platform/upstream/gstreamer.git
9 years agoglimagesink: only attempt a resize when the window has already resized
Matthew Waters [Thu, 27 Nov 2014 02:01:19 +0000 (13:01 +1100)]
glimagesink: only attempt a resize when the window has already resized

fixes a black startup screen on wayland

9 years agogl/wayland: implement resizing the window using the right mouse button
Matthew Waters [Thu, 27 Nov 2014 00:47:38 +0000 (11:47 +1100)]
gl/wayland: implement resizing the window using the right mouse button

9 years agobuild: Use newer AX_PTHREAD instead of ACX_PTHREAD
Arun Raghavan [Wed, 26 Nov 2014 16:35:10 +0000 (22:05 +0530)]
build: Use newer AX_PTHREAD instead of ACX_PTHREAD

9 years agoRemove unported and obsolete qtwrapper plugin code
Tim-Philipp Müller [Wed, 26 Nov 2014 14:36:42 +0000 (14:36 +0000)]
Remove unported and obsolete qtwrapper plugin code

QuickTime is deprecated, we want new plugins based
on avfoundation (osx) or dxva/mediafoundation (windows).

9 years agoglfilter: support fixed dimensions on both sides of the element
Matthew Waters [Wed, 26 Nov 2014 14:27:19 +0000 (01:27 +1100)]
glfilter: support fixed dimensions on both sides of the element

Fixes:

  width=320,height=240 ! glfilter ! width=800,height=600
  width=230,height=240 ! glfilter ! width=600
  ... ! glfilter ! width=800

9 years agogl: fix automagic disabling of desktop GL
Matthew Waters [Tue, 25 Nov 2014 02:33:07 +0000 (13:33 +1100)]
gl: fix automagic disabling of desktop GL

If configure finds GL + GLES2 but the user passes --enable-gles2 and
the two GL API's cannot be built against together, configure was still
allowing the desktop GL stack to be built.

9 years agovideoaggregator: Also sync pad properties to the controller if conversion is disabled
Sebastian Dröge [Wed, 26 Nov 2014 12:06:21 +0000 (13:06 +0100)]
videoaggregator: Also sync pad properties to the controller if conversion is disabled

9 years agocompositor: Remove unused zorder pad property
Sebastian Dröge [Wed, 26 Nov 2014 11:35:52 +0000 (12:35 +0100)]
compositor: Remove unused zorder pad property

It's handled in videoaggregator already.

9 years agocodecparsers: fix some compiler warnings
Julien Isorce [Sun, 23 Nov 2014 23:49:50 +0000 (23:49 +0000)]
codecparsers: fix some compiler warnings

i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1

"warning: comparison is always true due to limited
range of data type"

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=740673

9 years agocompositor: GstVideoMeta is supported just fine, tell upstream about that
Sebastian Dröge [Tue, 25 Nov 2014 17:53:55 +0000 (18:53 +0100)]
compositor: GstVideoMeta is supported just fine, tell upstream about that

Also provide a GstVideoBufferPool to upstream just in case.

9 years agowinks: fix debug message format issues
Tim-Philipp Müller [Mon, 24 Nov 2014 13:25:55 +0000 (13:25 +0000)]
winks: fix debug message format issues

And print last bits of GUID properly instead
of printing nonsense.

https://bugzilla.gnome.org/show_bug.cgi?id=733055

9 years agompegts: fix some compiler warnings
Julien Isorce [Tue, 25 Nov 2014 08:24:16 +0000 (08:24 +0000)]
mpegts: fix some compiler warnings

i686-apple-darwin11-llvm-gcc-4.2 (GCC) 4.2.1

"warning: comparison is always true due to limited
range of data type"

https://bugzilla.gnome.org/show_bug.cgi?id=740674

9 years agoconfigure: use pkg-config to detect X11
Tim-Philipp Müller [Mon, 24 Nov 2014 19:53:56 +0000 (19:53 +0000)]
configure: use pkg-config to detect X11

Like we do in -base and -good now.

