6 years agov4l2: Adding support for dmabuf memory for v4l2src master
Harinarayan Bhatta [Wed, 20 Aug 2014 09:01:33 +0000 (14:31 +0530)]
v4l2: Adding support for dmabuf memory for v4l2src

Adding code to use dmabuf buffers for video capture using v4l2src.
Dmabuf buffers can be consumed by vpe or ducati encoder elements
without a memory copy.

Signed-off-by: Harinarayan Bhatta <>
7 years agoDEBIAN: Debianization glsdk-6_00_00_07
Nikhil Devshatwar [Wed, 15 May 2013 11:11:37 +0000 (16:41 +0530)]
DEBIAN: Debianization

7 years agogit ring buffer null check.
Nikhil Devshatwar [Wed, 15 May 2013 11:10:46 +0000 (16:40 +0530)]
git ring buffer null check.

7 years agov4l2src: fix v4l2_munmap() for compressed formats
Oleksij Rempel (Alexey Fisher) [Thu, 1 Mar 2012 13:15:29 +0000 (13:15 +0000)]
v4l2src: fix v4l2_munmap() for compressed formats

Make sure we always call munmap() with the same size we called mmap()
with before.

Current v4l2src uses the same structure for VIDIOC_QUERYBUF, VIDIOC_QBUF
and v4l2_munmap calls. The problem is that the video buffer size (length)
may vary for compressed or emulated bufs. VIDIOC_QBUF will change it if
we pass the pointer of a v4l2_buffer. This is why we should avoid using
same variable for mmap and video buffers.

7 years agoqtmux: fix missing rebasing of known timestamps
Vincent Penquerc'h [Wed, 12 Sep 2012 11:46:33 +0000 (11:46 +0000)]
qtmux: fix missing rebasing of known timestamps

There are two places where the first timestamp known to a pad
is determined, and only one of them was then rebasing all known
timestamps (currently held buffers, and list of to-be-reordered
buffers), which would lead to mistakenly deducing input was out
of order when it was not.

7 years agoautovideosink: add video/x-raw-yuv-strided to the list of raw caps
Alessandro Decina [Thu, 22 Sep 2011 16:58:40 +0000 (18:58 +0200)]
autovideosink: add video/x-raw-yuv-strided to the list of raw caps

7 years agoPulsesink: Allow chunks up to bufsize instead of segsize
David Henningsson [Mon, 31 Jan 2011 04:58:36 +0000 (05:58 +0100)]
Pulsesink: Allow chunks up to bufsize instead of segsize

By allowing larger chunks to be sent, PulseAudio will have a
lower CPU usage. This is especially important on low-end machines,
where PulseAudio can crash if packets are coming in at a higher
rate than PulseAudio can process them.

Signed-off-by: David Henningsson <>
7 years agoqtmux: Store 1st TS when detect we get 1st buffer
Alejandro Gonzalez [Thu, 16 Dec 2010 15:14:06 +0000 (10:14 -0500)]
qtmux: Store 1st TS when detect we get 1st buffer

7 years agofixes to mux h264 stream utilizing ctts
Rob Clark [Tue, 28 Jun 2011 03:34:45 +0000 (00:34 -0300)]
fixes to mux h264 stream utilizing ctts

7 years agoeffectv: enable renegotiation flushing mid-way
Luciana Fujii Pontello [Fri, 17 Jun 2011 19:50:38 +0000 (16:50 -0300)]
effectv: enable renegotiation flushing mid-way

Re-enable re-negotiation for shagadelic, vertigo and warp effects.
It flushes when the re-negotiation happens to make the new media be
played as soon as possible.

7 years agov4l2: add rowstride support
Luciana Fujii Pontello [Thu, 10 May 2012 21:22:05 +0000 (17:22 -0400)]
v4l2: add rowstride support



7 years agov4l2: fix handling of RGB32/BGR32 formats
Rob Clark [Tue, 14 Sep 2010 12:42:50 +0000 (07:42 -0500)]
v4l2: fix handling of RGB32/BGR32 formats

bpp is 32, but depth is only 24..

