platform/upstream/gstreamer.git
11 years agocodec-utils: add a gst_codec_utils_aac_get_index_from_sample_rate
Vincent Penquerc'h [Thu, 25 Jul 2013 19:03:50 +0000 (20:03 +0100)]
codec-utils: add a gst_codec_utils_aac_get_index_from_sample_rate

It maps a sample rate to a well known AAC sample rate index.

11 years agovideodecoder/encoder: Call reset() always between start() and stop() and never outside
Sebastian Dröge [Fri, 26 Jul 2013 08:22:32 +0000 (10:22 +0200)]
videodecoder/encoder: Call reset() always between start() and stop() and never outside

11 years agovideoencoder/decoder: Call reset() before start() too
Sebastian Dröge [Thu, 25 Jul 2013 12:25:08 +0000 (14:25 +0200)]
videoencoder/decoder: Call reset() before start() too

11 years agotypefinding: don't detect mp3 based on just a few bits
Tim-Philipp Müller [Thu, 25 Jul 2013 10:56:07 +0000 (11:56 +0100)]
typefinding: don't detect mp3 based on just a few bits

Remove dodgy code that detects mp3 with as little as
a valid frame sync at the beginning. This was only used
in some unit tests in -good where there were only a few
bytes after the id3 tag. We now require at least two
frame headers.

Fixes mis-dection of text files with UTF-16 LE BOM as mp3.

https://bugzilla.gnome.org/show_bug.cgi?id=681368

11 years agoaudio/videodecoder: Rename variable in macro from dec to __dec
Sebastian Dröge [Thu, 25 Jul 2013 12:11:28 +0000 (14:11 +0200)]
audio/videodecoder: Rename variable in macro from dec to __dec

Otherwise it might shadow another variable in the outside scope
and cause interesting side effects.

11 years agotheoraenc: Clean up handling of reset/flushing/start/stop
Sebastian Dröge [Thu, 25 Jul 2013 11:31:07 +0000 (13:31 +0200)]
theoraenc: Clean up handling of reset/flushing/start/stop

11 years agotheoradec: Clean up handling of reset/flushing/start/stop
Sebastian Dröge [Thu, 25 Jul 2013 11:29:22 +0000 (13:29 +0200)]
theoradec: Clean up handling of reset/flushing/start/stop

11 years agovideoencoder: There's no point in resetting the encoder when the caps change
Sebastian Dröge [Thu, 25 Jul 2013 08:53:14 +0000 (10:53 +0200)]
videoencoder: There's no point in resetting the encoder when the caps change

The subclass will be called with set_format() and there it can drain
if necessary and reset whatever is necessary. This is the same behaviour
as for the video decoder.

11 years agovideoencoder: Reset internal state and segments on FLUSH_STOP
Sebastian Dröge [Thu, 25 Jul 2013 08:46:04 +0000 (10:46 +0200)]
videoencoder: Reset internal state and segments on FLUSH_STOP

https://bugzilla.gnome.org/show_bug.cgi?id=656007

11 years agovideoencoder: Refactor GstVideoEncoder::reset() handling a bit
Sebastian Dröge [Thu, 25 Jul 2013 08:42:56 +0000 (10:42 +0200)]
videoencoder: Refactor GstVideoEncoder::reset() handling a bit

Let gst_video_encoder_reset() call it as would be intuitive and
only call it indirectly from gst_video_encoder_drain(). Now it
actually makes sense.

11 years agovideodecoder: Refactor GstVideoDecoder::reset() handling a bit
Sebastian Dröge [Thu, 25 Jul 2013 08:20:01 +0000 (10:20 +0200)]
videodecoder: Refactor GstVideoDecoder::reset() handling a bit

Let gst_video_decoder_reset() call it as would be intuitive and
only call it indirectly from gst_video_decoder_flush(). Now it
actually makes sense.