https://bugzilla.gnome.org/show_bug.cgi?id=731047

9 years agoopenh264enc: Use a fixed number of slices by default
Sebastian Dröge [Mon, 24 Nov 2014 18:21:29 +0000 (19:21 +0100)]
openh264enc: Use a fixed number of slices by default

9 years agoopenh264enc: Add some more encoding properties
Sebastian Dröge [Mon, 24 Nov 2014 18:16:20 +0000 (19:16 +0100)]
openh264enc: Add some more encoding properties

Merged from https://github.com/EricssonResearch/openwebrtc-gst-plugins
based on a patchset by Robert Swain <robert.swain@ericsson.com>

9 years agoglcontext: Fix unused variable warning by moving declaration where it is actually...
Ramprakash Jelari [Mon, 24 Nov 2014 13:10:53 +0000 (18:40 +0530)]
glcontext: Fix unused variable warning by moving declaration where it is actually used

9 years agodecklink: Add support for 2k modes
Sebastian Dröge [Mon, 24 Nov 2014 13:07:19 +0000 (14:07 +0100)]
decklink: Add support for 2k modes

9 years agodecklink: HD1080i50 is 50 frames per second
Sebastian Dröge [Mon, 24 Nov 2014 12:56:56 +0000 (13:56 +0100)]
decklink: HD1080i50 is 50 frames per second

9 years agompegtspacketizer: Fix format string compiler warning
Ramprakash Jelari [Mon, 24 Nov 2014 12:28:42 +0000 (17:58 +0530)]
mpegtspacketizer: Fix format string compiler warning

9 years agodvbsection: Fix format string compiler warning
Ramprakash Jelari [Mon, 24 Nov 2014 11:48:44 +0000 (17:18 +0530)]
dvbsection: Fix format string compiler warning

9 years agompegtsdescriptor: Fix format string compiler warning
Ramprakash Jelari [Mon, 24 Nov 2014 11:24:42 +0000 (16:54 +0530)]
mpegtsdescriptor: Fix format string compiler warning

gstmpegtsdescriptor.c:778:16: warning: format specifies type 'unsigned long' but the argument has type 'int' [-Wformat]

9 years agoglupload: Fix valid compiler warning
Sebastian Dröge [Mon, 24 Nov 2014 10:32:33 +0000 (11:32 +0100)]
glupload: Fix valid compiler warning

gstglupload.c:442:32: error: if statement has empty body [-Werror,-Wempty-body]
    if (upload->texture_ids[i]);
                               ^

9 years agogl: ship the gstglcontext_egl.h header
Philippe Normand [Mon, 24 Nov 2014 09:43:23 +0000 (10:43 +0100)]
gl: ship the gstglcontext_egl.h header

It is required by gsteglimagememory.h.

https://bugzilla.gnome.org/show_bug.cgi?id=740611

9 years agogl: shadervariables: make parsing of floats locale-independent
Tim-Philipp Müller [Sun, 23 Nov 2014 21:43:20 +0000 (21:43 +0000)]
gl: shadervariables: make parsing of floats locale-independent

Floating point numbers are written differently in different
locales, e.g. in many countries 1/2 = 0,5 instead of 0.5, and
strtod will not be able to parse "0.5" correctly in such a
locale.

9 years agoglshadervariables: also trim \r as well as \n
Matthew Waters [Sun, 23 Nov 2014 15:52:24 +0000 (02:52 +1100)]
glshadervariables: also trim \r as well as \n

9 years agoglupload: rearchitecture for non GLMemory inputs/outputs
Matthew Waters [Wed, 5 Nov 2014 09:18:06 +0000 (20:18 +1100)]
glupload: rearchitecture for non GLMemory inputs/outputs

Allows other memory types to be implemented/returned/used by the caller.

