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>
31 #include "gstsubparse.h"
32 #include "gstssaparse.h"
33 #include "samiparse.h"
34 #include "tmplayerparse.h"
35 #include "mpl2parse.h"
37 GST_DEBUG_CATEGORY (sub_parse_debug);
39 #define DEFAULT_ENCODING NULL
48 gst_sub_parse_set_property (GObject * object, guint prop_id,
49 const GValue * value, GParamSpec * pspec);
51 gst_sub_parse_get_property (GObject * object, guint prop_id,
52 GValue * value, GParamSpec * pspec);
55 static const GstElementDetails sub_parse_details =
56 GST_ELEMENT_DETAILS ("Subtitle parser",
57 "Codec/Parser/Subtitle",
58 "Parses subtitle (.sub) files into text streams",
59 "Gustavo J. A. M. Carneiro <gjc@inescporto.pt>\n"
60 "GStreamer maintainers <gstreamer-devel@lists.sourceforge.net>");
62 #ifndef GST_DISABLE_XML
63 static GstStaticPadTemplate sink_templ = GST_STATIC_PAD_TEMPLATE ("sink",
66 GST_STATIC_CAPS ("application/x-subtitle; application/x-subtitle-sami; "
67 "application/x-subtitle-tmplayer; application/x-subtitle-mpl2")
70 static GstStaticPadTemplate sink_templ = GST_STATIC_PAD_TEMPLATE ("sink",
73 GST_STATIC_CAPS ("application/x-subtitle")
77 static GstStaticPadTemplate src_templ = GST_STATIC_PAD_TEMPLATE ("src",
80 GST_STATIC_CAPS ("text/plain; text/x-pango-markup")
83 static void gst_sub_parse_base_init (GstSubParseClass * klass);
84 static void gst_sub_parse_class_init (GstSubParseClass * klass);
85 static void gst_sub_parse_init (GstSubParse * subparse);
87 static gboolean gst_sub_parse_src_event (GstPad * pad, GstEvent * event);
88 static gboolean gst_sub_parse_src_query (GstPad * pad, GstQuery * query);
89 static gboolean gst_sub_parse_sink_event (GstPad * pad, GstEvent * event);
91 static GstStateChangeReturn gst_sub_parse_change_state (GstElement * element,
92 GstStateChange transition);
94 static GstFlowReturn gst_sub_parse_chain (GstPad * sinkpad, GstBuffer * buf);
96 static GstElementClass *parent_class = NULL;
99 gst_sub_parse_get_type (void)
101 static GType sub_parse_type = 0;
103 if (!sub_parse_type) {
104 static const GTypeInfo sub_parse_info = {
105 sizeof (GstSubParseClass),
106 (GBaseInitFunc) gst_sub_parse_base_init,
108 (GClassInitFunc) gst_sub_parse_class_init,
111 sizeof (GstSubParse),
113 (GInstanceInitFunc) gst_sub_parse_init,
116 sub_parse_type = g_type_register_static (GST_TYPE_ELEMENT,
117 "GstSubParse", &sub_parse_info, 0);
120 return sub_parse_type;
124 gst_sub_parse_base_init (GstSubParseClass * klass)
126 GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
128 gst_element_class_add_pad_template (element_class,
129 gst_static_pad_template_get (&sink_templ));
130 gst_element_class_add_pad_template (element_class,
131 gst_static_pad_template_get (&src_templ));
132 gst_element_class_set_details (element_class, &sub_parse_details);
136 gst_sub_parse_dispose (GObject * object)
138 GstSubParse *subparse = GST_SUBPARSE (object);
140 GST_DEBUG_OBJECT (subparse, "cleaning up subtitle parser");
142 if (subparse->encoding) {
143 g_free (subparse->encoding);
144 subparse->encoding = NULL;
147 if (subparse->detected_encoding) {
148 g_free (subparse->detected_encoding);
149 subparse->detected_encoding = NULL;
152 if (subparse->adapter) {
153 gst_object_unref (subparse->adapter);
154 subparse->adapter = NULL;
157 if (subparse->textbuf) {
158 g_string_free (subparse->textbuf, TRUE);
159 subparse->textbuf = NULL;
161 #ifndef GST_DISABLE_XML
162 sami_context_deinit (&subparse->state);
165 GST_CALL_PARENT (G_OBJECT_CLASS, dispose, (object));
169 gst_sub_parse_class_init (GstSubParseClass * klass)
171 GObjectClass *object_class = G_OBJECT_CLASS (klass);
172 GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
174 parent_class = g_type_class_peek_parent (klass);
176 object_class->dispose = gst_sub_parse_dispose;
177 object_class->set_property = gst_sub_parse_set_property;
178 object_class->get_property = gst_sub_parse_get_property;
180 element_class->change_state = gst_sub_parse_change_state;
182 g_object_class_install_property (object_class, PROP_ENCODING,
183 g_param_spec_string ("subtitle-encoding", "subtitle charset encoding",
184 "Encoding to assume if input subtitles are not in UTF-8 or any other "
185 "Unicode encoding. If not set, the GST_SUBTITLE_ENCODING environment "
186 "variable will be checked for an encoding to use. If that is not set "
187 "either, ISO-8859-15 will be assumed.", DEFAULT_ENCODING,
188 G_PARAM_READWRITE | G_PARAM_STATIC_STRINGS));
192 gst_sub_parse_init (GstSubParse * subparse)
194 subparse->sinkpad = gst_pad_new_from_static_template (&sink_templ, "sink");
195 gst_pad_set_chain_function (subparse->sinkpad,
196 GST_DEBUG_FUNCPTR (gst_sub_parse_chain));
197 gst_pad_set_event_function (subparse->sinkpad,
198 GST_DEBUG_FUNCPTR (gst_sub_parse_sink_event));
199 gst_element_add_pad (GST_ELEMENT (subparse), subparse->sinkpad);
201 subparse->srcpad = gst_pad_new_from_static_template (&src_templ, "src");
202 gst_pad_set_event_function (subparse->srcpad,
203 GST_DEBUG_FUNCPTR (gst_sub_parse_src_event));
204 gst_pad_set_query_function (subparse->srcpad,
205 GST_DEBUG_FUNCPTR (gst_sub_parse_src_query));
206 gst_element_add_pad (GST_ELEMENT (subparse), subparse->srcpad);
208 subparse->textbuf = g_string_new (NULL);
209 subparse->parser_type = GST_SUB_PARSE_FORMAT_UNKNOWN;
210 subparse->flushing = FALSE;
211 gst_segment_init (&subparse->segment, GST_FORMAT_TIME);
212 subparse->need_segment = TRUE;
213 subparse->encoding = g_strdup (DEFAULT_ENCODING);
214 subparse->detected_encoding = NULL;
215 subparse->adapter = gst_adapter_new ();
219 * Source pad functions.
