skip_frames, &context );
# endif
}
#elif FOUNDATION_PLATFORM_APPLE
//TODO: Implement
#elif FOUNDATION_PLATFORM_POSIX
// Add 1 skip frames for this function call
skip_frames += 1;
void* localframes[BUILD_SIZE_STACKTRACE_DEPTH];
num_frames = (unsigned int)backtrace( localframes, BUILD_SIZE_STACKTRACE_DEPTH );
if( num_frames > skip_frames )
{
num_frames -= skip_frames;
if( num_frames > max_depth )
num_frames = max_depth;
memcpy( trace, localframes + skip_frames, sizeof( void* ) * num_frames );
}
else
trace[0] = 0;
#endif
return num_frames;
}
static bool _symbol_resolve_initialized = false;
static bool _initialize_symbol_resolve()
{
if( _symbol_resolve_initialized )
return true;
#if FOUNDATION_PLATFORM_WINDOWS
{
unsigned int options;
void* dll = LoadLibraryA( "PSAPI.DLL" );
if( !dll )
return _symbol_resolve_initialized;
CallEnumProcesses = (EnumProcessesFn)GetProcAddress( dll, "EnumProcesses" );
CallEnumProcessModules = (EnumProcessModulesFn)GetProcAddress( dll, "EnumProcessModules" );
CallGetModuleFileNameEx = (GetModuleFileNameExFn)GetProcAddress( dll, "GetModuleFileNameExA" );
CallGetModuleBaseName = (GetModuleBaseNameFn)GetProcAddress( dll, "GetModuleBaseNameA" );
CallGetModuleInformation = (GetModuleInformationFn)GetProcAddress( dll, "GetModuleInformation" );
if( !CallEnumProcesses || !CallEnumProcessModules || !CallGetModuleFileNameEx || !CallGetModuleBaseName || !CallGetModuleInformation )
return _symbol_resolve_initialized;
dll = LoadLibraryA( "DBGHELP.DLL" );
if( !dll )
return _symbol_resolve_initialized;
CallSymInitialize = (SymInitializeFn)GetProcAddress( dll, "SymInitialize" );
CallSymSetOptions = (SymSetOptionsFn)GetProcAddress( dll, "SymSetOptions" );
CallSymGetOptions = (SymGetOptionsFn)GetProcAddress( dll, "SymGetOptions" );
CallSymLoadModule64 = (SymLoadModule64Fn)GetProcAddress( dll, "SymLoadModule64" );
CallSymSetSearchPath = (SymSetSearchPathFn)GetProcAddress( dll, "SymSetSearchPath" );
CallSymGetModuleInfo64 = (SymGetModuleInfo64Fn)GetProcAddress( dll, "SymGetModuleInfo64" );
CallSymGetLineFromAddr64 = (SymGetLineFromAddr64Fn)GetProcAddress( dll, "SymGetLineFromAddr64" );
CallSymGetSymFromAddr64 = (SymGetSymFromAddr64Fn)GetProcAddress( dll, "SymGetSymFromAddr64" );
CallSymGetModuleBase64 = (SymGetModuleBase64Fn)GetProcAddress( dll, "SymGetModuleBase64" );
CallSymFunctionTableAccess64 = (SymFunctionTableAccess64Fn)GetProcAddress( dll, "SymFunctionTableAccess64" );
if( !CallSymInitialize || !CallSymSetOptions || !CallSymGetOptions || !CallSymLoadModule64 || !CallSymSetSearchPath || !CallSymGetModuleInfo64 || !CallSymGetLineFromAddr64 || !CallSymGetSymFromAddr64 || !CallSymGetModuleBase64 || !CallSymFunctionTableAccess64 )
return _symbol_resolve_initialized;
options = CallSymGetOptions();
options |= SYMOPT_LOAD_LINES;
options |= SYMOPT_DEBUG;
options |= SYMOPT_UNDNAME;
options |= SYMOPT_LOAD_LINES;
options |= SYMOPT_FAIL_CRITICAL_ERRORS;
options |= SYMOPT_DEFERRED_LOADS;
options |= SYMOPT_ALLOW_ABSOLUTE_SYMBOLS;
options |= SYMOPT_EXACT_SYMBOLS;
options |= SYMOPT_CASE_INSENSITIVE;
CallSymSetOptions( options );
CallSymInitialize( GetCurrentProcess(), 0, TRUE );
}
_load_process_modules();
_symbol_resolve_initialized = true;
#else
_symbol_resolve_initialized = true;
#endif
return _symbol_resolve_initialized;
}
static NOINLINE char** _resolve_stack_frames( void** frames, unsigned int max_frames )
{
#if FOUNDATION_PLATFORM_WINDOWS
char** lines = 0;
char symbol_buffer[ sizeof( IMAGEHLP_SYMBOL64 ) + 512 ];
PIMAGEHLP_SYMBOL64 symbol;
DWORD displacement = 0;
uint64_t displacement64 = 0;
unsigned int iaddr = 0;
unsigned int last_error;
bool found = false;
HANDLE process_handle = GetCurrentProcess();
int buffer_offset = 0;
bool last_was_main = false;
IMAGEHLP_LINE64 line64;
IMAGEHLP_MODULE64 module64;
for( iaddr = 0; ( iaddr < max_frames ) && !last_was_main; ++iaddr )
{
char* resolved = 0;
const char* function_name = "??";
const char* file_name = "??";
const char* module_name = "??";
unsigned int line_number = 0;
//Allow first frame to be null in case of a function call to a null pointer
if( iaddr && !frames[iaddr] )
break;
// Initialize symbol.