9 years agorawparse: send SEEKING query upstream first
Wonchul Lee [Fri, 14 Nov 2014 01:55:14 +0000 (10:55 +0900)]
rawparse: send SEEKING query upstream first

Sometimes rawparse does not handle the seeking query
properly, the rawparse should send the query upstream
first. For example, upstream could support seeking in
TIME format (but not in BYTE format), so the BYTE format
seeking query that rawparse sends in push mode would
fail.

https://bugzilla.gnome.org/show_bug.cgi?id=722764

9 years agovtenc: let VT pick the bitrate by default
Alessandro Decina [Fri, 21 Nov 2014 14:46:49 +0000 (15:46 +0100)]
vtenc: let VT pick the bitrate by default

9 years agosrtpdec: fix rtp/rtcp sink event memory leak
Aleix Conchillo Flaqué [Thu, 20 Nov 2014 18:23:30 +0000 (10:23 -0800)]
srtpdec: fix rtp/rtcp sink event memory leak

https://bugzilla.gnome.org/show_bug.cgi?id=740451

9 years agovtenc: set H264_Baseline_AutoLevel on OSX as well
Alessandro Decina [Thu, 20 Nov 2014 15:04:34 +0000 (16:04 +0100)]
vtenc: set H264_Baseline_AutoLevel on OSX as well

9 years agovtdec: disable the reorder queue for baseline streams
Alessandro Decina [Thu, 20 Nov 2014 14:50:18 +0000 (15:50 +0100)]
vtdec: disable the reorder queue for baseline streams

9 years agovtdec: Fix divide by zero in the picture buffer length calculation
Sebastian Dröge [Thu, 20 Nov 2014 08:41:43 +0000 (09:41 +0100)]
vtdec: Fix divide by zero in the picture buffer length calculation

9 years agoglimagesink: critical error while seek playback-test(stop state)
Vineeth T M [Thu, 20 Nov 2014 03:43:58 +0000 (09:13 +0530)]
glimagesink: critical error while seek playback-test(stop state)

If we seek when media is in stop state, playback-test gives
critical error, since context of glimagesink is destroyed during stop.
But since context is not present, we need not handle send_event in glimagesink
Hence adding a condition to check if context is valid.

https://bugzilla.gnome.org/show_bug.cgi?id=740305

9 years agoaggregator: Unblock events/queries immediately if the pad is flushing
Sebastian Dröge [Wed, 19 Nov 2014 16:17:06 +0000 (17:17 +0100)]
aggregator: Unblock events/queries immediately if the pad is flushing

https://bugzilla.gnome.org/show_bug.cgi?id=740376

9 years agoaggregator: Drop serialized events/queries if the pad is flushing
Sebastian Dröge [Wed, 19 Nov 2014 16:15:02 +0000 (17:15 +0100)]
aggregator: Drop serialized events/queries if the pad is flushing

https://bugzilla.gnome.org/show_bug.cgi?id=740376

9 years agoaggregator: Block serialized events/queries until the pad has consumed all buffers
Sebastian Dröge [Wed, 19 Nov 2014 16:03:41 +0000 (17:03 +0100)]
aggregator: Block serialized events/queries until the pad has consumed all buffers

Otherwise the caps of the pad might change while the subclass still works with
a buffer of the old caps, assuming the the current pad caps apply to that
buffer. Which then leads to crashes and other nice effects.

https://bugzilla.gnome.org/show_bug.cgi?id=740376

9 years agovideoaggregator: Don't output 0-duration buffers at the segment end
Sebastian Dröge [Wed, 19 Nov 2014 16:02:40 +0000 (17:02 +0100)]
videoaggregator: Don't output 0-duration buffers at the segment end

https://bugzilla.gnome.org/show_bug.cgi?id=740376

9 years agoaggregator: Fix typo in debug output
Sebastian Dröge [Wed, 19 Nov 2014 16:03:33 +0000 (17:03 +0100)]
aggregator: Fix typo in debug output

9 years agovideoaggregator: fix up QoS handling for live sources
Matthew Waters [Mon, 17 Nov 2014 03:05:01 +0000 (14:05 +1100)]
videoaggregator: fix up QoS handling for live sources

Only attempt adaptive drop when we are not live

https://bugzilla.gnome.org/show_bug.cgi?id=739996

9 years agoaggregator: add _get_latency() for subclass usage
Matthew Waters [Mon, 17 Nov 2014 03:00:10 +0000 (14:00 +1100)]
aggregator: add _get_latency() for subclass usage

