platform/upstream/gstreamer.git
9 years agohlsdemux: drop TODO that doesn't need a solution
Thiago Santos [Fri, 5 Jun 2015 12:22:58 +0000 (09:22 -0300)]
hlsdemux: drop TODO that doesn't need a solution

Connection speed is only checked at that point in hlsdemux so there
is no real need to refactor it.

https://bugzilla.gnome.org/show_bug.cgi?id=749328

9 years agoRevert "hlsdemux: Simplify logic in process_manifest"
Thiago Santos [Fri, 5 Jun 2015 12:15:34 +0000 (09:15 -0300)]
Revert "hlsdemux: Simplify logic in process_manifest"

This reverts commit 4ca3a22b6b33ad8be4383063e76f79c4d346535d.

The connection-speed=0 is used as a special value in the property
of hlsdemux to mean 'automatic' selection, m3u8.c doesn't need
to know about that as it should be as simple as possible.

So this patch hides this automatic selection documented in hlsdemux
into m3u8 logic and I think the gets harder to understand the code.

It also makes the hlsdemux unit tests work again

https://bugzilla.gnome.org/show_bug.cgi?id=749328

9 years agompdparser: Fix inverted logic introduced in last commit
Sebastian Dröge [Fri, 5 Jun 2015 12:30:03 +0000 (14:30 +0200)]
mpdparser: Fix inverted logic introduced in last commit

9 years agosimplevideomarkdetect: fix detect of videomark partially or fully outside video
Vineeth TM [Thu, 4 Jun 2015 23:58:03 +0000 (08:58 +0900)]
simplevideomarkdetect: fix detect of videomark partially or fully outside video

In case of the videomark being partially or fully outside,
an error was bein thrown saying, mark width is more than video width.
And when the width, offset properties are set to maximum it resulted in crash.
Instead of throwing error, added logic to detect the mark
in case of partial visibility or dont show the mark when it is outside.

https://bugzilla.gnome.org/show_bug.cgi?id=743908

9 years agompdparser: Calculate the number of segments we have when a segment template instead...
Sebastian Dröge [Fri, 5 Jun 2015 12:10:43 +0000 (14:10 +0200)]
mpdparser: Calculate the number of segments we have when a segment template instead of a list is used

Otherwise we would just continue downloading new files forever until we get
404 for the first one, and then error out instead of going EOS.

9 years agosimplevideomark: Add Error logs
Vineeth TM [Thu, 4 Jun 2015 23:53:30 +0000 (08:53 +0900)]
simplevideomark: Add Error logs

When the pattern offset is outside the video, the print error message

https://bugzilla.gnome.org/show_bug.cgi?id=743908

9 years agoRevert "mpdparser: Don't consider streams with a known media presentation duration...
Sebastian Dröge [Fri, 5 Jun 2015 11:28:39 +0000 (13:28 +0200)]
Revert "mpdparser: Don't consider streams with a known media presentation duration as live"

This reverts commit 37011e519830e0786fa6e307ed26003a2a8774f1.

This change was actually completely unnecessary, the streams in question are
marked as static and are not considered live anyway.

9 years agompdparser: Don't consider streams with a known media presentation duration as live
Sebastian Dröge [Fri, 5 Jun 2015 11:21:55 +0000 (13:21 +0200)]
mpdparser: Don't consider streams with a known media presentation duration as live

9 years agoadaptivedemux: Error out if downloading a segment failed too often in live mode
Sebastian Dröge [Fri, 5 Jun 2015 09:26:11 +0000 (11:26 +0200)]
adaptivedemux: Error out if downloading a segment failed too often in live mode

It's true that we shouldn't consider errors fatal immediately, but if we
always ignore them we will loop infinitely on live streams with segments
that can't be downloaded at all.