11 years agovideodecoder: Take DTS as PTS for keyframes as a last resort if we can't calculate...
Wim Taymans [Wed, 24 Jul 2013 07:24:45 +0000 (09:24 +0200)]
videodecoder: Take DTS as PTS for keyframes as a last resort if we can't calculate any PTS

https://bugzilla.gnome.org/show_bug.cgi?id=704193

11 years agoplaysink: Remove two unused variables
Sebastian Dröge [Tue, 23 Jul 2013 11:42:40 +0000 (13:42 +0200)]
playsink: Remove two unused variables

11 years agosubtitleoverlay: Remove some more unneeded segment tracking
Sebastian Dröge [Tue, 23 Jul 2013 11:38:33 +0000 (13:38 +0200)]
subtitleoverlay: Remove some more unneeded segment tracking

11 years agoplaysinkconvertbin: Remove obsolete segment tracking
Sebastian Dröge [Tue, 23 Jul 2013 11:36:09 +0000 (13:36 +0200)]
playsinkconvertbin: Remove obsolete segment tracking

This is now all handled automatically with sticky events.

11 years agoplaybin: Ensure that everything in a GstSourceGroup gets the same group-id
Sebastian Dröge [Tue, 23 Jul 2013 10:40:57 +0000 (12:40 +0200)]
playbin: Ensure that everything in a GstSourceGroup gets the same group-id

Including streams from external subtitle files. If not everything implements
the group-ids playbin invents its own.

11 years agoplaysink: Fix handling of colorbalance element if the sink does not implement it
Sebastian Dröge [Tue, 23 Jul 2013 10:35:46 +0000 (12:35 +0200)]
playsink: Fix handling of colorbalance element if the sink does not implement it

11 years agoplaysink: Improve segment handling in the custom flushing for subtitle stream switches
Sebastian Dröge [Tue, 23 Jul 2013 09:13:19 +0000 (11:13 +0200)]
playsink: Improve segment handling in the custom flushing for subtitle stream switches

This code needs a lot more work to be improved for 1.0.

11 years agosubtitleoverlay: Segment events are sticky now and propagated automatically when...
Sebastian Dröge [Tue, 23 Jul 2013 09:11:33 +0000 (11:11 +0200)]
subtitleoverlay: Segment events are sticky now and propagated automatically when needed

11 years agostreamcombiner: Proxy all sink events downstream
Sebastian Dröge [Tue, 23 Jul 2013 07:14:23 +0000 (09:14 +0200)]
streamcombiner: Proxy all sink events downstream

Thanks to Mathieu Duponchelle for noticing this regression
introduced with the last change.

https://bugzilla.gnome.org/show_bug.cgi?id=704706

11 years agotagdemux: Add support for group-id in the stream-start event
Sebastian Dröge [Mon, 22 Jul 2013 13:24:50 +0000 (15:24 +0200)]
tagdemux: Add support for group-id in the stream-start event

11 years agoogg: Add support for group-id in the stream-start event
Sebastian Dröge [Mon, 22 Jul 2013 13:24:29 +0000 (15:24 +0200)]
ogg: Add support for group-id in the stream-start event

11 years agostreamsynchronizer: Implement grouping of streams via the group-id
Sebastian Dröge [Mon, 22 Jul 2013 11:15:09 +0000 (13:15 +0200)]
streamsynchronizer: Implement grouping of streams via the group-id

https://bugzilla.gnome.org/show_bug.cgi?id=704427
https://bugzilla.gnome.org/show_bug.cgi?id=704408

11 years agostreamcombiner: Fix locking
Sebastian Dröge [Mon, 22 Jul 2013 06:08:27 +0000 (08:08 +0200)]
streamcombiner: Fix locking

We have to hold the streams-lock when iterating over all pads,
also the stream-lock of the pad is already locked when we receive
EOS.

Call gst_pad_event_default() for the correct default handling of
events.

11 years agoencoding: fix EOS handling in streamsplitter / combiner.
Mathieu Duponchelle [Sun, 21 Jul 2013 22:48:54 +0000 (00:48 +0200)]
encoding: fix EOS handling in streamsplitter / combiner.

This commit adds a streamcombinerpad with an is_eos field.
When streamcombiner receives an EOS on one of its pads, it
forwards it all its other pads are EOS.
This commit also removes the notion of "stream-switching-eos".

