platform/upstream/gst-plugins-bad.git
7 years agokmssink: include drm.h after stdlib.h
Víctor Manuel Jáquez Leal [Wed, 23 Nov 2016 19:13:59 +0000 (20:13 +0100)]
kmssink: include drm.h after stdlib.h

As drm.h is internal to libdrm, it expects to have already included
stdlib.h.

7 years agokms: include stdlib.h
Víctor Manuel Jáquez Leal [Wed, 23 Nov 2016 18:41:43 +0000 (19:41 +0100)]
kms: include stdlib.h

drm.h does not include all what it needs to compile, in particular
stdlib.h which defines size_t

7 years agompdparser: Don't set variable just to overwrite it again later
Sebastian Dröge [Wed, 23 Nov 2016 17:01:04 +0000 (19:01 +0200)]
mpdparser: Don't set variable just to overwrite it again later

We set it to TRUE here, but later we set it to TRUE again anyway if the
parsing actually succeeded at this point. Let's keep the second one.

CID 1374360.

7 years agowldisplay: Don't compare a guint with < 0
Sebastian Dröge [Wed, 23 Nov 2016 16:45:42 +0000 (18:45 +0200)]
wldisplay: Don't compare a guint with < 0

-1 is the "invalid" value, compare to it directly.

CID 1394490.

7 years agoaudiobuffersplit: New element that splits raw audio buffers into equal-sized buffers
Sebastian Dröge [Fri, 18 Nov 2016 19:00:03 +0000 (21:00 +0200)]
audiobuffersplit: New element that splits raw audio buffers into equal-sized buffers

This is useful e.g. if audio buffers should be exactly the duration of a
video frame, or if a audio buffers should never be too large because of
latency constraints.

The element is taking a fractional buffer duration, to allow working
with e.g. 1001/30000 as output duration and it accumulates rounding
errors in the buffer durations and compensates for them by making some
buffers one sample larger than the others.

https://bugzilla.gnome.org/show_bug.cgi?id=774689

7 years agokmssink: remove dependency on libkms
Michael Tretter [Wed, 19 Oct 2016 10:39:36 +0000 (12:39 +0200)]
kmssink: remove dependency on libkms

libkms should not be used, because it imposes limitations on the DRM
API, especially regarding bpp and stride. Instead the DRM IOCTL should
be used directly.

Switch from libkms to the IOCTL interface. Set bpp and height for
framebuffer allocation to properly handle planar video formats.

https://bugzilla.gnome.org/show_bug.cgi?id=773473

Signed-off-by: Víctor Jáquez <vjaquez@igalia.com>
7 years agomimic: remove ancient codec
Tim-Philipp Müller [Wed, 23 Nov 2016 10:31:29 +0000 (10:31 +0000)]
mimic: remove ancient codec

This was used by MSN messenger in prehistoric times, it's safe
to say no one needs or wants this any more these days. For
decoding old recordings there's still a decoder in ffmpeg.

https://bugzilla.gnome.org/show_bug.cgi?id=597616

7 years agoh265parse: Ensure codec_data has the required size when reading number of NAL arrays
Sebastian Dröge [Wed, 23 Nov 2016 08:52:05 +0000 (10:52 +0200)]
h265parse: Ensure codec_data has the required size when reading number of NAL arrays

https://bugzilla.gnome.org/show_bug.cgi?id=774896

7 years agoh264parse: Ensure codec_data has the required size when reading number of SPS
Sebastian Dröge [Wed, 23 Nov 2016 08:51:17 +0000 (10:51 +0200)]
h264parse: Ensure codec_data has the required size when reading number of SPS

https://bugzilla.gnome.org/show_bug.cgi?id=774896

7 years agogl/contextglx: error out on more invalid display/window combinations
Matthew Waters [Wed, 23 Nov 2016 06:04:34 +0000 (17:04 +1100)]
gl/contextglx: error out on more invalid display/window combinations

https://bugzilla.gnome.org/show_bug.cgi?id=774851

7 years agogl: add necessary context_egl.h #includes to the source files
Matthew Waters [Wed, 23 Nov 2016 04:57:05 +0000 (15:57 +1100)]
gl: add necessary context_egl.h #includes to the source files

