platform/upstream/gst-plugins-good.git
11 years agov4l2: fix and cleanup VIDIOC_EXPBUF handling
Michael Olbrich [Tue, 30 Apr 2013 11:07:37 +0000 (13:07 +0200)]
v4l2: fix and cleanup VIDIOC_EXPBUF handling

clear the struct, and provide a correct error message

https://bugzilla.gnome.org/show_bug.cgi?id=699337

11 years agov4l2: handle return value -ENOTTY for unimplemented VIDIOC_G_PARM
Philipp Zabel [Thu, 5 Jul 2012 16:02:27 +0000 (18:02 +0200)]
v4l2: handle return value -ENOTTY for unimplemented VIDIOC_G_PARM

Newer kernels return -ENOTTY, older kernels return -EINVAL if the ioctl
is not implemented. With this patch, GStreamer handles both cases.

https://bugzilla.gnome.org/show_bug.cgi?id=698825

11 years agov4l2: fix broken boolean expression to detect non-frame buffers
Michael Olbrich [Tue, 30 Apr 2013 07:16:07 +0000 (09:16 +0200)]
v4l2: fix broken boolean expression to detect non-frame buffers

https://bugzilla.gnome.org/show_bug.cgi?id=699294

11 years agopulsesink: Better error message when server version is too old
Arun Raghavan [Mon, 29 Apr 2013 05:37:56 +0000 (11:07 +0530)]
pulsesink: Better error message when server version is too old

We check for the library version at configure time, but the server
version can only really be checked at run-time.

https://bugzilla.gnome.org/show_bug.cgi?id=698768

11 years agoudp: log WARNING debug message if UDP multicast is likely to be broken
Tim-Philipp Müller [Sat, 27 Apr 2013 10:24:38 +0000 (11:24 +0100)]
udp: log WARNING debug message if UDP multicast is likely to be broken

11 years agoudpsrc: add includes to get socklen_t defined on Windows
Tim-Philipp Müller [Sat, 27 Apr 2013 10:16:54 +0000 (11:16 +0100)]
udpsrc: add includes to get socklen_t defined on Windows

https://bugzilla.gnome.org/show_bug.cgi?id=692400

11 years agoqtdemux: add support for VP6F VP6 flash codec
Yury Delendik [Sat, 27 Apr 2013 08:39:45 +0000 (09:39 +0100)]
qtdemux: add support for VP6F VP6 flash codec

https://bugzilla.gnome.org/show_bug.cgi?id=699010

11 years agov4l2: also poll for output devices
Michael Olbrich [Wed, 5 Sep 2012 14:39:31 +0000 (16:39 +0200)]
v4l2: also poll for output devices

Note that the V4L2 API defines that for output devices POLLOUT
indicates that a buffer is ready to be dequeued.

https://bugzilla.gnome.org/show_bug.cgi?id=698992

11 years agov4l2: fix copying of encoded buffers
Philipp Zabel [Mon, 20 Aug 2012 07:52:34 +0000 (09:52 +0200)]
v4l2: fix copying of encoded buffers

The existence of a GstVideoFormatInfo does not guarantee, that
the buffer contains video frames, so the format must be checked.

Also, for encoded buffers the length is variable and must be set.

https://bugzilla.gnome.org/show_bug.cgi?id=698949

11 years agov4l2: add support for mpeg4 and H.263
Michael Olbrich [Tue, 10 Jul 2012 13:29:40 +0000 (15:29 +0200)]
v4l2: add support for mpeg4 and H.263

https://bugzilla.gnome.org/show_bug.cgi?id=698826

11 years agomonoscope: Fix debug statement
Edward Hervey [Fri, 26 Apr 2013 10:16:49 +0000 (12:16 +0200)]
monoscope: Fix debug statement

11 years agomulawdec: change base class to GstAudioDecoder
Alexander Schrab [Thu, 25 Apr 2013 19:50:33 +0000 (21:50 +0200)]
mulawdec: change base class to GstAudioDecoder

https://bugzilla.gnome.org/show_bug.cgi?id=698894

11 years agovideomixer: send stream-start event.
Mathieu Duponchelle [Thu, 25 Apr 2013 18:59:52 +0000 (20:59 +0200)]
videomixer: send stream-start event.

