platform/upstream/gstreamer.git
14 years agodecodebin2: Add simple unit test, mainly a copy of the decodebin unit test
Sebastian Dröge [Mon, 15 Feb 2010 00:17:49 +0000 (01:17 +0100)]
decodebin2: Add simple unit test, mainly a copy of the decodebin unit test

The only difference between the two unit tests right now is,
that the decodebin2 test resets the element to READY before trying
to reuse it instead of NULL. decodebin2 guarantees to be reusable
without going back to NULL.

14 years agoogg: theora PAR of 0:N, N:0 or 0:0 is allowed and maps to 1:1
Tim-Philipp Müller [Mon, 15 Feb 2010 00:11:17 +0000 (00:11 +0000)]
ogg: theora PAR of 0:N, N:0 or 0:0 is allowed and maps to 1:1

See #609252.

14 years agoAutomatic update of common submodule
Sebastian Dröge [Sun, 14 Feb 2010 22:16:32 +0000 (23:16 +0100)]
Automatic update of common submodule

From 96dc793 to 44ecce7

14 years agoplaybin2: Enable playbin2 unit test
Sebastian Dröge [Sun, 14 Feb 2010 22:10:06 +0000 (23:10 +0100)]
playbin2: Enable playbin2 unit test

It now contains a single working unit test and can be enabled.
The other more useful unit tests still need fixing.

14 years agoplaybin: Fix indention in the unit test
Sebastian Dröge [Sun, 14 Feb 2010 21:16:31 +0000 (22:16 +0100)]
playbin: Fix indention in the unit test

14 years agovolume: Replace this variables by self
Sebastian Dröge [Sat, 13 Feb 2010 00:08:05 +0000 (01:08 +0100)]
volume: Replace this variables by self

14 years agoplaysink: Reset the sink's state to NULL before unreffing it unless it's the same...
Josep Torra Valles [Fri, 12 Feb 2010 18:43:13 +0000 (19:43 +0100)]
playsink: Reset the sink's state to NULL before unreffing it unless it's the same instance again

This makes sure that we don't destroy the last reference before the
element gets back to NULL state. Fixes assertion failures if a playbin2
instance is reused but different sinks are automatically chosen because
of different caps.

14 years agoappsrc: fix Since tag
Wim Taymans [Fri, 12 Feb 2010 17:00:40 +0000 (18:00 +0100)]
appsrc: fix Since tag

14 years agoriff: treat JUNQ chunks like JUNK chunks
Tim-Philipp Müller [Fri, 12 Feb 2010 14:19:33 +0000 (14:19 +0000)]
riff: treat JUNQ chunks like JUNK chunks

14 years agoappsrc: Update basesrc segment duration and post duration messages from the streaming...
Sebastian Dröge [Fri, 12 Feb 2010 13:29:18 +0000 (14:29 +0100)]
appsrc: Update basesrc segment duration and post duration messages from the streaming thread

14 years agotags: improve docs about determining the encoding
Stefan Kost [Thu, 11 Feb 2010 12:10:02 +0000 (14:10 +0200)]
tags: improve docs about determining the encoding

14 years agocomment: fix wrong header comment
Stefan Kost [Thu, 11 Feb 2010 12:09:05 +0000 (14:09 +0200)]
comment: fix wrong header comment

14 years agoriff: add a variant of the JUNK tag that several adobe products produce
Stefan Kost [Mon, 1 Feb 2010 11:50:14 +0000 (13:50 +0200)]
riff: add a variant of the JUNK tag that several adobe products produce

JUNQ has same semantics as JUNK.

14 years agoappsrc: add min-percent property
Wim Taymans [Mon, 1 Feb 2010 18:01:33 +0000 (19:01 +0100)]
appsrc: add min-percent property

Emit need-data when the amount of data in the internal queue drops below
min-percent.

Fixes #608309

14 years agoappsrc: cleanups
Wim Taymans [Mon, 1 Feb 2010 17:56:34 +0000 (18:56 +0100)]
appsrc: cleanups

Avoid some typechecks.
Avoid dereferencing appsrc->priv all the time.