223 gst_sub_parse_src_query (GstPad * pad, GstQuery * query)
225 GstSubParse *self = GST_SUBPARSE (gst_pad_get_parent (pad));
226 gboolean ret = FALSE;
228 GST_DEBUG ("Handling %s query", GST_QUERY_TYPE_NAME (query));
230 switch (GST_QUERY_TYPE (query)) {
231 case GST_QUERY_POSITION:{
234 gst_query_parse_position (query, &fmt, NULL);
235 if (fmt != GST_FORMAT_TIME) {
236 ret = gst_pad_peer_query (self->sinkpad, query);
239 gst_query_set_position (query, GST_FORMAT_TIME,
240 self->segment.last_stop);
243 case GST_QUERY_SEEKING:
246 gboolean seekable = FALSE;
250 gst_query_parse_seeking (query, &fmt, NULL, NULL, NULL);
251 if (fmt == GST_FORMAT_TIME) {
252 GstQuery *peerquery = gst_query_new_seeking (GST_FORMAT_BYTES);
254 seekable = gst_pad_peer_query (self->sinkpad, peerquery);
256 gst_query_parse_seeking (peerquery, NULL, &seekable, NULL, NULL);
257 gst_query_unref (peerquery);
260 gst_query_set_seeking (query, fmt, seekable, seekable ? 0 : -1, -1);
265 ret = gst_pad_peer_query (self->sinkpad, query);
269 gst_object_unref (self);
275 gst_sub_parse_src_event (GstPad * pad, GstEvent * event)
277 GstSubParse *self = GST_SUBPARSE (gst_pad_get_parent (pad));
278 gboolean ret = FALSE;
280 GST_DEBUG ("Handling %s event", GST_EVENT_TYPE_NAME (event));
282 switch (GST_EVENT_TYPE (event)) {
286 GstSeekType start_type, stop_type;
291 gst_event_parse_seek (event, &rate, &format, &self->segment_flags,
292 &start_type, &start, &stop_type, &stop);
294 if (format != GST_FORMAT_TIME) {
295 GST_WARNING_OBJECT (self, "we only support seeking in TIME format");
296 gst_event_unref (event);
300 /* Convert that seek to a seeking in bytes at position 0,
301 FIXME: could use an index */
302 ret = gst_pad_push_event (self->sinkpad,
303 gst_event_new_seek (rate, GST_FORMAT_BYTES, self->segment_flags,
304 GST_SEEK_TYPE_SET, 0, GST_SEEK_TYPE_NONE, 0));
307 /* Apply the seek to our segment */
308 gst_segment_set_seek (&self->segment, rate, format, self->segment_flags,
309 start_type, start, stop_type, stop, &update);
311 GST_DEBUG_OBJECT (self, "segment after seek: %" GST_SEGMENT_FORMAT,
314 self->next_offset = 0;
316 self->need_segment = TRUE;
318 GST_WARNING_OBJECT (self, "seek to 0 bytes failed");
321 gst_event_unref (event);
325 ret = gst_pad_event_default (pad, event);
330 gst_object_unref (self);
336 gst_sub_parse_set_property (GObject * object, guint prop_id,
337 const GValue * value, GParamSpec * pspec)
339 GstSubParse *subparse = GST_SUBPARSE (object);
341 GST_OBJECT_LOCK (subparse);
344 g_free (subparse->encoding);
345 subparse->encoding = g_value_dup_string (value);
346 GST_LOG_OBJECT (object, "subtitle encoding set to %s",
347 GST_STR_NULL (subparse->encoding));
350 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
353 GST_OBJECT_UNLOCK (subparse);
357 gst_sub_parse_get_property (GObject * object, guint prop_id,
358 GValue * value, GParamSpec * pspec)
360 GstSubParse *subparse = GST_SUBPARSE (object);
362 GST_OBJECT_LOCK (subparse);
365 g_value_set_string (value, subparse->encoding);
368 G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
371 GST_OBJECT_UNLOCK (subparse);
375 gst_sub_parse_get_format_description (GstSubParseFormat format)
378 case GST_SUB_PARSE_FORMAT_MDVDSUB:
380 case GST_SUB_PARSE_FORMAT_SUBRIP:
382 case GST_SUB_PARSE_FORMAT_MPSUB:
384 case GST_SUB_PARSE_FORMAT_SAMI:
386 case GST_SUB_PARSE_FORMAT_TMPLAYER:
388 case GST_SUB_PARSE_FORMAT_MPL2:
390 case GST_SUB_PARSE_FORMAT_SUBVIEWER:
393 case GST_SUB_PARSE_FORMAT_UNKNOWN:
400 gst_convert_to_utf8 (const gchar * str, gsize len, const gchar * encoding,
401 gsize * consumed, GError ** err)
407 g_convert_with_fallback (str, len, "UTF-8", encoding, "*", consumed, NULL,
412 /* + 3 to skip UTF-8 BOM if it was added */
414 if (len >= 3 && (guint8) ret[0] == 0xEF && (guint8) ret[1] == 0xBB
415 && (guint8) ret[2] == 0xBF)
416 g_memmove (ret, ret + 3, len + 1 - 3);
422 detect_encoding (const gchar * str, gsize len)
424 if (len >= 3 && (guint8) str[0] == 0xEF && (guint8) str[1] == 0xBB
425 && (guint8) str[2] == 0xBF)
426 return g_strdup ("UTF-8");
428 if (len >= 2 && (guint8) str[0] == 0xFE && (guint8) str[1] == 0xFF)
429 return g_strdup ("UTF-16BE");
431 if (len >= 2 && (guint8) str[0] == 0xFF && (guint8) str[1] == 0xFE)
432 return g_strdup ("UTF-16LE");
434 if (len >= 4 && (guint8) str[0] == 0x00 && (guint8) str[1] == 0x00
435 && (guint8) str[2] == 0xFE && (guint8) str[3] == 0xFF)
436 return g_strdup ("UTF-32BE");
438 if (len >= 4 && (guint8) str[0] == 0xFF && (guint8) str[1] == 0xFE
439 && (guint8) str[2] == 0x00 && (guint8) str[3] == 0x00)
440 return g_strdup ("UTF-32LE");
446 convert_encoding (GstSubParse * self, const gchar * str, gsize len,
449 const gchar *encoding;
455 /* First try any detected encoding */
456 if (self->detected_encoding) {
458 gst_convert_to_utf8 (str, len, self->detected_encoding, consumed, &err);
463 GST_WARNING_OBJECT (self, "could not convert string from '%s' to UTF-8: %s",
464 self->detected_encoding, err->message);
465 g_free (self->detected_encoding);
466 self->detected_encoding = NULL;
470 /* Otherwise check if it's UTF8 */
471 if (self->valid_utf8) {
472 if (g_utf8_validate (str, len, NULL)) {
473 GST_LOG_OBJECT (self, "valid UTF-8, no conversion needed");