11 years agovideo-info: respect stride alignment
Wim Taymans [Fri, 19 Jul 2013 08:47:27 +0000 (10:47 +0200)]
video-info: respect stride alignment

Increase the left padding so that we don't cause stride alignments later when we
apply the padding.

https://bugzilla.gnome.org/show_bug.cgi?id=694299

11 years agoRevert "video: respect stride alignment when calculating planes offsets"
Wim Taymans [Fri, 19 Jul 2013 08:43:38 +0000 (10:43 +0200)]
Revert "video: respect stride alignment when calculating planes offsets"

This reverts commit 28e1dadbfaa403679e69f8173d1aa2c7500fd556.

Incrementing the offset to make the plane aligned causes the image to be
incompatible with what Xv expects. Rather that forcing a memcpy in the
xvimagesink we would like to do adjust the left padding instead.

11 years agovideo: respect stride alignment when calculating planes offsets
Arnaud Vrac [Thu, 18 Jul 2013 12:13:33 +0000 (14:13 +0200)]
video: respect stride alignment when calculating planes offsets

https://bugzilla.gnome.org/show_bug.cgi?id=694299

11 years agoid3: Use debug category and show FIXMEs
Edward Hervey [Thu, 18 Jul 2013 05:45:47 +0000 (07:45 +0200)]
id3: Use debug category and show FIXMEs

Allows spotting faster un-parsed tags

11 years agovideometa: Add docs to the region of interest meta functions
Sebastian Dröge [Wed, 17 Jul 2013 09:42:48 +0000 (11:42 +0200)]
videometa: Add docs to the region of interest meta functions

11 years agosubparse: use g_strdup() and friends
Tim-Philipp Müller [Wed, 17 Jul 2013 08:04:47 +0000 (09:04 +0100)]
subparse: use g_strdup() and friends

Fixes build issue on windows, but is also better seeing that
these string are going to get freed with g_free() and not free().

11 years agotagdemux: Put the modified time back in the time part of the segment
Olivier Crête [Tue, 16 Jul 2013 02:27:20 +0000 (22:27 -0400)]
tagdemux: Put the modified time back in the time part of the segment

https://bugzilla.gnome.org/show_bug.cgi?id=704301

11 years agotests: fix videoscale test after video format addition
Tim-Philipp Müller [Tue, 16 Jul 2013 17:50:09 +0000 (18:50 +0100)]
tests: fix videoscale test after video format addition

11 years agoconfigure: remove obsolete libxml checks
Tim-Philipp Müller [Tue, 16 Jul 2013 17:42:19 +0000 (18:42 +0100)]
configure: remove obsolete libxml checks

https://bugzilla.gnome.org/show_bug.cgi?id=693056

11 years agosubparse: don't leak parse context for sami and qttext
Tim-Philipp Müller [Tue, 16 Jul 2013 17:30:50 +0000 (18:30 +0100)]
subparse: don't leak parse context for sami and qttext

In gst_sub_parse_dispose() parser_type will be UNKNOWN,
so these deinit calls were never executed. And we should
clean up the parser state in the downwards state change
anyway.

11 years agotests: update sami parser testcases
Young-Ho Cha [Tue, 28 May 2013 07:56:28 +0000 (16:56 +0900)]
tests: update sami parser testcases

Remove libxml dependency for sami parser
and add more testcases.

https://bugzilla.gnome.org/show_bug.cgi?id=693056

11 years agosubparse: remove libxml dependency for sami parser and re-enable sami parser
Young-Ho Cha [Sat, 25 May 2013 08:10:14 +0000 (17:10 +0900)]
subparse: remove libxml dependency for sami parser and re-enable sami parser

To celebrate 2013.gnome.asia, updated sami parser for gstreamer 1.x. :D

Remove conditional block for check libxml usage and
implement a simple html markup parser for the sami
parser.

https://bugzilla.gnome.org/show_bug.cgi?id=693056

11 years agometa: fix ROI meta getter
Wim Taymans [Tue, 16 Jul 2013 14:54:10 +0000 (16:54 +0200)]
meta: fix ROI meta getter

11 years agoplaybin: Don't prefer decoders for which we found a matching sink
Sebastian Dröge [Tue, 16 Jul 2013 10:21:44 +0000 (12:21 +0200)]
playbin: Don't prefer decoders for which we found a matching sink

It doesn't make much sense.