9 years agosrtp: Allow getting ssrc of more package types than RR and SR
Jose Antonio Santos Cadenas [Wed, 3 Jun 2015 10:17:19 +0000 (12:17 +0200)]
srtp: Allow getting ssrc of more package types than RR and SR

This allows decrypting reduce size packages. See RFC 5506

https://bugzilla.gnome.org/show_bug.cgi?id=750326

9 years agotsdemux: refactor pad creation
Thiago Santos [Thu, 4 Jun 2015 16:36:32 +0000 (13:36 -0300)]
tsdemux: refactor pad creation

Avoid repeating the same pad creation code everywhere

9 years agompdparser: Don't crash in debug output if stream->segments is NULL
Sebastian Dröge [Thu, 4 Jun 2015 13:57:57 +0000 (15:57 +0200)]
mpdparser: Don't crash in debug output if stream->segments is NULL

9 years agoksvideosrc: fix logic and timestamp non-muxed streams again
Tim-Philipp Müller [Thu, 4 Jun 2015 12:31:56 +0000 (13:31 +0100)]
ksvideosrc: fix logic and timestamp non-muxed streams again

https://bugzilla.gnome.org/show_bug.cgi?id=750381

9 years agolibde265dec: fix up plugin name and decoder description
Tim-Philipp Müller [Thu, 4 Jun 2015 11:53:19 +0000 (12:53 +0100)]
libde265dec: fix up plugin name and decoder description

9 years agompdparse: Clamp seek times to the availabilityStartTime
Sebastian Dröge [Thu, 4 Jun 2015 11:36:05 +0000 (13:36 +0200)]
mpdparse: Clamp seek times to the availabilityStartTime

Otherwise we will seek to negative times, which are interpreted as unsigned
integers later.

9 years agompdparser: Improve debug output a bit when advancing segments
Sebastian Dröge [Thu, 4 Jun 2015 11:13:39 +0000 (13:13 +0200)]
mpdparser: Improve debug output a bit when advancing segments

9 years agoadaptivedemux: Release the manifest mutex before waiting for a download
Sebastian Dröge [Thu, 4 Jun 2015 10:36:41 +0000 (12:36 +0200)]
adaptivedemux: Release the manifest mutex before waiting for a download

Otherwise we will hurt parallel downloads in the best case, and often cause
deadlocks.

9 years agoopusdec: If channel/rate negotiation fails, fall back to stereo and 48kHz
Sebastian Dröge [Thu, 4 Jun 2015 09:54:24 +0000 (11:54 +0200)]
opusdec: If channel/rate negotiation fails, fall back to stereo and 48kHz

9 years agoopusparse: Set up default header with 48kHz sample rate instead of 0
Sebastian Dröge [Thu, 4 Jun 2015 09:52:40 +0000 (11:52 +0200)]
opusparse: Set up default header with 48kHz sample rate instead of 0

9 years agoopusdec: gst_structure_fixate_field_nearest_int() only works if the structure has...
Sebastian Dröge [Thu, 4 Jun 2015 09:45:05 +0000 (11:45 +0200)]
opusdec: gst_structure_fixate_field_nearest_int() only works if the structure has this field

Just set the rate/channels directly if the caps don't have this field.

9 years agompdparser: Use 64 bit integer format strings when passing a 64 bit integer
Sebastian Dröge [Thu, 4 Jun 2015 09:05:07 +0000 (11:05 +0200)]
mpdparser: Use 64 bit integer format strings when passing a 64 bit integer

Otherwise we'll only get half of its bits printed on 32 bit architectures.

For this, promote the %d-style format strings to something that accepts
64 bit integers with G_GINT64_MODIFIER.

9 years agompdparse: Validate format strings before using them
Sebastian Dröge [Thu, 4 Jun 2015 08:47:07 +0000 (10:47 +0200)]
mpdparse: Validate format strings before using them

Using format strings from an untrusted source without validation is
calling for problems, and at least allows to remotely crash your application.
If not worse.