9 years agoRelease 0.10.31 freedesktop-RELEASE-0.10.31
Tim-Philipp Müller [Tue, 21 Feb 2012 00:51:42 +0000 (00:51 +0000)]
Release 0.10.31

9 years agodocs: update docs
Tim-Philipp Müller [Tue, 21 Feb 2012 00:46:55 +0000 (00:46 +0000)]
docs: update docs

9 years agortph264pay: force baseline profile is profile-level-id is unspecified
Olivier Crête [Mon, 20 Feb 2012 17:22:12 +0000 (12:22 -0500)]
rtph264pay: force baseline profile is profile-level-id is unspecified

If profile-level-id is missing or invalid, we want any upstream
encoder to default to baseline profile, so specify that in the
caps we pass upstream. If the caps contain no profile restriction,
an encoder may default to high or main profile.

9 years agoequalizer: fix switching from passthrough to non-passthrough when parameters change
Tim-Philipp Müller [Fri, 17 Feb 2012 17:21:53 +0000 (17:21 +0000)]
equalizer: fix switching from passthrough to non-passthrough when parameters change

commit b5bf0294 moved the if(need_new_coefficients) set_passthrough(equ)
after the if(is_passthrough) return FLOW_OK shortcut, so the passthrough
mode would never get updated even if the coefficients change.

Fixes equalizer-test doing .. nothing.

9 years agov4l2src: failure to query some optional controls is not a fatal error
Gary Ching-Pang Lin [Thu, 16 Feb 2012 09:14:20 +0000 (17:14 +0800)]
v4l2src: failure to query some optional controls is not a fatal error

Don't post a (fatal) error message on the bus just because we
failed to query some control. Fixes issue with built-in
Suyin Corp webcam for HP notebook (usbid 064e:e28a) on
OpenSuse 12.1, where querying red/blue balance fails.

9 years agov4l2src: fix for webcamstudio vloopback
Tuukka Pasanen [Thu, 16 Feb 2012 12:59:10 +0000 (12:59 +0000)]
v4l2src: fix for webcamstudio vloopback

Because vlooback emits 25 - ENOTTY and no EINVAL v4l2src thought it
can't handle this and does not work.

9 years agotests: flacparse: check and compare intended data
Mark Nauwelaerts [Mon, 13 Feb 2012 11:06:37 +0000 (12:06 +0100)]
tests: flacparse: check and compare intended data

9 years agotests: mpegaudioparse: remove stray declaration
Mark Nauwelaerts [Thu, 9 Feb 2012 21:12:14 +0000 (22:12 +0100)]
tests: mpegaudioparse: remove stray declaration

9 years agomultiudpsink: typo fix (bytes send -> bytes sent)
Marc Leeman [Thu, 9 Feb 2012 09:11:48 +0000 (10:11 +0100)]
multiudpsink: typo fix (bytes send -> bytes sent)

9 years agoshout2send: send video/webm through libshout.
Ralph Giles [Tue, 7 Feb 2012 22:10:44 +0000 (14:10 -0800)]
shout2send: send video/webm through libshout.

This requires SHOUT_FORMAT_WEBM, added in libshout 2.3.0,
so video/webm support is contingent on that symbol being

Also an indentation change required by the pre-commit hook.

9 years agomatroskademux: avoid posting invalid duration for each frame
Nicola Murino [Sat, 28 Jan 2012 10:13:16 +0000 (11:13 +0100)]
matroskademux: avoid posting invalid duration for each frame

9 years ago0.10.30.3 pre-release
Tim-Philipp Müller [Sun, 5 Feb 2012 13:40:13 +0000 (13:40 +0000)] pre-release

9 years agopulseaudiosink: Lower rank to prevent autoplugging
Arun Raghavan [Fri, 3 Feb 2012 16:35:59 +0000 (22:05 +0530)]
pulseaudiosink: Lower rank to prevent autoplugging

pulseaudiosink breaks visualisations in its current form, so let's
prevent it from being autoplugged for the time being.

The best we can hope to do in the 0.10 series is query the list of
available sinks and their formats, and expose these as the bin's sinkpad
caps. While this is not a comprehensive solution, it will make sure that
we're only trying to support compressed formats if we're certain that
one exists.

