platform/upstream/gstreamer.git
12 years agodebug: change rndbuffersize properties from long to int
Tim-Philipp Müller [Sat, 9 Jun 2012 15:53:54 +0000 (16:53 +0100)]
debug: change rndbuffersize properties from long to int

These should all be int instead of long, to avoid bugs
when passing these as varargs with g_object_set(), and
there was no reason to use long in the first place here.
Fixes FIXME.

12 years agoelements: Use gst_pad_set_caps() instead of manual event fiddling
Sebastian Dröge [Fri, 8 Jun 2012 13:54:42 +0000 (15:54 +0200)]
elements: Use gst_pad_set_caps() instead of manual event fiddling

12 years agoAutomatic update of common submodule
Edward Hervey [Fri, 8 Jun 2012 13:04:59 +0000 (15:04 +0200)]
Automatic update of common submodule

From 03a0e57 to 98e386f

12 years agoupdate for audio api change
Wim Taymans [Fri, 8 Jun 2012 08:11:12 +0000 (10:11 +0200)]
update for audio api change

12 years agoBack to development
Sebastian Dröge [Thu, 7 Jun 2012 14:12:34 +0000 (16:12 +0200)]
Back to development

12 years agoRelease 0.11.92
Sebastian Dröge [Thu, 7 Jun 2012 14:12:22 +0000 (16:12 +0200)]
Release 0.11.92

12 years agoUpdate .po files
Sebastian Dröge [Thu, 7 Jun 2012 14:11:17 +0000 (16:11 +0200)]
Update .po files

12 years agopulsesrc: improve clock handling
Wim Taymans [Thu, 7 Jun 2012 13:03:02 +0000 (15:03 +0200)]
pulsesrc: improve clock handling

Post the notify outside of the pa_lock to avoid a deadlock caused by basesrc
calling get_time with the object lock.
Reset the clock on connect.
Post clock-lost and clock-provide messages.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=673977

Conflicts:

ext/pulse/pulsesrc.c

12 years agoBetter GstClock for pulsesrc
Mohammed Sameer [Thu, 12 Apr 2012 10:21:17 +0000 (13:21 +0300)]
Better GstClock for pulsesrc

This clock uses the actual stream time (pa_stream_get_time) to get a more accurate timestamp.

Conflicts:

ext/pulse/pulsesrc.c

12 years agopng: fix video state leaks
Vincent Penquerc'h [Thu, 7 Jun 2012 10:16:50 +0000 (11:16 +0100)]
png: fix video state leaks

12 years agojpegdec: fix video state leak
Vincent Penquerc'h [Thu, 7 Jun 2012 10:16:37 +0000 (11:16 +0100)]
jpegdec: fix video state leak

12 years agortspsrc: only reset the manager object when we did a seek
Wim Taymans [Thu, 7 Jun 2012 10:11:14 +0000 (12:11 +0200)]
rtspsrc: only reset the manager object when we did a seek

Only reset the manager object when we used a Range header, ie. when we did a
seek. Otherwise we just paused and we can resume just fine.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=677475

12 years agotests: add test for rtpsession cleanup
Wim Taymans [Wed, 6 Jun 2012 14:13:29 +0000 (16:13 +0200)]
tests: add test for rtpsession cleanup

12 years agoAutomatic update of common submodule
Edward Hervey [Wed, 6 Jun 2012 16:18:41 +0000 (18:18 +0200)]
Automatic update of common submodule

From 1fab359 to 03a0e57

12 years agomatroskademux: Update for TOC event API change
Sebastian Dröge [Wed, 6 Jun 2012 12:17:08 +0000 (14:17 +0200)]
matroskademux: Update for TOC event API change

12 years agoupdate for tag event change
Wim Taymans [Wed, 6 Jun 2012 11:02:12 +0000 (13:02 +0200)]
update for tag event change

12 years agofix Y800 format
Wim Taymans [Wed, 6 Jun 2012 11:00:58 +0000 (13:00 +0200)]
fix Y800 format

12 years agoosxvideo: straightforward port to 0.11
Thiago Santos [Fri, 1 Jun 2012 04:19:35 +0000 (01:19 -0300)]
osxvideo: straightforward port to 0.11