11 years agovideo: Add support for NV24 color format
Sebastian Dröge [Tue, 16 Jul 2013 09:47:59 +0000 (11:47 +0200)]
video: Add support for NV24 color format

This is semi-planar 4:4:4 YUV.

https://bugzilla.gnome.org/show_bug.cgi?id=703259

11 years agoplaybin: Also consider possible converters for raw streams when selecting compatible...
Sebastian Dröge [Tue, 16 Jul 2013 09:22:35 +0000 (11:22 +0200)]
playbin: Also consider possible converters for raw streams when selecting compatible sink/source combinations

https://bugzilla.gnome.org/show_bug.cgi?id=704285

11 years agovideometa: Add to the docs and make function names more consistent with others
Sebastian Dröge [Tue, 16 Jul 2013 08:09:27 +0000 (10:09 +0200)]
videometa: Add to the docs and make function names more consistent with others

11 years agovideometa: Add Region Of Interest meta
Miguel Casas-Sanchez [Tue, 16 Jul 2013 08:04:00 +0000 (10:04 +0200)]
videometa: Add Region Of Interest meta

https://bugzilla.gnome.org/show_bug.cgi?id=704070

11 years agoplaybin: Fix sorting for decoder factories for which we didn't find a compatible...
Sebastian Dröge [Tue, 16 Jul 2013 07:30:33 +0000 (09:30 +0200)]
playbin: Fix sorting for decoder factories for which we didn't find a compatible sink

They might just need some converters for raw audio/video.

https://bugzilla.gnome.org/show_bug.cgi?id=704285

11 years agoriff-media: Add 'png ' fourcc
Olivier Crête [Mon, 15 Jul 2013 21:09:16 +0000 (17:09 -0400)]
riff-media: Add 'png ' fourcc

On top of mpng, MPNG, PNG, there is also png it seems

https://bugzilla.gnome.org/show_bug.cgi?id=704291

11 years agoallocators: dmabuf: allow testing allocator type
Benjamin Gaignard [Mon, 15 Jul 2013 13:23:17 +0000 (15:23 +0200)]
allocators: dmabuf: allow testing allocator type

In decide_allocation function some element may when to test the proposed allocator.
For example like this:
if (gst_query_get_n_allocation_params (query) > 0) {
GstAllocator * allocator;
GstAllocationParams params;
gst_query_parse_nth_allocation_param (query, 0, &allocator, &params);
if (g_strcmp0(allocator->mem_type, GST_ALLOCATOR_DMABUF) == 0)
GST_DEBUG("got dmabuf allocator");
else
GST_DEBUG("got an other allocator");
}

https://bugzilla.gnome.org/show_bug.cgi?id=703659

11 years agooggdemux: Make bisecting fully accurate
Mathieu Duponchelle [Sat, 13 Jul 2013 23:42:52 +0000 (01:42 +0200)]
oggdemux: Make bisecting fully accurate

When bisecting after an earliest time has been found, we need
to only consider the stream for which the earliest time was found.

Before, the following scenario could be and was encountered:

a) Find the earliest time for stream X
b) bisect and find a page which granuletime is indeed < target, but
   contains another stream.
c) decide to seek at the wrong offset, sometimes inferior to
   the real one, in which case the error was undected or
d) the offset was superior, and thus the actual target keyframe was
   not processed, and packets were skipped waiting
   for a granulepos.

https://bugzilla.gnome.org/show_bug.cgi?id=700537

11 years agoRevert "oggdemux: fix seeking with negative rate with skeleton"
Mathieu Duponchelle [Sat, 13 Jul 2013 18:45:01 +0000 (20:45 +0200)]
Revert "oggdemux: fix seeking with negative rate with skeleton"

This reverts commit b41cd0428956f3ade9b428149e38be8e788556fe.