The long-term fix for this will be in the form of proper upstream
renegotiation support in the 0.11/1.0 series.

9 years agoflacenc: fix event leak when there is no peer on the src pad
Vincent Penquerc'h [Fri, 3 Feb 2012 14:53:31 +0000 (14:53 +0000)]
flacenc: fix event leak when there is no peer on the src pad

9 years agoflvmux: specify we only accept raw AAC in template caps
Vincent Penquerc'h [Thu, 2 Feb 2012 12:27:09 +0000 (12:27 +0000)]
flvmux: specify we only accept raw AAC in template caps

No header seems to be added, and the codec ID is the same as used
for raw by flvdemux, so raw seems the only supported case.

9 years agoflvdemux: specify we only output raw AAC in template caps
Vincent Penquerc'h [Thu, 2 Feb 2012 12:25:21 +0000 (12:25 +0000)]
flvdemux: specify we only output raw AAC in template caps

9 years agortpmp2tpay: do not try to flush a packet when no data is available
Vincent Penquerc'h [Mon, 30 Jan 2012 14:52:37 +0000 (14:52 +0000)]
rtpmp2tpay: do not try to flush a packet when no data is available

9 years agortph264depay: Exclude NALu size from payload length on truncated packets.
Pascal Buhler [Fri, 11 Jun 2010 06:36:33 +0000 (08:36 +0200)]
rtph264depay: Exclude NALu size from payload length on truncated packets.

9 years agovideobox: avoid wrapping opaque to transparent
Vincent Penquerc'h [Sat, 28 Jan 2012 13:05:09 +0000 (13:05 +0000)]
videobox: avoid wrapping opaque to transparent

9 years agocairo: fix build, make sure libgstvideo can be found
Jayakrishnan M [Wed, 25 Jan 2012 15:21:44 +0000 (15:21 +0000)]
cairo: fix build, make sure libgstvideo can be found

9 years agortpmanager: don't pretend our random hostnames are fully-qualified domain names
Tim-Philipp Müller [Wed, 25 Jan 2012 13:19:12 +0000 (13:19 +0000)]
rtpmanager: don't pretend our random hostnames are fully-qualified domain names

9 years agortpmanager: don't reveal the user's username, hostname or real name by default
Tim-Philipp Müller [Mon, 23 Jan 2012 13:15:46 +0000 (13:15 +0000)]
rtpmanager: don't reveal the user's username, hostname or real name by default

Send a randomly made-up user@hostname as CNAME and don't
send a NAME at all by default.

9 years agortspsrc: simplify internal src event debug logging
Mark Nauwelaerts [Fri, 20 Jan 2012 16:06:42 +0000 (17:06 +0100)]
rtspsrc: simplify internal src event debug logging

... which avoids almost superfluous obtaining of rtsp element.

9 years agortspsrc: avoid NULL string comparison
Mark Nauwelaerts [Fri, 20 Jan 2012 16:03:50 +0000 (17:03 +0100)]
rtspsrc: avoid NULL string comparison

9 years agortpmp4adepay: prevent out-of-bound array access
Mark Nauwelaerts [Fri, 20 Jan 2012 16:02:15 +0000 (17:02 +0100)]
rtpmp4adepay: prevent out-of-bound array access

9 years agoisomp4: recovery: add sanity check
Mark Nauwelaerts [Fri, 20 Jan 2012 16:01:37 +0000 (17:01 +0100)]
isomp4: recovery: add sanity check

... on possibly bogus/corrupt input data.

9 years agomatroska-demux: remove redundant variable
Mark Nauwelaerts [Fri, 20 Jan 2012 15:58:28 +0000 (16:58 +0100)]
matroska-demux: remove redundant variable

9 years agodeinterlace: fix arithmetic for unsigned comparison
Mark Nauwelaerts [Fri, 20 Jan 2012 15:57:52 +0000 (16:57 +0100)]
deinterlace: fix arithmetic for unsigned comparison

9 years agoimagefreeze: add various missing break
Mark Nauwelaerts [Fri, 20 Jan 2012 15:55:06 +0000 (16:55 +0100)]
imagefreeze: add various missing break

9 years agoalphacolor: remove redundant statement
Mark Nauwelaerts [Fri, 20 Jan 2012 15:49:14 +0000 (16:49 +0100)]
alphacolor: remove redundant statement

9 years agoflacdec: improve upstream peer duration querying
Mark Nauwelaerts [Fri, 20 Jan 2012 15:48:49 +0000 (16:48 +0100)]
flacdec: improve upstream peer duration querying

... to avoid accepting unhandled duration query result.