12 years agoSome printf variable format fixes
Thiago Santos [Thu, 31 May 2012 21:39:25 +0000 (18:39 -0300)]
Some printf variable format fixes

The osx compiler complains about those

12 years agoaudioparsers: Fix GstBaseParse::get_sink_caps() implementations
Sebastian Dröge [Tue, 5 Jun 2012 07:18:12 +0000 (09:18 +0200)]
audioparsers: Fix GstBaseParse::get_sink_caps() implementations

They should take the filter caps into account and always return
the template caps appended to the actual caps. Otherwise the
parsers stop to accept unparsed streams where upstream does not
know about channels, rate, etc.

Fixes bug #677401.

12 years agojpegdec: set colorimetry on output info
Wim Taymans [Mon, 4 Jun 2012 14:17:51 +0000 (16:17 +0200)]
jpegdec: set colorimetry on output info

12 years agoosxaudiosink: Handle endianness correctly
Josep Torra [Mon, 4 Jun 2012 06:10:15 +0000 (08:10 +0200)]
osxaudiosink: Handle endianness correctly

12 years agoosxaudiosink: Add support for int audio
Josep Torra [Fri, 1 Jun 2012 14:37:00 +0000 (16:37 +0200)]
osxaudiosink: Add support for int audio

12 years agoAutomatic update of common submodule
Edward Hervey [Fri, 1 Jun 2012 08:28:53 +0000 (10:28 +0200)]
Automatic update of common submodule

From f1b5a96 to 1fab359

12 years agoqtdemux: set the palette size correctly
Wim Taymans [Thu, 31 May 2012 11:36:32 +0000 (13:36 +0200)]
qtdemux: set the palette size correctly

12 years agov4l2: add missing G_END_DECLS
Michael Jones [Thu, 31 May 2012 08:15:43 +0000 (10:15 +0200)]
v4l2: add missing G_END_DECLS

G_BEGIN_DECLS didn't have matching G_END_DECLS

https://bugzilla.gnome.org/show_bug.cgi?id=677165

12 years agoAutomatic update of common submodule
Sebastian Dröge [Thu, 31 May 2012 11:08:19 +0000 (13:08 +0200)]
Automatic update of common submodule

From 92b7266 to f1b5a96

12 years agoosxvideosink: Really fix the build on 10.5
Josep Torra [Thu, 31 May 2012 08:26:27 +0000 (10:26 +0200)]
osxvideosink: Really fix the build on 10.5

The API that we use to run the Cocoa loop in another
thread does not exist in 10.5 or earlier.

12 years agoosxvideosink: fix race in starting the runloop thread
Alessandro Decina [Sat, 26 May 2012 10:21:18 +0000 (12:21 +0200)]
osxvideosink: fix race in starting the runloop thread

Block gst_osx_video_sink_run_cocoa_loop until the loop thread has started and
finished initializing NSApp. Fixes occasional warnings/crashes due to two
threads going inside NSApp before finishLaunching had completed.

12 years agoosxvideosink: Fix last commit to actually work
Josep Torra [Wed, 30 May 2012 14:03:55 +0000 (16:03 +0200)]
osxvideosink: Fix last commit to actually work

MAC_OS_X_VERSION_10_6 is obviously not defined on 10.5.

12 years agoosxvideosink: Put the right flags in the right variable
Sebastian Dröge [Wed, 30 May 2012 11:51:35 +0000 (13:51 +0200)]
osxvideosink: Put the right flags in the right variable

12 years agoconfigure: Fix GST_OBJCFLAGS
Sebastian Dröge [Wed, 30 May 2012 11:24:03 +0000 (13:24 +0200)]
configure: Fix GST_OBJCFLAGS

12 years agoAutomatic update of common submodule
Sebastian Dröge [Wed, 30 May 2012 10:45:23 +0000 (12:45 +0200)]
Automatic update of common submodule

From ec1c4a8 to 92b7266

12 years agoosxvideosink: NSWindowDelegate is available in all OSX versions newer than 10.6
Sebastian Dröge [Wed, 30 May 2012 10:43:37 +0000 (12:43 +0200)]
osxvideosink: NSWindowDelegate is available in all OSX versions newer than 10.6

12 years agoosxvideosink: Fix build with older OSX versions
Josep Torra [Wed, 30 May 2012 10:40:57 +0000 (12:40 +0200)]
osxvideosink: Fix build with older OSX versions

12 years agoconfigure: Add OBJC specific compiler flags
Sebastian Dröge [Wed, 30 May 2012 09:09:25 +0000 (11:09 +0200)]
configure: Add OBJC specific compiler flags

See bug #643939.