API: gst_aggregator_get_latency

https://bugzilla.gnome.org/show_bug.cgi?id=739996

9 years agoglcontext: build errors when GST_GL_HAVE_OPENGL is 0
Vineeth T M [Mon, 17 Nov 2014 04:06:52 +0000 (09:36 +0530)]
glcontext: build errors when GST_GL_HAVE_OPENGL is 0

Moving variables within #if GST_GL_HAVE_OPENGL, which are not going to be used otherwise.
and which are needed only when OpenGl is present

https://bugzilla.gnome.org/show_bug.cgi?id=740235

9 years agoexamples: relicense opencv gsthanddetect_test to LGPLv2+
Sreerenj Balachandran [Mon, 17 Nov 2014 09:10:35 +0000 (11:10 +0200)]
examples: relicense opencv gsthanddetect_test to LGPLv2+

The GPL3 license header was copy'n'pasted from a
neighbouring source file by mistake, the original
code was not GPL3 licensed, but fell under the
default GStreamer license, which is LGPLv2+.

https://bugzilla.gnome.org/show_bug.cgi?id=685655

9 years agoglfilter: add read-only context property
Matthew Waters [Mon, 17 Nov 2014 07:50:28 +0000 (18:50 +1100)]
glfilter: add read-only context property

9 years agoglmixer: add read-only context property
Matthew Waters [Mon, 17 Nov 2014 07:50:04 +0000 (18:50 +1100)]
glmixer: add read-only context property

9 years agoopencv-example: Use g_snprintf instead of sprintf
Hyunjun Ko [Tue, 11 Nov 2014 08:08:25 +0000 (17:08 +0900)]
opencv-example: Use g_snprintf instead of sprintf

https://bugzilla.gnome.org/show_bug.cgi?id=739846

9 years agodvbsrc: drop explicit comparisons against 0 for true/false
Reynaldo H. Verdejo Pinochet [Sun, 16 Nov 2014 23:32:56 +0000 (20:32 -0300)]
dvbsrc: drop explicit comparisons against 0 for true/false

9 years agogl/cocoa: Only use convertRectToBacking on OSX >= 10.7
Sebastian Dröge [Sun, 16 Nov 2014 10:00:14 +0000 (11:00 +0100)]
gl/cocoa: Only use convertRectToBacking on OSX >= 10.7

It does not exist before and older versions also don't have
support for HiDPI displays anyway.

https://bugzilla.gnome.org/show_bug.cgi?id=740201

9 years agogl: Use numeric OSX version instead of the macro
Sebastian Dröge [Sun, 16 Nov 2014 09:57:55 +0000 (10:57 +0100)]
gl: Use numeric OSX version instead of the macro

The macro is not defined on older OSX versions and evaluates to 0.

https://bugzilla.gnome.org/show_bug.cgi?id=740201

9 years agodvbbasebin: fix possible crash by passing 64 bits for 64-bit queue property
Tim-Philipp Müller [Sat, 15 Nov 2014 21:59:48 +0000 (21:59 +0000)]
dvbbasebin: fix possible crash by passing 64 bits for 64-bit queue property

https://bugzilla.gnome.org/show_bug.cgi?id=740191

9 years agodecklink: Fix indention
Sebastian Dröge [Fri, 14 Nov 2014 09:40:31 +0000 (10:40 +0100)]
decklink: Fix indention

9 years agocodecparsers: h264: fix derivation of MaxPicNum variable.
Gwenole Beauchesne [Mon, 10 Nov 2014 09:01:24 +0000 (10:01 +0100)]
codecparsers: h264: fix derivation of MaxPicNum variable.

The logic to compute the MaxPicNum variable was reversed. In paricular,
MaxPicNum is double MaxFrameNum only if field_pic_flag is set to 1.

https://bugzilla.gnome.org/show_bug.cgi?id=739291

9 years agogl/cocoa: Don't override the application delegate
Sebastian Dröge [Thu, 13 Nov 2014 11:01:25 +0000 (12:01 +0100)]
gl/cocoa: Don't override the application delegate

Otherwise interesting things will happen in Cocoa applications, like
infinite event loops that block the NSApplication loop forever.