9 years agopulsesrc: additional error condition checking
Mark Nauwelaerts [Fri, 20 Jan 2012 15:47:36 +0000 (16:47 +0100)]
pulsesrc: additional error condition checking

9 years agopulsesink: additional error condition checking
Mark Nauwelaerts [Fri, 20 Jan 2012 15:46:21 +0000 (16:46 +0100)]
pulsesink: additional error condition checking

9 years agojpegenc: check _alloc_buffer result and perform fallback alloc if needed
Mark Nauwelaerts [Fri, 20 Jan 2012 15:44:21 +0000 (16:44 +0100)]
jpegenc: check _alloc_buffer result and perform fallback alloc if needed

... rather than carrying on with NULL buffer.

9 years agopulsesrc: fix wrong error check
Vincent Penquerc'h [Fri, 13 Jan 2012 18:11:36 +0000 (18:11 +0000)]
pulsesrc: fix wrong error check

pa_stream_* functions return negative on error, despite the defines
for error codes being positive.

I only got to repro the error twice, so I'm not sure 100% sure this
fixes the issue (the negative var being uninitialized after returning
from pa_stream_get_latency).

9 years agocutter: fix leak of unused GValue
Vincent Penquerc'h [Mon, 16 Jan 2012 17:51:18 +0000 (17:51 +0000)]
cutter: fix leak of unused GValue

9 years agotests: fix autodetect test not testing correctly for state change success
Vincent Penquerc'h [Mon, 16 Jan 2012 16:10:08 +0000 (16:10 +0000)]
tests: fix autodetect test not testing correctly for state change success

State change to PAUSED can be done async, so if this happens, we need
to wait for the change to be done (or failed).

9 years agortph263ppay: fix caps leak
Vincent Penquerc'h [Mon, 16 Jan 2012 15:42:46 +0000 (15:42 +0000)]
rtph263ppay: fix caps leak

9 years agodeinterlace: make interlacedness test deterministic
Vincent Penquerc'h [Mon, 16 Jan 2012 12:13:50 +0000 (12:13 +0000)]
deinterlace: make interlacedness test deterministic

If the interlaced flag is not present in the caps, we assume the
data is not interlaced, instead of leaving the boolean uninitialized.

9 years agooss4: fix caps leaks
Vincent Penquerc'h [Fri, 13 Jan 2012 17:43:49 +0000 (17:43 +0000)]
oss4: fix caps leaks

9 years agov4l2src: fix caps leak
Vincent Penquerc'h [Fri, 13 Jan 2012 17:25:59 +0000 (17:25 +0000)]
v4l2src: fix caps leak

9 years agotests: fix caps leak in videocrop test
Vincent Penquerc'h [Fri, 13 Jan 2012 15:57:20 +0000 (15:57 +0000)]
tests: fix caps leak in videocrop test

9 years agortpptdemux: plug pad leak in error code path
Tim-Philipp Müller [Fri, 13 Jan 2012 10:32:59 +0000 (10:32 +0000)]
rtpptdemux: plug pad leak in error code path

Based on patch by: Stig Sandnes <>

Don't leak srcpad if there are no caps.

9 years agoosxvideo: Fix leak of NSOpenGLPixelFormat object
Stig Sandnes [Tue, 4 Oct 2011 08:00:02 +0000 (10:00 +0200)]
osxvideo: Fix leak of NSOpenGLPixelFormat object

9 years agov4l2src: Don't assert when the interface is not implemented.
Havard Graff [Mon, 5 Sep 2011 08:43:19 +0000 (10:43 +0200)]
v4l2src: Don't assert when the interface is not implemented.

