2 * Copyright (C) <1999> Erik Walthinsen <omega@cse.ogi.edu>
3 * Copyright (C) 2004 Ronald S. Bultje <rbultje@ronald.bitfreak.net>
4 * Copyright (C) 2006 Tim-Philipp Müller <tim centricular net>
6 * This library is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Library General Public
8 * License as published by the Free Software Foundation; either
9 * version 2 of the License, or (at your option) any later version.
11 * This library is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * Library General Public License for more details.
16 * You should have received a copy of the GNU Library General Public
17 * License along with this library; if not, write to the
18 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 * Boston, MA 02111-1307, USA.
28 #include <sys/types.h>
32 #include "gstsubparse.h"
33 #include "gstssaparse.h"
34 #include "samiparse.h"
35 #include "tmplayerparse.h"
36 #include "mpl2parse.h"
38 GST_DEBUG_CATEGORY (sub_parse_debug);
40 #define DEFAULT_ENCODING NULL
49 gst_sub_parse_set_property (GObject * object, guint prop_id,
50 const GValue * value, GParamSpec * pspec);
52 gst_sub_parse_get_property (GObject * object, guint prop_id,
53 GValue * value, GParamSpec * pspec);
56 static const GstElementDetails sub_parse_details =
57 GST_ELEMENT_DETAILS ("Subtitle parser",
58 "Codec/Parser/Subtitle",
59 "Parses subtitle (.sub) files into text streams",
60 "Gustavo J. A. M. Carneiro <gjc@inescporto.pt>\n"
61 "Ronald S. Bultje <rbultje@ronald.bitfreak.net>");
63 #ifndef GST_DISABLE_XML
64 static GstStaticPadTemplate sink_templ = GST_STATIC_PAD_TEMPLATE ("sink",
67 GST_STATIC_CAPS ("application/x-subtitle; application/x-subtitle-sami; "
68 "application/x-subtitle-tmplayer; application/x-subtitle-mpl2")
71 static GstStaticPadTemplate sink_templ = GST_STATIC_PAD_TEMPLATE ("sink",
74 GST_STATIC_CAPS ("application/x-subtitle")
78 static GstStaticPadTemplate src_templ = GST_STATIC_PAD_TEMPLATE ("src",
81 GST_STATIC_CAPS ("text/plain; text/x-pango-markup")
84 static void gst_sub_parse_base_init (GstSubParseClass * klass);
85 static void gst_sub_parse_class_init (GstSubParseClass * klass);
86 static void gst_sub_parse_init (GstSubParse * subparse);
88 static gboolean gst_sub_parse_src_event (GstPad * pad, GstEvent * event);
89 static gboolean gst_sub_parse_src_query (GstPad * pad, GstQuery * query);
90 static gboolean gst_sub_parse_sink_event (GstPad * pad, GstEvent * event);
92 static GstStateChangeReturn gst_sub_parse_change_state (GstElement * element,
93 GstStateChange transition);
95 static GstFlowReturn gst_sub_parse_chain (GstPad * sinkpad, GstBuffer * buf);
97 static GstElementClass *parent_class = NULL;
100 gst_sub_parse_get_type (void)
102 static GType sub_parse_type = 0;
104 if (!sub_parse_type) {
105 static const GTypeInfo sub_parse_info = {
106 sizeof (GstSubParseClass),
107 (GBaseInitFunc) gst_sub_parse_base_init,
109 (GClassInitFunc) gst_sub_parse_class_init,
112 sizeof (GstSubParse),
114 (GInstanceInitFunc) gst_sub_parse_init,
117 sub_parse_type = g_type_register_static (GST_TYPE_ELEMENT,
118 "GstSubParse", &sub_parse_info, 0);
121 return sub_parse_type;
125 gst_sub_parse_base_init (GstSubParseClass * klass)
127 GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
129 gst_element_class_add_pad_template (element_class,
130 gst_static_pad_template_get (&sink_templ));
131 gst_element_class_add_pad_template (element_class,
132 gst_static_pad_template_get (&src_templ));
133 gst_element_class_set_details (element_class, &sub_parse_details);
137 gst_sub_parse_dispose (GObject * object)
139 GstSubParse *subparse = GST_SUBPARSE (object);
141 GST_DEBUG_OBJECT (subparse, "cleaning up subtitle parser");
143 if (subparse->encoding) {
144 g_free (subparse->encoding);
145 subparse->encoding = NULL;
148 if (subparse->detected_encoding) {
149 g_free (subparse->detected_encoding);
150 subparse->detected_encoding = NULL;
153 if (subparse->adapter) {
154 gst_object_unref (subparse->adapter);
155 subparse->adapter = NULL;
158 if (subparse->textbuf) {
159 g_string_free (subparse->textbuf, TRUE);
160 subparse->textbuf = NULL;
162 #ifndef GST_DISABLE_XML
163 sami_context_deinit (&subparse->state);
166 GST_CALL_PARENT (G_OBJECT_CLASS, dispose, (object));
170 gst_sub_parse_class_init (GstSubParseClass * klass)
172 GObjectClass *object_class = G_OBJECT_CLASS (klass);
173 GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
175 parent_class = g_type_class_peek_parent (klass);
177 object_class->dispose = gst_sub_parse_dispose;
178 object_class->set_property = gst_sub_parse_set_property;
179 object_class->get_property = gst_sub_parse_get_property;
181 element_class->change_state = gst_sub_parse_change_state;
183 g_object_class_install_property (object_class, PROP_ENCODING,
184 g_param_spec_string ("subtitle-encoding", "subtitle charset encoding",
185 "Encoding to assume if input subtitles are not in UTF-8 or any other "
186 "Unicode encoding. If not set, the GST_SUBTITLE_ENCODING environment "
187 "variable will be checked for an encoding to use. If that is not set "
188 "either, ISO-8859-15 will be assumed.", DEFAULT_ENCODING,
189 G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
193 gst_sub_parse_init (GstSubParse * subparse)
195 subparse->sinkpad = gst_pad_new_from_static_template (&sink_templ, "sink");
196 gst_pad_set_chain_function (subparse->sinkpad,
197 GST_DEBUG_FUNCPTR (gst_sub_parse_chain));
198 gst_pad_set_event_function (subparse->sinkpad,
199 GST_DEBUG_FUNCPTR (gst_sub_parse_sink_event));
200 gst_element_add_pad (GST_ELEMENT (subparse), subparse->sinkpad);
202 subparse->srcpad = gst_pad_new_from_static_template (&src_templ, "src");
203 gst_pad_set_event_function (subparse->srcpad,
204 GST_DEBUG_FUNCPTR (gst_sub_parse_src_event));
205 gst_pad_set_query_function (subparse->srcpad,
206 GST_DEBUG_FUNCPTR (gst_sub_parse_src_query));
207 gst_element_add_pad (GST_ELEMENT (subparse), subparse->srcpad);
209 subparse->textbuf = g_string_new (NULL);
210 subparse->parser_type = GST_SUB_PARSE_FORMAT_UNKNOWN;
211 subparse->flushing = FALSE;
212 gst_segment_init (&subparse->segment, GST_FORMAT_TIME);
213 subparse->need_segment = TRUE;
214 subparse->encoding = g_strdup (DEFAULT_ENCODING);
215 subparse->detected_encoding = NULL;
216 subparse->adapter = gst_adapter_new ();
220 * Source pad functions.