11 years agoplaybin: Don't print a warning when setting a sink to NULL
Sebastian Dröge [Mon, 15 Jul 2013 07:10:30 +0000 (09:10 +0200)]
playbin: Don't print a warning when setting a sink to NULL

https://bugzilla.gnome.org/show_bug.cgi?id=704194

11 years agortspconnection: Create a new write GSource after removing it
Sebastian Dröge [Sun, 14 Jul 2013 16:11:59 +0000 (18:11 +0200)]
rtspconnection: Create a new write GSource after removing it

After removal, a GSource is destroyed and can never be attached
again to a main context. We need to create a new one instead.

https://bugzilla.gnome.org/show_bug.cgi?id=704198

11 years agoplaybin: Properly destroy and set to NULL sinks that don't work
Sebastian Dröge [Fri, 12 Jul 2013 10:05:37 +0000 (12:05 +0200)]
playbin: Properly destroy and set to NULL sinks that don't work

11 years agoplaybin: Fix logic to detect if a stream-change is currently pending
Alban Browaeys [Mon, 8 Jul 2013 21:49:39 +0000 (23:49 +0200)]
playbin: Fix logic to detect if a stream-change is currently pending

Fixes duration reporting in gapless playback between files.

https://bugzilla.gnome.org/show_bug.cgi?id=585969

11 years agovideometa: fix header formatting
Wim Taymans [Fri, 12 Jul 2013 07:37:51 +0000 (09:37 +0200)]
videometa: fix header formatting

11 years agoriff: Provide correct media type for XSub
Reynaldo H. Verdejo Pinochet [Wed, 10 Jul 2013 17:27:21 +0000 (13:27 -0400)]
riff: Provide correct media type for XSub

Xsub (fourcc DXSB) is a subpicture stream used for embeded
subtitles on divx files. This provides a correct media type
for them instead of just video/x-avi-unknown.

11 years agoBack to development
Sebastian Dröge [Thu, 11 Jul 2013 14:57:11 +0000 (16:57 +0200)]
Back to development

11 years agoRelease 1.1.2
Sebastian Dröge [Thu, 11 Jul 2013 13:30:23 +0000 (15:30 +0200)]
Release 1.1.2

11 years agoUpdate .po files
Sebastian Dröge [Thu, 11 Jul 2013 13:29:57 +0000 (15:29 +0200)]
Update .po files

11 years agoplaybin: Only give sinks a new bus if they have no parent yet
Sebastian Dröge [Wed, 10 Jul 2013 15:16:14 +0000 (17:16 +0200)]
playbin: Only give sinks a new bus if they have no parent yet

Otherwise we will remove the bus that would proxy messages to playsink
and never set it again. If the sink is already in playsink, all failures
are fatal anyway as it's either a sink that worked before or one that
was set by the user.

https://bugzilla.gnome.org/show_bug.cgi?id=701997

11 years agoplaybin: Store a/v/t sinks locally too, not just in playsink
Sebastian Dröge [Wed, 10 Jul 2013 11:22:04 +0000 (13:22 +0200)]
playbin: Store a/v/t sinks locally too, not just in playsink

11 years agoplaysink: ref_sink() any sinks that are set on playsink
Sebastian Dröge [Wed, 10 Jul 2013 11:21:29 +0000 (13:21 +0200)]
playsink: ref_sink() any sinks that are set on playsink

Otherwise the behaviour of the properties is inconsistent.

11 years agoplaybin: Fix assumptions in the unit test
Sebastian Dröge [Wed, 10 Jul 2013 11:20:34 +0000 (13:20 +0200)]
playbin: Fix assumptions in the unit test

Unused sinks are still set to READY now during autoplugging
to check their caps. Also playsink owns a ref to the sinks too.

11 years agostreamsynchronizer: Non-TIME segment streams are not waiting automatically
Sebastian Dröge [Wed, 10 Jul 2013 11:00:21 +0000 (13:00 +0200)]
streamsynchronizer: Non-TIME segment streams are not waiting automatically

This was leftover code from porting to 1.0 and fixes the playbin
unit test.

https://bugzilla.gnome.org/show_bug.cgi?id=701943

11 years agowin32: add missing rtp buffer methods
Branko Subasic [Tue, 9 Jul 2013 21:04:49 +0000 (23:04 +0200)]
win32: add missing rtp buffer methods

11 years agoplaybin: Change sink ownership handling to be a bit more sane
Sebastian Dröge [Tue, 9 Jul 2013 12:55:57 +0000 (14:55 +0200)]
playbin: Change sink ownership handling to be a bit more sane

playbin will now only activate the sinks in a single place and
will never change the states of any sinks that are owned by
playsink.