9 years agogl/dispmanx: Fix build
Philippe Normand [Thu, 4 Jun 2015 07:40:19 +0000 (09:40 +0200)]
gl/dispmanx: Fix build

9 years agogl/example: fix build error when compiling cocoa-videooverlay
Julien Isorce [Wed, 3 Jun 2015 21:21:30 +0000 (22:21 +0100)]
gl/example: fix build error when compiling cocoa-videooverlay

libtool: error: ignoring unknown tag OBJC
and
clang: error: argument unused during compilation: '-pthread'

9 years agogl/window/dispmanx: fix compiler warning
Matthew Waters [Thu, 4 Jun 2015 05:01:16 +0000 (15:01 +1000)]
gl/window/dispmanx: fix compiler warning

9 years agogl/window: use the default main loop implementation for all backends
Matthew Waters [Thu, 4 Jun 2015 02:16:35 +0000 (12:16 +1000)]
gl/window: use the default main loop implementation for all backends

fixes glimagsink being unable to display.

https://bugzilla.gnome.org/show_bug.cgi?id=750337

9 years agoglwindow_cocoa: use parent default implementation
Julien Isorce [Wed, 3 Jun 2015 20:22:09 +0000 (21:22 +0100)]
glwindow_cocoa: use parent default implementation

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=750337

9 years agompdparser: Add the startNumber as offset to the segment index when using a template
Sebastian Dröge [Wed, 3 Jun 2015 18:03:37 +0000 (20:03 +0200)]
mpdparser: Add the startNumber as offset to the segment index when using a template

9 years agodtlsdec: Fix critical warning "got data flow before stream-start event"
Santiago Carot-Nemesio [Wed, 3 Jun 2015 14:59:44 +0000 (16:59 +0200)]
dtlsdec: Fix critical warning "got data flow before stream-start event"

Forward sticky events on requested src pad.

https://bugzilla.gnome.org/show_bug.cgi?id=750348

9 years agovc1parse: remove useless value assignments
Luis de Bethencourt [Wed, 3 Jun 2015 14:25:10 +0000 (15:25 +0100)]
vc1parse: remove useless value assignments

In all these cases ret is set but overwritten before the return of the function

9 years agoexamples: gst-camerabin2-test: protect from division by zero
Luis de Bethencourt [Wed, 3 Jun 2015 13:44:31 +0000 (14:44 +0100)]
examples: gst-camerabin2-test: protect from division by zero

Highly unlikely to have 0 captures, but protect from crashes in the future by
doing none of the math if there is no data.

9 years agoconfigure: lower curl requirement to 7.35.0
Tim-Philipp Müller [Tue, 2 Jun 2015 19:57:34 +0000 (20:57 +0100)]
configure: lower curl requirement to 7.35.0

Build and unit tests were tested against 7.35 too now,
which is the version in Ubuntu 14.04.

9 years agompegtsbase: do not access variable after unref
Thiago Santos [Tue, 2 Jun 2015 19:40:25 +0000 (16:40 -0300)]
mpegtsbase: do not access variable after unref

Data might not exist anymore

9 years agotests: audiointerleave: test not setting positions
Olivier Crête [Tue, 2 Jun 2015 19:45:13 +0000 (15:45 -0400)]
tests: audiointerleave: test not setting positions

Disable "channel-positions-from-input", but without actually giving
a position table, so every position should be NONE

9 years agotests: Fix indentation in audiointerleave test
Olivier Crête [Tue, 2 Jun 2015 19:44:57 +0000 (15:44 -0400)]
tests: Fix indentation in audiointerleave test

9 years agozbar: don't ignore zbar_scan_image() errors
Reynaldo H. Verdejo Pinochet [Tue, 2 Jun 2015 14:31:00 +0000 (11:31 -0300)]
zbar: don't ignore zbar_scan_image() errors

9 years agoglsyncmeta: Set new sync point when copying
Jan Schmidt [Tue, 2 Jun 2015 14:31:42 +0000 (00:31 +1000)]
glsyncmeta: Set new sync point when copying

Set a new sync point when copying instead of
transferring the (non-refcounted) GL sync object
from the source meta.

https://bugzilla.gnome.org/show_bug.cgi?id=750279

9 years agoglsyncmeta: Add some debug output
Jan Schmidt [Tue, 2 Jun 2015 13:59:50 +0000 (23:59 +1000)]
glsyncmeta: Add some debug output

9 years agocheck: Use GST_CHECK_MAIN () macro everywhere
Edward Hervey [Tue, 2 Jun 2015 14:02:37 +0000 (16:02 +0200)]
check: Use GST_CHECK_MAIN () macro everywhere

Makes source code smaller, and ensures we go through common initialization
path (like the one that sets up XML unit test output ...)