475 return g_strndup (str, len);
477 GST_INFO_OBJECT (self, "invalid UTF-8!");
478 self->valid_utf8 = FALSE;
481 /* Else try fallback */
482 encoding = self->encoding;
483 if (encoding == NULL || *encoding == '\0') {
484 encoding = g_getenv ("GST_SUBTITLE_ENCODING");
486 if (encoding == NULL || *encoding == '\0') {
487 /* if local encoding is UTF-8 and no encoding specified
488 * via the environment variable, assume ISO-8859-15 */
489 if (g_get_charset (&encoding)) {
490 encoding = "ISO-8859-15";
494 ret = gst_convert_to_utf8 (str, len, encoding, consumed, &err);
497 GST_WARNING_OBJECT (self, "could not convert string from '%s' to UTF-8: %s",
498 encoding, err->message);
501 /* invalid input encoding, fall back to ISO-8859-15 (always succeeds) */
502 ret = gst_convert_to_utf8 (str, len, "ISO-8859-15", consumed, NULL);
505 GST_LOG_OBJECT (self,
506 "successfully converted %" G_GSIZE_FORMAT " characters from %s to UTF-8"
507 "%s", len, encoding, (err) ? " , using ISO-8859-15 as fallback" : "");
513 get_next_line (GstSubParse * self)
516 const char *line_end;
518 gboolean have_r = FALSE;
520 line_end = strchr (self->textbuf->str, '\n');
523 /* end-of-line not found; return for more data */
527 /* get rid of '\r' */
528 if (line_end != self->textbuf->str && *(line_end - 1) == '\r') {
533 line_len = line_end - self->textbuf->str;
534 line = g_strndup (self->textbuf->str, line_len);
535 self->textbuf = g_string_erase (self->textbuf, 0,
536 line_len + (have_r ? 2 : 1));
541 parse_mdvdsub (ParserState * state, const gchar * line)
543 const gchar *line_split;
545 guint start_frame, end_frame;
546 gint64 clip_start = 0, clip_stop = 0;
547 gboolean in_seg = FALSE;
551 /* style variables */
556 if (sscanf (line, "{%u}{%u}", &start_frame, &end_frame) != 2) {
557 g_warning ("Parse of the following line, assumed to be in microdvd .sub"
558 " format, failed:\n%s", line);
562 /* skip the {%u}{%u} part */
563 line = strchr (line, '}') + 1;
564 line = strchr (line, '}') + 1;
566 /* see if there's a first line with a framerate */
567 if (state->fps == 0.0 && start_frame == 1 && end_frame == 1) {
568 gchar *rest, *end = NULL;
570 rest = g_strdup (line);
571 g_strdelimit (rest, ",", '.');
572 state->fps = g_ascii_strtod (rest, &end);
575 GST_INFO ("framerate from file: %f ('%s')", state->fps, rest);
580 if (state->fps == 0.0) {
581 /* FIXME: hardcoded for now, is there a better way/assumption? */
582 state->fps = 24000.0 / 1001.0;
583 GST_INFO ("no framerate specified, assuming %f", state->fps);
586 state->start_time = start_frame / state->fps * GST_SECOND;
587 state->duration = (end_frame - start_frame) / state->fps * GST_SECOND;
589 /* Check our segment start/stop */
590 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
591 state->start_time, state->start_time + state->duration, &clip_start,
594 /* No need to parse that text if it's out of segment */
596 state->start_time = clip_start;
597 state->duration = clip_stop - clip_start;
602 markup = g_string_new (NULL);
607 /* parse style markup */
608 if (strncmp (line, "{y:i}", 5) == 0) {
610 line = strchr (line, '}') + 1;
612 if (strncmp (line, "{y:b}", 5) == 0) {
614 line = strchr (line, '}') + 1;
616 if (sscanf (line, "{s:%u}", &fontsize) == 1) {
617 line = strchr (line, '}') + 1;
619 /* forward slashes at beginning/end signify italics too */
620 if (g_str_has_prefix (line, "/")) {
624 if ((line_split = strchr (line, '|')))
625 line_chunk = g_markup_escape_text (line, line_split - line);
627 line_chunk = g_markup_escape_text (line, strlen (line));
629 /* Remove italics markers at end of line/stanza (CHECKME: are end slashes
630 * always at the end of a line or can they span multiple lines?) */
631 if (g_str_has_suffix (line_chunk, "/")) {
632 line_chunk[strlen (line_chunk) - 1] = '\0';
635 markup = g_string_append (markup, "<span");
637 g_string_append (markup, " style=\"italic\"");
639 g_string_append (markup, " weight=\"bold\"");
641 g_string_append_printf (markup, " size=\"%u\"", fontsize * 1000);
642 g_string_append_printf (markup, ">%s</span>", line_chunk);
645 g_string_append (markup, "\n");
646 line = line_split + 1;
652 g_string_free (markup, FALSE);
653 GST_DEBUG ("parse_mdvdsub returning (%f+%f): %s",
654 state->start_time / (double) GST_SECOND,
655 state->duration / (double) GST_SECOND, ret);
660 strip_trailing_newlines (gchar * txt)
666 while (len > 1 && txt[len - 1] == '\n') {
673 /* we want to escape text in general, but retain basic markup like
674 * <i></i>, <u></u>, and <b></b>. The easiest and safest way is to
675 * just unescape a white list of allowed markups again after
676 * escaping everything (the text between these simple markers isn't
677 * necessarily escaped, so it seems best to do it like this) */
679 subrip_unescape_formatting (gchar * txt)
683 for (pos = txt; pos != NULL && *pos != '\0'; ++pos) {
684 if (g_ascii_strncasecmp (pos, "<u>", 9) == 0 ||
685 g_ascii_strncasecmp (pos, "<i>", 9) == 0 ||
686 g_ascii_strncasecmp (pos, "<b>", 9) == 0) {
688 pos[1] = g_ascii_tolower (pos[4]);
690 /* move NUL terminator as well */
691 g_memmove (pos + 3, pos + 9, strlen (pos + 9) + 1);
696 for (pos = txt; pos != NULL && *pos != '\0'; ++pos) {