This was only needed for GNUStep and thus can safely be removed now.

9 years agogl/cocoa: Remove GNUStep support
Sebastian Dröge [Thu, 13 Nov 2014 10:58:07 +0000 (11:58 +0100)]
gl/cocoa: Remove GNUStep support

Until gcc and GNUStep properly support Objective-C blocks and other
"new" features of Objective-C we can't properly support them without
making the code much more ugly.

https://bugzilla.gnome.org/show_bug.cgi?id=739152

9 years agointeraudiosink: Make sure to store at least period-time samples at once in the adapter
Sebastian Dröge [Wed, 12 Nov 2014 17:06:45 +0000 (18:06 +0100)]
interaudiosink: Make sure to store at least period-time samples at once in the adapter

It's better to create some more silence than having the source pull multiple
short buffers out of the adapter and insert silence between them.

9 years agovtenc: fix build on iOS
Alessandro Decina [Wed, 12 Nov 2014 15:45:12 +0000 (16:45 +0100)]
vtenc: fix build on iOS

9 years agocoremediabuffer: set ->pixel_buf for CVPixelBuffer buffers
Alessandro Decina [Wed, 12 Nov 2014 14:35:36 +0000 (15:35 +0100)]
coremediabuffer: set ->pixel_buf for CVPixelBuffer buffers

Fix a bug for which we never actually set ->pixel_buf to non-NULL for
CVPixelBuffer buffers.

9 years agofaad: set channel positions using the appropriate API
Vincent Penquerc'h [Fri, 6 Jun 2014 12:57:30 +0000 (13:57 +0100)]
faad: set channel positions using the appropriate API

https://bugzilla.gnome.org/show_bug.cgi?id=731038

9 years agointervideosrc: Stop showing the current frame after 1 second, not after 30 frames
Sebastian Dröge [Wed, 12 Nov 2014 12:58:56 +0000 (13:58 +0100)]
intervideosrc: Stop showing the current frame after 1 second, not after 30 frames

9 years agotests: add CFLAGS and LADD for glimagesink unit test
Julien Isorce [Wed, 12 Nov 2014 09:41:53 +0000 (09:41 +0000)]
tests: add CFLAGS and LADD for glimagesink unit test

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732694

9 years agoopenh264enc: Add new enable-frame-skip property and disable it by default
Sebastian Dröge [Wed, 12 Nov 2014 10:52:58 +0000 (11:52 +0100)]
openh264enc: Add new enable-frame-skip property and disable it by default

Otherwise e.g. videotestsrc ! openh264enc ! ... will drop every second frame
because otherwise the target bitrate can't be reached without loosing too
much quality.

9 years agovtenc: fix compilation on < 10.9
Alessandro Decina [Wed, 12 Nov 2014 08:59:53 +0000 (09:59 +0100)]
vtenc: fix compilation on < 10.9

9 years agoglimagesink: don't set the upload to NULL on the drain query
Matthew Waters [Wed, 12 Nov 2014 00:44:07 +0000 (11:44 +1100)]
glimagesink: don't set the upload to NULL on the drain query

https://bugzilla.gnome.org/show_bug.cgi?id=732694

9 years agogl: Correct invalid comment text
Hyunjun Ko [Tue, 11 Nov 2014 08:04:48 +0000 (17:04 +0900)]
gl: Correct invalid comment text

9 years agogl: add unit test that checks for glimagesink drain query handling
Julien Isorce [Tue, 11 Nov 2014 23:54:44 +0000 (23:54 +0000)]
gl: add unit test that checks for glimagesink drain query handling

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732694

9 years agoglimagesink: release stored buffers on drain query
Julien Isorce [Tue, 11 Nov 2014 23:49:59 +0000 (23:49 +0000)]
glimagesink: release stored buffers on drain query

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=732694

9 years agogl: do not raise a critical msg if the backend does not handle window events
Julien Isorce [Tue, 11 Nov 2014 23:43:42 +0000 (23:43 +0000)]
gl: do not raise a critical msg if the backend does not handle window events

Fix "assertion 'window_class->handle_events != NULL' failed"
if not using a X11 window.