11 years agov4l2: handle ENODATA return value for VIDIOC_ENUMSTD
Philipp Zabel [Thu, 18 Oct 2012 08:37:35 +0000 (10:37 +0200)]
v4l2: handle ENODATA return value for VIDIOC_ENUMSTD

In kernel v3.7-rc1, VIDIOC_ENUMSTD returns ENODATA if the current input
does not support the STD API.

https://bugzilla.gnome.org/show_bug.cgi?id=698827

11 years agodocs: add some pay/depayloaders
Wim Taymans [Thu, 25 Apr 2013 11:19:35 +0000 (13:19 +0200)]
docs: add some pay/depayloaders

See https://bugzilla.gnome.org/show_bug.cgi?id=551631

11 years agomulaw: Some minor memleak fixes and cleanup
Sebastian Dröge [Thu, 25 Apr 2013 10:44:15 +0000 (12:44 +0200)]
mulaw: Some minor memleak fixes and cleanup

11 years agomulawenc: change to gstaudioencoder base, added bitrate tags
Alexander Schrab [Wed, 24 Apr 2013 11:56:56 +0000 (13:56 +0200)]
mulawenc: change to gstaudioencoder base, added bitrate tags

11 years agov4l2: bufferpool: reset buffer size in release_buffer
Michael Olbrich [Thu, 3 May 2012 14:07:27 +0000 (16:07 +0200)]
v4l2: bufferpool: reset buffer size in release_buffer

The buffer might still be in use elsewhere when dequeuing buffers for
outputs.

https://bugzilla.gnome.org/show_bug.cgi?id=698822

11 years agov4l2: bufferpool: remove unused includes
Michael Olbrich [Fri, 20 Apr 2012 07:53:35 +0000 (09:53 +0200)]
v4l2: bufferpool: remove unused includes

The hacks that needed these are long gone.

https://bugzilla.gnome.org/show_bug.cgi?id=698821

11 years ago(multi)udpsink: Use separate sockets for IPv4 and IPv6
Sebastian Dröge [Thu, 25 Apr 2013 10:12:23 +0000 (12:12 +0200)]
(multi)udpsink: Use separate sockets for IPv4 and IPv6

https://bugzilla.gnome.org/show_bug.cgi?id=534243

11 years agodynudpsink: Use separate sockets for IPv4 and IPv6
Sebastian Dröge [Thu, 25 Apr 2013 08:44:44 +0000 (10:44 +0200)]
dynudpsink: Use separate sockets for IPv4 and IPv6

https://bugzilla.gnome.org/show_bug.cgi?id=534243

11 years agoudp: Don't include removed gstudp.h in noinst_HEADERS
Sebastian Dröge [Thu, 25 Apr 2013 08:43:56 +0000 (10:43 +0200)]
udp: Don't include removed gstudp.h in noinst_HEADERS

11 years agoosxaudio: Use gst_audio_channel_positions_to_mask() to create mask
Todd Agulnick [Wed, 17 Apr 2013 23:47:31 +0000 (16:47 -0700)]
osxaudio: Use gst_audio_channel_positions_to_mask() to create mask

https://bugzilla.gnome.org/show_bug.cgi?id=698807

11 years agoosxaudio: Remove unused code
Todd Agulnick [Wed, 17 Apr 2013 23:12:26 +0000 (16:12 -0700)]
osxaudio: Remove unused code

11 years agoudp: Remove unused enum type
Sebastian Dröge [Thu, 25 Apr 2013 07:16:14 +0000 (09:16 +0200)]
udp: Remove unused enum type

11 years agoudp: Use the generic marshaller instead of generating marshallers
Sebastian Dröge [Thu, 25 Apr 2013 07:13:51 +0000 (09:13 +0200)]
udp: Use the generic marshaller instead of generating marshallers

11 years agoudpsrc: Rename instance variable from host to multi_group
Sebastian Dröge [Thu, 25 Apr 2013 07:07:41 +0000 (09:07 +0200)]
udpsrc: Rename instance variable from host to multi_group

This is more consistent as it's used for the multicast-group property.