7 years agogl/memoryegl: remove access to private header gstglcontext_egl.h
Matthew Waters [Wed, 23 Nov 2016 02:32:38 +0000 (13:32 +1100)]
gl/memoryegl: remove access to private header gstglcontext_egl.h

It's been removed and thus compiling anything against GstGLMemoryEGL
would error with:

In file included from gstomxvideodec.c:41:0:
usr/include/gstreamer-1.0/gst/gl/egl/gstglmemoryegl.h:32:41: fatal error: gst/gl/egl/gstglcontext_egl.h: No such file or directory
 #include <gst/gl/egl/gstglcontext_egl.h>
                                         ^

https://bugzilla.gnome.org/show_bug.cgi?id=774886

7 years agompegtspacketizer: Reset values before parsing each section
Sebastian Dröge [Tue, 22 Nov 2016 10:58:44 +0000 (12:58 +0200)]
mpegtspacketizer: Reset values before parsing each section

7 years agompegtspacketizer: Make sure in all cases that we have enough data left for reading...
Sebastian Dröge [Tue, 22 Nov 2016 10:42:45 +0000 (12:42 +0200)]
mpegtspacketizer: Make sure in all cases that we have enough data left for reading a section start

In some cases we ended up parsing sections without.

7 years agoopencv: forward declare opencv types
Philippe Renon [Mon, 21 Nov 2016 16:19:46 +0000 (17:19 +0100)]
opencv: forward declare opencv types

Forward declare opencv types to avoid exposing them in the API.

https://bugzilla.gnome.org/show_bug.cgi?id=774223

7 years agointerfaces: photography: clean up header indentation
Tim-Philipp Müller [Mon, 21 Nov 2016 10:52:39 +0000 (10:52 +0000)]
interfaces: photography: clean up header indentation

7 years agompdparser: Modify return of the function for loading external resources to void
Seungha Yang [Sun, 20 Nov 2016 06:10:42 +0000 (15:10 +0900)]
mpdparser: Modify return of the function for loading external resources to void

gst_mpd_client_fetch_on_load_external_resources() never ever return
FALSE due to modified external xml loading functions.

https://bugzilla.gnome.org/show_bug.cgi?id=774463

7 years agompdparser: Handle invalid external xml link for SegmentList element
Seungha Yang [Sat, 19 Nov 2016 14:04:04 +0000 (23:04 +0900)]
mpdparser: Handle invalid external xml link for SegmentList element

Ignore invalid xml link for SegmentList likewise external Period without error.

https://bugzilla.gnome.org/show_bug.cgi?id=774463

7 years agompdparser: Handle invalid external xml link for AdaptationSet element
Seungha Yang [Sat, 19 Nov 2016 13:38:20 +0000 (22:38 +0900)]
mpdparser: Handle invalid external xml link for AdaptationSet element

Ignore invalid xml link for AdaptationSet likewise external Period without error.

https://bugzilla.gnome.org/show_bug.cgi?id=774463

7 years agompdparser: Handle invalid external xml link for Period element
Seungha Yang [Sat, 19 Nov 2016 09:37:54 +0000 (18:37 +0900)]
mpdparser: Handle invalid external xml link for Period element

Section 5.3.3 in ISO/IEC 23009-1:2014 defines that invalid references
(e.g., invalide URI or cannot be resolved) specified by "@xlink:href" attribute
shall be removed. That means, we should play it without error,
and just ignore the corresponding element.
It's similar to "urn:mpeg:dash:resolve-to-zero:2013".