224 gst_sub_parse_src_query (GstPad * pad, GstQuery * query)
226 GstSubParse *self = GST_SUBPARSE (gst_pad_get_parent (pad));
227 gboolean ret = FALSE;
229 GST_DEBUG ("Handling %s query", GST_QUERY_TYPE_NAME (query));
231 switch (GST_QUERY_TYPE (query)) {
232 case GST_QUERY_SEEKING:
235 gboolean seekable = FALSE;
239 gst_query_parse_seeking (query, &fmt, NULL, NULL, NULL);
240 if (fmt == GST_FORMAT_TIME) {
241 GstQuery *peerquery = gst_query_new_seeking (GST_FORMAT_BYTES);
243 seekable = gst_pad_peer_query (self->sinkpad, peerquery);
245 gst_query_parse_seeking (peerquery, NULL, &seekable, NULL, NULL);
246 gst_query_unref (peerquery);
249 gst_query_set_seeking (query, fmt, seekable, seekable ? 0 : -1, -1);
254 ret = gst_pad_peer_query (self->sinkpad, query);
258 gst_object_unref (self);
264 gst_sub_parse_src_event (GstPad * pad, GstEvent * event)
266 GstSubParse *self = GST_SUBPARSE (gst_pad_get_parent (pad));
267 gboolean ret = FALSE;
269 GST_DEBUG ("Handling %s event", GST_EVENT_TYPE_NAME (event));
271 switch (GST_EVENT_TYPE (event)) {
275 GstSeekType start_type, stop_type;
280 gst_event_parse_seek (event, &rate, &format, &self->segment_flags,
281 &start_type, &start, &stop_type, &stop);
283 if (format != GST_FORMAT_TIME) {
284 GST_WARNING_OBJECT (self, "we only support seeking in TIME format");
285 gst_event_unref (event);
289 /* Convert that seek to a seeking in bytes at position 0,
290 FIXME: could use an index */
291 ret = gst_pad_push_event (self->sinkpad,
292 gst_event_new_seek (rate, GST_FORMAT_BYTES, self->segment_flags,
293 GST_SEEK_TYPE_SET, 0, GST_SEEK_TYPE_NONE, 0));
296 /* Apply the seek to our segment */
297 gst_segment_set_seek (&self->segment, rate, format, self->segment_flags,
298 start_type, start, stop_type, stop, &update);
300 GST_DEBUG_OBJECT (self, "segment after seek: %" GST_SEGMENT_FORMAT,
303 self->next_offset = 0;
305 self->need_segment = TRUE;
307 GST_WARNING_OBJECT (self, "seek to 0 bytes failed");
310 gst_event_unref (event);
314 ret = gst_pad_event_default (pad, event);
319 gst_object_unref (self);
325 gst_sub_parse_set_property (GObject * object, guint prop_id,
326 const GValue * value, GParamSpec * pspec)
328 GstSubParse *subparse = GST_SUBPARSE (object);
330 GST_OBJECT_LOCK (subparse);
333 g_free (subparse->encoding);
334 subparse->encoding = g_value_dup_string (value);
335 GST_LOG_OBJECT (object, "subtitle encoding set to %s",
336 GST_STR_NULL (subparse->encoding));
339 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
342 GST_OBJECT_UNLOCK (subparse);
346 gst_sub_parse_get_property (GObject * object, guint prop_id,
347 GValue * value, GParamSpec * pspec)
349 GstSubParse *subparse = GST_SUBPARSE (object);
351 GST_OBJECT_LOCK (subparse);
354 g_value_set_string (value, subparse->encoding);
357 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
360 GST_OBJECT_UNLOCK (subparse);
364 gst_sub_parse_get_format_description (GstSubParseFormat format)
367 case GST_SUB_PARSE_FORMAT_MDVDSUB:
369 case GST_SUB_PARSE_FORMAT_SUBRIP:
371 case GST_SUB_PARSE_FORMAT_MPSUB:
373 case GST_SUB_PARSE_FORMAT_SAMI:
375 case GST_SUB_PARSE_FORMAT_TMPLAYER:
377 case GST_SUB_PARSE_FORMAT_MPL2:
379 case GST_SUB_PARSE_FORMAT_SUBVIEWER:
382 case GST_SUB_PARSE_FORMAT_UNKNOWN:
389 gst_convert_to_utf8 (const gchar * str, gsize len, const gchar * encoding,
390 gsize * consumed, GError ** err)
396 g_convert_with_fallback (str, len, "UTF-8", encoding, "*", consumed, NULL,
401 /* + 3 to skip UTF-8 BOM if it was added */
403 if (len >= 3 && (guint8) ret[0] == 0xEF && (guint8) ret[1] == 0xBB
404 && (guint8) ret[2] == 0xBF)
405 g_memmove (ret, ret + 3, len + 1 - 3);
411 detect_encoding (const gchar * str, gsize len)
413 if (len >= 3 && (guint8) str[0] == 0xEF && (guint8) str[1] == 0xBB
414 && (guint8) str[2] == 0xBF)
415 return g_strdup ("UTF-8");
417 if (len >= 2 && (guint8) str[0] == 0xFE && (guint8) str[1] == 0xFF)
418 return g_strdup ("UTF-16BE");
420 if (len >= 2 && (guint8) str[0] == 0xFF && (guint8) str[1] == 0xFE)
421 return g_strdup ("UTF-16LE");
423 if (len >= 4 && (guint8) str[0] == 0x00 && (guint8) str[1] == 0x00
424 && (guint8) str[2] == 0xFE && (guint8) str[3] == 0xFF)
425 return g_strdup ("UTF-32BE");
427 if (len >= 4 && (guint8) str[0] == 0xFF && (guint8) str[1] == 0xFE
428 && (guint8) str[2] == 0x00 && (guint8) str[3] == 0x00)
429 return g_strdup ("UTF-32LE");
435 convert_encoding (GstSubParse * self, const gchar * str, gsize len,
438 const gchar *encoding;
444 /* First try any detected encoding */
445 if (self->detected_encoding) {
447 gst_convert_to_utf8 (str, len, self->detected_encoding, consumed, &err);
452 GST_WARNING_OBJECT (self, "could not convert string from '%s' to UTF-8: %s",
453 self->detected_encoding, err->message);
454 g_free (self->detected_encoding);
455 self->detected_encoding = NULL;
459 /* Otherwise check if it's UTF8 */
460 if (self->valid_utf8) {
461 if (g_utf8_validate (str, len, NULL)) {
462 GST_LOG_OBJECT (self, "valid UTF-8, no conversion needed");
464 return g_strndup (str, len);
466 GST_INFO_OBJECT (self, "invalid UTF-8!");
467 self->valid_utf8 = FALSE;
470 /* Else try fallback */