11 years agoudpsrc: Add bind-address property
Sebastian Dröge [Thu, 25 Apr 2013 07:03:56 +0000 (09:03 +0200)]
udpsrc: Add bind-address property

This is equivalent to multicast-group currently for backwards compatibility.
In 2.0 this should be handled separately, the former only being the multicast
group and the latter always being the address the socket is bound to, even if
a multicast group is given.

11 years agovrawdepay: return output buffer from process
Wim Taymans [Wed, 24 Apr 2013 14:24:25 +0000 (16:24 +0200)]
vrawdepay: return output buffer from process

Return the output buffer from the process function instead of pushing
it ourselves. This way, the subclass can actually deal with the return
value of the push.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=693727

11 years agov4l2: save the format correctly
Diogo Carbonera Luvizon [Mon, 1 Oct 2012 12:29:21 +0000 (09:29 -0300)]
v4l2: save the format correctly

If TRY_FMT is not implemented,  gst_v4l2_object_get_nearest_size will
use S_FMT and will change the device's operation mode. To save the
old device mode we need to set the type field or else it will fail
to save the previous format.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=685209

11 years agortp: a marker bit should translate to RESYNC
Wim Taymans [Wed, 24 Apr 2013 13:38:50 +0000 (15:38 +0200)]
rtp: a marker bit should translate to RESYNC

A marker bit on an audio packet does not mean a DISCONT (in the GStreamer sense
of missing data) but it means that the packet is the end of a talkspurt and thus
a good opportunity to resync to the clock. Use the RESYNC buffer flag to note
this.
Real discontinuities are marked with DISCONT still when the seqnum has a GAP or
when the input buffer has the DISCONT flag set.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=627204

11 years agoAutomatic update of common submodule
Tim-Philipp Müller [Mon, 22 Apr 2013 22:51:38 +0000 (23:51 +0100)]
Automatic update of common submodule

From 3cb3d3c to 5edcd85

11 years agortpjpegdepay: Drop frame if it's less than 2 bytes large
Sebastian Dröge [Mon, 22 Apr 2013 08:19:29 +0000 (10:19 +0200)]
rtpjpegdepay: Drop frame if it's less than 2 bytes large

https://bugzilla.gnome.org/show_bug.cgi?id=677560

11 years agoautodetect: use _plugin_feature_rank_compare API instead of duplicating the code.
Sreerenj Balachandran [Thu, 18 Apr 2013 09:20:08 +0000 (12:20 +0300)]
autodetect: use _plugin_feature_rank_compare API instead of duplicating the code.

11 years agoosxaudio: Include gstaudioringbuffer.h to fix compilation in 1.0
Sebastian Dröge [Thu, 18 Apr 2013 07:37:30 +0000 (09:37 +0200)]
osxaudio: Include gstaudioringbuffer.h to fix compilation in 1.0

11 years agoosxaudiosink: channel-mask configuration fixes
Philippe Normand [Wed, 17 Apr 2013 19:05:14 +0000 (21:05 +0200)]
osxaudiosink: channel-mask configuration fixes

Set channel-mask according to sink's layout in case of stereo layout.
Also initialize and reset the mask when an unrecognized channel is detected.

https://bugzilla.gnome.org/show_bug.cgi?id=698224

11 years agov4l2src: Disable renegotiation in the negotiate method
Olivier Crête [Mon, 15 Apr 2013 23:53:28 +0000 (19:53 -0400)]
v4l2src: Disable renegotiation in the negotiate method

This way, we don't block the initial negotiation.

Thanks to Jeremy Whiting for doing all the testing.

https://bugzilla.gnome.org/show_bug.cgi?id=695981

11 years agoRevert "v4l2: disable renegotiation"
Olivier Crête [Mon, 15 Apr 2013 23:46:12 +0000 (19:46 -0400)]
Revert "v4l2: disable renegotiation"

This reverts commit d1b26e1d594ab2b63324e43a36330475e98cdf18.

This causes the initial negotiation to never happen if a reconfigure
event is received after gst_base_src_start_complete() but before the loop
starts.

https://bugzilla.gnome.org/show_bug.cgi?id=695981

11 years agoflactag: forward caps event
Stefan Sauer [Wed, 17 Apr 2013 19:12:55 +0000 (21:12 +0200)]
flactag: forward caps event

This ensures that the downstream element will get the event and negotiates. Add
a FIXME for updating the streamheader field on th caps.

