Original commit message from CVS:
* ext/dvdread/dvdreadsrc.c: (gst_dvd_read_src_init),
(gst_dvd_read_src_class_init), (gst_dvd_read_src_start),
(gst_dvd_read_src_stop), (gst_dvd_read_src_goto_title),
(gst_dvd_read_src_set_property), (gst_dvd_read_src_get_property),
(gst_dvd_read_src_uri_get_uri), (gst_dvd_read_src_uri_set_uri):
* ext/dvdread/dvdreadsrc.h:
Let's try to play the title the user set via the URI handler or
via properties instead of always playing the first title. Also,
Also, count title/chapter/angle number in URI and properties
from 1 rather than 0.
Original commit message from CVS:
* ext/dvdread/dvdreadsrc.c: (gst_dvd_read_src_class_init),
(gst_dvd_read_src_get_size), (gst_dvd_read_src_do_seek),
(gst_dvd_read_src_do_duration_query):
Don't implement GstBaseSrc::get_size or GstBaseSrc::is_seekable,
otherwise GstBaseSrc will think we can operate pull_range based,
which we don't really, and typefinding will fail miserably.
Also, make seeking work somewhat (only works with flumpegdemux
at the moment, mpegstream needs fixing for that first).
Original commit message from CVS:
* ext/dvdread/dvdreadsrc.c: (gst_dvd_read_src_read),
(gst_dvd_read_src_create), (gst_dvd_read_src_src_event):
Only allocate buffer once we know exactly how much we need,
rather than gratuitously allocating 2MB-buffers all the time
even if we usually need much less than that. Also, demote
a debug message from DEBUG to LOG level.
Original commit message from CVS:
* ext/dvdread/dvdreadsrc.c: (gst_dvd_read_src_class_init),
(gst_dvd_read_src_seekable), (gst_dvd_read_src_get_size),
(gst_dvd_read_src_do_seek), (gst_dvd_read_src_do_duration_query),
(gst_dvd_read_src_do_position_query):
Add some more debugging and fix duration query in BYTES.
Original commit message from CVS:
* configure.ac:
* ext/Makefile.am:
* ext/dvdread/Makefile.am:
* ext/dvdread/dvdreadsrc.c:
* ext/dvdread/dvdreadsrc.h:
Half-baked port to 0.10. Needs some love
in the seeking department, but at least
it does something.
* ext/dvdread/stream_labels.c:
* ext/dvdread/stream_labels.h:
Remove these (we use ISO-639 language codes internally; applications
that want to translate those into language names for display to the
user should rely on the iso-codes package for that).
Original commit message from CVS:
* ext/amrnb/amrnbdec.c: (gst_amrnbdec_init),
(gst_amrnbdec_setcaps), (gst_amrnbdec_event), (gst_amrnbdec_chain),
(gst_amrnbdec_state_change):
* ext/amrnb/amrnbdec.h:
Fix amrnbdec, handle events, take copy from adapter since the decoder
apparently writes in the source data.
Use some _scale_int, and precalc duration.
Fix some leaks, post ERROR messages.
Original commit message from CVS:
* gst/realmedia/rmdemux.c: (gst_rmdemux_add_stream):
One source pad not being linked is not an error condition when we're
still parsing the header. In this case (e.g. where we don't have a
suitable decoder installed) just pretend everything is fine, so that
the demuxer will actually go on to signal no-more-pads when done
parsing the header, otherwise decodebin/playbin will never post the
appropriate error message if decoders are not available.
Original commit message from CVS:
* gst/mpegstream/gstmpegparse.c: (gst_mpeg_parse_send_buffer):
If we haven't set caps on a source pad yet, the caps on the
pad are NULL, not un-fixed. Set caps on outgoing buffers.
Original commit message from CVS:
* ext/lame/gstlame.c: (gst_lame_finalize), (gst_lame_class_init),
(gst_lame_init), (add_one_tag), (gst_lame_set_metadata):
* ext/lame/gstlame.h:
Contrary to what the const char in the lame API might suggest,
lame expects us to keep the strings we pass to id3tag_set_foo()
around; it doesn't free them either though, so we have to store
them somewhere and free them later when we can be sure lame
doesn't need them any longer.
Original commit message from CVS:
* ext/lame/gstlame.c: (add_one_tag):
Fix handling of GST_TAG_DATE (#311679), don't pass an
uninitialised string pointer to lame if we don't know
how to handle the tag type, and fix minor memory leak.