471 encoding = self->encoding;
472 if (encoding == NULL || *encoding == '\0') {
473 encoding = g_getenv ("GST_SUBTITLE_ENCODING");
475 if (encoding == NULL || *encoding == '\0') {
476 /* if local encoding is UTF-8 and no encoding specified
477 * via the environment variable, assume ISO-8859-15 */
478 if (g_get_charset (&encoding)) {
479 encoding = "ISO-8859-15";
483 ret = gst_convert_to_utf8 (str, len, encoding, consumed, &err);
486 GST_WARNING_OBJECT (self, "could not convert string from '%s' to UTF-8: %s",
487 encoding, err->message);
490 /* invalid input encoding, fall back to ISO-8859-15 (always succeeds) */
491 ret = gst_convert_to_utf8 (str, len, "ISO-8859-15", consumed, NULL);
494 GST_LOG_OBJECT (self,
495 "successfully converted %" G_GSIZE_FORMAT " characters from %s to UTF-8"
496 "%s", len, encoding, (err) ? " , using ISO-8859-15 as fallback" : "");
502 get_next_line (GstSubParse * self)
505 const char *line_end;
507 gboolean have_r = FALSE;
509 line_end = strchr (self->textbuf->str, '\n');
512 /* end-of-line not found; return for more data */
516 /* get rid of '\r' */
517 if (line_end != self->textbuf->str && *(line_end - 1) == '\r') {
522 line_len = line_end - self->textbuf->str;
523 line = g_strndup (self->textbuf->str, line_len);
524 self->textbuf = g_string_erase (self->textbuf, 0,
525 line_len + (have_r ? 2 : 1));
530 parse_mdvdsub (ParserState * state, const gchar * line)
532 const gchar *line_split;
534 guint start_frame, end_frame;
535 gint64 clip_start = 0, clip_stop = 0;
536 gboolean in_seg = FALSE;
540 /* style variables */
545 if (sscanf (line, "{%u}{%u}", &start_frame, &end_frame) != 2) {
546 g_warning ("Parse of the following line, assumed to be in microdvd .sub"
547 " format, failed:\n%s", line);
551 /* skip the {%u}{%u} part */
552 line = strchr (line, '}') + 1;
553 line = strchr (line, '}') + 1;
555 /* see if there's a first line with a framerate */
556 if (state->fps == 0.0 && start_frame == 1 && end_frame == 1) {
557 gchar *rest, *end = NULL;
559 rest = g_strdup (line);
560 g_strdelimit (rest, ",", '.');
561 state->fps = g_ascii_strtod (rest, &end);
564 GST_INFO ("framerate from file: %f ('%s')", state->fps, rest);
569 if (state->fps == 0.0) {
570 /* FIXME: hardcoded for now, is there a better way/assumption? */
571 state->fps = 24000.0 / 1001.0;
572 GST_INFO ("no framerate specified, assuming %f", state->fps);
575 state->start_time = start_frame / state->fps * GST_SECOND;
576 state->duration = (end_frame - start_frame) / state->fps * GST_SECOND;
578 /* Check our segment start/stop */
579 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
580 state->start_time, state->start_time + state->duration, &clip_start,
583 /* No need to parse that text if it's out of segment */
585 state->start_time = clip_start;
586 state->duration = clip_stop - clip_start;
591 markup = g_string_new (NULL);
596 /* parse style markup */
597 if (strncmp (line, "{y:i}", 5) == 0) {
599 line = strchr (line, '}') + 1;
601 if (strncmp (line, "{y:b}", 5) == 0) {
603 line = strchr (line, '}') + 1;
605 if (sscanf (line, "{s:%u}", &fontsize) == 1) {
606 line = strchr (line, '}') + 1;
608 /* forward slashes at beginning/end signify italics too */
609 if (g_str_has_prefix (line, "/")) {
613 if ((line_split = strchr (line, '|')))
614 line_chunk = g_markup_escape_text (line, line_split - line);
616 line_chunk = g_markup_escape_text (line, strlen (line));
618 /* Remove italics markers at end of line/stanza (CHECKME: are end slashes
619 * always at the end of a line or can they span multiple lines?) */
620 if (g_str_has_suffix (line_chunk, "/")) {
621 line_chunk[strlen (line_chunk) - 1] = '\0';
624 markup = g_string_append (markup, "<span");
626 g_string_append (markup, " style=\"italic\"");
628 g_string_append (markup, " weight=\"bold\"");
630 g_string_append_printf (markup, " size=\"%u\"", fontsize * 1000);
631 g_string_append_printf (markup, ">%s</span>", line_chunk);
634 g_string_append (markup, "\n");
635 line = line_split + 1;
641 g_string_free (markup, FALSE);
642 GST_DEBUG ("parse_mdvdsub returning (%f+%f): %s",
643 state->start_time / (double) GST_SECOND,
644 state->duration / (double) GST_SECOND, ret);
649 strip_trailing_newlines (gchar * txt)
655 while (len > 1 && txt[len - 1] == '\n') {
662 /* we want to escape text in general, but retain basic markup like
663 * <i></i>, <u></u>, and <b></b>. The easiest and safest way is to
664 * just unescape a white list of allowed markups again after
665 * escaping everything (the text between these simple markers isn't
666 * necessarily escaped, so it seems best to do it like this) */
668 subrip_unescape_formatting (gchar * txt)
672 for (pos = txt; pos != NULL && *pos != '\0'; ++pos) {
673 if (g_ascii_strncasecmp (pos, "<u>", 9) == 0 ||
674 g_ascii_strncasecmp (pos, "<i>", 9) == 0 ||
675 g_ascii_strncasecmp (pos, "<b>", 9) == 0) {
677 pos[1] = g_ascii_tolower (pos[4]);
679 /* move NUL terminator as well */
680 g_memmove (pos + 3, pos + 9, strlen (pos + 9) + 1);
685 for (pos = txt; pos != NULL && *pos != '\0'; ++pos) {
686 if (g_ascii_strncasecmp (pos, "</u>", 10) == 0 ||