12 years agoAutomatic update of common submodule
Sebastian Dröge [Wed, 30 May 2012 09:23:36 +0000 (11:23 +0200)]
Automatic update of common submodule

From 3429ba6 to ec1c4a8

12 years agovideo: remove duplicate format
Wim Taymans [Tue, 29 May 2012 15:50:21 +0000 (17:50 +0200)]
video: remove duplicate format

12 years agoflvdemux: Post error message if EOS before pads were created
Edward Hervey [Tue, 29 May 2012 14:52:02 +0000 (16:52 +0200)]
flvdemux: Post error message if EOS before pads were created

Happens with some files with only headers

12 years agopng: Port to 0.11 again
Sebastian Dröge [Mon, 28 May 2012 13:22:26 +0000 (15:22 +0200)]
png: Port to 0.11 again

12 years agosoup: Drop transferMode.dlna.org header
Jens Georg [Mon, 14 May 2012 10:46:57 +0000 (12:46 +0200)]
soup: Drop transferMode.dlna.org header

Leave it to the application to decide on the header. No header at all
is better than having the wrong header as DLNA mandates that a missing
header has to be tolerated while a wrong header is an error.

https://bugzilla.gnome.org/show_bug.cgi?id=676020

12 years agopng: Port to base video classes
Edward Hervey [Sat, 7 Apr 2012 07:52:09 +0000 (09:52 +0200)]
png: Port to base video classes

Conflicts:

ext/libpng/gstpngdec.c
ext/libpng/gstpngdec.h
ext/libpng/gstpngenc.c
ext/libpng/gstpngenc.h

Reverted to 0.10, needs to be ported again.

12 years agoflv, matroska: don't use GstStructure API on tag lists
Tim-Philipp Müller [Sat, 26 May 2012 23:02:08 +0000 (00:02 +0100)]
flv, matroska: don't use GstStructure API on tag lists

12 years agortpmp2tdepay: Only output integral mpeg-ts packets
Edward Hervey [Sat, 26 May 2012 09:57:16 +0000 (11:57 +0200)]
rtpmp2tdepay: Only output integral mpeg-ts packets

From RFC 2250

2. Encapsulation of MPEG System and Transport Streams
...
   For MPEG2 Transport Streams the RTP payload will contain an integral
   number of MPEG transport packets.  To avoid end system
   inefficiencies, data from multiple small MTS packets (normally fixed
   in size at 188 bytes) are aggregated into a single RTP packet.  The
   number of transport packets contained is computed by dividing RTP
   payload length by the length of an MTS packet (188).
....

Since it needs to contain "an integral number of MPEG transport packets", a
simple fix is to check that's the case, and strip off any leftover data.

Fixes #676799

Conflicts:

gst/rtp/gstrtpmp2tdepay.c

12 years agoosxvideosink: make sure all selectors are performed on the same thread
Andoni Morales Alastruey [Thu, 24 May 2012 18:43:16 +0000 (20:43 +0200)]
osxvideosink: make sure all selectors are performed on the same thread

When we are using a dedicated thread to run the main run loop we
must make sure that all selectors are performed on this same thread.
For instance if performSelectorOnMainThread is called from the real
main thread, it will not go through the message queue and will be
executed from the real main thread. By forcing the target thread,
we ensure that all functions will be called either from the real
main thread when the main run loop is running or from our thread
spinning the main loop.