Also handle text-sinks the same way as audio/video sinks inside
playbin.

11 years agopo: update POTFILES.in
Piotr Drąg [Fri, 5 Jul 2013 19:55:26 +0000 (21:55 +0200)]
po: update POTFILES.in

https://bugzilla.gnome.org/show_bug.cgi?id=703684

11 years agocolorbalance: Fix the typo in base_init().
Sreerenj Balachandran [Thu, 4 Jul 2013 14:09:00 +0000 (17:09 +0300)]
colorbalance: Fix the typo in base_init().

11 years agoadder: Do not send flush_start event with the stream lock taken
Thibault Saunier [Thu, 4 Jul 2013 16:54:59 +0000 (12:54 -0400)]
adder: Do not send flush_start event with the stream lock taken

FLUSH_START is not serialized, so the lock should not be taken when
sending it.

11 years agotag: ignore malformed ID3v2 TDAT frames
Marcin Lewandowski [Thu, 4 Jul 2013 23:47:08 +0000 (00:47 +0100)]
tag: ignore malformed ID3v2 TDAT frames

Just skip them, don't cause criticals.

https://bugzilla.gnome.org/show_bug.cgi?id=703283

11 years agoaudioresample: make explicit that neon is disabled and why
Tim-Philipp Müller [Wed, 3 Jul 2013 08:44:32 +0000 (09:44 +0100)]
audioresample: make explicit that neon is disabled and why

https://bugzilla.gnome.org/show_bug.cgi?id=703477

11 years agoaudioresample: disable 16-bit integer NEON support
Carlos Rafael Giani [Tue, 2 Jul 2013 16:20:39 +0000 (18:20 +0200)]
audioresample: disable 16-bit integer NEON support

it seems to be broken (produces no audio), plus the performance gain
is small

Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
11 years agoplaybin: If we had a previous autoplugged sink, try to reuse it
Sebastian Dröge [Tue, 2 Jul 2013 12:25:28 +0000 (14:25 +0200)]
playbin: If we had a previous autoplugged sink, try to reuse it

https://bugzilla.gnome.org/show_bug.cgi?id=701997

11 years agoplaysink: If we switch sinks, make sure that the old sink is set to NULL
Sebastian Dröge [Tue, 2 Jul 2013 12:18:20 +0000 (14:18 +0200)]
playsink: If we switch sinks, make sure that the old sink is set to NULL

11 years agoplaybin: Don't change the state of sinks that we passed to playsink already
Sebastian Dröge [Tue, 2 Jul 2013 12:02:57 +0000 (14:02 +0200)]
playbin: Don't change the state of sinks that we passed to playsink already

11 years agoplaysink: Consider new audio/video sinks when reconfiguring
Sebastian Dröge [Tue, 2 Jul 2013 12:01:52 +0000 (14:01 +0200)]
playsink: Consider new audio/video sinks when reconfiguring

11 years agoplaybin: Improve debug output regarding sink selection
Sebastian Dröge [Tue, 2 Jul 2013 10:27:03 +0000 (12:27 +0200)]
playbin: Improve debug output regarding sink selection

11 years agoplaybin: Post an error message if a stream combiner doesn't return a request pad.
Brendan Long [Mon, 1 Jul 2013 18:52:43 +0000 (12:52 -0600)]
playbin: Post an error message if a stream combiner doesn't return a request pad.

11 years agoplaybin: Only intersect to check if a sink can handle raw caps
Sebastian Dröge [Mon, 1 Jul 2013 11:45:25 +0000 (13:45 +0200)]
playbin: Only intersect to check if a sink can handle raw caps

Doing a subset check requires fixed caps, which we might not have here.

https://bugs.webkit.org/show_bug.cgi?id=116042

11 years agopbutils: allow describing unfixed caps if they share the same media type
Vincent Penquerc'h [Mon, 1 Jul 2013 09:39:02 +0000 (10:39 +0100)]
pbutils: allow describing unfixed caps if they share the same media type

Caps description and missing plugin code does not really need caps to
be fixed, and indeed they may not be if giving encodebin unfixed caps
that correspond to an unknown encoder or muxer.