14 years agoappsink: cleanups
Wim Taymans [Mon, 1 Feb 2010 17:55:39 +0000 (18:55 +0100)]
appsink: cleanups

Avoid some typecasting.
Avoid dereferencing appsink->priv all the time.

14 years agouridecodebin: avoid some typecasts
Wim Taymans [Mon, 1 Feb 2010 14:09:56 +0000 (15:09 +0100)]
uridecodebin: avoid some typecasts

14 years agortsp: ignore \n and \r as the first line
Wim Taymans [Fri, 29 Jan 2010 15:34:44 +0000 (16:34 +0100)]
rtsp: ignore \n and \r as the first line

Be more forgiving for bad servers and ignore \r and \n when we are looking for
the response/request line.

See #608417

14 years agortsp: fail gracefully on bad Content-Length headers
Wim Taymans [Wed, 10 Feb 2010 15:05:29 +0000 (16:05 +0100)]
rtsp: fail gracefully on bad Content-Length headers

Be careful when allocating the amount of bytes specified in the Content-Length
because it can be an insanely huge value. Try to allocate the memory but fail
gracefully with a nice error when the allocation failed.

14 years agoffmpegcolorspace: Add conversions from all ARGB formats to AYUV and back
Sebastian Dröge [Wed, 10 Feb 2010 09:12:18 +0000 (10:12 +0100)]
ffmpegcolorspace: Add conversions from all ARGB formats to AYUV and back

14 years agoappsrc: Update segment duration and post a duration message if the duration changes
Sebastian Dröge [Tue, 9 Feb 2010 16:39:21 +0000 (17:39 +0100)]
appsrc: Update segment duration and post a duration message if the duration changes

Fixes bug #609423.

14 years agobuild: link to libm in examples that use it
Benjamin Otte [Thu, 11 Feb 2010 09:56:17 +0000 (10:56 +0100)]
build: link to libm in examples that use it

This fixes build failure in Fedora 13.

14 years agoUpdate MAINTAINERS, add myself
Tim-Philipp Müller [Thu, 11 Feb 2010 01:11:30 +0000 (01:11 +0000)]
Update MAINTAINERS, add myself

14 years agoconfigure: back to development
Tim-Philipp Müller [Thu, 11 Feb 2010 23:57:38 +0000 (23:57 +0000)]
configure: back to development

Slushy freeze remains in effect.

14 years agoRelease 0.10.26
Tim-Philipp Müller [Wed, 10 Feb 2010 20:17:36 +0000 (20:17 +0000)]
Release 0.10.26

14 years agoUpdate .po files
Tim-Philipp Müller [Wed, 10 Feb 2010 20:16:37 +0000 (20:16 +0000)]
Update .po files

14 years agotheoradec: PARs of 0:x, x:0 and 0:0 are all allowed and map to 1:1
Benjamin M. Schwartz [Mon, 8 Feb 2010 10:21:35 +0000 (11:21 +0100)]
theoradec: PARs of 0:x, x:0 and 0:0 are all allowed and map to 1:1

Fixes #609252.

14 years agooggdemux: use the default granpos functions for kate streams
Vincent Penquerc'h [Sun, 24 Jan 2010 12:31:04 +0000 (12:31 +0000)]
oggdemux: use the default granpos functions for kate streams

Set timestamps on kate packets. See bug #600929.

14 years ago0.10.25.3 pre-release
Tim-Philipp Müller [Fri, 5 Feb 2010 01:18:43 +0000 (01:18 +0000)]
0.10.25.3 pre-release

14 years agopo: update translations
Tim-Philipp Müller [Thu, 4 Feb 2010 18:52:59 +0000 (18:52 +0000)]
po: update translations

14 years agoRevert "playbin2: Only allow to set the URIs in states <= READY or from an about...
Tim-Philipp Müller [Thu, 4 Feb 2010 18:32:48 +0000 (18:32 +0000)]
Revert "playbin2: Only allow to set the URIs in states <= READY or from an about-to-finish signal handler"

This reverts commit 7335ce5d3e03c126a417a721571cb6f3af136ecf.