12 years agojpegdec: remove framerate
Mathias Hasselmann [Thu, 24 May 2012 14:09:54 +0000 (16:09 +0200)]
jpegdec: remove framerate

The jpeg decoder doesn't need/care about the framerate to so it should
not be in the caps.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=676302

12 years agoosxvideosink: start the loop before calling [gstview haveSuperview]
Alessandro Decina [Thu, 24 May 2012 11:08:35 +0000 (13:08 +0200)]
osxvideosink: start the loop before calling [gstview haveSuperview]

...as haveSuperview requires the mainloop to be running

12 years agoosxvideosink: fix indentation
Alessandro Decina [Thu, 24 May 2012 11:08:13 +0000 (13:08 +0200)]
osxvideosink: fix indentation

12 years agoosxvideosink: enable running the cocoa main runloop in a thread
Alessandro Decina [Tue, 22 May 2012 14:47:36 +0000 (16:47 +0200)]
osxvideosink: enable running the cocoa main runloop in a thread

12 years agoosxvideosink: add code to optionally run the cocoa main runloop in a separate thread
Alessandro Decina [Tue, 22 May 2012 14:45:28 +0000 (16:45 +0200)]
osxvideosink: add code to optionally run the cocoa main runloop in a separate thread

Add a little hack to run the cocoa main runloop from a separate thread _when_
the main runloop is not being run (which means that the app doesn't use cocoa).
Runloops are thread specific, so the hack boils down to getting the runloop for
the main thread and setting it as the runloop for our dedicated thread.

12 years agoosxvideosink: reset app_started to FALSE when shutting down
Alessandro Decina [Tue, 22 May 2012 14:32:53 +0000 (16:32 +0200)]
osxvideosink: reset app_started to FALSE when shutting down

12 years agoosxvideosink: rename cocoa runloop helper funcs
Alessandro Decina [Tue, 22 May 2012 12:49:17 +0000 (14:49 +0200)]
osxvideosink: rename cocoa runloop helper funcs

12 years agoosxvideosink: don't create application menus
Alessandro Decina [Tue, 22 May 2012 12:26:13 +0000 (14:26 +0200)]
osxvideosink: don't create application menus

12 years agoosxvideosink: reset the embed property for backward compatilibity
Andoni Morales Alastruey [Wed, 16 May 2012 19:52:45 +0000 (21:52 +0200)]
osxvideosink: reset the embed property for backward compatilibity

12 years agoosxvideosink: fix navigation when force-aspect-ratio is activated
Andoni Morales Alastruey [Wed, 16 May 2012 19:12:22 +0000 (21:12 +0200)]
osxvideosink: fix navigation when force-aspect-ratio is activated

12 years agoosxvideosink: add force-aspect-ratio property
Andoni Morales Alastruey [Wed, 16 May 2012 16:52:45 +0000 (18:52 +0200)]
osxvideosink: add force-aspect-ratio property

12 years agoosxvideosink: start internal window if no view is provided
Andoni Morales Alastruey [Mon, 14 May 2012 16:01:02 +0000 (18:01 +0200)]
osxvideosink: start internal window if no view is provided

12 years agoosxvideosink: implement the navigation interface
Andoni Morales Alastruey [Mon, 14 May 2012 12:27:58 +0000 (14:27 +0200)]
osxvideosink: implement the navigation interface

12 years agoosvideosink: create, destroy, resize and draw from the main thread
Andoni Morales Alastruey [Fri, 11 May 2012 16:24:08 +0000 (18:24 +0200)]
osvideosink: create, destroy, resize and draw from the main thread

12 years agomatroskademux: increase NEWSEGMENT accuracy after seeking
Alessandro Decina [Thu, 19 Apr 2012 06:37:28 +0000 (08:37 +0200)]
matroskademux: increase NEWSEGMENT accuracy after seeking

demux->common.segment is populated during seek handling with the target
start/stop positions. Don't override them when sending out a NEWSEGMENT.

Conflicts:

gst/matroska/matroska-demux.c

12 years agomatroskademux: don't discard the incoming seek segment on push based seeking
Alessandro Decina [Thu, 19 Apr 2012 06:31:00 +0000 (08:31 +0200)]
matroskademux: don't discard the incoming seek segment on push based seeking

The incoming seek segment was being discarded leading to push based seeking
being potentially inaccurate.