11 years agoflac: add more logging
Stefan Sauer [Wed, 17 Apr 2013 05:50:27 +0000 (07:50 +0200)]
flac: add more logging

11 years agoosxaudio: Fix merge conflicts
Sebastian Dröge [Wed, 17 Apr 2013 18:24:48 +0000 (20:24 +0200)]
osxaudio: Fix merge conflicts

11 years agoosxaudio: Fix configure check for osxaudio plugin
Sebastian Dröge [Wed, 17 Apr 2013 08:10:46 +0000 (10:10 +0200)]
osxaudio: Fix configure check for osxaudio plugin

11 years agoosxaudioringbuffer: First check the type, then cast
Sebastian Dröge [Wed, 17 Apr 2013 07:50:43 +0000 (09:50 +0200)]
osxaudioringbuffer: First check the type, then cast

11 years agoosxaudio: use GST_IS_OSX_AUDIO_SINK in ring buffer.
Takashi Nakajima [Tue, 16 Apr 2013 13:46:00 +0000 (22:46 +0900)]
osxaudio: use GST_IS_OSX_AUDIO_SINK in ring buffer.

11 years agoosxaudio: call set_channel_positions() in osxaudioringbuffer acquire()
Takashi Nakajima [Wed, 10 Apr 2013 12:06:16 +0000 (21:06 +0900)]
osxaudio: call set_channel_positions() in osxaudioringbuffer acquire()

11 years agoosxaudio: use GST_AUDIO_INFO_* accessors
Todd Agulnick [Fri, 12 Apr 2013 19:18:04 +0000 (12:18 -0700)]
osxaudio: use GST_AUDIO_INFO_* accessors

Changes include the following:

 * Update classname references
 * Replace GST_BOILERPLATE_FULL with G_DEFINE_TYPE
 * Use new GstAudioInfo struct and methods
 * Use new buffer memory allocation scheme

Conflicts:
sys/osxaudio/gstosxaudioringbuffer.c

11 years agoosxaudio: adjust for changes to glib mutex api.
Todd Agulnick [Fri, 12 Apr 2013 18:51:46 +0000 (11:51 -0700)]
osxaudio: adjust for changes to glib mutex api.

11 years agoosxaudio: try to fix up according to Sebastian's comments
Takashi Nakajima [Tue, 9 Apr 2013 16:21:49 +0000 (01:21 +0900)]
osxaudio: try to fix up according to Sebastian's comments

11 years agoosxaudio: build fixes
Philippe Normand [Fri, 5 Apr 2013 08:02:38 +0000 (10:02 +0200)]
osxaudio: build fixes

Enable the osxaudio plugin build in configure.ac and fix some
include directive order issues.

11 years agoosxaudio: fix layout for osxaudiosrc
ted-n [Tue, 2 Apr 2013 13:28:09 +0000 (22:28 +0900)]
osxaudio: fix layout for osxaudiosrc

11 years agoosxaudio: port to v.1.0
ted-n [Sat, 30 Mar 2013 13:49:34 +0000 (22:49 +0900)]
osxaudio: port to v.1.0

11 years agovideomixer: Don't unref query, we don't own it
Olivier Crête [Tue, 16 Apr 2013 23:29:48 +0000 (19:29 -0400)]
videomixer: Don't unref query, we don't own it

Fixes double-unref bug. Bug found by Youness Alaoui

11 years agosouphttpsrc: fix SCHEDULING query support
Philippe Normand [Tue, 16 Apr 2013 18:41:10 +0000 (20:41 +0200)]
souphttpsrc: fix SCHEDULING query support

Chain the query up to parent before adding _BANDWIDTH_LIMITED flag,
so that all the other flags get set, and push mode gets added as
supported activation mode.

https://bugzilla.gnome.org/show_bug.cgi?id=693484
https://bugzilla.gnome.org/show_bug.cgi?id=698156

11 years agosouphttpsrc: basic scheduling query support
Philippe Normand [Sun, 31 Mar 2013 10:05:49 +0000 (12:05 +0200)]
souphttpsrc: basic scheduling query support

Answer to scheduling queries with default parameters and the new
_BANDWIDTH_LIMITED_FLAG so that downstream is advised to minimize seek
operations and perform on-disk buffering if possible.