Support abusing the uri property to configure the next uri to play
outside of the about-to-finish handler for the time being after all.
We also shouldn't use thread private structures for this, since it
should be possible to block the thread that emitted about-to-finish
while the main thread sets the uri property. See #607226.

14 years agooggdemux: Don't leak allocated buffers
Sebastian Dröge [Tue, 2 Feb 2010 09:18:05 +0000 (10:18 +0100)]
oggdemux: Don't leak allocated buffers

This can happen if the combined flow return is not OK although the
allocation succeeded or if the packet in question is a BOS and we're
not going to push headers.

Fixes bug #608699.

14 years agouridecodebin: clean up decodebin properties
Wim Taymans [Mon, 1 Feb 2010 10:44:34 +0000 (11:44 +0100)]
uridecodebin: clean up decodebin properties

When reusing a decodebin2 element, clear the properties we might have changed,
to their default values or else we might end up with old configuration.

Fixes #608484

14 years agoplaybin2: when no uri is set, post an error message
Tim-Philipp Müller [Fri, 29 Jan 2010 13:56:05 +0000 (13:56 +0000)]
playbin2: when no uri is set, post an error message

When no uri is set, don't just return STATE_CHANGE_FAILURE from the
state change function, but actually post an error message.

14 years agoAutomatic update of common submodule
Tim-Philipp Müller [Sat, 30 Jan 2010 15:18:13 +0000 (15:18 +0000)]
Automatic update of common submodule

From 15d47a6 to 96dc793

14 years agoadder: don't hold object lock when calling peer elements
Wim Taymans [Thu, 28 Jan 2010 16:12:34 +0000 (17:12 +0100)]
adder: don't hold object lock when calling peer elements

Do not hold the object lock while we call methods on peer elements as this can
lead to deadlocks.

Fixes #608179

14 years ago0.10.25.2 pre-release
Tim-Philipp Müller [Wed, 27 Jan 2010 01:12:49 +0000 (01:12 +0000)]
0.10.25.2 pre-release

14 years agowin32: update generated files for non-autotools win32 builds
Tim-Philipp Müller [Wed, 27 Jan 2010 01:07:55 +0000 (01:07 +0000)]
win32: update generated files for non-autotools win32 builds

14 years agopo: update translation files
Tim-Philipp Müller [Wed, 27 Jan 2010 00:56:00 +0000 (00:56 +0000)]
po: update translation files

14 years agoaudiosrc: add gratuitious FIXME for use of generic G_TYPE_POINTER type
Tim-Philipp Müller [Wed, 27 Jan 2010 00:41:24 +0000 (00:41 +0000)]
audiosrc: add gratuitious FIXME for use of generic G_TYPE_POINTER type

14 years agodecodebin2: Don't skip an element when getting the topology
Edward Hervey [Tue, 26 Jan 2010 15:47:40 +0000 (16:47 +0100)]
decodebin2: Don't skip an element when getting the topology

Fixes #608167

14 years agooggdemux: sparse streams aren't timed by end time, and their duration isn't implicit
Vincent Penquerc'h [Sun, 24 Jan 2010 14:41:44 +0000 (14:41 +0000)]
oggdemux: sparse streams aren't timed by end time, and their duration isn't implicit

Fixes timestamps and durations on Kate subtitle streams.

See http://www.xiph.org/ogg/doc/ogg-multiplex.html section 'start-time and
end-time positioning' for some more details, and bug #600929.

14 years agooggdemux: properly set up the media type for kate streams
Vincent Penquerc'h [Sat, 23 Jan 2010 20:15:08 +0000 (20:15 +0000)]
oggdemux: properly set up the media type for kate streams

See #600929.

14 years agosubtitleoverlay: relax caps template on sink pads
Julien Moutte [Mon, 25 Jan 2010 17:57:52 +0000 (18:57 +0100)]
subtitleoverlay: relax caps template on sink pads

Allow any caps on sink pad templates as we could do passthrough with non raw
video caps.

14 years agooggdemux: use right type for the serialno
Wim Taymans [Mon, 25 Jan 2010 14:14:56 +0000 (15:14 +0100)]
oggdemux: use right type for the serialno

Use a consistent type for the serialno to avoid problems when comparing between
signed and unsigned variants.