Original commit message from CVS:
2006-01-22 Martin Soto <martinsoto@users.sourceforge.net>
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_process_event):
Erase spurious call to gst_segment_set_newsegment.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_process_event): Call
the superclass method when handling NEWSEGMENT events.
(gst_dvd_demux_handle_dvd_event): Get rid of dvd-audio-shutdown
and dvd-audio-restart event handling. There are currently less
hackish ways of handling the sparse audio stream problem.
Original commit message from CVS:
2006-01-20 Thomas Vander Stichele <thomas at apestaart dot org>
* ext/dvdnav/dvdnavsrc.c: (if):
* ext/dvdread/stream_labels.c:
* gst/asfdemux/gstasfdemux.c: (gst_asf_demux_process_segment):
* gst/realmedia/rmdemux.c: (gst_rmdemux_loop):
fix up error domains, error strings, and use of translation
* po/POTFILES.in:
fix up this file, even though none of them are actually marked
for build yet.
Original commit message from CVS:
2006-01-19 Martin Soto <martinsoto@users.sourceforge.net>
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParse):
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_reset)
(gst_mpeg_parse_process_event, gst_mpeg_parse_parse_packhead)
(gst_mpeg_parse_change_state):
Make timestamp adjustment somewhat milder. Actual timestamps are
now sent unmodified unless an actual gap is found in the
stream. This should fix time display when playing most MPEG
files.
Original commit message from CVS:
* ext/amrnb/amrnbparse.c: (gst_amrnbparse_query):
* ext/mad/gstid3tag.c: (gst_id3_tag_src_query):
* ext/mad/gstmad.c: (gst_mad_src_query):
* ext/mpeg2dec/gstmpeg2dec.c: (gst_mpeg2dec_src_query):
* ext/sidplay/gstsiddec.cc:
* gst/mpegstream/gstmpegparse.c: (gst_mpeg_parse_handle_src_query):
* gst/realmedia/rmdemux.c: (gst_rmdemux_src_query):
Pass unhandled queries upstream (useful e.g. for SEEKING query)
(fixes#325652; based on patch by: Philippe); make rmdemux return
FALSE for position queries, instead of setting -1 as value and
returning TRUE.
Original commit message from CVS:
2006-01-05 Martin Soto <martinsoto@users.sourceforge.net>
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_adjust_ts)
(gst_mpeg_parse_process_event, gst_mpeg_parse_pad_added): Don't
rewrite timestamps in the case segments are being set from
upstream, but use timestamps unmodified. Also send proper position
values. This allows for correct time display and makes queries
work in sink elements.
* gst/mpegstream/gstdvddemux.h:
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_init)
(gst_dvd_demux_handle_dvd_event, gst_dvd_demux_send_subbuffer):
Rename flush_filter to segment_filter, which is better represents
what the arreibute does.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_process_event):
Activate segment filtering when a timestamp discontinuity is seen.
Original commit message from CVS:
* ext/mad/gstmad.c: (gst_mad_chain):
Don't forget that we need to send out a newsegment event after a
restart even if we don't have enough data to decode a frame right
now.
Original commit message from CVS:
2005-12-31 Martin Soto <martinsoto@users.sourceforge.net>
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_handle_dvd_event):
Use the new "audio-shutdown" and "audio-restart" DVD events
instead of the "spu-still-frame" event to shutdown and restart
the audio pipeline.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_adjust_ts):
Check for cases where the segment-based adjustment calculation
would produce negative values (which result in an overflow) and
return GST_CLOCK_TIME_NONE instead.
* gst/mpegstream/gstdvddemux.h:
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_init)
(gst_dvd_demux_process_event, gst_dvd_demux_send_subbuffer):
Add a mechanism to discard audio buffers with timestamps outside
the currently set segment. This was causing (sometimes serious)
synchronization problems after seeking in DVDs with LPCM audio,
since VOBUs usually contain audio material that lies outside the
timestamp range specified by the header.
Original commit message from CVS:
* ext/mad/gstid3tag.c: (gst_id3_tag_get_type):
Remove lingering reference to GstID3Demux
* gst/mpegstream/gstmpegdemux.c: (gst_mpeg_demux_parse_pes):
Don't output debug saying padding streams are unknown type,
because they're not
Original commit message from CVS:
2005-12-27 Martin Soto <martinsoto@users.sourceforge.net>
* gst/mpegstream/gstdvddemux.c (AUDIO_CAPS)
(gst_dvd_demux_get_audio_stream): Use audio/x-lpcm as mimetype for
LPCM.