9 years agosimplevideomark: fix display of videomark partially or fully outside video
Vineeth TM [Tue, 2 Jun 2015 01:21:42 +0000 (10:21 +0900)]
simplevideomark: fix display of videomark partially or fully outside video

In case of the videomark being partially or fully outside, an error was being
thrown saying the mark width is more than video width. And when the width,
offset properties are set to maximum it resulted in crash. Instead of throwing
an error, add logic to detect the mark in case of partial visibility or don't
show the mark when it is outside.

https://bugzilla.gnome.org/show_bug.cgi?id=743908

9 years agotsdemux: enable Chinese AVS video stream
Aurélien Zanelli [Tue, 24 Mar 2015 14:17:16 +0000 (15:17 +0100)]
tsdemux: enable Chinese AVS video stream

Chinese broadcaster encapsulate AVS video codec into MPEG2-TS. They
use the stream_id 0x42 to identify AVS video streams. It should be noted
that this id is currently within the ISO reserved range, hence it's
utilisation is unofficial.

https://bugzilla.gnome.org/show_bug.cgi?id=727731

9 years agoglmemory: provide compatibility definition for GLES2 for GL_RGBA8
Matthew Waters [Tue, 2 Jun 2015 07:24:18 +0000 (17:24 +1000)]
glmemory: provide compatibility definition for GLES2 for GL_RGBA8

9 years agogl: consolidate internal_rgba_format into glmemory
Matthew Waters [Tue, 2 Jun 2015 06:32:03 +0000 (16:32 +1000)]
gl: consolidate internal_rgba_format into glmemory

Expose some useful value format conversion functions available in
GstGLMemory.

9 years agoaudiointerleave: Always have "channels" be the actual pad count
Olivier Crête [Mon, 1 Jun 2015 23:43:20 +0000 (19:43 -0400)]
audiointerleave: Always have "channels" be the actual pad count

Don't force it anywhere

https://bugzilla.gnome.org/show_bug.cgi?id=750252

9 years agoaudiointerleave: Use the channel count from the set caps
Olivier Crête [Mon, 1 Jun 2015 23:42:49 +0000 (19:42 -0400)]
audiointerleave: Use the channel count from the set caps

This is the same number that was used to allocate the buffer

9 years agoaggregator: Document that the latency is in ns
Olivier Crête [Mon, 1 Jun 2015 22:50:14 +0000 (18:50 -0400)]
aggregator: Document that the latency is in ns

9 years agoamc: Only lower ranks of OMX.Exynos. audio codecs, the video codecs are actually...
Sebastian Dröge [Mon, 1 Jun 2015 18:03:22 +0000 (20:03 +0200)]
amc: Only lower ranks of OMX.Exynos. audio codecs, the video codecs are actually working

9 years agoamc: Give marginal rank to codecs that start with OMX.Exynos.
Sebastian Dröge [Mon, 1 Jun 2015 17:40:11 +0000 (19:40 +0200)]
amc: Give marginal rank to codecs that start with OMX.Exynos.

OMX.Exynos. codecs are existing on some devices like the
Galaxy S5 mini, and cause random crashes (of the device,
not the app!) and generally misbehave. That specific device
has other codecs that work with a different name, but let's
just give them marginal rank in case there are devices that
have no other codecs and these are actually the only working
ones

9 years agoandroidmedia: Give lower ranks to codecs not starting with OMX.
Sebastian Dröge [Mon, 1 Jun 2015 17:23:01 +0000 (19:23 +0200)]
androidmedia: Give lower ranks to codecs not starting with OMX.

On some devices there are codecs that don't start with OMX., while
there are also some that do. And on some of these devices the ones
that don't start with OMX. just crash during initialization while
the others work. To make things even more complicated other devices
have codecs with the same name that work and no alternatives.
So just give a lower rank to these non-OMX codecs and hope that
there's an alternative with a higher rank.