Fixes #607926

14 years agooggdemux: don't push headers twice
Wim Taymans [Mon, 25 Jan 2010 13:00:52 +0000 (14:00 +0100)]
oggdemux: don't push headers twice

Don't push the stream headers twice but only in the activation of a chain.

Fixes #607929

14 years agoMerge branch 'master' of ssh://git.freedesktop.org/git/gstreamer/gst-plugins-base
Wim Taymans [Mon, 25 Jan 2010 12:18:18 +0000 (13:18 +0100)]
Merge branch 'master' of ssh://git.freedesktop.org/git/gstreamer/gst-plugins-base

14 years agooggdemux: rename a variable
Wim Taymans [Mon, 25 Jan 2010 11:31:24 +0000 (12:31 +0100)]
oggdemux: rename a variable

Rename the 'seekable' variable to 'pullmode'. We might be able to seek in push
mode too eventually.

14 years agoRevert "inputselector: Protect g_object_notify() with the object's mutex"
Sebastian Dröge [Mon, 25 Jan 2010 11:22:17 +0000 (12:22 +0100)]
Revert "inputselector: Protect g_object_notify() with the object's mutex"

This reverts commit a37426c41c80fd21e5017fea01a786c05bcd9661, it's
causing deadlocks with playbin2.

14 years agoinputselector: Protect g_object_notify() with the object's mutex
Kipp Cannon [Sun, 24 Jan 2010 19:55:26 +0000 (20:55 +0100)]
inputselector: Protect g_object_notify() with the object's mutex

This works around the thread unsafety of g_object_notify()

Fixes bug #607513.

14 years agotypefindfunctions: Add typefinder for ISO MP4 files
Sebastian Dröge [Sun, 24 Jan 2010 19:46:58 +0000 (20:46 +0100)]
typefindfunctions: Add typefinder for ISO MP4 files

Fixes bug #607848.

14 years agooggdemux: fix crash when freeing headers
Tim-Philipp Müller [Sun, 24 Jan 2010 13:29:07 +0000 (13:29 +0000)]
oggdemux: fix crash when freeing headers

Use _ogg_packet_free() instead of gst_mini_object_unref in one more
place now that the header list contains ogg packets and not buffers.

file: Stephen_Fry-Happy_Birthday_GNU-nq_600px_425kbit.ogv

14 years agooggdemux: Strip trailing \0 for subtitle OGM streams
Sebastian Dröge [Sun, 24 Jan 2010 07:57:13 +0000 (08:57 +0100)]
oggdemux: Strip trailing \0 for subtitle OGM streams

Fixes bug #607870.

14 years agooggdemux: Correctly set DELTA_UNIT flag for OGM streams
Sebastian Dröge [Sat, 23 Jan 2010 21:09:45 +0000 (22:09 +0100)]
oggdemux: Correctly set DELTA_UNIT flag for OGM streams

14 years agooggdemux: Don't strip all 0-bytes from the end of OGM packets
Sebastian Dröge [Sat, 23 Jan 2010 21:05:40 +0000 (22:05 +0100)]
oggdemux: Don't strip all 0-bytes from the end of OGM packets

This fixes broken packets pushed downstream by oggdemux for
MPEG4 streams for example.

14 years agooggdemux: Extract tags from OGM text streams and don't push them downstream
Sebastian Dröge [Sat, 23 Jan 2010 21:03:18 +0000 (22:03 +0100)]
oggdemux: Extract tags from OGM text streams and don't push them downstream

14 years agooggdemux: Store header/queued packets as ogg_packet and use normal peer chaining...
Sebastian Dröge [Sat, 23 Jan 2010 13:46:19 +0000 (14:46 +0100)]
oggdemux: Store header/queued packets as ogg_packet and use normal peer chaining functions to pass them downstream

14 years agotypefinding: optimise AC-3 typefinder a bit
Tim-Philipp Müller [Sat, 23 Jan 2010 15:25:42 +0000 (15:25 +0000)]
typefinding: optimise AC-3 typefinder a bit

Make AC-3 typefinder use the DataScanCtx stuff so we don't have to
do gst_type_find_peek() in the inner loop all the time. Also return
when we've suggested AC3 caps, instead of continuing with the loop.