9 years agovtenc: enable the hw encoder on OSX
Alessandro Decina [Tue, 11 Nov 2014 17:01:54 +0000 (18:01 +0100)]
vtenc: enable the hw encoder on OSX

9 years agodecklinksink: Check the correct variable for failure
Sebastian Dröge [Tue, 11 Nov 2014 11:00:30 +0000 (12:00 +0100)]
decklinksink: Check the correct variable for failure

If we can't get an output, we would otherwise just use it later and crash.

9 years agointeraudiosrc: Make sure we have a writable buffer before we start changing metadata
Sebastian Dröge [Tue, 11 Nov 2014 10:18:19 +0000 (11:18 +0100)]
interaudiosrc: Make sure we have a writable buffer before we start changing metadata

9 years agointervideosrc: Set the GAP flag on repeated buffers
Sebastian Dröge [Tue, 11 Nov 2014 10:15:35 +0000 (11:15 +0100)]
intervideosrc: Set the GAP flag on repeated buffers

videorate does the same and it allows optimizations downstream.

9 years agointeraudiosrc: Set the GAP flag on buffers that only contain silence
Sebastian Dröge [Tue, 11 Nov 2014 10:12:55 +0000 (11:12 +0100)]
interaudiosrc: Set the GAP flag on buffers that only contain silence

9 years agogl: remove the width/height fields from the caps to support frame resizing
Matthew Waters [Tue, 11 Nov 2014 03:23:55 +0000 (14:23 +1100)]
gl: remove the width/height fields from the caps to support frame resizing

It was previously only occuring with sysmem caps features

https://bugzilla.gnome.org/show_bug.cgi?id=739334

9 years agoh264parser: fix GstH264ParserResult documentation typo
Aurélien Zanelli [Mon, 10 Nov 2014 16:42:20 +0000 (17:42 +0100)]
h264parser: fix GstH264ParserResult documentation typo

https://bugzilla.gnome.org/show_bug.cgi?id=739906

9 years agotests: Add x265enc unit test to .gitignore
Sebastian Dröge [Mon, 10 Nov 2014 15:01:25 +0000 (16:01 +0100)]
tests: Add x265enc unit test to .gitignore

9 years agopngparse: optimise reading of png files
Tim-Philipp Müller [Sun, 9 Nov 2014 20:53:34 +0000 (20:53 +0000)]
pngparse: optimise reading of png files

Read PNG data chunk in one go by letting the parser
base class know the size we need, so that it doesn't
drip-feed us small chunks of data (causing a lot of
reallocs and memcpy in the process) until we have
everything.

Improves parsing performance of very large PNG files
(65MB) from ~13 seconds to a couple of millisecs.

https://bugzilla.gnome.org/show_bug.cgi?id=736176

9 years agoexamples: remove pointless mpegtsmux example
Tim-Philipp Müller [Sun, 9 Nov 2014 19:23:47 +0000 (19:23 +0000)]
examples: remove pointless mpegtsmux example

Serves no purpose, is not even hooked up to the
build system, has hard coded file names and paths,
and can easily be replaced with a gst-launch line.

https://bugzilla.gnome.org/show_bug.cgi?id=739844

9 years agoavfvideosrc: Don't pretend to support reconfiguration
Sebastian Dröge [Sun, 9 Nov 2014 09:25:25 +0000 (10:25 +0100)]
avfvideosrc: Don't pretend to support reconfiguration

We will run into an assertion in set_caps() if we try to change
caps while the source is already running. Don't try to find new
caps in GstBaseSrc::negotiate() to prevent caps changes.