Also stagefright gives codecs starting with OMX. a higher rank too and
considers other codecs that don't start with OMX. as software codecs.

9 years agotests: zbar: check for frame field on attach-frame=true
Reynaldo H. Verdejo Pinochet [Mon, 1 Jun 2015 14:12:22 +0000 (11:12 -0300)]
tests: zbar: check for frame field on attach-frame=true

Also check for the sample to have a buffer and caps
associated with it.

Related to:

https://bugzilla.gnome.org/show_bug.cgi?id=747557

9 years agovtdec: Require width and height field for H264
Nicolas Dufresne [Wed, 20 May 2015 19:04:40 +0000 (15:04 -0400)]
vtdec: Require width and height field for H264

This decoder does not work if width and height field are not set
in the sinkpad caps. Let's make this explicit by adding them to
the template caps.

https://bugzilla.gnome.org/show_bug.cgi?id=749655

9 years agoglcontext_gpu_process: close the window when done
Julien Isorce [Mon, 1 Jun 2015 13:07:37 +0000 (14:07 +0100)]
glcontext_gpu_process: close the window when done

9 years agoglwindow_gpu_process: avoid empty struct
Julien Isorce [Mon, 1 Jun 2015 13:05:58 +0000 (14:05 +0100)]
glwindow_gpu_process: avoid empty struct

9 years agoglwindow: move main loop/context creation back to init/finalize
Julien Isorce [Mon, 1 Jun 2015 11:58:11 +0000 (12:58 +0100)]
glwindow: move main loop/context creation back to init/finalize

Fixes https://bugzilla.gnome.org/show_bug.cgi?id=750199

9 years agoexamples: Fix gl usage without wayland support
Edward Hervey [Mon, 1 Jun 2015 12:05:06 +0000 (14:05 +0200)]
examples: Fix gl usage without wayland support

Not all platforms have wayland support. Handle that gracefully at
compile time

9 years agortmpsink: Initialize GstMapInfo
Edward Hervey [Mon, 1 Jun 2015 11:55:20 +0000 (13:55 +0200)]
rtmpsink: Initialize GstMapInfo

Avoids doing a call to unmap with it uninitialized

CID #1302834

9 years agoaudiovisualizer: clean dereferences of private structure
Luis de Bethencourt [Mon, 1 Jun 2015 11:51:46 +0000 (12:51 +0100)]
audiovisualizer: clean dereferences of private structure

https://bugzilla.gnome.org/show_bug.cgi?id=742875

9 years agovtenc: fix keyframe request race condition
Ilya Konstantinov [Sat, 25 Apr 2015 19:55:28 +0000 (22:55 +0300)]
vtenc: fix keyframe request race condition

It is incorrect to modify the frame properties after passing them, since
VTCompressionSessionEncodeFrame takes reference and we have no control
over when it's being used.

In fact, the code can be simplified. We just preallocate the frame
properties for keyframe requests, and pass NULL otherwise.

https://bugzilla.gnome.org/show_bug.cgi?id=748467

9 years agoaudiovisualizer: make private all variable subclasses don't need
Luis de Bethencourt [Mon, 1 Jun 2015 10:56:13 +0000 (11:56 +0100)]
audiovisualizer: make private all variable subclasses don't need

https://bugzilla.gnome.org/show_bug.cgi?id=742875

9 years agozbar: add frame sample to barcode message
Reynaldo H. Verdejo Pinochet [Sat, 2 May 2015 02:20:30 +0000 (23:20 -0300)]
zbar: add frame sample to barcode message

New attach-frame property enables barcode frame
dumping when set to true.

https://bugzilla.gnome.org/show_bug.cgi?id=747557

9 years agogl: Fix compiler warning
Sebastian Dröge [Sun, 31 May 2015 19:30:23 +0000 (21:30 +0200)]
gl: Fix compiler warning

gstglwindow_wayland_egl.c:246:41: error: unused variable 'surface_listener' [-Werror,-Wunused-const-variable]
static const struct wl_surface_listener surface_listener = {
                                        ^

9 years agoandroidmedia: Conditionally use get_{input,output}_buffer() Android 21 APIs
Sebastian Dröge [Wed, 27 May 2015 21:34:14 +0000 (23:34 +0200)]
androidmedia: Conditionally use get_{input,output}_buffer() Android 21 APIs

Also properly set limit/position on byte buffer, some codecs prefer to have
correct values there.