Bug 693484

11 years agoosxvideosink: fix segfault accessing osxwindow when not set yet
Andoni Morales Alastruey [Mon, 15 Apr 2013 14:32:46 +0000 (14:32 +0000)]
osxvideosink: fix segfault accessing osxwindow when not set yet

11 years agogst: Add better support for static plugins
Sebastian Dröge [Wed, 24 Oct 2012 10:14:50 +0000 (12:14 +0200)]
gst: Add better support for static plugins

11 years agogoom2k1: fix duplicated symbol with goom
Andoni Morales Alastruey [Fri, 12 Apr 2013 19:26:11 +0000 (19:26 +0000)]
goom2k1: fix duplicated symbol with goom

11 years agoosxaudio: Fixes error: "GST_LEVEL_DEFAULT" redefined
Josep Torra [Sun, 10 Mar 2013 17:17:17 +0000 (17:17 +0000)]
osxaudio: Fixes error: "GST_LEVEL_DEFAULT" redefined

11 years agoosxaudio: fixes implicit declaration of function 'getpid'
Josep Torra [Sun, 10 Mar 2013 17:27:30 +0000 (17:27 +0000)]
osxaudio: fixes implicit declaration of function 'getpid'

11 years agoAutomatic update of common submodule
Tim-Philipp Müller [Sun, 14 Apr 2013 16:55:02 +0000 (17:55 +0100)]
Automatic update of common submodule

From aed87ae to 3cb3d3c

11 years agosouphttpsrc: add back "iradio-mode" property to disable sending of icecast request...
Tim-Philipp Müller [Sun, 14 Apr 2013 11:32:06 +0000 (12:32 +0100)]
souphttpsrc: add back "iradio-mode" property to disable sending of icecast request headers

In 1.0 we now always send the icecast request headers by default, which
makes the server send icecasts metadata inserted into the stream if it
supports that. However, there are some use cases where this is not
desirable, like when just saving a radio stream to disk, so add back
the "iradio-mode" property to allow people to disable this.

https://bugzilla.gnome.org/show_bug.cgi?id=697984

11 years agortp: register tag image types
Wim Taymans [Fri, 12 Apr 2013 15:16:41 +0000 (16:16 +0100)]
rtp: register tag image types

The rtpgstdepay needs the type to be available in order to deserialize the
event.

11 years agortpgstdepay: handle event parse failures better
Wim Taymans [Fri, 12 Apr 2013 15:08:58 +0000 (16:08 +0100)]
rtpgstdepay: handle event parse failures better

11 years agowavenc: add TOC setter support
Anton Belka [Thu, 11 Apr 2013 19:25:05 +0000 (22:25 +0300)]
wavenc: add TOC setter support

11 years agowavenc: small cleanups for toc handling
Stefan Sauer [Fri, 12 Apr 2013 10:31:30 +0000 (12:31 +0200)]
wavenc: small cleanups for toc handling

Don't add empty labl/note chunks. Always pass instance as the first param. Add more logging.

11 years agortspsrc: Proxy the ntp-sync property of rtpbin
Sebastian Dröge [Fri, 12 Apr 2013 10:58:50 +0000 (12:58 +0200)]
rtspsrc: Proxy the ntp-sync property of rtpbin

11 years agortspsrc: Give the manager always the name "manager"
Sebastian Dröge [Fri, 12 Apr 2013 10:51:05 +0000 (12:51 +0200)]
rtspsrc: Give the manager always the name "manager"

This allows to use the GstChildProxy interface to adjust
properties on it.

11 years agotests: fix some printf format issues in debug messages
Tim-Philipp Müller [Thu, 11 Apr 2013 21:53:28 +0000 (22:53 +0100)]
tests: fix some printf format issues in debug messages

11 years agowavenc: add 'note' chunk support
Anton Belka [Thu, 11 Apr 2013 16:27:15 +0000 (19:27 +0300)]
wavenc: add 'note' chunk support

11 years agopulsesink: add a little more docs to the audioclock
Stefan Sauer [Thu, 11 Apr 2013 18:46:26 +0000 (20:46 +0200)]
pulsesink: add a little more docs to the audioclock

11 years agortspsrc: add support for NetClientClock
Wim Taymans [Thu, 11 Apr 2013 14:00:05 +0000 (15:00 +0100)]
rtspsrc: add support for NetClientClock

When the server suggests a GstNetTimeProvider in the SDP, set up a
GstNetClientClock that slaves to the remote clock and suggest this clock in
provide_clock.