https://bugzilla.gnome.org/show_bug.cgi?id=774463

7 years agogl: add gstreamer-video to pkg-config path
David Evans [Thu, 17 Nov 2016 15:41:00 +0000 (15:41 +0000)]
gl: add gstreamer-video to pkg-config path

https://bugzilla.gnome.org/show_bug.cgi?id=774624

7 years agotests: fix lib and cflags order in Makefile.am
Tim-Philipp Müller [Sun, 20 Nov 2016 12:40:07 +0000 (12:40 +0000)]
tests: fix lib and cflags order in Makefile.am

pnm: add misssing continuation slash at end of line
vp8parser: no need to link codecparsers twice
dash_demux, mssdemux, hlsdemux: fix lib/flags ordering - local
libs and flags must come first, then base before core before rest

7 years agodashdemux: Fix leak in gst_dash_demux_stream_free
Seungha Yang [Sat, 19 Nov 2016 07:12:06 +0000 (16:12 +0900)]
dashdemux: Fix leak in gst_dash_demux_stream_free

Free parsed moof and sync samples with _stream_free().

https://bugzilla.gnome.org/show_bug.cgi?id=774702

7 years agowayland: avoid prefix clash in combined pkg-config check
Reynaldo H. Verdejo Pinochet [Thu, 17 Nov 2016 19:19:54 +0000 (11:19 -0800)]
wayland: avoid prefix clash in combined pkg-config check

Adittionally, remove leftover WAYLAND_DRM_ include flags var

Fixes building of ext/wayland/* broken by dfd86de67a8c04f6d1

Related to:

https://bugzilla.gnome.org/show_bug.cgi?id=773927

7 years agovc1parse: Mark pad as needing reconfiguration again if it failed
Sebastian Dröge [Thu, 17 Nov 2016 18:21:17 +0000 (20:21 +0200)]
vc1parse: Mark pad as needing reconfiguration again if it failed

And return FLUSHING instead of NOT_NEGOTIATED on flushing pads.

https://bugzilla.gnome.org/show_bug.cgi?id=774623

7 years agovtdec: Mark pad as needing reconfiguration again if it failed
Sebastian Dröge [Thu, 17 Nov 2016 18:20:52 +0000 (20:20 +0200)]
vtdec: Mark pad as needing reconfiguration again if it failed

And return FLUSHING instead of NOT_NEGOTIATED on flushing pads.

https://bugzilla.gnome.org/show_bug.cgi?id=774623

7 years agovideoaggregator: Mark pad as needing reconfiguration again if it failed
Sebastian Dröge [Thu, 17 Nov 2016 18:20:15 +0000 (20:20 +0200)]
videoaggregator: Mark pad as needing reconfiguration again if it failed

And return FLUSHING instead of NOT_NEGOTIATED on flushing pads.

https://bugzilla.gnome.org/show_bug.cgi?id=774623

7 years agottmlrender: Mark pad as needing reconfiguration again if it failed
Sebastian Dröge [Thu, 17 Nov 2016 18:19:44 +0000 (20:19 +0200)]
ttmlrender: Mark pad as needing reconfiguration again if it failed

And return FLUSHING instead of NOT_NEGOTIATED on flushing pads.

https://bugzilla.gnome.org/show_bug.cgi?id=774623

7 years agoassrender: Mark pad as needing reconfiguration again if reconfiguration failed
Sebastian Dröge [Thu, 17 Nov 2016 18:16:49 +0000 (20:16 +0200)]
assrender: Mark pad as needing reconfiguration again if reconfiguration failed

Also return FLUSHING instead of NOT_NEGOTIATED on flushing pads.

https://bugzilla.gnome.org/show_bug.cgi?id=774623

7 years agoiqa: Add to DIST_SUBDIRS
Sebastian Dröge [Fri, 18 Nov 2016 10:21:22 +0000 (12:21 +0200)]
iqa: Add to DIST_SUBDIRS

7 years agomeson: add_global_arguments -> add_project_arguments
Scott D Phillips [Fri, 18 Nov 2016 01:17:22 +0000 (17:17 -0800)]
meson: add_global_arguments -> add_project_arguments

https://bugzilla.gnome.org/show_bug.cgi?id=774656

7 years agomeson: wayland: wl_scanner is required: false
Scott D Phillips [Thu, 17 Nov 2016 18:54:45 +0000 (10:54 -0800)]
meson: wayland: wl_scanner is required: false

https://bugzilla.gnome.org/show_bug.cgi?id=774637

7 years agopo: Remove remaining reference to videomeasure
Thibault Saunier [Thu, 17 Nov 2016 18:11:20 +0000 (15:11 -0300)]
po: Remove remaining reference to videomeasure