9 years agoglwindow: fix compiler error
Matthew Waters [Sun, 31 May 2015 06:53:14 +0000 (16:53 +1000)]
glwindow: fix compiler error

gstglwindow.c:1118:544: error: 'return' with no value, in function returning non-void [-Werror]
   g_return_if_fail (GST_GL_IS_WINDOW (window));

9 years agogl/examples: update gtk examples for glupload
Matthew Waters [Sun, 31 May 2015 05:42:48 +0000 (15:42 +1000)]
gl/examples: update gtk examples for glupload

9 years agogl/examples: update gtk examples for wayland
Matthew Waters [Sun, 31 May 2015 05:40:15 +0000 (15:40 +1000)]
gl/examples: update gtk examples for wayland

9 years agogl/wayland: implement setting the render rectangle
Matthew Waters [Fri, 29 May 2015 08:06:27 +0000 (18:06 +1000)]
gl/wayland: implement setting the render rectangle

Places our subsurface at the rectangle provided position

9 years agoglwindow: handle gst_video_overlay_set_render_rectangle
Matthew Waters [Fri, 29 May 2015 08:03:52 +0000 (18:03 +1000)]
glwindow: handle gst_video_overlay_set_render_rectangle

9 years agogl/wayland: don't block the event loop after poll
Matthew Waters [Fri, 29 May 2015 08:01:29 +0000 (18:01 +1000)]
gl/wayland: don't block the event loop after poll

Use the dispatch_pending set of functions which just run the currently
queued up events instead of potentially waiting for an event to occur.

9 years agotests/gl/gtk: implement setting a wayland display/surface
Matthew Waters [Wed, 27 May 2015 06:42:55 +0000 (16:42 +1000)]
tests/gl/gtk: implement setting a wayland display/surface

9 years agogl/wayland: implement basic video overlay support via subsurfaces
Matthew Waters [Wed, 27 May 2015 06:39:06 +0000 (16:39 +1000)]
gl/wayland: implement basic video overlay support via subsurfaces

Currently does not position the subsurface relative to the parent surface at all

9 years agogl/wayland: allow a NULL wl_event_queue
Matthew Waters [Wed, 27 May 2015 05:44:33 +0000 (15:44 +1000)]
gl/wayland: allow a NULL wl_event_queue

perform operations on the default wl_display event queue in that case

9 years agogl/utils: implement wayland display GstContext
Matthew Waters [Wed, 27 May 2015 05:43:06 +0000 (15:43 +1000)]
gl/utils: implement wayland display GstContext

requried for sharing GL contexts or subsurface support

9 years agogl/wayland: run each window on a separate queue
Matthew Waters [Fri, 22 May 2015 06:07:49 +0000 (16:07 +1000)]
gl/wayland: run each window on a separate queue

Based on patch by Julien Isorce <julien.isorce@collabora.co.uk>

https://bugzilla.gnome.org/show_bug.cgi?id=709747

9 years agogl: Don't leak temp strings in _RGB_pixel_order()
Jan Schmidt [Fri, 29 May 2015 16:19:25 +0000 (02:19 +1000)]
gl: Don't leak temp strings in _RGB_pixel_order()

Fix a memory leak of temporary strings when computing
swizzling of RGB formats.

9 years agoglfilter: Add transform_internal_caps() vfunc
Jan Schmidt [Fri, 29 May 2015 16:29:47 +0000 (02:29 +1000)]
glfilter: Add transform_internal_caps() vfunc

Add a vfunc that is called by glfilter before it sets
caps features and intersects with the peer caps, and
move removing the size from caps into its default
implementation. Allows sub-classes to do more
sophisticated management of the size fields in case they
don't support arbitrary resizing or have distinct
preferences.