So we relax the check, and allow unfixed caps if all the structures
refer to the same media type.

11 years agovideodecoder: Send all pending events with type < CAPS before sending caps
Sebastian Dröge [Mon, 1 Jul 2013 09:16:34 +0000 (11:16 +0200)]
videodecoder: Send all pending events with type < CAPS before sending caps

11 years agovideoencoder: Send all pending events with type < CAPS before sending caps.
Mathieu Duponchelle [Thu, 27 Jun 2013 14:33:15 +0000 (16:33 +0200)]
videoencoder: Send all pending events with type < CAPS before sending caps.

https://bugzilla.gnome.org/show_bug.cgi?id=703196

11 years agotypefind: avoid too low mpeg/ts probability on small amount of data
Vincent Penquerc'h [Fri, 28 Jun 2013 13:48:19 +0000 (14:48 +0100)]
typefind: avoid too low mpeg/ts probability on small amount of data

With the current test, we get into problems when we try to typefind
a MPEG stream from a small amount of data, which can happen when
we get data pushed from a HTTP source. We thus make a second test
to give higher probability if all the potential headers were either
pack or pes headers (ie, no potential header was unrecognized).

This fixes an issue with a MPEG1/MP2 stream being properly discovered
as video/mpeg from a file, but as audio/mpeg from souphttpsrc.

https://bugzilla.gnome.org/show_bug.cgi?id=703256

11 years agovideo(enc|dec)oder: Don't return not-negotiated if flushing
Sebastian Dröge [Sun, 30 Jun 2013 16:17:15 +0000 (18:17 +0200)]
video(enc|dec)oder: Don't return not-negotiated if flushing

If the pad is flushing after a failed negotiation, return
GST_FLOW_FLUSHING instead from finish_frame().

https://bugzilla.gnome.org/show_bug.cgi?id=701763

11 years agoaudioencoder: Don't return not-negotiated if flushing
Sebastian Dröge [Sun, 30 Jun 2013 16:16:35 +0000 (18:16 +0200)]
audioencoder: Don't return not-negotiated if flushing

If the pad is flushing after a failed negotiation, return
GST_FLOW_FLUSHING instead from finish_frame().

https://bugzilla.gnome.org/show_bug.cgi?id=701763

11 years agopbutils: descriptions: Allow smart codec tag handling
Edward Hervey [Fri, 14 Jun 2013 05:23:40 +0000 (07:23 +0200)]
pbutils: descriptions: Allow smart codec tag handling

We already have internally the information on what type of stream (audio,
video, container, subtitle, ...) a certain caps is.
Instead of forcing callers to specify which CODEC_TAG category a certain
caps is, use that information to make a smart choice.

Does not break previous behaviour of gst_pb_utils_add_codec_description_to_tag_list
(if tag is specified it will be used, if caps is invalid it will be rejected,
...).

https://bugzilla.gnome.org/show_bug.cgi?id=702215

11 years agoxmptag: Add a debug category
Edward Hervey [Wed, 19 Jun 2013 07:25:48 +0000 (09:25 +0200)]
xmptag: Add a debug category

Instead of using the default category

11 years agovideotestsrc: do not leak lines
Patricia Muscalu [Thu, 27 Jun 2013 10:23:27 +0000 (12:23 +0200)]
videotestsrc: do not leak lines

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=703177

11 years agortpbasepayload: Do not leak the event when segment is delayed
Ognyan Tonchev [Wed, 26 Jun 2013 12:36:17 +0000 (14:36 +0200)]
rtpbasepayload: Do not leak the event when segment is delayed

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=703119

11 years agortsp: make read uncancelable when reading a message
Wim Taymans [Wed, 26 Jun 2013 13:03:05 +0000 (15:03 +0200)]
rtsp: make read uncancelable when reading a message

When we start to read a message, we need to continue reading until the end of
the message or else we lose track and cause parse errors. Use a variable
may_cancel to avoid cancelation after we read the first byte until we have
the complete message.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=703088