14 years agoRevert "typefind: Reduce number of calls to gst_type_find_peek."
Tim-Philipp Müller [Sat, 23 Jan 2010 14:31:15 +0000 (14:31 +0000)]
Revert "typefind: Reduce number of calls to gst_type_find_peek."

This reverts commit c661bfaa991c58f1fbd9fbc0dae90b8b2c27f92b.

This breaks AC-3 typefinding for all cases where the first frame
is at an offset > 0.

14 years agopbutils: Add description for Zip Block Motion Video
Edward Hervey [Sat, 23 Jan 2010 14:35:05 +0000 (15:35 +0100)]
pbutils: Add description for Zip Block Motion Video

14 years agoriff: Add mapping for Zip Block Motion Video
Edward Hervey [Sat, 23 Jan 2010 14:34:54 +0000 (15:34 +0100)]
riff: Add mapping for Zip Block Motion Video

14 years agoriff: YUNV is a fourcc which is also used for YUY2 raw video
Edward Hervey [Sat, 23 Jan 2010 14:26:37 +0000 (15:26 +0100)]
riff: YUNV is a fourcc which is also used for YUY2 raw video

14 years agoriff: vp61 and VP61 are also valid On2 VP6 fourcc
Edward Hervey [Sat, 23 Jan 2010 14:13:45 +0000 (15:13 +0100)]
riff: vp61 and VP61 are also valid On2 VP6 fourcc

14 years agoriff: Add mapping for On2 VP5
Edward Hervey [Sat, 23 Jan 2010 14:10:45 +0000 (15:10 +0100)]
riff: Add mapping for On2 VP5

14 years agoriff: Add mapping for Sigma-Designs MPEG4
Edward Hervey [Sat, 23 Jan 2010 14:04:35 +0000 (15:04 +0100)]
riff: Add mapping for Sigma-Designs MPEG4

It's actually a xvid-compatible stream. both xviddec and ffmpeg handle it.

14 years agopbutils: Add description for LOCO Lossless codec
Edward Hervey [Sat, 23 Jan 2010 13:35:28 +0000 (14:35 +0100)]
pbutils: Add description for LOCO Lossless codec

14 years agoriff: Add mapping for LOCO Lossless codec
Edward Hervey [Sat, 23 Jan 2010 13:35:16 +0000 (14:35 +0100)]
riff: Add mapping for LOCO Lossless codec

14 years agoriff: Add support for YV12 / Uncompressed packed YVU 4:2:2
Edward Hervey [Sat, 23 Jan 2010 13:08:39 +0000 (14:08 +0100)]
riff: Add support for YV12 / Uncompressed packed YVU 4:2:2

14 years agopbutils: add description for Autodesk Animator codec
Edward Hervey [Sat, 23 Jan 2010 12:50:26 +0000 (13:50 +0100)]
pbutils: add description for Autodesk Animator codec

14 years agoriff: Add mapping for Autodesk Animator Codec
Edward Hervey [Sat, 23 Jan 2010 12:50:09 +0000 (13:50 +0100)]
riff: Add mapping for Autodesk Animator Codec

14 years agooggdemux: ...and set caps on queued packet buffers too
Sebastian Dröge [Sat, 23 Jan 2010 12:20:46 +0000 (13:20 +0100)]
oggdemux: ...and set caps on queued packet buffers too

14 years agooggdemux: Set caps on header buffers
Sebastian Dröge [Sat, 23 Jan 2010 12:19:08 +0000 (13:19 +0100)]
oggdemux: Set caps on header buffers

14 years agouridecodebin: handle raw sources about-to-finish signals
Wim Taymans [Fri, 22 Jan 2010 15:23:43 +0000 (16:23 +0100)]
uridecodebin: handle raw sources about-to-finish signals