Simply return FALSE instead.

9 years agowaveformsink: Fix mingw warnings
Raimo Järvi [Wed, 11 Jan 2012 22:18:39 +0000 (00:18 +0200)]
waveformsink: Fix mingw warnings

9 years agogstrtpssrcdemux: fix element leak
Vincent Penquerc'h [Thu, 12 Jan 2012 18:23:42 +0000 (18:23 +0000)]
gstrtpssrcdemux: fix element leak

9 years agomatroska: do not leak attachment buffers
Vincent Penquerc'h [Thu, 12 Jan 2012 14:19:22 +0000 (14:19 +0000)]
matroska: do not leak attachment buffers

9 years agoflacenc: do not drop the first data buffer on the floor (and leak it either)
Vincent Penquerc'h [Thu, 12 Jan 2012 10:30:11 +0000 (10:30 +0000)]
flacenc: do not drop the first data buffer on the floor (and leak it either)

9 years agoTemporarily disabling multifile for the Android build
Reynaldo H. Verdejo Pinochet [Wed, 11 Jan 2012 21:45:33 +0000 (18:45 -0300)]
Temporarily disabling multifile for the Android build

There is a hard dependency on inotify comming from gio. We
are not currently bundling inotify with the Android dist so
I'm disabling multifile for now until someone gets around
to sort this out.

This change fixes building on Android

9 years agotests: fix wavenc test on big endian
Tim-Philipp Müller [Wed, 11 Jan 2012 01:45:34 +0000 (01:45 +0000)]
tests: fix wavenc test on big endian

wavenc only accepts little-endian PCM, but most of our
elements such as audiotestsrc only produce or process
audio in native endianness, so we need to plug a
converter before wavenc on big endian systems.

9 years agoisomp4: fix caps leak
Vincent Penquerc'h [Thu, 5 Jan 2012 19:25:33 +0000 (19:25 +0000)]
isomp4: fix caps leak

9 years agoisomp4: remove dead assignment
Vincent Penquerc'h [Thu, 5 Jan 2012 19:08:03 +0000 (19:08 +0000)]
isomp4: remove dead assignment

9 years agoAutomatic update of common submodule
Tim-Philipp Müller [Wed, 4 Jan 2012 19:40:14 +0000 (19:40 +0000)]
Automatic update of common submodule

From 11f0cd5 to cb5da59

9 years agotests: fix some leaks and remove files when done in qtmux test
Tim-Philipp Müller [Wed, 4 Jan 2012 17:59:55 +0000 (17:59 +0000)]
tests: fix some leaks and remove files when done in qtmux test

9 years agomultifilesink: post better error message when we run out of disk space
Peter Seiderer [Wed, 14 Dec 2011 09:14:20 +0000 (10:14 +0100)]
multifilesink: post better error message when we run out of disk space


9 years agoudpsrc: fix valgrind warning
Tim-Philipp Müller [Tue, 27 Dec 2011 11:50:03 +0000 (11:50 +0000)]
udpsrc: fix valgrind warning

9 years agoudpsrc: drop dataless UDP packets
John Ogness [Wed, 21 Dec 2011 12:22:03 +0000 (13:22 +0100)]
udpsrc: drop dataless UDP packets

It is allowed to send/receive UDP packets with no data. When such
a packet is available, select() will return with success but
ioctl(FIONREAD) will return 0. But a read() must still occur in
order to clear off the UDP packet from the queue.

This patch will read the dataless packet from the socket. If
select() was woken for other reasons (and FIONREAD returns 0),
this may result in a UDP packet getting accidentally dropped.
But since UDP is not reliable, this is acceptable.

NOTE: This patch fixes a nasty bug where sending a dataless
      UDP packet to a udpsrc instance will cause an infinite

Signed-off-by: John Ogness <>
9 years agojpegdec: fix peer_caps leak
Nicola Murino [Wed, 21 Dec 2011 19:50:21 +0000 (20:50 +0100)]
jpegdec: fix peer_caps leak

9 years agoflvmux: don't try to push already-freed buffers
Tim-Philipp Müller [Sun, 25 Dec 2011 14:23:29 +0000 (14:23 +0000)]
flvmux: don't try to push already-freed buffers

Fixes unit test.