11 years agoaudiodecoder: Don't return not-negotiated if flushing
Mathieu Duponchelle [Fri, 21 Jun 2013 18:41:15 +0000 (20:41 +0200)]
audiodecoder: Don't return not-negotiated if flushing

If the pad is flushing after a failed negotiation, return GST_FLOW_FLUSHING.

https://bugzilla.gnome.org/show_bug.cgi?id=701763

11 years agoogg: The Daala headers are little endian, not big endian
Sebastian Dröge [Sun, 23 Jun 2013 10:07:41 +0000 (12:07 +0200)]
ogg: The Daala headers are little endian, not big endian

11 years agoogg: Add Daala support
Sebastian Dröge [Sun, 23 Jun 2013 08:30:02 +0000 (10:30 +0200)]
ogg: Add Daala support

11 years agopbutils: Add VP9 description
Sebastian Dröge [Fri, 21 Jun 2013 17:04:43 +0000 (19:04 +0200)]
pbutils: Add VP9 description

11 years agovideodecoder: Fix drop frame handling at startup
Edward Hervey [Mon, 17 Jun 2013 06:58:13 +0000 (08:58 +0200)]
videodecoder: Fix drop frame handling at startup

In the unlikely case that the decoder drops a frame before the first
input frame is outputted, use the input segment (since it wasn't
carried over to the output segment yet)

https://bugzilla.gnome.org/show_bug.cgi?id=702502

11 years agortsp: dispatch when initial buffer has data
Wim Taymans [Fri, 21 Jun 2013 09:50:33 +0000 (11:50 +0200)]
rtsp: dispatch when initial buffer has data

When we have data in the inital buffer, dispath the read function to read it
even if the socket has no data to read.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702652

11 years agortsp: manage writer child source better
Wim Taymans [Thu, 20 Jun 2013 15:28:46 +0000 (17:28 +0200)]
rtsp: manage writer child source better

Only add the write child source when we have something to write or else
we will dispatch forever without doing anything.

11 years agoaudioencoder: unref before memset
Jonas Holmberg [Wed, 19 Jun 2013 11:21:45 +0000 (13:21 +0200)]
audioencoder: unref before memset

Unref allocator and input_caps in encoder context before memsetting the
context.

11 years agoxmptag: More efficient GSList usage
Edward Hervey [Wed, 19 Jun 2013 07:22:50 +0000 (09:22 +0200)]
xmptag: More efficient GSList usage

Instead of constantly appending (which gets more and more expensive), just
prepend to the list (O(1)) and reverse the list before usage.

https://bugzilla.gnome.org/show_bug.cgi?id=702545

11 years agortpbuffer: add gst_rtp_buffer_get_payload_bytes
Branko Subasic [Sun, 16 Jun 2013 20:39:30 +0000 (22:39 +0200)]
rtpbuffer: add gst_rtp_buffer_get_payload_bytes

The function gst_rtp_buffer_get_payload can not be used in Python
because it lacks necessary length parameter. This patch adds a new
function, gst_rtp_buffer_get_payload_bytes, to use from Python
bindings. The new function has the advisory "Rename to:" annotation
so it can replace the gst_rtp_buffer_get_payload whan creating
bindings.

The function gst_rtp_buffer_get_extension_bytes is also added. It wraps
gst_rtp_buffer_get_extension_data which doesn't work in Python due to
incomplete annotation and because it returns the length as number of
32-bit words.

https://bugzilla.gnome.org/show_bug.cgi?id=698562

11 years agoaudiobasesrc: add 2 missing gst_buffer_unmap () calls
Ognyan Tonchev [Mon, 17 Jun 2013 14:34:26 +0000 (16:34 +0200)]
audiobasesrc: add 2 missing gst_buffer_unmap () calls

There are 2 missing calls to gst_buffer_unmap () in the error handling in
create ().

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702467

11 years agoplaysink: Fix the block diagram of deinterlace bin.
Sreerenj Balachandran [Mon, 17 Jun 2013 13:02:41 +0000 (16:02 +0300)]
playsink: Fix the block diagram of deinterlace bin.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=702465