12 years agocommon: Update so the plugin scanner changes are included
Sebastian Rasmussen [Wed, 23 May 2012 16:12:24 +0000 (18:12 +0200)]
common: Update so the plugin scanner changes are included

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=676674

12 years agoconfigure: suppress some warnings when debug is disabled
Sebastian Rasmussen [Wed, 23 May 2012 16:07:35 +0000 (18:07 +0200)]
configure: suppress some warnings when debug is disabled

Warnings about unused variables should be suppressed if core has the
debug system disabled.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=676671

12 years agortp: fix build issue in gstrtph264pay.c
Luis de Bethencourt [Thu, 24 May 2012 08:29:25 +0000 (09:29 +0100)]
rtp: fix build issue in gstrtph264pay.c

12 years agortph264pay: Add unrestricted caps
Jonas Holmberg [Mon, 21 May 2012 10:17:35 +0000 (12:17 +0200)]
rtph264pay: Add unrestricted caps

If there are no profile restrictions downstream, return caps with
profile=constrained-baseline in the first structure and append
unrestricted caps as the last structure.

Fixes bug #672019

12 years agortsp: add the Scale header when needed
Maria Giovanna Chiossa [Thu, 24 May 2012 07:57:31 +0000 (09:57 +0200)]
rtsp: add the Scale header when needed

Setting GST_SEEK_FLAG_SKIP when sending a seek event in rtspsrc should
set the "Scale" field in the rtsp PLAY header.
Because the boolean "src->skip" is set after the call, "Speed" instead
of "Scale" is always set. Move the assignment before issuing the _play
request.

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=676618

12 years agovideobox: Fix the sample pipeline.
Sreerenj Balachandran [Thu, 17 May 2012 13:23:59 +0000 (16:23 +0300)]
videobox: Fix the sample pipeline.

12 years agoicydemux: warning if setting srcpad caps fails
Anton Novikov [Tue, 22 May 2012 08:35:04 +0000 (12:35 +0400)]
icydemux: warning if setting srcpad caps fails

12 years agoicydemux: activate srcpad before setting caps
Anton Novikov [Tue, 22 May 2012 08:35:29 +0000 (12:35 +0400)]
icydemux: activate srcpad before setting caps

Before gst_pad_set_active() is called, the pad has
FLUSHING flag set, so setting the caps fails

12 years agomng: remove ext/libmng
Luis de Bethencourt [Tue, 22 May 2012 12:46:27 +0000 (13:46 +0100)]
mng: remove ext/libmng

Port to 0.10 was never finished.
Interest was lost.

https://bugzilla.gnome.org/show_bug.cgi?id=324364

12 years agoavimux: fix assertion when handling a date tag as a string
Thiago Santos [Fri, 18 May 2012 19:37:04 +0000 (16:37 -0300)]
avimux: fix assertion when handling a date tag as a string

Date tags are GDate, not strings. Add a special case to convert
it to the exif date format representation in string to avoid
the assertion

12 years agopulsesrc: Listen to source output events, not sink input
Sjoerd Simons [Mon, 21 May 2012 09:47:07 +0000 (11:47 +0200)]
pulsesrc: Listen to source output events, not sink input

12 years agortpmp2tpay: respect mtu and packet boundaries
Mark Nauwelaerts [Fri, 18 May 2012 10:53:44 +0000 (12:53 +0200)]
rtpmp2tpay: respect mtu and packet boundaries

See #659915.

12 years agojpeg: Remove dead code
Edward Hervey [Fri, 18 May 2012 09:10:46 +0000 (11:10 +0200)]
jpeg: Remove dead code

Conflicts:

ext/jpeg/gstjpegdec.c

12 years agojpegdec: Fix compilation
Sebastian Dröge [Fri, 18 May 2012 09:05:35 +0000 (11:05 +0200)]
jpegdec: Fix compilation

12 years agojpegdec: When dropping frames on EOS, flush out data
Edward Hervey [Fri, 18 May 2012 09:02:52 +0000 (11:02 +0200)]
jpegdec: When dropping frames on EOS, flush out data