697 if (g_ascii_strncasecmp (pos, "</u>", 10) == 0 ||
698 g_ascii_strncasecmp (pos, "</i>", 10) == 0 ||
699 g_ascii_strncasecmp (pos, "</b>", 10) == 0) {
702 pos[2] = g_ascii_tolower (pos[5]);
704 /* move NUL terminator as well */
705 g_memmove (pos + 4, pos + 10, strlen (pos + 10) + 1);
713 subrip_remove_unhandled_tag (gchar * start, gchar * stop)
717 tag = start + strlen ("<");
721 if (g_ascii_tolower (*tag) < 'a' || g_ascii_tolower (*tag) > 'z')
726 GST_LOG ("removing unhandled tag '%s'", start);
728 g_memmove (start, stop, strlen (stop) + 1);
732 /* remove tags we haven't explicitly allowed earlier on, like font tags
735 subrip_remove_unhandled_tags (gchar * txt)
739 for (pos = txt; pos != NULL && *pos != '\0'; ++pos) {
740 if (strncmp (pos, "<", 4) == 0 && (gt = strstr (pos + 4, ">"))) {
741 if (subrip_remove_unhandled_tag (pos, gt + strlen (">")))
747 /* we only allow <i>, <u> and <b>, so let's take a simple approach. This code
748 * assumes the input has been escaped and subrip_unescape_formatting() has then
749 * been run over the input! This function adds missing closing markup tags and
750 * removes broken closing tags for tags that have never been opened. */
752 subrip_fix_up_markup (gchar ** p_txt)
754 gchar *cur, *next_tag;
756 guint num_open_tags = 0;
758 g_assert (*p_txt != NULL);
761 while (*cur != '\0') {
762 next_tag = strchr (cur, '<');
763 if (next_tag == NULL)
769 if (num_open_tags == 0 || open_tags[num_open_tags - 1] != *next_tag) {
770 GST_LOG ("broken input, closing tag '%c' is not open", *next_tag);
771 g_memmove (next_tag - 2, next_tag + 2, strlen (next_tag + 2) + 1);
774 /* it's all good, closing tag which is open */
782 if (num_open_tags == G_N_ELEMENTS (open_tags))
783 return; /* something dodgy is going on, stop parsing */
784 open_tags[num_open_tags] = *next_tag;
788 GST_ERROR ("unexpected tag '%c' (%s)", *next_tag, next_tag);
789 g_assert_not_reached ();
795 if (num_open_tags > 0) {
798 s = g_string_new (*p_txt);
799 while (num_open_tags > 0) {
800 GST_LOG ("adding missing closing tag '%c'", open_tags[num_open_tags - 1]);
801 g_string_append_c (s, '<');
802 g_string_append_c (s, '/');
803 g_string_append_c (s, open_tags[num_open_tags - 1]);
804 g_string_append_c (s, '>');
808 *p_txt = g_string_free (s, FALSE);
813 parse_subrip_time (const gchar * ts_string, GstClockTime * t)
815 gchar s[128] = { '\0', };
817 guint hour, min, sec, msec, len;
819 while (*ts_string == ' ')
822 g_strlcpy (s, ts_string, sizeof (s));
823 if ((end = strstr (s, "-->")))
827 /* ms may be in these formats:
828 * hh:mm:ss,500 = 500ms
831 * hh:mm:ss, 50 = 50ms
833 * and the same with . instead of ,.
834 * sscanf() doesn't differentiate between ' 5' and '5' so munge
835 * the white spaces within the timestamp to '0' (I'm sure there's a
836 * way to make sscanf() do this for us, but how?)
838 g_strdelimit (s, " ", '0');
839 g_strdelimit (s, ".", ',');
841 /* make sure we have exactly three digits after he comma */
843 g_assert (p != NULL);
850 g_strlcat (&p[len], "0", 2);
854 GST_LOG ("parsing timestamp '%s'", s);
855 if (sscanf (s, "%u:%u:%u,%u", &hour, &min, &sec, &msec) != 4) {
856 GST_WARNING ("failed to parse subrip timestamp string '%s'", s);
860 *t = ((hour * 3600) + (min * 60) + sec) * GST_SECOND + msec * GST_MSECOND;
865 parse_subrip (ParserState * state, const gchar * line)
870 switch (state->state) {
872 /* looking for a single integer */
873 if (sscanf (line, "%u", &subnum) == 1)
878 GstClockTime ts_start, ts_end;
881 /* looking for start_time --> end_time */
882 if ((end_time = strstr (line, " --> ")) &&
883 parse_subrip_time (line, &ts_start) &&
884 parse_subrip_time (end_time + strlen (" --> "), &ts_end) &&
885 state->start_time <= ts_end) {
887 state->start_time = ts_start;
888 state->duration = ts_end - ts_start;
890 GST_DEBUG ("error parsing subrip time line '%s'", line);
897 /* No need to parse that text if it's out of segment */
898 gint64 clip_start = 0, clip_stop = 0;
899 gboolean in_seg = FALSE;
901 /* Check our segment start/stop */
902 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
903 state->start_time, state->start_time + state->duration,
904 &clip_start, &clip_stop);
907 state->start_time = clip_start;
908 state->duration = clip_stop - clip_start;
914 /* looking for subtitle text; empty line ends this subtitle entry */
916 g_string_append_c (state->buf, '\n');
917 g_string_append (state->buf, line);
918 if (strlen (line) == 0) {
919 ret = g_markup_escape_text (state->buf->str, state->buf->len);
920 g_string_truncate (state->buf, 0);
922 subrip_unescape_formatting (ret);
923 subrip_remove_unhandled_tags (ret);
924 strip_trailing_newlines (ret);
925 subrip_fix_up_markup (&ret);
930 g_return_val_if_reached (NULL);
935 subviewer_unescape_newlines (gchar * read)
939 /* Replace all occurences of '[br]' with a newline as version 2
940 * of the subviewer format uses this for newlines */
942 if (read[0] == '\0' || read[1] == '\0' || read[2] == '\0' || read[3] == '\0')
946 if (strncmp (read, "[br]", 4) == 0) {
960 parse_subviewer (ParserState * state, const gchar * line)
962 guint h1, m1, s1, ms1;
963 guint h2, m2, s2, ms2;
966 /* TODO: Maybe also parse the fields in the header, especially DELAY.