687 g_ascii_strncasecmp (pos, "</i>", 10) == 0 ||
688 g_ascii_strncasecmp (pos, "</b>", 10) == 0) {
691 pos[2] = g_ascii_tolower (pos[5]);
693 /* move NUL terminator as well */
694 g_memmove (pos + 4, pos + 10, strlen (pos + 10) + 1);
702 subrip_remove_unhandled_tag (gchar * start, gchar * stop)
706 tag = start + strlen ("<");
710 if (g_ascii_tolower (*tag) < 'a' || g_ascii_tolower (*tag) > 'z')
715 GST_LOG ("removing unhandled tag '%s'", start);
717 g_memmove (start, stop, strlen (stop) + 1);
721 /* remove tags we haven't explicitly allowed earlier on, like font tags
724 subrip_remove_unhandled_tags (gchar * txt)
728 for (pos = txt; pos != NULL && *pos != '\0'; ++pos) {
729 if (strncmp (pos, "<", 4) == 0 && (gt = strstr (pos + 4, ">"))) {
730 if (subrip_remove_unhandled_tag (pos, gt + strlen (">")))
736 /* we only allow <i>, <u> and <b>, so let's take a simple approach. This code
737 * assumes the input has been escaped and subrip_unescape_formatting() has then
738 * been run over the input! This function adds missing closing markup tags and
739 * removes broken closing tags for tags that have never been opened. */
741 subrip_fix_up_markup (gchar ** p_txt)
743 gchar *cur, *next_tag;
745 guint num_open_tags = 0;
747 g_assert (*p_txt != NULL);
750 while (*cur != '\0') {
751 next_tag = strchr (cur, '<');
752 if (next_tag == NULL)
758 if (num_open_tags == 0 || open_tags[num_open_tags - 1] != *next_tag) {
759 GST_LOG ("broken input, closing tag '%c' is not open", *next_tag);
760 g_memmove (next_tag - 2, next_tag + 2, strlen (next_tag + 2) + 1);
763 /* it's all good, closing tag which is open */
771 if (num_open_tags == G_N_ELEMENTS (open_tags))
772 return; /* something dodgy is going on, stop parsing */
773 open_tags[num_open_tags] = *next_tag;
777 GST_ERROR ("unexpected tag '%c' (%s)", *next_tag, next_tag);
778 g_assert_not_reached ();
784 if (num_open_tags > 0) {
787 s = g_string_new (*p_txt);
788 while (num_open_tags > 0) {
789 GST_LOG ("adding missing closing tag '%c'", open_tags[num_open_tags - 1]);
790 g_string_append_c (s, '<');
791 g_string_append_c (s, '/');
792 g_string_append_c (s, open_tags[num_open_tags - 1]);
793 g_string_append_c (s, '>');
797 *p_txt = g_string_free (s, FALSE);
802 parse_subrip_time (const gchar * ts_string, GstClockTime * t)
804 gchar s[128] = { '\0', };
806 guint hour, min, sec, msec, len;
808 while (*ts_string == ' ')
811 g_strlcpy (s, ts_string, sizeof (s));
812 if ((end = strstr (s, "-->")))
816 /* ms may be in these formats:
817 * hh:mm:ss,500 = 500ms
820 * hh:mm:ss, 50 = 50ms
822 * and sscanf() doesn't differentiate between ' 5' and '5' so munge
823 * the white spaces within the timestamp to '0' (I'm sure there's a
824 * way to make sscanf() do this for us, but how?)
826 g_strdelimit (s, " ", '0');
828 /* make sure we have exactly three digits after he comma */
830 g_assert (p != NULL);
837 g_strlcat (&p[len], "0", 2);
841 GST_LOG ("parsing timestamp '%s'", s);
842 if (sscanf (s, "%u:%u:%u,%u", &hour, &min, &sec, &msec) != 4) {
843 GST_WARNING ("failed to parse subrip timestamp string '%s'", s);
847 *t = ((hour * 3600) + (min * 60) + sec) * GST_SECOND + msec * GST_MSECOND;
852 parse_subrip (ParserState * state, const gchar * line)
857 switch (state->state) {
859 /* looking for a single integer */
860 if (sscanf (line, "%u", &subnum) == 1)
865 GstClockTime ts_start, ts_end;
868 /* looking for start_time --> end_time */
869 if ((end_time = strstr (line, " --> ")) &&
870 parse_subrip_time (line, &ts_start) &&
871 parse_subrip_time (end_time + strlen (" --> "), &ts_end) &&
872 state->start_time <= ts_end) {
874 state->start_time = ts_start;
875 state->duration = ts_end - ts_start;
877 GST_DEBUG ("error parsing subrip time line '%s'", line);
884 /* No need to parse that text if it's out of segment */
885 gint64 clip_start = 0, clip_stop = 0;
886 gboolean in_seg = FALSE;
888 /* Check our segment start/stop */
889 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
890 state->start_time, state->start_time + state->duration,
891 &clip_start, &clip_stop);
894 state->start_time = clip_start;
895 state->duration = clip_stop - clip_start;
901 /* looking for subtitle text; empty line ends this subtitle entry */
903 g_string_append_c (state->buf, '\n');
904 g_string_append (state->buf, line);
905 if (strlen (line) == 0) {
906 ret = g_markup_escape_text (state->buf->str, state->buf->len);
907 g_string_truncate (state->buf, 0);
909 subrip_unescape_formatting (ret);
910 subrip_remove_unhandled_tags (ret);
911 strip_trailing_newlines (ret);
912 subrip_fix_up_markup (&ret);
917 g_return_val_if_reached (NULL);
922 subviewer_unescape_newlines (gchar * read)
926 /* Replace all occurences of '[br]' with a newline as version 2
927 * of the subviewer format uses this for newlines */
929 if (read[0] == '\0' || read[1] == '\0' || read[2] == '\0' || read[3] == '\0')
933 if (strncmp (read, "[br]", 4) == 0) {
947 parse_subviewer (ParserState * state, const gchar * line)
949 guint h1, m1, s1, ms1;
950 guint h2, m2, s2, ms2;
953 /* TODO: Maybe also parse the fields in the header, especially DELAY.