9 years agovc1parse: set seq_layer_sent to FALSE on reset()
Aurélien Zanelli [Thu, 9 Oct 2014 11:11:48 +0000 (13:11 +0200)]
vc1parse: set seq_layer_sent to FALSE on reset()

https://bugzilla.gnome.org/show_bug.cgi?id=738526

9 years agovc1parse: implement asf to *-frame-layer stream-format
Aurélien Zanelli [Thu, 9 Oct 2014 11:03:56 +0000 (13:03 +0200)]
vc1parse: implement asf to *-frame-layer stream-format

This commit add an helper to convert a frame to frame-layer format and
use it to implement these two stream-format conversion:
- asf --> sequence-layer-frame-layer
- asf --> frame-layer

In simple/main profile, we basically have a raw frame, so building a
frame layer isn't too complicated. But in advanced profile, the first
frame-layer should contain sequence-header, entrypoint, and frame and
each keyframe should contain entrypoint, so we have to handle these
carefully.

https://bugzilla.gnome.org/show_bug.cgi?id=738526

9 years agovc1parse: check output format at negotiation time
Aurélien Zanelli [Thu, 9 Oct 2014 08:38:40 +0000 (10:38 +0200)]
vc1parse: check output format at negotiation time

Add an helper to check that output stream-format is coherent with
profile and header-format. It also check if we know how to do the
conversion if the input stream-format differs from selected
output-format.
So, in case output stream-format is not allowed, it will now fail at
negotiation rather than in pre_push_frame.

https://bugzilla.gnome.org/show_bug.cgi?id=738526

9 years agovc1parse: add some asf related stream-format conversions
Aurélien Zanelli [Tue, 30 Sep 2014 13:35:38 +0000 (15:35 +0200)]
vc1parse: add some asf related stream-format conversions

This commit introduces an helper to convert an ASF frame to BDUs format with
startcodes and use this helper to implements following stream-format
conversions:
- asf --> bdu
- asf --> sequence-layer-bdu
- asf --> sequence-layer-raw-frame

https://bugzilla.gnome.org/show_bug.cgi?id=738526

9 years agovc1parse: add some simple stream-format conversion
Aurélien Zanelli [Tue, 30 Sep 2014 13:28:55 +0000 (15:28 +0200)]
vc1parse: add some simple stream-format conversion

It add the support of following stream-format conversion:
- bdu --> sequence-layer-bdu
- bdu-frame --> sequence-layer-bdu-frame
- frame-layer --> sequence-layer-frame-layer

For these conversion, the only requirements is to push a sequence-layer
buffer prior to data.

https://bugzilla.gnome.org/show_bug.cgi?id=738526

9 years agovc1parse: prepare the stream-format conversion code and add the simplest ones
Aurélien Zanelli [Tue, 30 Sep 2014 12:53:25 +0000 (14:53 +0200)]
vc1parse: prepare the stream-format conversion code and add the simplest ones

It prepares the template for stream-format conversion and it implements
the following conversion:
- sequence-layer-bdu --> bdu
- sequence-layer-bdu-frame --> bdu-frame
- sequence-layer-frame-layer --> frame-layer

Work is done in the pre_push_frame() method.

https://bugzilla.gnome.org/show_bug.cgi?id=738526

9 years agodashdemux: do not override last known stream error
Matthieu Bouron [Fri, 5 Sep 2014 15:57:47 +0000 (17:57 +0200)]
dashdemux: do not override last known stream error

https://bugzilla.gnome.org/show_bug.cgi?id=736141

9 years agomssdemux: seeks in the manifest never fail
Thiago Santos [Fri, 7 Nov 2014 22:30:18 +0000 (19:30 -0300)]
mssdemux: seeks in the manifest never fail

Set void as the return and remove error handling as it could
never fail

9 years agodashdemux: make sure flushes have the same seqnum as the seek event
Thiago Santos [Fri, 7 Nov 2014 22:16:47 +0000 (19:16 -0300)]
dashdemux: make sure flushes have the same seqnum as the seek event

They are part of the same operation and should preserve the seqnum

9 years agodashdemux: simplify seek to eos handling
Thiago Santos [Fri, 7 Nov 2014 22:09:15 +0000 (19:09 -0300)]
dashdemux: simplify seek to eos handling

Remove redundant flush stop pushing and remove a goto to make
the code flow simpler