967 * For examples see the unit test or
968 * http://www.doom9.org/index.html?/sub.htm */
970 switch (state->state) {
972 /* looking for start_time,end_time */
973 if (sscanf (line, "%u:%u:%u.%u,%u:%u:%u.%u",
974 &h1, &m1, &s1, &ms1, &h2, &m2, &s2, &ms2) == 8) {
977 (((guint64) h1) * 3600 + m1 * 60 + s1) * GST_SECOND +
980 (((guint64) h2) * 3600 + m2 * 60 + s2) * GST_SECOND +
981 ms2 * GST_MSECOND - state->start_time;
986 /* No need to parse that text if it's out of segment */
987 gint64 clip_start = 0, clip_stop = 0;
988 gboolean in_seg = FALSE;
990 /* Check our segment start/stop */
991 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
992 state->start_time, state->start_time + state->duration,
993 &clip_start, &clip_stop);
996 state->start_time = clip_start;
997 state->duration = clip_stop - clip_start;
1003 /* looking for subtitle text; empty line ends this subtitle entry */
1004 if (state->buf->len)
1005 g_string_append_c (state->buf, '\n');
1006 g_string_append (state->buf, line);
1007 if (strlen (line) == 0) {
1008 ret = g_strdup (state->buf->str);
1009 subviewer_unescape_newlines (ret);
1010 strip_trailing_newlines (ret);
1011 g_string_truncate (state->buf, 0);
1017 g_assert_not_reached ();
1023 parse_mpsub (ParserState * state, const gchar * line)
1028 switch (state->state) {
1030 /* looking for two floats (offset, duration) */
1031 if (sscanf (line, "%f %f", &t1, &t2) == 2) {
1033 state->start_time += state->duration + GST_SECOND * t1;
1034 state->duration = GST_SECOND * t2;
1038 { /* No need to parse that text if it's out of segment */
1039 gint64 clip_start = 0, clip_stop = 0;
1040 gboolean in_seg = FALSE;
1042 /* Check our segment start/stop */
1043 in_seg = gst_segment_clip (state->segment, GST_FORMAT_TIME,
1044 state->start_time, state->start_time + state->duration,
1045 &clip_start, &clip_stop);
1048 state->start_time = clip_start;
1049 state->duration = clip_stop - clip_start;
1055 /* looking for subtitle text; empty line ends this
1057 if (state->buf->len)
1058 g_string_append_c (state->buf, '\n');
1059 g_string_append (state->buf, line);
1060 if (strlen (line) == 0) {
1061 ret = g_strdup (state->buf->str);
1062 g_string_truncate (state->buf, 0);
1068 g_assert_not_reached ();
1074 parser_state_init (ParserState * state)
1076 GST_DEBUG ("initialising parser");
1079 g_string_truncate (state->buf, 0);
1081 state->buf = g_string_new (NULL);
1084 state->start_time = 0;
1085 state->duration = 0;
1086 state->max_duration = 0; /* no limit */
1088 state->segment = NULL;
1092 parser_state_dispose (ParserState * state)
1095 g_string_free (state->buf, TRUE);
1098 #ifndef GST_DISABLE_XML
1099 if (state->user_data) {
1100 sami_context_reset (state);
1105 /* regex type enum */
1108 GST_SUB_PARSE_REGEX_UNKNOWN = 0,
1109 GST_SUB_PARSE_REGEX_MDVDSUB = 1,
1110 GST_SUB_PARSE_REGEX_SUBRIP = 2,
1114 gst_sub_parse_data_format_autodetect_regex_once (GstSubParseRegex regtype)
1116 gpointer result = NULL;
1117 GError *gerr = NULL;
1119 case GST_SUB_PARSE_REGEX_MDVDSUB:
1121 (gpointer) g_regex_new ("^\\{[0-9]+\\}\\{[0-9]+\\}", 0, 0, &gerr);
1122 if (result == NULL) {
1123 g_warning ("Compilation of mdvd regex failed: %s", gerr->message);
1124 g_error_free (gerr);
1127 case GST_SUB_PARSE_REGEX_SUBRIP:
1128 result = (gpointer) g_regex_new ("^([ 0-9]){0,3}[0-9]\\s*(\x0d)?\x0a"
1129 "[ 0-9][0-9]:[ 0-9][0-9]:[ 0-9][0-9][,.][ 0-9]{0,2}[0-9]"
1130 " +--> +([ 0-9])?[0-9]:[ 0-9][0-9]:[ 0-9][0-9][,.][ 0-9]{0,2}[0-9]",
1132 if (result == NULL) {
1133 g_warning ("Compilation of subrip regex failed: %s", gerr->message);
1134 g_error_free (gerr);
1138 GST_WARNING ("Trying to allocate regex of unknown type %u", regtype);
1144 * FIXME: maybe we should pass along a second argument, the preceding
1145 * text buffer, because that is how this originally worked, even though
1146 * I don't really see the use of that.