954 * For examples see the unit test or
955 * http://www.doom9.org/index.html?/sub.htm */
957 switch (state->state) {
959 /* looking for start_time,end_time */
960 if (sscanf (line, "%u:%u:%u.%u,%u:%u:%u.%u",
961 &h1, &m1, &s1, &ms1, &h2, &m2, &s2, &ms2) == 8) {
964 (((guint64) h1) * 3600 + m1 * 60 + s1) * GST_SECOND +
967 (((guint64) h2) * 3600 + m2 * 60 + s2) * GST_SECOND +
968 ms2 * GST_MSECOND - state->start_time;
973 /* No need to parse that text if it's out of segment */
974 gint64 clip_start = 0, clip_stop = 0;
975 gboolean in_seg = FALSE;
977 /* Check our segment start/stop */
978 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
979 state->start_time, state->start_time + state->duration,
980 &clip_start, &clip_stop);
983 state->start_time = clip_start;
984 state->duration = clip_stop - clip_start;
990 /* looking for subtitle text; empty line ends this subtitle entry */
992 g_string_append_c (state->buf, '\n');
993 g_string_append (state->buf, line);
994 if (strlen (line) == 0) {
995 ret = g_strdup (state->buf->str);
996 subviewer_unescape_newlines (ret);
997 strip_trailing_newlines (ret);
998 g_string_truncate (state->buf, 0);
1004 g_assert_not_reached ();
1010 parse_mpsub (ParserState * state, const gchar * line)
1015 switch (state->state) {
1017 /* looking for two floats (offset, duration) */
1018 if (sscanf (line, "%f %f", &t1, &t2) == 2) {
1020 state->start_time += state->duration + GST_SECOND * t1;
1021 state->duration = GST_SECOND * t2;
1025 { /* No need to parse that text if it's out of segment */
1026 gint64 clip_start = 0, clip_stop = 0;
1027 gboolean in_seg = FALSE;
1029 /* Check our segment start/stop */
1030 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
1031 state->start_time, state->start_time + state->duration,
1032 &clip_start, &clip_stop);
1035 state->start_time = clip_start;
1036 state->duration = clip_stop - clip_start;
1042 /* looking for subtitle text; empty line ends this
1044 if (state->buf->len)
1045 g_string_append_c (state->buf, '\n');
1046 g_string_append (state->buf, line);
1047 if (strlen (line) == 0) {
1048 ret = g_strdup (state->buf->str);
1049 g_string_truncate (state->buf, 0);
1055 g_assert_not_reached ();
1061 parser_state_init (ParserState * state)
1063 GST_DEBUG ("initialising parser");
1066 g_string_truncate (state->buf, 0);
1068 state->buf = g_string_new (NULL);
1071 state->start_time = 0;
1072 state->duration = 0;
1073 state->max_duration = 0; /* no limit */
1075 state->segment = NULL;
1079 parser_state_dispose (ParserState * state)
1082 g_string_free (state->buf, TRUE);
1085 #ifndef GST_DISABLE_XML
1086 if (state->user_data) {
1087 sami_context_reset (state);
1092 /* regex type enum */
1095 GST_SUB_PARSE_REGEX_UNKNOWN = 0,
1096 GST_SUB_PARSE_REGEX_MDVDSUB = 1,
1097 GST_SUB_PARSE_REGEX_SUBRIP = 2,
1101 gst_sub_parse_data_format_autodetect_regex_once (GstSubParseRegex regtype)
1103 gpointer result = NULL;
1104 GError *gerr = NULL;
1106 case GST_SUB_PARSE_REGEX_MDVDSUB:
1108 (gpointer) g_regex_new ("^\\{[0-9]+\\}\\{[0-9]+\\}", 0, 0, &gerr);
1109 if (result == NULL) {
1110 g_warning ("Compilation of mdvd regex failed: %s", gerr->message);
1111 g_error_free (gerr);
1114 case GST_SUB_PARSE_REGEX_SUBRIP:
1115 result = (gpointer) g_regex_new ("^([ 0-9]){0,3}[0-9]\\s*(\x0d)?\x0a"
1116 "[ 0-9][0-9]:[ 0-9][0-9]:[ 0-9][0-9],[ 0-9]{0,2}[0-9]"
1117 " +--> +([ 0-9])?[0-9]:[ 0-9][0-9]:[ 0-9][0-9],[ 0-9]{0,2}[0-9]",
1119 if (result == NULL) {
1120 g_warning ("Compilation of subrip regex failed: %s", gerr->message);
1121 g_error_free (gerr);
1125 GST_WARNING ("Trying to allocate regex of unknown type %u", regtype);
1131 * FIXME: maybe we should pass along a second argument, the preceding
1132 * text buffer, because that is how this originally worked, even though
1133 * I don't really see the use of that.