When we are dealing with a source that produces raw audio/video, we don't use a
decodebin2 to decode the data and we thus don't have the drained/about-to-finish
signal emited. To fix this, we add a padprobe on the source pads and emit the
drained signal ourselves. This then makes playbin2 emit the about-to-finish
signal for raw sources such as cdda://

Fixes #607116

14 years agotypefind: include stdio.h for sscanf
Stefan Kost [Fri, 22 Jan 2010 14:15:54 +0000 (16:15 +0200)]
typefind: include stdio.h for sscanf

14 years agotypefinding: add PNM typefinder
Tim-Philipp Müller [Fri, 22 Jan 2010 01:49:38 +0000 (01:49 +0000)]
typefinding: add PNM typefinder

Add PNM typefinder, so we can remove the one that's in the PNM plugin
in -bad (which btw uses different/wrong media types that don't match
the ones used by gdkpixbufdec) and people don't make fun of us for
loading image decoders when typefinding and playing back audio files.

14 years agoffmpegcolorspace: rename performance category
Thijs Vermeir [Thu, 21 Jan 2010 18:31:23 +0000 (19:31 +0100)]
ffmpegcolorspace: rename performance category

rename the performance category to ffmpegcolorspace_performance
as there is already a global GST_CAT_PERFORMANCE in core

14 years agooggdemux: keep track of added pads
Wim Taymans [Thu, 21 Jan 2010 16:32:33 +0000 (17:32 +0100)]
oggdemux: keep track of added pads

Keep track of the pads we added and removed.
Remove some unused fields.
Don't add pads for which we don't have caps.

14 years agooggstream: don't call NULL setup functions
Wim Taymans [Thu, 21 Jan 2010 16:31:13 +0000 (17:31 +0100)]
oggstream: don't call NULL setup functions

If we find a known mapper but it doesn't have a setup function, simply skip it
instead of crashing.

14 years agooggstream: avoid division by 0 on bad annodex streams
Wim Taymans [Thu, 21 Jan 2010 16:30:15 +0000 (17:30 +0100)]
oggstream: avoid division by 0 on bad annodex streams

14 years agopbutils: Add description for y4m container
Edward Hervey [Thu, 21 Jan 2010 12:47:01 +0000 (13:47 +0100)]
pbutils: Add description for y4m container

14 years agobasertppayload: ptime/maxptime should be unsigned
Olivier Crête [Tue, 19 Jan 2010 19:31:34 +0000 (14:31 -0500)]
basertppayload: ptime/maxptime should be unsigned

https://bugzilla.gnome.org/show_bug.cgi?id=607403

14 years agobasertppayload: ptime should be in nanoseconds
Olivier Crête [Tue, 19 Jan 2010 02:16:32 +0000 (21:16 -0500)]
basertppayload: ptime should be in nanoseconds

https://bugzilla.gnome.org/show_bug.cgi?id=607403

14 years agoAutomatic update of common submodule
Tim-Philipp Müller [Wed, 20 Jan 2010 00:53:20 +0000 (00:53 +0000)]
Automatic update of common submodule

From 14cec89 to 15d47a6

14 years agotypefind: rewrite h.264 detection
David Schleef [Tue, 19 Jan 2010 21:33:06 +0000 (13:33 -0800)]
typefind: rewrite h.264 detection

Make detection simpler: check for NALs, check that they make
sense, and report how certain we are that it's a raw H.264 stream.
Fixes: #583376.

14 years agobasertppayload: Reject empty caps
Olivier Crête [Mon, 18 Jan 2010 19:33:30 +0000 (14:33 -0500)]
basertppayload: Reject empty caps

https://bugzilla.gnome.org/show_bug.cgi?id=607353

14 years agooggdemux: No need to subtract begin time
Thiago Santos [Tue, 19 Jan 2010 11:39:14 +0000 (08:39 -0300)]
oggdemux: No need to subtract begin time

Last stop is already based on the chain start and there is no need
to subtract the chain start as it may lead to a negative overflow.
This was causing seeking issues when the target chain was not
the first one (that has chain start = 0)

Fixes #606382

14 years agoaudio: Use rounding scaling functions for GST_CLOCK_TIME_TO_FRAMES and _FRAMES_TO_CLO...
Sebastian Dröge [Tue, 19 Jan 2010 08:25:35 +0000 (09:25 +0100)]
audio: Use rounding scaling functions for GST_CLOCK_TIME_TO_FRAMES and _FRAMES_TO_CLOCK_TIME

Fixes bug #607381.