1149 static GstSubParseFormat
1150 gst_sub_parse_data_format_autodetect (gchar * match_str)
1154 static GOnce mdvd_rx_once = G_ONCE_INIT;
1155 static GOnce subrip_rx_once = G_ONCE_INIT;
1160 g_once (&mdvd_rx_once,
1161 (GThreadFunc) gst_sub_parse_data_format_autodetect_regex_once,
1162 (gpointer) GST_SUB_PARSE_REGEX_MDVDSUB);
1163 g_once (&subrip_rx_once,
1164 (GThreadFunc) gst_sub_parse_data_format_autodetect_regex_once,
1165 (gpointer) GST_SUB_PARSE_REGEX_SUBRIP);
1167 mdvd_grx = (GRegex *) mdvd_rx_once.retval;
1168 subrip_grx = (GRegex *) subrip_rx_once.retval;
1170 if (g_regex_match (mdvd_grx, match_str, 0, NULL) == TRUE) {
1171 GST_LOG ("MicroDVD (frame based) format detected");
1172 return GST_SUB_PARSE_FORMAT_MDVDSUB;
1174 if (g_regex_match (subrip_grx, match_str, 0, NULL) == TRUE) {
1175 GST_LOG ("SubRip (time based) format detected");
1176 return GST_SUB_PARSE_FORMAT_SUBRIP;
1179 if (!strncmp (match_str, "FORMAT=TIME", 11)) {
1180 GST_LOG ("MPSub (time based) format detected");
1181 return GST_SUB_PARSE_FORMAT_MPSUB;
1183 #ifndef GST_DISABLE_XML
1184 if (strstr (match_str, "<SAMI>") != NULL ||
1185 strstr (match_str, "<sami>") != NULL) {
1186 GST_LOG ("SAMI (time based) format detected");
1187 return GST_SUB_PARSE_FORMAT_SAMI;
1190 /* we're boldly assuming the first subtitle appears within the first hour */
1191 if (sscanf (match_str, "0:%02u:%02u:", &n1, &n2) == 2 ||
1192 sscanf (match_str, "0:%02u:%02u=", &n1, &n2) == 2 ||
1193 sscanf (match_str, "00:%02u:%02u:", &n1, &n2) == 2 ||
1194 sscanf (match_str, "00:%02u:%02u=", &n1, &n2) == 2 ||
1195 sscanf (match_str, "00:%02u:%02u,%u=", &n1, &n2, &n3) == 3) {
1196 GST_LOG ("TMPlayer (time based) format detected");
1197 return GST_SUB_PARSE_FORMAT_TMPLAYER;
1199 if (sscanf (match_str, "[%u][%u]", &n1, &n2) == 2) {
1200 GST_LOG ("MPL2 (time based) format detected");
1201 return GST_SUB_PARSE_FORMAT_MPL2;
1203 if (strstr (match_str, "[INFORMATION]") != NULL) {
1204 GST_LOG ("SubViewer (time based) format detected");
1205 return GST_SUB_PARSE_FORMAT_SUBVIEWER;
1208 GST_DEBUG ("no subtitle format detected");
1209 return GST_SUB_PARSE_FORMAT_UNKNOWN;
1213 gst_sub_parse_format_autodetect (GstSubParse * self)
1216 GstSubParseFormat format;
1218 if (strlen (self->textbuf->str) < 30) {
1219 GST_DEBUG ("File too small to be a subtitles file");
1223 data = g_strndup (self->textbuf->str, 35);
1224 format = gst_sub_parse_data_format_autodetect (data);
1227 self->parser_type = format;
1228 self->subtitle_codec = gst_sub_parse_get_format_description (format);
1229 parser_state_init (&self->state);
1232 case GST_SUB_PARSE_FORMAT_MDVDSUB:
1233 self->parse_line = parse_mdvdsub;
1234 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1235 case GST_SUB_PARSE_FORMAT_SUBRIP:
1236 self->parse_line = parse_subrip;
1237 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1238 case GST_SUB_PARSE_FORMAT_MPSUB:
1239 self->parse_line = parse_mpsub;
1240 return gst_caps_new_simple ("text/plain", NULL);
1241 #ifndef GST_DISABLE_XML
1242 case GST_SUB_PARSE_FORMAT_SAMI:
1243 self->parse_line = parse_sami;
1244 sami_context_init (&self->state);
1245 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1247 case GST_SUB_PARSE_FORMAT_TMPLAYER:
1248 self->parse_line = parse_tmplayer;
1249 self->state.max_duration = 5 * GST_SECOND;
1250 return gst_caps_new_simple ("text/plain", NULL);
1251 case GST_SUB_PARSE_FORMAT_MPL2:
1252 self->parse_line = parse_mpl2;
1253 return gst_caps_new_simple ("text/x-pango-markup", NULL);
1254 case GST_SUB_PARSE_FORMAT_SUBVIEWER:
1255 self->parse_line = parse_subviewer;
1256 return gst_caps_new_simple ("text/plain", NULL);
1257 case GST_SUB_PARSE_FORMAT_UNKNOWN:
1259 GST_DEBUG ("no subtitle format detected");
1260 GST_ELEMENT_ERROR (self, STREAM, WRONG_TYPE,
1261 ("The input is not a valid/supported subtitle file"), (NULL));
1267 feed_textbuf (GstSubParse * self, GstBuffer * buf)
1271 gchar *input = NULL;
1273 discont = GST_BUFFER_IS_DISCONT (buf);
1275 if (GST_BUFFER_OFFSET_IS_VALID (buf) &&
1276 GST_BUFFER_OFFSET (buf) != self->offset) {
1277 self->offset = GST_BUFFER_OFFSET (buf);
1282 GST_INFO ("discontinuity");
1283 /* flush the parser state */
1284 parser_state_init (&self->state);
1285 g_string_truncate (self->textbuf, 0);
1286 gst_adapter_clear (self->adapter);
1287 #ifndef GST_DISABLE_XML
1288 sami_context_reset (&self->state);
1290 /* we could set a flag to make sure that the next buffer we push out also
1291 * has the DISCONT flag set, but there's no point really given that it's
1292 * subtitles which are discontinuous by nature. */
1295 self->offset = GST_BUFFER_OFFSET (buf) + GST_BUFFER_SIZE (buf);
1296 self->next_offset = self->offset;
1298 gst_adapter_push (self->adapter, buf);
1301 convert_encoding (self, (const gchar *) gst_adapter_peek (self->adapter,
1302 gst_adapter_available (self->adapter)),
1303 (gsize) gst_adapter_available (self->adapter), &consumed);
1305 if (input && consumed > 0) {
1306 self->textbuf = g_string_append (self->textbuf, input);
1307 gst_adapter_flush (self->adapter, consumed);
1313 static GstFlowReturn
1314 handle_buffer (GstSubParse * self, GstBuffer * buf)
1316 GstFlowReturn ret = GST_FLOW_OK;
1317 GstCaps *caps = NULL;