1136 static GstSubParseFormat
1137 gst_sub_parse_data_format_autodetect (gchar * match_str)
1141 static GOnce mdvd_rx_once = G_ONCE_INIT;
1142 static GOnce subrip_rx_once = G_ONCE_INIT;
1147 g_once (&mdvd_rx_once,
1148 (GThreadFunc) gst_sub_parse_data_format_autodetect_regex_once,
1149 (gpointer) GST_SUB_PARSE_REGEX_MDVDSUB);
1150 g_once (&subrip_rx_once,
1151 (GThreadFunc) gst_sub_parse_data_format_autodetect_regex_once,
1152 (gpointer) GST_SUB_PARSE_REGEX_SUBRIP);
1154 mdvd_grx = (GRegex *) mdvd_rx_once.retval;
1155 subrip_grx = (GRegex *) subrip_rx_once.retval;
1157 if (g_regex_match (mdvd_grx, match_str, 0, NULL) == TRUE) {
1158 GST_LOG ("MicroDVD (frame based) format detected");
1159 return GST_SUB_PARSE_FORMAT_MDVDSUB;
1161 if (g_regex_match (subrip_grx, match_str, 0, NULL) == TRUE) {
1162 GST_LOG ("SubRip (time based) format detected");
1163 return GST_SUB_PARSE_FORMAT_SUBRIP;
1166 if (!strncmp (match_str, "FORMAT=TIME", 11)) {
1167 GST_LOG ("MPSub (time based) format detected");
1168 return GST_SUB_PARSE_FORMAT_MPSUB;
1170 #ifndef GST_DISABLE_XML
1171 if (strstr (match_str, "<SAMI>") != NULL ||
1172 strstr (match_str, "<sami>") != NULL) {
1173 GST_LOG ("SAMI (time based) format detected");
1174 return GST_SUB_PARSE_FORMAT_SAMI;
1177 /* we're boldly assuming the first subtitle appears within the first hour */
1178 if (sscanf (match_str, "0:%02u:%02u:", &n1, &n2) == 2 ||
1179 sscanf (match_str, "0:%02u:%02u=", &n1, &n2) == 2 ||
1180 sscanf (match_str, "00:%02u:%02u:", &n1, &n2) == 2 ||
1181 sscanf (match_str, "00:%02u:%02u=", &n1, &n2) == 2 ||
1182 sscanf (match_str, "00:%02u:%02u,%u=", &n1, &n2, &n3) == 3) {
1183 GST_LOG ("TMPlayer (time based) format detected");
1184 return GST_SUB_PARSE_FORMAT_TMPLAYER;
1186 if (sscanf (match_str, "[%u][%u]", &n1, &n2) == 2) {
1187 GST_LOG ("MPL2 (time based) format detected");
1188 return GST_SUB_PARSE_FORMAT_MPL2;
1190 if (strstr (match_str, "[INFORMATION]") != NULL) {
1191 GST_LOG ("SubViewer (time based) format detected");
1192 return GST_SUB_PARSE_FORMAT_SUBVIEWER;
1195 GST_DEBUG ("no subtitle format detected");
1196 return GST_SUB_PARSE_FORMAT_UNKNOWN;
1200 gst_sub_parse_format_autodetect (GstSubParse * self)
1203 GstSubParseFormat format;
1205 if (strlen (self->textbuf->str) < 30) {
1206 GST_DEBUG ("File too small to be a subtitles file");
1210 data = g_strndup (self->textbuf->str, 35);
1211 format = gst_sub_parse_data_format_autodetect (data);
1214 self->parser_type = format;
1215 self->subtitle_codec = gst_sub_parse_get_format_description (format);
1216 parser_state_init (&self->state);
1219 case GST_SUB_PARSE_FORMAT_MDVDSUB:
1220 self->parse_line = parse_mdvdsub;
1221 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1222 case GST_SUB_PARSE_FORMAT_SUBRIP:
1223 self->parse_line = parse_subrip;
1224 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1225 case GST_SUB_PARSE_FORMAT_MPSUB:
1226 self->parse_line = parse_mpsub;
1227 return gst_caps_new_simple ("text/plain", NULL);
1228 #ifndef GST_DISABLE_XML
1229 case GST_SUB_PARSE_FORMAT_SAMI:
1230 self->parse_line = parse_sami;
1231 sami_context_init (&self->state);
1232 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1234 case GST_SUB_PARSE_FORMAT_TMPLAYER:
1235 self->parse_line = parse_tmplayer;
1236 self->state.max_duration = 5 * GST_SECOND;
1237 return gst_caps_new_simple ("text/plain", NULL);
1238 case GST_SUB_PARSE_FORMAT_MPL2:
1239 self->parse_line = parse_mpl2;
1240 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1241 case GST_SUB_PARSE_FORMAT_SUBVIEWER:
1242 self->parse_line = parse_subviewer;
1243 return gst_caps_new_simple ("text/plain", NULL);
1244 case GST_SUB_PARSE_FORMAT_UNKNOWN:
1246 GST_DEBUG ("no subtitle format detected");
1247 GST_ELEMENT_ERROR (self, STREAM, WRONG_TYPE,
1248 ("The input is not a valid/supported subtitle file"), (NULL));
1254 feed_textbuf (GstSubParse * self, GstBuffer * buf)
1258 gchar *input = NULL;
1260 discont = GST_BUFFER_IS_DISCONT (buf);
1262 if (GST_BUFFER_OFFSET_IS_VALID (buf) &&
1263 GST_BUFFER_OFFSET (buf) != self->offset) {
1264 self->offset = GST_BUFFER_OFFSET (buf);
1269 GST_INFO ("discontinuity");
1270 /* flush the parser state */
1271 parser_state_init (&self->state);
1272 g_string_truncate (self->textbuf, 0);
1273 gst_adapter_clear (self->adapter);
1274 #ifndef GST_DISABLE_XML
1275 sami_context_reset (&self->state);
1277 /* we could set a flag to make sure that the next buffer we push out also
1278 * has the DISCONT flag set, but there's no point really given that it's
1279 * subtitles which are discontinuous by nature. */
1282 self->offset = GST_BUFFER_OFFSET (buf) + GST_BUFFER_SIZE (buf);
1283 self->next_offset = self->offset;
1285 gst_adapter_push (self->adapter, buf);
1288 convert_encoding (self, (const gchar *) gst_adapter_peek (self->adapter,
1289 gst_adapter_available (self->adapter)),
1290 (gsize) gst_adapter_available (self->adapter), &consumed);
1292 if (input && consumed > 0) {
1293 self->textbuf = g_string_append (self->textbuf, input);
1294 gst_adapter_flush (self->adapter, consumed);
1300 static GstFlowReturn
1301 handle_buffer (GstSubParse * self, GstBuffer * buf)
1303 GstFlowReturn ret = GST_FLOW_OK;
1304 GstCaps *caps = NULL;