* gstmpegdemux.h:
* gstmpegdemux.c (gst_mpeg_demux_get_video_stream)
(gst_mpeg_demux_get_audio_stream):
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_get_video_stream)
(gst_dvd_demux_get_audio_stream)
(gst_dvd_demux_get_subpicture_stream)
(gst_dvd_demux_send_subbuffer):
Send current* pad buffers with appropriate caps.
Original commit message from CVS:
2005-12-26 Martin Soto <martinsoto@users.sourceforge.net>
* gst/iec958/ac3iec.h:
* gst/iec958/ac3iec.c:
(NORMAL_CAPS_DEF, RAW_AUDIO_CAPS_DEF, ac3iec_class_init)
(ac3iec_init, ac3iec_set_property, ac3iec_get_property): Add a
raw-audio property to ac3iec958 that allows setting the source pad
caps to raw audio instead of audio/x-iec958. This makes it
possible to use ac3iec958 together with the normal alsasink
element to drive an external receiver that autodetects AC3
content.
Original commit message from CVS:
* gst/iec958/ac3iec.c: (ac3iec_init), (ac3iec_setcaps),
(ac3iec_chain_dvd):
* gst/iec958/ac3iec.h:
Accept audio/x-ac3 and audio/ac3 to ac3iec958 element.
Original commit message from CVS:
* ext/mpeg2dec/gstmpeg2dec.c: (gst_mpeg2dec_src_event):
When getting a seek event, first check if the upstream element
can handle it and only do our own seek stuff when it can't
(should fix#322856).
Original commit message from CVS:
Reviewed by: Tim-Philipp Müller <tim at centricular dot net>
* ext/mpeg2dec/gstmpeg2dec.c: (normal_seek):
Really convert to BYTES format when we want to convert
to bytes (use right variable; fixes#322856).
Original commit message from CVS:
2005-12-19 Martin Soto <martinsoto@users.sourceforge.net>
Big mpegparse clean up, second round:
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_handle_dvd_event):
Send and EOS event down the audio pipeline when an still frame
event arrives. This prevents the pipeline from locking when a
still menu comes directly after a flush.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_reset):
Don't send a newsegment in reset.
(gst_mpeg_parse_adjust_ts): Check for invalid timestamps.
(gst_mpeg_parse_handle_newsegment, gst_mpeg_parse_process_event):
Move the code of handle_newsegment to process_event. Send a
NEWSEGMENT after FLUSH_STOP.
(gst_mpeg_parse_change_state): Send a NEWSEGMENT right after
moving to PAUSED.
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_send_event)
(gst_mpeg_demux_class_init): Don't override send_event.
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_init)
(gst_mpeg_demux_send_event, gst_mpeg_demux_send_subbuffer)
(gst_mpeg_demux_reset):
* gst/mpegstream/gstmpegdemux.h: Get rid of just_flushed
attribute.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_process_event):
Reset the mpegparse element after a flush.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_handle_newsegment):
Don't forward events.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_process_event):
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParseClass):
handle_newsegment is not a virtual method anymore.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_send_newsegment)
(gst_mpeg_parse_reset, gst_mpeg_parse_class_init):
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParseClass): Get
rid of send_newsegment virtual method.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_process_event): Only
handle DVD events and call the superclass method for other event
types.
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_send_event): Don't
override process_event anymore.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_event)
(gst_mpeg_parse_process_event): Move actual event processing to
process event so that subclasses can properly override or extend
it.
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParseClass):
Eliminate time parameter in process event.
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_init)
(gst_mpeg_demux_parse_packet, gst_mpeg_demux_parse_pes)
(gst_mpeg_demux_send_subbuffer):
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_class_init)
(gst_mpeg_parse_parse_packhead, gst_mpeg_parse_event)
(gst_mpeg_parse_chain): Use the new adjust_ts method instead of
adding the value of the adjust attribute.
* gst/mpegstream/gstmpegdemux.h (struct _GstMPEGVideoStream): Get
rid of the adjust attribute. Now all timestamp adjustments are
performed by mpegparse using the current segment.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_adjust_ts)
(gst_mpeg_parse_class_init): Implement the adjust_ts method based
on the adjust attribute for SCR values and the current segment.