1318 gchar *line, *subtitle;
1320 if (self->first_buffer) {
1321 self->detected_encoding =
1322 detect_encoding ((gchar *) GST_BUFFER_DATA (buf),
1323 GST_BUFFER_SIZE (buf));
1324 self->first_buffer = FALSE;
1327 feed_textbuf (self, buf);
1329 /* make sure we know the format */
1330 if (G_UNLIKELY (self->parser_type == GST_SUB_PARSE_FORMAT_UNKNOWN)) {
1331 if (!(caps = gst_sub_parse_format_autodetect (self))) {
1332 return GST_FLOW_UNEXPECTED;
1334 if (!gst_pad_set_caps (self->srcpad, caps)) {
1335 gst_caps_unref (caps);
1336 return GST_FLOW_UNEXPECTED;
1338 gst_caps_unref (caps);
1341 if (self->subtitle_codec != NULL) {
1344 tags = gst_tag_list_new ();
1345 gst_tag_list_add (tags, GST_TAG_MERGE_APPEND, GST_TAG_SUBTITLE_CODEC,
1346 self->subtitle_codec, NULL);
1347 gst_element_found_tags_for_pad (GST_ELEMENT (self), self->srcpad, tags);
1351 while (!self->flushing && (line = get_next_line (self))) {
1354 /* Set segment on our parser state machine */
1355 self->state.segment = &self->segment;
1356 /* Now parse the line, out of segment lines will just return NULL */
1357 GST_LOG_OBJECT (self, "Parsing line '%s'", line + offset);
1358 subtitle = self->parse_line (&self->state, line + offset);
1362 guint subtitle_len = strlen (subtitle);
1364 /* +1 for terminating NUL character */
1365 ret = gst_pad_alloc_buffer_and_set_caps (self->srcpad,
1366 GST_BUFFER_OFFSET_NONE, subtitle_len + 1,
1367 GST_PAD_CAPS (self->srcpad), &buf);
1369 if (ret == GST_FLOW_OK) {
1370 /* copy terminating NUL character as well */
1371 memcpy (GST_BUFFER_DATA (buf), subtitle, subtitle_len + 1);
1372 GST_BUFFER_SIZE (buf) = subtitle_len;
1373 GST_BUFFER_TIMESTAMP (buf) = self->state.start_time;
1374 GST_BUFFER_DURATION (buf) = self->state.duration;
1376 /* in some cases (e.g. tmplayer) we can only determine the duration
1377 * of a text chunk from the timestamp of the next text chunk; in those
1378 * cases, we probably want to limit the duration to something
1379 * reasonable, so we don't end up showing some text for e.g. 40 seconds
1380 * just because nothing else is being said during that time */
1381 if (self->state.max_duration > 0 && GST_BUFFER_DURATION_IS_VALID (buf)) {
1382 if (GST_BUFFER_DURATION (buf) > self->state.max_duration)
1383 GST_BUFFER_DURATION (buf) = self->state.max_duration;
1386 gst_segment_set_last_stop (&self->segment, GST_FORMAT_TIME,
1387 self->state.start_time);
1389 GST_DEBUG_OBJECT (self, "Sending text '%s', %" GST_TIME_FORMAT " + %"
1390 GST_TIME_FORMAT, subtitle, GST_TIME_ARGS (self->state.start_time),
1391 GST_TIME_ARGS (self->state.duration));
1393 ret = gst_pad_push (self->srcpad, buf);
1396 /* move this forward (the tmplayer parser needs this) */
1397 if (self->state.duration != GST_CLOCK_TIME_NONE)
1398 self->state.start_time += self->state.duration;
1403 if (ret != GST_FLOW_OK) {
1404 GST_DEBUG_OBJECT (self, "flow: %s", gst_flow_get_name (ret));
1413 static GstFlowReturn
1414 gst_sub_parse_chain (GstPad * sinkpad, GstBuffer * buf)
1419 self = GST_SUBPARSE (GST_PAD_PARENT (sinkpad));
1421 /* Push newsegment if needed */
1422 if (self->need_segment) {
1423 GST_LOG_OBJECT (self, "pushing newsegment event with %" GST_SEGMENT_FORMAT,
1426 gst_pad_push_event (self->srcpad, gst_event_new_new_segment (FALSE,
1427 self->segment.rate, self->segment.format,
1428 self->segment.last_stop, self->segment.stop, self->segment.time));
1429 self->need_segment = FALSE;
1432 ret = handle_buffer (self, buf);
1438 gst_sub_parse_sink_event (GstPad * pad, GstEvent * event)
1440 GstSubParse *self = GST_SUBPARSE (gst_pad_get_parent (pad));
1441 gboolean ret = FALSE;
1443 GST_DEBUG ("Handling %s event", GST_EVENT_TYPE_NAME (event));
1445 switch (GST_EVENT_TYPE (event)) {
1446 case GST_EVENT_EOS:{
1447 /* Make sure the last subrip chunk is pushed out even
1448 * if the file does not have an empty line at the end */
1449 if (self->parser_type == GST_SUB_PARSE_FORMAT_SUBRIP ||
1450 self->parser_type == GST_SUB_PARSE_FORMAT_TMPLAYER ||
1451 self->parser_type == GST_SUB_PARSE_FORMAT_MPL2) {
1452 GstBuffer *buf = gst_buffer_new_and_alloc (2 + 1);
1454 GST_DEBUG ("EOS. Pushing remaining text (if any)");
1455 GST_BUFFER_DATA (buf)[0] = '\n';
1456 GST_BUFFER_DATA (buf)[1] = '\n';
1457 GST_BUFFER_DATA (buf)[2] = '\0'; /* play it safe */
1458 GST_BUFFER_SIZE (buf) = 2;
1459 GST_BUFFER_OFFSET (buf) = self->offset;
1460 gst_sub_parse_chain (pad, buf);
1462 ret = gst_pad_event_default (pad, event);
1465 case GST_EVENT_NEWSEGMENT:
1469 gint64 start, stop, time;
1472 gst_event_parse_new_segment (event, &update, &rate, &format, &start,
1475 GST_DEBUG_OBJECT (self, "newsegment (%s)", gst_format_get_name (format));
1477 if (format == GST_FORMAT_TIME) {
1478 gst_segment_set_newsegment (&self->segment, update, rate, format,
1481 /* if not time format, we'll either start with a 0 timestamp anyway or
1482 * it's following a seek in which case we'll have saved the requested
1483 * seek segment and don't want to overwrite it (remember that on a seek
1484 * we always just seek back to the start in BYTES format and just throw
1485 * away all text that's before the requested position; if the subtitles
1486 * come from an upstream demuxer, it won't be able to handle our BYTES