Cleaner way of handling stray data

12 years agojpegdec: Remove unused variable
Edward Hervey [Thu, 17 May 2012 07:34:03 +0000 (09:34 +0200)]
jpegdec: Remove unused variable

Conflicts:

ext/jpeg/gstjpegdec.c

12 years agojpegdec: Only parse for SOI when we didn't see it before
Edward Hervey [Thu, 17 May 2012 07:33:18 +0000 (09:33 +0200)]
jpegdec: Only parse for SOI when we didn't see it before

12 years agojpegdec: Remember if we saw SOI and handle stray data on EOS
Edward Hervey [Thu, 17 May 2012 07:31:41 +0000 (09:31 +0200)]
jpegdec: Remember if we saw SOI and handle stray data on EOS

12 years agortpjpegpay: Allow U and V components to use different quant tables if they contain...
Youness Alaoui [Tue, 15 May 2012 20:58:25 +0000 (20:58 +0000)]
rtpjpegpay: Allow U and V components to use different quant tables if they contain the same data

This allows some cameras (Logitech C920) that specify different quant
tables but both with the same data, to work.
Bug reported by Robert Krakora

12 years agosouphttpsrc: fix possible data corruption after seeking
Tim-Philipp Müller [Mon, 14 May 2012 14:51:29 +0000 (15:51 +0100)]
souphttpsrc: fix possible data corruption after seeking

Consider a downstream element that may issue seeks in very short
succession (e.g. queue2), depending on the access pattern of
the downstream element (e.g. qtdemux with audio/video chunks
interleaved so that there's always a sizeable gap between the
current chunks for each stream). In this case, queue2 will maintain
two ranges, and even when it serves a chunk from memory, it will
switch ranges and make souphttpsrc seek to the end of the available
data for that range, assuming that that's where we'll want to
continue reading from next.

This may lead to the following seek request pattern:

 - source reading position A
 - seek to B
 - now reading position still A, requested_postion is B
 - streaming thread to be restarted to continue from B
 - seek to A, before streaming thread had time to do the seek
 - do_seek() now sees reading position == seek position and
   returns early.
 - however, requested position is still B from the earlier
   seek request
 - streaming thread starts up, sees that a seek to B is pending
   and requests data from B from the server, while the GstBaseSrc
   segment has of course been updated/reset to position A, which
   was the last seek request.
 - we will now send data for position B and pretend that's the
   data from position A (via the newsegment event, etc.)
 - this causes data corruption

Reproducible doing seek-emulated fast-forward/backward on 006648.

12 years agoconfigure: Require core/base 0.11.91
Sebastian Dröge [Wed, 16 May 2012 07:12:55 +0000 (09:12 +0200)]
configure: Require core/base 0.11.91

12 years ago.gitignore: add visual studio IDE files and OS X .DS_Store files
Matej Knopp [Fri, 13 Jan 2012 23:09:50 +0000 (18:09 -0500)]
.gitignore: add visual studio IDE files and OS X .DS_Store files

https://bugzilla.gnome.org/show_bug.cgi?id=667899

12 years agojpeg: Port to 0.11 again
Sebastian Dröge [Thu, 3 May 2012 07:32:50 +0000 (09:32 +0200)]
jpeg: Port to 0.11 again

12 years agojpeg: Port jpegdec/jpegenc to base video classes
Edward Hervey [Fri, 6 Apr 2012 10:13:24 +0000 (12:13 +0200)]
jpeg: Port jpegdec/jpegenc to base video classes

Conflicts:

ext/jpeg/gstjpegdec.c
ext/jpeg/gstjpegdec.h
ext/jpeg/gstjpegenc.c
ext/jpeg/gstjpegenc.h

Reverted to 0.10 versions for now, next port again.

12 years agoannodex: remove annodex plugin and CMML elements
Tim-Philipp Müller [Sun, 13 May 2012 18:21:19 +0000 (19:21 +0100)]
annodex: remove annodex plugin and CMML elements

This never really took off and is most likely completely
unused. If there is still a need for this, it should
probably be done differently, perhaps inside oggdemux/mux.