9 years agoac3parse: let bsid 9 and 10 through
Vincent Penquerc'h [Fri, 9 Sep 2011 10:42:09 +0000 (11:42 +0100)]
ac3parse: let bsid 9 and 10 through

Files with 9 and 10 happen, and seem to comply with the <= 8
format, so let them through.
The spec says nothing about 9 and 10.

9 years agoflvmux: properly determine final duration
Mark Nauwelaerts [Fri, 16 Dec 2011 18:15:38 +0000 (19:15 +0100)]
flvmux: properly determine final duration

... which can be authoratively obtained from our own written timestamps.

9 years agoflvmux: only write full metadata at start
Mark Nauwelaerts [Mon, 19 Dec 2011 12:56:30 +0000 (13:56 +0100)]
flvmux: only write full metadata at start

... rather than having (potentially) unnecessary duplicates written all over,
or even contradictory varying filesize info, or duration info that will not
be rewritten upon header rewrite.

9 years agomatroskademux: do not consider duration of non-finalized file
Branko Subasic [Wed, 21 Dec 2011 16:43:10 +0000 (17:43 +0100)]
matroskademux: do not consider duration of non-finalized file

... to avoid it clamping requested seek position.

Non-finalized file case, determined by whether
_parse_blockgroup_or_simpleblock ever updates the segment duration.

Fixes #652195.

9 years agomatroskademux: improve decision to fall back to scanning when seeking
Mark Nauwelaerts [Wed, 21 Dec 2011 14:06:57 +0000 (15:06 +0100)]
matroskademux: improve decision to fall back to scanning when seeking

... which is basically iff not streaming and no entry found in index

9 years agomatroskademux: filter bogus index entries with missing block number
Mark Nauwelaerts [Tue, 13 Dec 2011 17:18:45 +0000 (18:18 +0100)]
matroskademux: filter bogus index entries with missing block number

... to avoid contradictory information resulting in seeks sending more
downstream than needed for the corresponding segment.

9 years agomatroskademux: cater for safer arithmetic with global start time
Mark Nauwelaerts [Tue, 13 Dec 2011 17:15:18 +0000 (18:15 +0100)]
matroskademux: cater for safer arithmetic with global start time

9 years agomatroskademux: tweak final closing segment sending
Mark Nauwelaerts [Tue, 13 Dec 2011 16:02:01 +0000 (17:02 +0100)]
matroskademux: tweak final closing segment sending

... to avoid it interfering with (sparse) stream syncing.

9 years agoglib-compat: Add license boilerplate for LGPL
Sebastian Dröge [Mon, 12 Dec 2011 10:54:56 +0000 (11:54 +0100)]
glib-compat: Add license boilerplate for LGPL

9 years agomatroskademux: mind (un)signed in some timestamp arithmetic
Mark Nauwelaerts [Mon, 12 Dec 2011 14:15:46 +0000 (15:15 +0100)]
matroskademux: mind (un)signed in some timestamp arithmetic

... to avoid ending up with invalid (negative) duration.

9 years agoqtdemux: increase parse tolerance for fuzzy file cases
Mark Nauwelaerts [Wed, 9 Feb 2011 14:31:22 +0000 (15:31 +0100)]
qtdemux: increase parse tolerance for fuzzy file cases

9 years agobuild: dist glib-compat-private.h properly
Tim-Philipp Müller [Mon, 12 Dec 2011 10:38:20 +0000 (10:38 +0000)]
build: dist glib-compat-private.h properly

Add missing slash.

9 years agotests: use atexit, g_atexit has been deprecated in glib master
Tim-Philipp Müller [Mon, 12 Dec 2011 10:18:14 +0000 (10:18 +0000)]
tests: use atexit, g_atexit has been deprecated in glib master

9 years agoSuppress deprecation warnings in selected files, for g_static_rec_mutex_* mostly
Tim-Philipp Müller [Mon, 12 Dec 2011 02:52:13 +0000 (02:52 +0000)]
Suppress deprecation warnings in selected files, for g_static_rec_mutex_* mostly

GStaticRecMutex is part of our API/ABI, not much we can do here
in 0.10 for most of these.