1305 gchar *line, *subtitle;
1307 if (self->first_buffer) {
1308 self->detected_encoding =
1309 detect_encoding ((gchar *) GST_BUFFER_DATA (buf),
1310 GST_BUFFER_SIZE (buf));
1311 self->first_buffer = FALSE;
1314 feed_textbuf (self, buf);
1316 /* make sure we know the format */
1317 if (G_UNLIKELY (self->parser_type == GST_SUB_PARSE_FORMAT_UNKNOWN)) {
1318 if (!(caps = gst_sub_parse_format_autodetect (self))) {
1319 return GST_FLOW_UNEXPECTED;
1321 if (!gst_pad_set_caps (self->srcpad, caps)) {
1322 gst_caps_unref (caps);
1323 return GST_FLOW_UNEXPECTED;
1325 gst_caps_unref (caps);
1328 if (self->subtitle_codec != NULL) {
1331 tags = gst_tag_list_new ();
1332 gst_tag_list_add (tags, GST_TAG_MERGE_APPEND, GST_TAG_SUBTITLE_CODEC,
1333 self->subtitle_codec, NULL);
1334 gst_element_found_tags_for_pad (GST_ELEMENT (self), self->srcpad, tags);
1338 while (!self->flushing && (line = get_next_line (self))) {
1341 /* Set segment on our parser state machine */
1342 self->state.segment = &self->segment;
1343 /* Now parse the line, out of segment lines will just return NULL */
1344 GST_LOG_OBJECT (self, "Parsing line '%s'", line + offset);
1345 subtitle = self->parse_line (&self->state, line + offset);
1349 guint subtitle_len = strlen (subtitle);
1351 /* +1 for terminating NUL character */
1352 ret = gst_pad_alloc_buffer_and_set_caps (self->srcpad,
1353 GST_BUFFER_OFFSET_NONE, subtitle_len + 1,
1354 GST_PAD_CAPS (self->srcpad), &buf);
1356 if (ret == GST_FLOW_OK) {
1357 /* copy terminating NUL character as well */
1358 memcpy (GST_BUFFER_DATA (buf), subtitle, subtitle_len + 1);
1359 GST_BUFFER_SIZE (buf) = subtitle_len;
1360 GST_BUFFER_TIMESTAMP (buf) = self->state.start_time;
1361 GST_BUFFER_DURATION (buf) = self->state.duration;
1363 /* in some cases (e.g. tmplayer) we can only determine the duration
1364 * of a text chunk from the timestamp of the next text chunk; in those
1365 * cases, we probably want to limit the duration to something
1366 * reasonable, so we don't end up showing some text for e.g. 40 seconds
1367 * just because nothing else is being said during that time */
1368 if (self->state.max_duration > 0 && GST_BUFFER_DURATION_IS_VALID (buf)) {
1369 if (GST_BUFFER_DURATION (buf) > self->state.max_duration)
1370 GST_BUFFER_DURATION (buf) = self->state.max_duration;
1373 gst_segment_set_last_stop (&self->segment, GST_FORMAT_TIME,
1374 self->state.start_time);
1376 GST_DEBUG_OBJECT (self, "Sending text '%s', %" GST_TIME_FORMAT " + %"
1377 GST_TIME_FORMAT, subtitle, GST_TIME_ARGS (self->state.start_time),
1378 GST_TIME_ARGS (self->state.duration));
1380 ret = gst_pad_push (self->srcpad, buf);
1383 /* move this forward (the tmplayer parser needs this) */
1384 if (self->state.duration != GST_CLOCK_TIME_NONE)
1385 self->state.start_time += self->state.duration;
1390 if (ret != GST_FLOW_OK) {
1391 GST_DEBUG_OBJECT (self, "flow: %s", gst_flow_get_name (ret));
1400 static GstFlowReturn
1401 gst_sub_parse_chain (GstPad * sinkpad, GstBuffer * buf)
1406 self = GST_SUBPARSE (GST_PAD_PARENT (sinkpad));
1408 /* Push newsegment if needed */
1409 if (self->need_segment) {
1410 GST_LOG_OBJECT (self, "pushing newsegment event with %" GST_SEGMENT_FORMAT,
1413 gst_pad_push_event (self->srcpad, gst_event_new_new_segment (FALSE,
1414 self->segment.rate, self->segment.format,
1415 self->segment.last_stop, self->segment.stop, self->segment.time));
1416 self->need_segment = FALSE;
1419 ret = handle_buffer (self, buf);
1425 gst_sub_parse_sink_event (GstPad * pad, GstEvent * event)
1427 GstSubParse *self = GST_SUBPARSE (gst_pad_get_parent (pad));
1428 gboolean ret = FALSE;
1430 GST_DEBUG ("Handling %s event", GST_EVENT_TYPE_NAME (event));
1432 switch (GST_EVENT_TYPE (event)) {
1433 case GST_EVENT_EOS:{
1434 /* Make sure the last subrip chunk is pushed out even
1435 * if the file does not have an empty line at the end */
1436 if (self->parser_type == GST_SUB_PARSE_FORMAT_SUBRIP ||
1437 self->parser_type == GST_SUB_PARSE_FORMAT_TMPLAYER ||
1438 self->parser_type == GST_SUB_PARSE_FORMAT_MPL2) {
1439 GstBuffer *buf = gst_buffer_new_and_alloc (2 + 1);
1441 GST_DEBUG ("EOS. Pushing remaining text (if any)");
1442 GST_BUFFER_DATA (buf)[0] = '\n';
1443 GST_BUFFER_DATA (buf)[1] = '\n';
1444 GST_BUFFER_DATA (buf)[2] = '\0'; /* play it safe */
1445 GST_BUFFER_SIZE (buf) = 2;
1446 GST_BUFFER_OFFSET (buf) = self->offset;
1447 gst_sub_parse_chain (pad, buf);
1449 ret = gst_pad_event_default (pad, event);
1452 case GST_EVENT_NEWSEGMENT:
1456 gint64 start, stop, time;
1459 gst_event_parse_new_segment (event, &update, &rate, &format, &start,
1462 GST_DEBUG_OBJECT (self, "newsegment (%s)", gst_format_get_name (format));
1464 if (format == GST_FORMAT_TIME) {
1465 gst_segment_set_newsegment (&self->segment, update, rate, format,
1468 /* if not time format, we'll either start with a 0 timestamp anyway or
1469 * it's following a seek in which case we'll have saved the requested
1470 * seek segment and don't want to overwrite it (remember that on a seek
1471 * we always just seek back to the start in BYTES format and just throw
1472 * away all text that's before the requested position; if the subtitles
1473 * come from an upstream demuxer, it won't be able to handle our BYTES