symbol = (PIMAGEHLP_SYMBOL64)symbol_buffer;
memset( symbol, 0, sizeof( symbol_buffer ) );
symbol->SizeOfStruct = sizeof( symbol_buffer );
symbol->MaxNameLength = 512;
// Get symbol from address.
if( CallSymGetSymFromAddr64 && CallSymGetSymFromAddr64( process_handle, (uint64_t)((uintptr_t)frames[iaddr]), &displacement64, symbol ) )
{
int offset = 0;
while( symbol->Name[offset] < 32 )
++offset;
function_name = symbol->Name + offset;
}
else
{
// No symbol found for this address.
last_error = GetLastError();
}
memset( &line64, 0, sizeof( line64 ) );
line64.SizeOfStruct = sizeof( line64 );
if( CallSymGetLineFromAddr64 && CallSymGetLineFromAddr64( process_handle, (uint64_t)((uintptr_t)frames[iaddr]), &displacement, &line64 ) )
{
file_name = line64.FileName;
line_number = line64.LineNumber;
}
memset( &module64, 0, sizeof( module64 ) );
module64.SizeOfStruct = sizeof( module64 );
if( CallSymGetModuleInfo64 && CallSymGetModuleInfo64( process_handle, (uint64_t)((uintptr_t)frames[iaddr]), &module64 ) )
{
int last_slash = STRING_NPOS;
module_name = module64.ImageName;
last_slash = string_rfind( module_name, '\\', STRING_NPOS );
if( last_slash != STRING_NPOS )
module_name += last_slash + 1;
}
resolved = string_format( "[" STRING_FORMAT_POINTER "] %s (%s:%d +%d bytes) [in %s]", frames[iaddr], function_name, file_name, line_number, displacement, module_name );
array_push( lines, resolved );
if( string_equal( function_name, "main" ) )
last_was_main = true;
}
return lines;
#elif FOUNDATION_PLATFORM_LINUX
char** addrs = 0;
char** lines = 0;
const char** args = 0;
process_t* proc = process_allocate();
unsigned int num_frames = 0;
unsigned int requested_frames = 0;
bool last_was_main = false;
if( !string_length( environment_executable_path() ) )
{
for( unsigned int iaddr = 0; iaddr < max_frames; ++iaddr )
{
//Allow first frame to be null in case of a function call to a null pointer
if( iaddr && !frames[iaddr] )
break;
array_push( lines, string_format( "[" STRING_FORMAT_POINTER "]", frames[iaddr] ) );
}
return lines;
}
array_push( args, "-e" );
array_push( args, environment_executable_path() );
array_push( args, "-f" );
for( unsigned int iaddr = 0; iaddr < max_frames; ++iaddr )
{
//Allow first frame to be null in case of a function call to a null pointer
if( iaddr && !frames[iaddr] )
break;
char* addr = string_format( STRING_FORMAT_POINTER, frames[iaddr] );
array_push( addrs, addr );
array_push( args, addr );
++requested_frames;
}
process_set_working_directory( proc, environment_initial_working_directory() );
process_set_executable_path( proc, "/usr/bin/addr2line" );
process_set_arguments( proc, args, array_size( args ) );
process_set_flags( proc, PROCESS_ATTACHED | PROCESS_STDSTREAMS );
process_spawn( proc );
stream_t* procout = process_stdout( proc );
while( !stream_eos( procout ) && ( num_frames < requested_frames ) && !last_was_main )
{
char* function = stream_read_line( procout, '\n' );
char* filename = stream_read_line( procout, '\n' );
array_push( lines, string_format( "[" STRING_FORMAT_POINTER "] %s (%s)",
frames[num_frames],
function && string_length( function ) ? function : "??",
filename && string_length( filename ) ? filename : "??"
) );
if( string_equal( function, "main" ) )
last_was_main = true;
string_deallocate( function );
string_deallocate( filename );
++num_frames;
}
process_wait( proc );
process_deallocate( proc );
string_array_deallocate( addrs );
array_deallocate( args );
return lines;
#else
char** lines = 0;
for( unsigned int iaddr = 0; iaddr < max_frames; ++iaddr )
{
//Allow first frame to be null in case of a function call to a null pointer
if( iaddr && !frames[iaddr] )
break;
array_push( lines, string_format( "[" STRING_FORMAT_POINTER "]\n", frames[iaddr] ) );
}
return lines;
#endif
}