12 years agoBack to development
Sebastian Dröge [Sun, 13 May 2012 14:59:58 +0000 (16:59 +0200)]
Back to development

12 years agoRelease 0.11.91
Sebastian Dröge [Sun, 13 May 2012 14:31:03 +0000 (16:31 +0200)]
Release 0.11.91

12 years agoUpdate .po files
Sebastian Dröge [Sun, 13 May 2012 14:30:03 +0000 (16:30 +0200)]
Update .po files

12 years agoAutomatic update of common submodule
Sebastian Dröge [Sun, 13 May 2012 13:56:05 +0000 (15:56 +0200)]
Automatic update of common submodule

From dc70203 to 3429ba6

12 years agorndbuffersize: only send flush-stop if it was a flushing seek
Tim-Philipp Müller [Wed, 9 May 2012 14:14:55 +0000 (15:14 +0100)]
rndbuffersize: only send flush-stop if it was a flushing seek

12 years agov4l2src: fix v4l2_std_id logging
Peter Seiderer [Wed, 9 May 2012 10:54:11 +0000 (12:54 +0200)]
v4l2src: fix v4l2_std_id logging

input.std is of type v4l2_std_id which is defined as 64-bit unsigned integer.
Casting to uint means the higher bits, wich are used for the private video
standards of the TI video capture/display driver for example, are lost.

12 years agorndbuffersize: must send flush-stop after acquiring the stream lock
Tim-Philipp Müller [Wed, 9 May 2012 11:24:37 +0000 (12:24 +0100)]
rndbuffersize: must send flush-stop after acquiring the stream lock

Otherwise the streaming thread might just keep on going and we
might never get the stream lock.

12 years agorndbuffersize: port seeking code to 0.11
Tim-Philipp Müller [Wed, 9 May 2012 10:15:21 +0000 (11:15 +0100)]
rndbuffersize: port seeking code to 0.11

12 years agorndbuffersize: add support for seeks
Tim-Philipp Müller [Tue, 8 May 2012 18:07:04 +0000 (19:07 +0100)]
rndbuffersize: add support for seeks

Useful for e.g. filesrc ! rndbuffersize ! queue2 ! ...

12 years agorndbuffersize: send SEGMENT event before pushing buffers
Tim-Philipp Müller [Tue, 8 May 2012 17:45:34 +0000 (18:45 +0100)]
rndbuffersize: send SEGMENT event before pushing buffers

Conflicts:

gst/debugutils/rndbuffersize.c

12 years agointerleave: fix compilation again
Wim Taymans [Wed, 9 May 2012 09:15:57 +0000 (11:15 +0200)]
interleave: fix compilation again

12 years agortpsession: creation should be signaled before validation
Pascal Buhler [Fri, 13 Jan 2012 09:49:43 +0000 (10:49 +0100)]
rtpsession: creation should be signaled before validation

https://bugzilla.gnome.org/show_bug.cgi?id=667850

12 years agojpegenc: do not proxy our filter caps downstream on caps queries
Thiago Santos [Fri, 4 May 2012 18:20:47 +0000 (15:20 -0300)]
jpegenc: do not proxy our filter caps downstream on caps queries

Downstream likely won't accept video/x-raw and the caps query
will return EMPTY caps. Instead, create a copy of the caps that
has all structure names replaced by 'image/jpeg'

Simple pipeline that shows the problem:
gst-launch-1.0 videotestsrc num-buffers=1 ! "video/x-raw, \
  width=(int)640, height=(int)480" ! videoscale ! jpegenc ! \
  "image/jpeg, width=(int)800, height=(int)600" ! filesink \
  location=/tmp/image.jpg

12 years agoisomp4: set layout=interleaved on raw audio caps
Alban Browaeys [Wed, 2 May 2012 19:17:43 +0000 (21:17 +0200)]
isomp4: set layout=interleaved on raw audio caps

This fixes a not-negotiated error at least on mov files with
twos audio with two channels and video dvcp. As playbin and gst-launch
sample coming from the qtdemux.c file uses audioconvert and the latter
require format interleaved.

https://bugzilla.gnome.org/show_bug.cgi?id=675326