11 years agoudpsink: avoid alloc and free in render function
Wim Taymans [Thu, 11 Apr 2013 13:57:11 +0000 (14:57 +0100)]
udpsink: avoid alloc and free in render function

Avoid doing alloc and free in the render function for each buffer. Instead,
allocate the needed arrays in _init and use those.

11 years agowaveparse: remove superfluous g_list_first() calls
Stefan Sauer [Wed, 10 Apr 2013 06:36:00 +0000 (08:36 +0200)]
waveparse: remove superfluous g_list_first() calls

The variables already point to the start of the list.

11 years agortpsbcdepay: fix sbc frame length calculation for mono and stereo modes
Andreas Fenkart [Tue, 9 Apr 2013 22:13:18 +0000 (23:13 +0100)]
rtpsbcdepay: fix sbc frame length calculation for mono and stereo modes

https://bugzilla.gnome.org/show_bug.cgi?id=697463

11 years agowavparse: add 'note' chunk support
Anton Belka [Mon, 25 Mar 2013 11:35:02 +0000 (14:35 +0300)]
wavparse: add 'note' chunk support

Add 'note' chunk support in TOC as GST_TAG_COMMENT

https://bugzilla.gnome.org/show_bug.cgi?id=696549

11 years agoqtdemux: check value inside enda to set endianness
David Schleef [Tue, 9 Apr 2013 00:53:09 +0000 (17:53 -0700)]
qtdemux: check value inside enda to set endianness

11 years agoAutomatic update of common submodule
Stefan Sauer [Tue, 9 Apr 2013 19:00:12 +0000 (21:00 +0200)]
Automatic update of common submodule

From 04c7a1e to aed87ae

11 years agoicydemux: avoid copy when we can
Wim Taymans [Tue, 9 Apr 2013 15:34:12 +0000 (17:34 +0200)]
icydemux: avoid copy when we can

11 years agogstpay: use bufferlist to avoid memcpy
Wim Taymans [Tue, 9 Apr 2013 14:52:21 +0000 (16:52 +0200)]
gstpay: use bufferlist to avoid memcpy

11 years agoudpsink: improve debug
Wim Taymans [Tue, 9 Apr 2013 14:50:56 +0000 (16:50 +0200)]
udpsink: improve debug

11 years agotests: refactor new wavparse test a little
Tim-Philipp Müller [Mon, 8 Apr 2013 23:28:54 +0000 (00:28 +0100)]
tests: refactor new wavparse test a little

Use fakesrc instead of filesrc with /dev/null.

https://bugzilla.gnome.org/show_bug.cgi?id=696684

11 years agowavparse: error out if we receive eos before any valid data
Alexander Schrab [Mon, 8 Apr 2013 09:38:33 +0000 (11:38 +0200)]
wavparse: error out if we receive eos before any valid data

https://bugzilla.gnome.org/show_bug.cgi?id=696684

11 years agodeinterlace: force deinterlacing in "interlaced" mode
Matej Knopp [Sat, 6 Apr 2013 23:47:56 +0000 (01:47 +0200)]
deinterlace: force deinterlacing in "interlaced" mode

https://bugzilla.gnome.org/show_bug.cgi?id=697467

11 years agogdkpixbufsink: Add timestamp/running-time/stream-time to the emited message
Thibault Saunier [Sat, 6 Apr 2013 15:45:28 +0000 (12:45 -0300)]
gdkpixbufsink: Add timestamp/running-time/stream-time to the emited message

11 years agortpsbcdepay: fix printf format compiler warnings
Nicola Murino [Fri, 5 Apr 2013 12:38:43 +0000 (14:38 +0200)]
rtpsbcdepay: fix printf format compiler warnings

https://bugzilla.gnome.org/show_bug.cgi?id=697343

11 years agoosxvideo: include pthread.h to fix compiler warning
Todd Agulnick [Fri, 5 Apr 2013 08:34:23 +0000 (09:34 +0100)]
osxvideo: include pthread.h to fix compiler warning

https://bugzilla.gnome.org/show_bug.cgi?id=697303

11 years agolevel: resync on discont
Stefan Sauer [Thu, 4 Apr 2013 20:48:45 +0000 (22:48 +0200)]
level: resync on discont

Drop pending data on discont and start a new cycle with a new base timestamp.
Cleanup some variables.