9 years agoglfilter: Don't remove size fields from caps
Jan Schmidt [Fri, 29 May 2015 16:23:44 +0000 (02:23 +1000)]
glfilter: Don't remove size fields from caps

Instead of removing size fields, set them to the full range
like videoscale does, so the caps are clearly unfixed in
all cases.

9 years agovideoaggregator: Add class property to disable caps scaling
Jan Schmidt [Sun, 10 May 2015 08:55:16 +0000 (18:55 +1000)]
videoaggregator: Add class property to disable caps scaling

Add preserve_update_caps_result boolean on the class to allow
sub-classes to disable videoaggregator removing sizes and framerate
from the update_caps() return result.

9 years agovideoaggregator: Catch errors, and allow sub-class to return NULL from get_output_buf...
Jan Schmidt [Mon, 4 May 2015 08:17:21 +0000 (18:17 +1000)]
videoaggregator: Catch errors, and allow sub-class to return NULL from get_output_buffer()

A return value of GST_FLOW_OK with a NULL buffer from get_output_buffer()
means the sub-class doesn't want to produce an output buffer, so
skip it.

If gst_videoaggregator_do_aggregate() generates an error, make sure
to propagate it - don't just ignore and discard the error by
over-writing it with the gst_pad_push() result.

9 years agortmpsink: Do not crash when receiving buffers after GST_FLOW_ERROR
Vivia Nikolaidou [Fri, 29 May 2015 13:02:31 +0000 (16:02 +0300)]
rtmpsink: Do not crash when receiving buffers after GST_FLOW_ERROR

If the RTMP URI is invalid, the rtmpsink will return GST_FLOW_ERROR.
If it still receives buffers after that, it shouldn't crash.

https://bugzilla.gnome.org/show_bug.cgi?id=750104

9 years agodashdemux: avoid assertion when splitting buffer
Thiago Santos [Fri, 29 May 2015 12:01:29 +0000 (09:01 -0300)]
dashdemux: avoid assertion when splitting buffer

Pass size=-1 as is if that is the case instead of subtracting
the offset. Otherwise we have an invalid size passed for that
buffer.

9 years agodtlssrtpdec: Release requested pads
Santiago Carot-Nemesio [Thu, 28 May 2015 13:59:02 +0000 (15:59 +0200)]
dtlssrtpdec: Release requested pads

https://bugzilla.gnome.org/show_bug.cgi?id=750036

9 years agodtlsdec: Remove unnecessary ref/unref operations
Santiago Carot-Nemesio [Thu, 28 May 2015 13:57:10 +0000 (15:57 +0200)]
dtlsdec: Remove unnecessary ref/unref operations

https://bugzilla.gnome.org/show_bug.cgi?id=750036

9 years agodtlsdec: Fix release request pad function
Santiago Carot-Nemesio [Thu, 28 May 2015 13:55:03 +0000 (15:55 +0200)]
dtlsdec: Fix release request pad function

Don't unref the pad reference we don't own and just remove the pad.

https://bugzilla.gnome.org/show_bug.cgi?id=750036

9 years agohlsdemux: Search more when advancing fragment
Edward Hervey [Thu, 28 May 2015 12:30:46 +0000 (14:30 +0200)]
hlsdemux: Search more when advancing fragment

In live situations, it is not uncommon for the current fragment to end
up out of the (updated) play range (lowest/highest sequence). But the next
fragment to play *is* present in the play range.

When advancing, if we can't find the current GstM3U8MediaFile, don't abort
straight away. Instead, look if a GstM3U8MediaFile with the next sequence value
is present, and if so switch to it.

https://bugzilla.gnome.org/show_bug.cgi?id=750028

9 years agogl: add GstGLContextGPUProcess backend
Julien Isorce [Wed, 27 May 2015 09:58:10 +0000 (10:58 +0100)]
gl: add GstGLContextGPUProcess backend

It builds its GL vtable from a proc address provided
by the application.