14 years agooggdemux: granulepos is relative to its chain
Thiago Santos [Mon, 18 Jan 2010 18:22:52 +0000 (15:22 -0300)]
oggdemux: granulepos is relative to its chain

When performing seeks, the granulepos should be offset by
its chain start time to avoid using wrong values to
update segment's last_stop. A sample file is indicated on
bug #606382

14 years agopbutils: Add description for MXF container format
Edward Hervey [Mon, 18 Jan 2010 16:57:16 +0000 (17:57 +0100)]
pbutils: Add description for MXF container format

14 years agoplaysink: re-use iterator callback to avoid code duplication
Tim-Philipp Müller [Mon, 18 Jan 2010 10:07:30 +0000 (10:07 +0000)]
playsink: re-use iterator callback to avoid code duplication

14 years agoplaysink: when looking for sink properties, make sure they have the right type
Tim-Philipp Müller [Mon, 18 Jan 2010 02:08:39 +0000 (02:08 +0000)]
playsink: when looking for sink properties, make sure they have the right type

We don't want to end up setting values on elements where the property is of
a different type than we expect. Can't transform the value either, since we
can't really make assumptions about the scale and transform function.

Fixes crashes when using playbin2 with apexsink (#606949).

14 years agoplaybin2: Only allow to set the URIs in states <= READY or from an about-to-finish...
Sebastian Dröge [Mon, 18 Jan 2010 08:30:18 +0000 (09:30 +0100)]
playbin2: Only allow to set the URIs in states <= READY or from an about-to-finish signal handler

Changing the URIs in a state > READY results in unexpected behaviour,
i.e. the new URIs are only used after the current track has finished.

Fixes bug #607226.

14 years agodecodebin2: sprinkle some more locking
Mark Nauwelaerts [Fri, 15 Jan 2010 18:52:29 +0000 (19:52 +0100)]
decodebin2: sprinkle some more locking

... to avoid races and ensure some data structure consistency.

See also #574289.

14 years agodecodebin2: mind blocked pads when shutting down
Mark Nauwelaerts [Thu, 14 Jan 2010 17:26:03 +0000 (18:26 +0100)]
decodebin2: mind blocked pads when shutting down

Fix regression in shutdown deadlock handling now that the
target of a ghostpad is blocked instead of ghostpad itself.

See also #574293.

14 years agoplaysink: Fix disabling of subtitles if subtitles were used before
Sebastian Dröge [Thu, 14 Jan 2010 12:36:23 +0000 (13:36 +0100)]
playsink: Fix disabling of subtitles if subtitles were used before

In this case the video still goes through the text chain and
subtitles are still going in there, in case subtitles are
enabled again. This makes sure that re-enabling subtitles
happens instantly.

Fixes hanging video when disabling subtitles, caused by an
unliked video pad.

14 years agoplaybin2: fix pad ref leak
Mark Nauwelaerts [Thu, 14 Jan 2010 09:43:59 +0000 (10:43 +0100)]
playbin2: fix pad ref leak

14 years agodocs: fix out-of-source build
Mark Nauwelaerts [Tue, 12 Jan 2010 20:42:59 +0000 (21:42 +0100)]
docs: fix out-of-source build

14 years agostress-playbin: fix error return check
Mark Nauwelaerts [Wed, 29 Apr 2009 09:50:03 +0000 (11:50 +0200)]
stress-playbin: fix error return check

14 years agotheora: Rename source files to have the same name as the headers
Sebastian Dröge [Thu, 14 Jan 2010 09:10:23 +0000 (10:10 +0100)]
theora: Rename source files to have the same name as the headers

14 years agovorbis: Rename source files to have the same name as the headers
Sebastian Dröge [Thu, 14 Jan 2010 09:07:22 +0000 (10:07 +0100)]
vorbis: Rename source files to have the same name as the headers