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParseClass): New
adjust_ts virtual method to adjust timestamps for outgoing
buffers.
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_send_newsegment)
(gst_mpeg_demux_parse_packet): Don't override send_newsegment.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_class_init)
(gst_dvd_demux_handle_newsegment): Don't override
handle_newsegment.
(gst_dvd_demux_process_event, gst_dvd_demux_handle_dvd_event):
Check for DVD events in process_event instead of
handle_dvd_event.
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParseClass):
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_handle_newsegment)
(gst_mpeg_parse_send_newsegment, gst_mpeg_parse_send_event):
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_process_event)
(gst_mpeg_demux_send_event):
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_process_event)
(gst_dvd_demux_handle_dvd_event):
Eliminate the time parameter in send_event.
Original commit message from CVS:
* ext/mad/gstid3tag.c: (gst_id3_tag_do_typefind),
(gst_id3_tag_do_caps_nego), (gst_id3_tag_chain), (plugin_init):
Fix typefinding in id3demux, and then remove it in favour
of the new LGPL id3demux in gst-plugins-good
* ext/mad/gstmad.c: (gst_mad_dispose):
dispose can run more than once.
Original commit message from CVS:
* ext/mad/gstid3tag.c: (gst_mad_id3_to_tag_list),
(gst_id3_tag_do_typefind):
Use the correct function to free typefind factory list.
Original commit message from CVS:
* ext/mad/gstmad.c: (gst_mad_src_query):
For position and duration queries in TIME format, try the peer
first (might be a demuxer).
Original commit message from CVS:
2005-12-08 Martin Soto <martinsoto@users.sourceforge.net>
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_handle_newsegment):
Properly handle non contiguous VOBUs by adding the segment accum
field to the adjust value.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_class_init)
(gst_dvd_demux_init, gst_dvd_demux_handle_newsegment): Extend
handle_newsegment to prevent sending actual newsegment events and
use tiemstamp rewriting instead.
(gst_dvd_demux_handle_dvd_event): Don't send a newsegment after
dvd-lang-codes.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_handle_newsegment)
(gst_mpeg_parse_event):
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParseClass):
Add a new parameter to handle_newsegment to allow controlling
whether newsegment events are forwarded or not.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_init)
(gst_dvd_demux_handle_dvd_event, gst_dvd_demux_send_subbuffer)
(gst_dvd_demux_reset):
last_end_ptm and discont_time aren't necessary anymore, since
timestamp adjustment is now replaced by newsegment events.
(gst_dvd_demux_init): Prevent MPEGParse from adjusting
timestamps.
* gst/mpegstream/gstdvddemux.h:
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_handle_dvd_event):
Don't handle dvd-nav-packet events anymore, since the are now
replaced by standard newsegment events.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_init)
(gst_dvd_demux_handle_dvd_event, gst_dvd_demux_send_subbuffer)
(gst_dvd_demux_change_state):
* gst/mpegstream/gstdvddemux.h:
Get rid of the ignore_next_newmedia_discont hack.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_class_init)
(gst_dvd_demux_handle_newsegment):
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_class_init)
(gst_mpeg_demux_handle_newsegment): Don't override
handle_newsegment anymore. It was only necessary to handle
NEWMEDIA events.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_process_event)
(gst_dvd_demux_handle_dvd_event, gst_dvd_demux_set_cur_audio)
(gst_dvd_demux_set_cur_subpicture):
Reactivate handling of DVD events.
* gst/mpegstream/gstmpegparse.c (normal_seek)
(gst_mpeg_parse_handle_src_event)
(gst_mpeg_parse_handle_src_query):
First attempt at reenabling seek.
* gst/mpegstream/gstmpegparse.h:
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_init)
(gst_mpeg_parse_update_streaminfo)
(gst_mpeg_parse_get_rate)
(gst_mpeg_parse_convert, gst_mpeg_parse_get_src_query_types)
(gst_mpeg_parse_handle_src_query):
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_new_output_pad):
Make queries work again.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_reset)
(gst_mpeg_parse_handle_newsegment)
(gst_mpeg_parse_send_newsegment, gst_mpeg_parse_pad_added)
(gst_mpeg_parse_chain, gst_mpeg_parse_handle_src_event):
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParse):
Get rid of the newsegment_pending attribute, and rely instead on
proper timestamp adjustment.