9 years agotests: g_thread_init() is deprecated in glib master
Tim-Philipp Müller [Mon, 12 Dec 2011 02:41:37 +0000 (02:41 +0000)]
tests: g_thread_init() is deprecated in glib master

It's not needed any longer.

9 years agoUse g_thread_try_new() instead of g_thread_crate() with newer glib versions
Tim-Philipp Müller [Mon, 12 Dec 2011 02:38:37 +0000 (02:38 +0000)]
Use g_thread_try_new() instead of g_thread_crate() with newer glib versions

9 years agoalpha: use new glib API for static mutex if available
Tim-Philipp Müller [Mon, 12 Dec 2011 02:31:36 +0000 (02:31 +0000)]
alpha: use new glib API for static mutex if available

9 years agoWork around deprecated thread API in glib master
Tim-Philipp Müller [Mon, 12 Dec 2011 02:30:45 +0000 (02:30 +0000)]
Work around deprecated thread API in glib master

Add private replacements for deprecated functions such as
g_mutex_new(), g_mutex_free(), g_cond_new() etc., mostly
to avoid the deprecation warnings. We'll change these
over to the new API once we depend on glib >= 2.32.

9 years agoconfigure: Require GLib >= 2.24
Sebastian Dröge [Mon, 12 Dec 2011 09:24:45 +0000 (10:24 +0100)]
configure: Require GLib >= 2.24

All other modules require this already and nobody is testing with
older versions anyway.

9 years agogdkpixbufsink: fix inverted pixel-aspect-ratio
Tim-Philipp Müller [Sun, 11 Dec 2011 18:40:31 +0000 (18:40 +0000)]
gdkpixbufsink: fix inverted pixel-aspect-ratio

Spotted by Mike Morrison.

9 years agopulseaudiosink: don't leak pad template
Tim-Philipp Müller [Sun, 11 Dec 2011 17:55:14 +0000 (17:55 +0000)]
pulseaudiosink: don't leak pad template

9 years ago0.10.30.2 pre-release
Tim-Philipp Müller [Sat, 10 Dec 2011 15:13:07 +0000 (15:13 +0000)] pre-release

9 years agosoup: fix start/stop race in souphttpclientsink
Tim-Philipp Müller [Sat, 10 Dec 2011 14:48:57 +0000 (14:48 +0000)]
soup: fix start/stop race in souphttpclientsink

Fix crash or hang in generic/states unit test when doing stop()
right after start(). Create main loop in the start function already
and not just in the thread function, so that stop() always has a
valid main loop to quit on. Also, calling g_main_loop_quit() before
g_main_loop_run() won't work and result in the stop function waiting
for the thread to join forever. Therefore, wait for the thread to
be ready and get the main loop running in the start() function, to
be sure stop() always works.

9 years agotests: dist test file used in matroskaparse unit test
Tim-Philipp Müller [Sat, 10 Dec 2011 13:35:08 +0000 (13:35 +0000)]
tests: dist test file used in matroskaparse unit test

9 years agotests: fix up rgvolume test for basetransform event caching
Tim-Philipp Müller [Sat, 10 Dec 2011 12:32:32 +0000 (12:32 +0000)]
tests: fix up rgvolume test for basetransform event caching

Some tests assumed that tag events would always pushed through
immediately, which isn't the case any longer, so push a newsegment
event and an empty buffer first.

9 years agopo: update translations
Tim-Philipp Müller [Sat, 10 Dec 2011 02:21:02 +0000 (02:21 +0000)]
po: update translations

9 years agojack: don't leak client name when freeing the element
Tim-Philipp Müller [Fri, 9 Dec 2011 15:50:28 +0000 (15:50 +0000)]
jack: don't leak client name when freeing the element

And add gtk-doc chunks for the new property.

9 years agojack: add "client-name" property to jackaudiosink and jackaudiosrc
Nicolas Baron [Fri, 9 Dec 2011 15:45:03 +0000 (15:45 +0000)]
jack: add "client-name" property to jackaudiosink and jackaudiosrc