9 years agogl: add and use gst_gl_internal_format_rgba
Julien Isorce [Wed, 27 May 2015 15:28:39 +0000 (16:28 +0100)]
gl: add and use gst_gl_internal_format_rgba

Previously when compiling GstGL with both GL and GLES2,
GL_RGBA8 was picked from GL/gl.h. But a clash may happen at
runtime when one is selecting GLES2.

gst_gl_internal_format_rgba allows to check at runtime
if it should use GL_RGBA or GL_RGBA8.

9 years agoglwindow_x11: use parent default implementation
Julien Isorce [Wed, 27 May 2015 13:10:16 +0000 (14:10 +0100)]
glwindow_x11: use parent default implementation

9 years agoglcontext_glx: rename variable from window to context
Julien Isorce [Wed, 27 May 2015 10:16:32 +0000 (11:16 +0100)]
glcontext_glx: rename variable from window to context

9 years agoglwindow: initialize debug in class_init
Julien Isorce [Wed, 27 May 2015 10:11:55 +0000 (11:11 +0100)]
glwindow: initialize debug in class_init

Useful if one uses gst_gl_window_x11_new directly
instead of gst_gl_window_new.

9 years agoglcontext: initialize debug in class_init
Julien Isorce [Wed, 27 May 2015 10:07:38 +0000 (11:07 +0100)]
glcontext: initialize debug in class_init

Useful if one uses gst_gl_context_egl_new directly
instead of gst_gl_context_new.

9 years agoglcontext: add gst_gl_context_set_display helper
Julien Isorce [Wed, 27 May 2015 09:55:20 +0000 (10:55 +0100)]
glcontext: add gst_gl_context_set_display helper

9 years agoglcontext: add missing _CAST after G_TYPE_CHECK_CLASS
Julien Isorce [Wed, 27 May 2015 14:25:40 +0000 (15:25 +0100)]
glcontext: add missing _CAST after G_TYPE_CHECK_CLASS

9 years agoglwindow: add missing _CAST after G_TYPE_CHECK_CLASS
Julien Isorce [Wed, 27 May 2015 09:51:54 +0000 (10:51 +0100)]
glwindow: add missing _CAST after G_TYPE_CHECK_CLASS

9 years agoglwindow: provides some default implementations to factorize with all backends
Julien Isorce [Tue, 26 May 2015 17:38:39 +0000 (18:38 +0100)]
glwindow: provides some default implementations to factorize with all backends

9 years agoaggregator: Push EOS on error return.
Jan Schmidt [Wed, 27 May 2015 14:59:39 +0000 (00:59 +1000)]
aggregator: Push EOS on error return.

Before shutting down the srcpad task due to a
downstream error, push an EOS to give downstream
a chance to shut down somewhat cleanly.

9 years agompegtsmux: don't overwrite value
Luis de Bethencourt [Wed, 27 May 2015 12:12:57 +0000 (13:12 +0100)]
mpegtsmux: don't overwrite value

Value of res is reset to FALSE in each iteration of the while loop. We want to
conserve TRUE if any pad event succeeded until we arrive to done.

Also, buf is set to the value of *outbuf twice. Removing the first assignment
since the second one is outside of a conditional.

9 years agoopencv: 'for' loop initial declaration
Luis de Bethencourt [Wed, 27 May 2015 12:08:46 +0000 (13:08 +0100)]
opencv: 'for' loop initial declaration

'for' loop initial declarations are not allowed in C89, moving the declarations
to before the 'for' loops.

9 years agogleffects: 'for' loop initial declaration
Luis de Bethencourt [Wed, 27 May 2015 12:05:11 +0000 (13:05 +0100)]
gleffects: 'for' loop initial declaration

'for' loop initial declarations are not allowed in C89, moving the declarations
to before the 'for' loops.

9 years agoresindvd: remove unused value
Luis de Bethencourt [Wed, 27 May 2015 11:03:42 +0000 (12:03 +0100)]
resindvd: remove unused value

The data variable is incremented but never read again. Remove the unused value
assignment.

9 years agohlsdemux: remove more unused attributes
Thiago Santos [Wed, 27 May 2015 09:17:49 +0000 (06:17 -0300)]
hlsdemux: remove more unused attributes

Those are not used or only read