1487 * seek request and instead send us a newsegment from the seek request
1488 * it received via its video pads instead, so all is fine then too) */
1492 gst_event_unref (event);
1495 case GST_EVENT_FLUSH_START:
1497 self->flushing = TRUE;
1499 ret = gst_pad_event_default (pad, event);
1502 case GST_EVENT_FLUSH_STOP:
1504 self->flushing = FALSE;
1506 ret = gst_pad_event_default (pad, event);
1510 ret = gst_pad_event_default (pad, event);
1514 gst_object_unref (self);
1520 static GstStateChangeReturn
1521 gst_sub_parse_change_state (GstElement * element, GstStateChange transition)
1523 GstStateChangeReturn ret = GST_STATE_CHANGE_SUCCESS;
1524 GstSubParse *self = GST_SUBPARSE (element);
1526 switch (transition) {
1527 case GST_STATE_CHANGE_READY_TO_PAUSED:
1528 /* format detection will init the parser state */
1530 self->next_offset = 0;
1531 self->parser_type = GST_SUB_PARSE_FORMAT_UNKNOWN;
1532 self->valid_utf8 = TRUE;
1533 self->first_buffer = TRUE;
1534 g_free (self->detected_encoding);
1535 self->detected_encoding = NULL;
1536 g_string_truncate (self->textbuf, 0);
1537 gst_adapter_clear (self->adapter);
1543 ret = parent_class->change_state (element, transition);
1544 if (ret == GST_STATE_CHANGE_FAILURE)
1547 switch (transition) {
1548 case GST_STATE_CHANGE_PAUSED_TO_READY:
1549 parser_state_dispose (&self->state);
1550 self->parser_type = GST_SUB_PARSE_FORMAT_UNKNOWN;
1563 /* FIXME 0.11: these caps are ugly, use app/x-subtitle + type field or so;
1564 * also, give different subtitle formats really different types */
1565 static GstStaticCaps mpl2_caps =
1566 GST_STATIC_CAPS ("application/x-subtitle-mpl2");
1567 #define SUB_CAPS (gst_static_caps_get (&sub_caps))
1569 static GstStaticCaps tmp_caps =
1570 GST_STATIC_CAPS ("application/x-subtitle-tmplayer");
1571 #define TMP_CAPS (gst_static_caps_get (&tmp_caps))
1573 static GstStaticCaps sub_caps = GST_STATIC_CAPS ("application/x-subtitle");
1574 #define MPL2_CAPS (gst_static_caps_get (&mpl2_caps))
1576 #ifndef GST_DISABLE_XML
1577 static GstStaticCaps smi_caps = GST_STATIC_CAPS ("application/x-subtitle-sami");
1578 #define SAMI_CAPS (gst_static_caps_get (&smi_caps))
1582 gst_subparse_type_find (GstTypeFind * tf, gpointer private)
1584 GstSubParseFormat format;
1588 gchar *encoding = NULL;
1591 if (!(data = gst_type_find_peek (tf, 0, 129)))
1594 /* make sure string passed to _autodetect() is NUL-terminated */
1595 str = g_malloc0 (129);
1596 memcpy (str, data, 128);
1598 if ((encoding = detect_encoding (str, 128)) != NULL) {
1599 gchar *converted_str;
1603 converted_str = gst_convert_to_utf8 (str, 128, encoding, &tmp, &err);
1604 if (converted_str == NULL) {
1605 GST_DEBUG ("Encoding '%s' detected but conversion failed: %s", encoding,
1611 str = converted_str;
1616 /* Check if at least the first 120 chars are valid UTF8,
1617 * otherwise convert as always */
1618 if (!g_utf8_validate (str, 128, &end) && (end - str) < 120) {
1619 gchar *converted_str;
1624 enc = g_getenv ("GST_SUBTITLE_ENCODING");
1625 if (enc == NULL || *enc == '\0') {
1626 /* if local encoding is UTF-8 and no encoding specified
1627 * via the environment variable, assume ISO-8859-15 */
1628 if (g_get_charset (&enc)) {
1629 enc = "ISO-8859-15";
1632 converted_str = gst_convert_to_utf8 (str, 128, enc, &tmp, &err);
1633 if (converted_str == NULL) {
1634 GST_DEBUG ("Charset conversion failed: %s", err->message);
1640 str = converted_str;
1644 format = gst_sub_parse_data_format_autodetect (str);
1648 case GST_SUB_PARSE_FORMAT_MDVDSUB:
1649 GST_DEBUG ("MicroDVD format detected");
1652 case GST_SUB_PARSE_FORMAT_SUBRIP:
1653 GST_DEBUG ("SubRip format detected");
1656 case GST_SUB_PARSE_FORMAT_MPSUB:
1657 GST_DEBUG ("MPSub format detected");
1660 #ifndef GST_DISABLE_XML
1661 case GST_SUB_PARSE_FORMAT_SAMI:
1662 GST_DEBUG ("SAMI (time-based) format detected");
1666 case GST_SUB_PARSE_FORMAT_TMPLAYER:
1667 GST_DEBUG ("TMPlayer (time based) format detected");
1670 /* FIXME: our MPL2 typefinding is not really good enough to warrant
1671 * returning a high probability (however, since we registered our
1672 * typefinder here with a rank of MARGINAL we should pretty much only
1673 * be called if most other typefinders have already run */
1674 case GST_SUB_PARSE_FORMAT_MPL2:
1675 GST_DEBUG ("MPL2 (time based) format detected");
1678 case GST_SUB_PARSE_FORMAT_SUBVIEWER:
1679 GST_DEBUG ("SubViewer format detected");
1683 case GST_SUB_PARSE_FORMAT_UNKNOWN:
1684 GST_DEBUG ("no subtitle format detected");
1688 /* if we're here, it's ok */
1689 gst_type_find_suggest (tf, GST_TYPE_FIND_MAXIMUM, caps);
1693 plugin_init (GstPlugin * plugin)
1695 static gchar *sub_exts[] = { "srt", "sub", "mpsub", "mdvd", "smi", "txt",
1699 GST_DEBUG_CATEGORY_INIT (sub_parse_debug, "subparse", 0, ".sub parser");
1701 if (!gst_type_find_register (plugin, "subparse_typefind", GST_RANK_MARGINAL,
1702 gst_subparse_type_find, sub_exts, SUB_CAPS, NULL, NULL))
1705 if (!gst_element_register (plugin, "subparse",
1706 GST_RANK_PRIMARY, GST_TYPE_SUBPARSE) ||
1707 !gst_element_register (plugin, "ssaparse",
1708 GST_RANK_PRIMARY, GST_TYPE_SSA_PARSE)) {
1715 GST_PLUGIN_DEFINE (GST_VERSION_MAJOR,
1719 plugin_init, VERSION, "LGPL", GST_PACKAGE_NAME, GST_PACKAGE_ORIGIN)