11 years agovp8dec: Improve logging when vpx_codec_peek_stream_info fails
Tom Greenwood [Wed, 3 Apr 2013 22:52:47 +0000 (23:52 +0100)]
vp8dec: Improve logging when vpx_codec_peek_stream_info fails

Decode failures and missing keyframes should get different debug
output.

https://bugzilla.gnome.org/show_bug.cgi?id=697232

11 years agortpsbcdepay: Rank as secondary
Olivier Crête [Wed, 3 Apr 2013 22:24:29 +0000 (18:24 -0400)]
rtpsbcdepay: Rank as secondary

This way, it will be selected by decodebin
Bug reported by andreas.fenkart@streamunlimited.com

https://bugzilla.gnome.org/show_bug.cgi?id=697227

11 years agolevel: subdivide buffers for sample accurate interval handling
Stefan Sauer [Wed, 3 Apr 2013 17:05:38 +0000 (19:05 +0200)]
level: subdivide buffers for sample accurate interval handling

Previously we would skip level message when processing buffers > the requested
interval. Also the message frequency would contain quite some jitter due to only
considering them at the end of buffers.

Cleanup the tests while we're at it.

11 years agoflacenc: remove old since comments and update logging
Stefan Sauer [Tue, 19 Mar 2013 07:23:25 +0000 (08:23 +0100)]
flacenc: remove old since comments and update logging

Don't pretend that we have a timestamp on a buffer when we never set one.

11 years agospectrum: remove old since comment
Stefan Sauer [Mon, 18 Mar 2013 19:59:23 +0000 (20:59 +0100)]
spectrum: remove old since comment

11 years agortspsrc: Proxy the multicast-iface property of udpsrc
Sebastian Dröge [Wed, 3 Apr 2013 15:53:13 +0000 (17:53 +0200)]
rtspsrc: Proxy the multicast-iface property of udpsrc

11 years agov4l2: free all queued buffers
Wim Taymans [Wed, 3 Apr 2013 09:09:50 +0000 (11:09 +0200)]
v4l2: free all queued buffers

Don't just loop over the first num_queued buffers but loop over
all the buffers and check if they need to be freed. It is possible that
not all buffers are queued and then the entry in our array will be NULL.
Those buffers that are not queued were freed in stop().

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=696651

11 years agov4l2: improve debug
Wim Taymans [Wed, 3 Apr 2013 09:09:37 +0000 (11:09 +0200)]
v4l2: improve debug

11 years agortpssrcdemux: Only forward stick events while holding the sinkpad stream lock
Olivier Crête [Wed, 3 Apr 2013 03:42:23 +0000 (23:42 -0400)]
rtpssrcdemux: Only forward stick events while holding the sinkpad stream lock

Otherwise we get a race where if the RTCP packet comes in first and while
it is added the pads, the segment event arrives on the RTP stream, the event
may be lost completely and never forwarded.

11 years agortpssrcdemux: No need to explicitely forward the caps
Olivier Crête [Wed, 3 Apr 2013 03:35:06 +0000 (23:35 -0400)]
rtpssrcdemux: No need to explicitely forward the caps

They are forwarded with the other events

11 years agortpssrcdemux: Remove unused GstSegment
Olivier Crête [Wed, 3 Apr 2013 02:29:38 +0000 (22:29 -0400)]
rtpssrcdemux: Remove unused GstSegment

11 years agortpssrcdemux: Simplify event forwarding
Olivier Crête [Wed, 3 Apr 2013 02:26:02 +0000 (22:26 -0400)]
rtpssrcdemux: Simplify event forwarding

Use the gst_pad_forward() mechanic, this way we won't miss pads that are
added while we are pushing

11 years agortpssrcdemux: Don't cross the internal links
Olivier Crête [Wed, 3 Apr 2013 01:53:10 +0000 (21:53 -0400)]
rtpssrcdemux: Don't cross the internal links

We had the wrong condition to check for the internal links, so RTP and RTCP
pads got crossed!