* gst/mpegstream/gstmpegparse.c
(gst_mpeg_parse_get_src_event_masks): Erase.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_init):
Initialize fields before creating pads to prevent the pad creation
callback from failing.
(gst_mpeg_parse_reset): Initialize new fields, and change
initialization order to match the order in the structure.
(gst_mpeg_parse_handle_newsegment): Forward new segment events
whenever possible, and update the current segment.
(gst_mpeg_parse_send_newsegment): Update the current segment and
pending_newsegment.
(gst_mpeg_parse_pad_added): Use the current segment to send
newsegment events to new pads.
(gst_mpeg_parse_chain): Properly add adjust time to sent buffers
and events. Properly update newsegment_pending.
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParse):
New fields do_adjust and current_segment.
* gst/mpegstream/gstmpegdemux.c:
* gst/mpegstream/gstdvddemux.c:
* gst/mpegstream/gstmpegparse.h:
Rename handle_discont virtual method to handle_newsegment. Erase
some (already commented out support) for old NEW_MEDIA events.
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParse):
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_init)
(gst_mpeg_parse_reset, gst_mpeg_parse_handle_discont)
(gst_mpeg_parse_pad_added, gst_mpeg_parse_parse_packhead)
(gst_mpeg_parse_event, gst_mpeg_parse_chain):
Erase the "pending_scr" field, and replace it by a slightly
different handling of the current SCR. Document code blocks in
parse_packhead and chain.
Original commit message from CVS:
* ext/mad/gstid3tag.c: (gst_mad_id3_to_tag_list):
g_warning() are only for really fatal warnings. If we can't decode a
tag, just ignore that tag and do a GST_WARNING.
Original commit message from CVS:
2005-12-05 Andy Wingo <wingo@pobox.com>
* ext/mad/gstid3tag.c (plugin_init): Remove id3tag -- there's just
id3demux and id3mux now. Fixes#323199.
Original commit message from CVS:
Reviewed by: Tim-Philipp Müller <tim at centricular dot net>
* gst/mpegstream/gstmpegdemux.c: (gst_mpeg_demux_send_event):
* gst/mpegstream/gstmpegparse.c: (gst_mpeg_parse_send_newsegment),
(gst_mpeg_parse_send_event):
Ref events before sending them to multiple pads, after all
gst_pad_send_event() takes ownership of events. Don't leak
events that have not been handled (fixes#322745).
Original commit message from CVS:
2005-11-28 Martin Soto <martinsoto@users.sourceforge.net>
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_handle_dvd_event)
(gst_dvd_demux_handle_dvd_event): Erase code to prevent mpegparse
from making timestamp adjustments. This will have to be re-added
in some form in the near future, but in order to do that, some nav
packet parsing will be necessary in mpegdemux.
* gst/mpegstream/gstmpegparse.h (struct _GstMPEGParse):
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_init)
(g_value_set_int, gst_mpeg_parse_set_property)
Get rid of do_adjust and use_adjust. Rename max_discont to
max_src_gap.
(gst_mpeg_parse_parse_packhead): When max_scr_gap has a value of
-1, no adjustment is made.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_class_init):
Rename max_discont property to max_scr_gap. Erase "adjust"
property.
* gst/mpegstream/gstdvddemux.c (gst_dvd_demux_send_event): Don't
override send_event anymore, base class does the job.
* gst/mpegstream/gstmpegdemux.c (gst_mpeg_demux_send_event):
Base class now does most of the work.
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_send_event):
Generalize to forwarding the event to all source pads in the
element.
* gst/mpegstream/gstmpegparse.h:
* gst/mpegstream/gstmpegparse.c (gst_mpeg_parse_class_init)
(gst_mpeg_parse_init, gst_mpeg_parse_set_clock)
(gst_mpeg_parse_chain, gst_mpeg_parse_get_property)
(gst_mpeg_parse_set_property):
Clock synchronization doesn't make sense anymore for a
demultiplexer.
Original commit message from CVS:
* ext/mad/gstmad.c: (gst_mad_chain):
Proper warning statements,
Don't error if the src pad isn't linked when pushing.
* gst/mpegstream/gstmpegparse.c: (gst_mpeg_parse_change_state):
Proper state_change, avoids borkage when going to READY