9 years agodashdemux: send missing FLUSH_STOP/EOS when no seeking period is found
Matthieu Bouron [Wed, 17 Sep 2014 13:22:59 +0000 (15:22 +0200)]
dashdemux: send missing FLUSH_STOP/EOS when no seeking period is found

https://bugzilla.gnome.org/show_bug.cgi?id=736800

9 years agoh264parse: set the HEADER flag on buffers containing SPS or PPS
Josep Torra [Fri, 7 Nov 2014 14:50:51 +0000 (15:50 +0100)]
h264parse: set the HEADER flag on buffers containing SPS or PPS

9 years agoh265parse: set the HEADER flag on buffers containing VPS, SPS or PPS
Josep Torra [Fri, 7 Nov 2014 14:41:15 +0000 (15:41 +0100)]
h265parse: set the HEADER flag on buffers containing VPS, SPS or PPS

9 years agogstglconfig: Put gstglconfig.h into $(libdir)/gstreamer-1.0/include
Sebastian Dröge [Fri, 7 Nov 2014 10:41:01 +0000 (11:41 +0100)]
gstglconfig: Put gstglconfig.h into $(libdir)/gstreamer-1.0/include

It's architecture dependent and should not be placed into the include
directory as the assumption is that all those headers are architecture
independent.

https://bugzilla.gnome.org/show_bug.cgi?id=739767

9 years agosmoothstreaming: initialise new GstH264NalUnit fields
Tim-Philipp Müller [Fri, 7 Nov 2014 00:16:12 +0000 (00:16 +0000)]
smoothstreaming: initialise new GstH264NalUnit fields

CID #1251107.

9 years agoglimagesink: don't make gst_glimage_sink_handle_events call _ensure_gl_setup
Alessandro Decina [Thu, 6 Nov 2014 22:58:06 +0000 (23:58 +0100)]
glimagesink: don't make gst_glimage_sink_handle_events call _ensure_gl_setup

gst_glimage_sink_handle_events can be called from the overlay interface and from
the main thread before GL is setup. Before this change, that would call
_ensure_gl_setup() and deadlock on OSX.

Change things so that it's always safe to call gst_glimage_sink_handle_events()
without stuff deadlocking.

9 years agoglimagesink: fix possible deadlock on osx
Alessandro Decina [Thu, 6 Nov 2014 22:55:20 +0000 (23:55 +0100)]
glimagesink: fix possible deadlock on osx

Remove gst_glimage_sink_handle_events call in gst_glimage_sink_init. It was
unnecessary and when the element was instantiated from the main thread, caused a
deadlock in OSX creating the context (thread).

9 years agovtenc: add realtime and allow-frame-reordering properties
Alessandro Decina [Thu, 6 Nov 2014 12:22:17 +0000 (13:22 +0100)]
vtenc: add realtime and allow-frame-reordering properties

9 years agoglimagesink: clamp the resize width/height to >= 1 to avoid a GL error
Matthew Waters [Thu, 6 Nov 2014 23:01:59 +0000 (10:01 +1100)]
glimagesink: clamp the resize width/height to >= 1 to avoid a GL error

9 years agointervideosrc: Print a warning into the debug log on generating black frames
Sebastian Dröge [Thu, 6 Nov 2014 16:36:15 +0000 (17:36 +0100)]
intervideosrc: Print a warning into the debug log on generating black frames

9 years agointervideosink: Add some debug output in the render function
Sebastian Dröge [Thu, 6 Nov 2014 16:21:39 +0000 (17:21 +0100)]
intervideosink: Add some debug output in the render function

9 years agoh265parse: don't unnecesarily set src_caps
Sebastian Dröge [Thu, 6 Nov 2014 10:53:04 +0000 (11:53 +0100)]
h265parse: don't unnecesarily set src_caps

https://bugzilla.gnome.org/show_bug.cgi?id=739374

9 years agoh264parse: don't unnecesarily set src_caps
Matej Knopp [Wed, 29 Oct 2014 21:55:47 +0000 (22:55 +0100)]
h264parse: don't unnecesarily set src_caps

https://bugzilla.gnome.org/show_bug.cgi?id=739374