1474 * seek request and instead send us a newsegment from the seek request
1475 * it received via its video pads instead, so all is fine then too) */
1479 gst_event_unref (event);
1482 case GST_EVENT_FLUSH_START:
1484 self->flushing = TRUE;
1486 ret = gst_pad_event_default (pad, event);
1489 case GST_EVENT_FLUSH_STOP:
1491 self->flushing = FALSE;
1493 ret = gst_pad_event_default (pad, event);
1497 ret = gst_pad_event_default (pad, event);
1501 gst_object_unref (self);
1507 static GstStateChangeReturn
1508 gst_sub_parse_change_state (GstElement * element, GstStateChange transition)
1510 GstStateChangeReturn ret = GST_STATE_CHANGE_SUCCESS;
1511 GstSubParse *self = GST_SUBPARSE (element);
1513 switch (transition) {
1514 case GST_STATE_CHANGE_READY_TO_PAUSED:
1515 /* format detection will init the parser state */
1517 self->next_offset = 0;
1518 self->parser_type = GST_SUB_PARSE_FORMAT_UNKNOWN;
1519 self->valid_utf8 = TRUE;
1520 self->first_buffer = TRUE;
1521 g_free (self->detected_encoding);
1522 self->detected_encoding = NULL;
1523 g_string_truncate (self->textbuf, 0);
1524 gst_adapter_clear (self->adapter);
1530 ret = parent_class->change_state (element, transition);
1531 if (ret == GST_STATE_CHANGE_FAILURE)
1534 switch (transition) {
1535 case GST_STATE_CHANGE_PAUSED_TO_READY:
1536 parser_state_dispose (&self->state);
1537 self->parser_type = GST_SUB_PARSE_FORMAT_UNKNOWN;
1550 /* FIXME 0.11: these caps are ugly, use app/x-subtitle + type field or so;
1551 * also, give different subtitle formats really different types */
1552 static GstStaticCaps mpl2_caps =
1553 GST_STATIC_CAPS ("application/x-subtitle-mpl2");
1554 #define SUB_CAPS (gst_static_caps_get (&sub_caps))
1556 static GstStaticCaps tmp_caps =
1557 GST_STATIC_CAPS ("application/x-subtitle-tmplayer");
1558 #define TMP_CAPS (gst_static_caps_get (&tmp_caps))
1560 static GstStaticCaps sub_caps = GST_STATIC_CAPS ("application/x-subtitle");
1561 #define MPL2_CAPS (gst_static_caps_get (&mpl2_caps))
1563 #ifndef GST_DISABLE_XML
1564 static GstStaticCaps smi_caps = GST_STATIC_CAPS ("application/x-subtitle-sami");
1565 #define SAMI_CAPS (gst_static_caps_get (&smi_caps))
1569 gst_subparse_type_find (GstTypeFind * tf, gpointer private)
1571 GstSubParseFormat format;
1575 gchar *encoding = NULL;
1578 if (!(data = gst_type_find_peek (tf, 0, 129)))
1581 /* make sure string passed to _autodetect() is NUL-terminated */
1582 str = g_malloc0 (129);
1583 memcpy (str, data, 128);
1585 if ((encoding = detect_encoding (str, 128)) != NULL) {
1586 gchar *converted_str;
1590 converted_str = gst_convert_to_utf8 (str, 128, encoding, &tmp, &err);
1591 if (converted_str == NULL) {
1592 GST_DEBUG ("Encoding '%s' detected but conversion failed: %s", encoding,
1598 str = converted_str;
1603 /* Check if at least the first 120 chars are valid UTF8,
1604 * otherwise convert as always */
1605 if (!g_utf8_validate (str, 128, &end) && (end - str) < 120) {
1606 gchar *converted_str;
1611 enc = g_getenv ("GST_SUBTITLE_ENCODING");
1612 if (enc == NULL || *enc == '\0') {
1613 /* if local encoding is UTF-8 and no encoding specified
1614 * via the environment variable, assume ISO-8859-15 */
1615 if (g_get_charset (&enc)) {
1616 enc = "ISO-8859-15";
1619 converted_str = gst_convert_to_utf8 (str, 128, enc, &tmp, &err);
1620 if (converted_str == NULL) {
1621 GST_DEBUG ("Charset conversion failed: %s", err->message);
1627 str = converted_str;
1631 format = gst_sub_parse_data_format_autodetect (str);
1635 case GST_SUB_PARSE_FORMAT_MDVDSUB:
1636 GST_DEBUG ("MicroDVD format detected");
1639 case GST_SUB_PARSE_FORMAT_SUBRIP:
1640 GST_DEBUG ("SubRip format detected");
1643 case GST_SUB_PARSE_FORMAT_MPSUB:
1644 GST_DEBUG ("MPSub format detected");
1647 #ifndef GST_DISABLE_XML
1648 case GST_SUB_PARSE_FORMAT_SAMI:
1649 GST_DEBUG ("SAMI (time-based) format detected");
1653 case GST_SUB_PARSE_FORMAT_TMPLAYER:
1654 GST_DEBUG ("TMPlayer (time based) format detected");
1657 /* FIXME: our MPL2 typefinding is not really good enough to warrant
1658 * returning a high probability (however, since we registered our
1659 * typefinder here with a rank of MARGINAL we should pretty much only
1660 * be called if most other typefinders have already run */
1661 case GST_SUB_PARSE_FORMAT_MPL2:
1662 GST_DEBUG ("MPL2 (time based) format detected");
1665 case GST_SUB_PARSE_FORMAT_SUBVIEWER:
1666 GST_DEBUG ("SubViewer format detected");
1670 case GST_SUB_PARSE_FORMAT_UNKNOWN:
1671 GST_DEBUG ("no subtitle format detected");
1675 /* if we're here, it's ok */
1676 gst_type_find_suggest (tf, GST_TYPE_FIND_MAXIMUM, caps);
1680 plugin_init (GstPlugin * plugin)
1682 static gchar *sub_exts[] = { "srt", "sub", "mpsub", "mdvd", "smi", "txt",
1686 GST_DEBUG_CATEGORY_INIT (sub_parse_debug, "subparse", 0, ".sub parser");
1688 if (!gst_type_find_register (plugin, "subparse_typefind", GST_RANK_MARGINAL,
1689 gst_subparse_type_find, sub_exts, SUB_CAPS, NULL, NULL))
1692 if (!gst_element_register (plugin, "subparse",
1693 GST_RANK_PRIMARY, GST_TYPE_SUBPARSE) ||
1694 !gst_element_register (plugin, "ssaparse",
1695 GST_RANK_PRIMARY, GST_TYPE_SSA_PARSE)) {
1702 GST_PLUGIN_DEFINE (GST_VERSION_MAJOR,
1706 plugin_init, VERSION, "LGPL", GST_PACKAGE_NAME, GST_PACKAGE_ORIGIN)