7 years agodecklinkvideosrc: Issue "no signal" warnings
Vivia Nikolaidou [Thu, 17 Nov 2016 16:23:04 +0000 (18:23 +0200)]
decklinkvideosrc: Issue "no signal" warnings

When a frame is found to not have an associated input source (cable
unplugged, wrong mode selected), an element warning will be issued. When
the next frame in the stream is found to have an input source selected
(e.g. cable replugged), an element info will be issued.

https://bugzilla.gnome.org/show_bug.cgi?id=774629

7 years agovideomeasure: Remove old unported videomeasure plugin as it is now replaced by IQA
Thibault Saunier [Wed, 16 Nov 2016 21:01:19 +0000 (18:01 -0300)]
videomeasure: Remove old unported videomeasure plugin as it is now replaced by IQA

7 years agoiqa: Error out when trying to compare video stream with different sizes
Thibault Saunier [Wed, 16 Nov 2016 19:30:50 +0000 (16:30 -0300)]
iqa: Error out when trying to compare video stream with different sizes

We do not handle it yet and we should let the user know.

7 years agoiqa: Add meson build support
Thibault Saunier [Wed, 16 Nov 2016 17:46:24 +0000 (14:46 -0300)]
iqa: Add meson build support

7 years agoAdds a new Image Quality Assessment plugin.
Mathieu Duponchelle [Wed, 16 Nov 2016 16:18:53 +0000 (13:18 -0300)]
Adds a new Image Quality Assessment plugin.

It only offers one metric for now, "dssim", available if
https://github.com/pornel/dssim was installed on the system
at the time the plugin was compiled.

The spearman correlation for dssim against the TID2008 dataset
is 0.81, against 0.70 for the old ssim implementation, and
it runs 15 times faster.

https://bugzilla.gnome.org/show_bug.cgi?id=751324

7 years agogl/tests: fix glimagesink element test on OSX
Matthew Waters [Thu, 17 Nov 2016 03:35:26 +0000 (14:35 +1100)]
gl/tests: fix glimagesink element test on OSX

It was suffering a race starting up the thread that could cause the main
loop to quit before the main loop had started.

7 years agoapplemedia: error out if the uri property is not set
Matthew Waters [Thu, 17 Nov 2016 03:21:23 +0000 (14:21 +1100)]
applemedia: error out if the uri property is not set

Fixes:

Terminating app due to uncaught exception 'NSInvalidArgumentException', reason: '*** +[NSString stringWithUTF8String:]: NULL cString

in the state change test.

7 years agomeson: wayland: move dep checks to the lib
Matthew Waters [Thu, 17 Nov 2016 04:15:18 +0000 (15:15 +1100)]
meson: wayland: move dep checks to the lib

Moves them closer to where they are used.  We don't want every check in
the main meson.build file.

7 years agomeson: gl/egl: add missing source file
Matthew Waters [Thu, 17 Nov 2016 03:32:10 +0000 (14:32 +1100)]
meson: gl/egl: add missing source file

d42145e8c1ba3bc0445506b92bb7ac04ae98f4dd didn't add the necessary
meson build changes.

7 years agogl/egl: remove EGLImage functions from egl context
Matthew Waters [Wed, 16 Nov 2016 15:41:14 +0000 (02:41 +1100)]
gl/egl: remove EGLImage functions from egl context

By adding the necessary GstEGLImage entry points to create a GstEGLImage
from a GstGLMemory.

https://bugzilla.gnome.org/show_bug.cgi?id=774518

7 years agoglcontext: add vfunc to retrieve the OpenGL platform version
Matthew Waters [Wed, 16 Nov 2016 14:45:38 +0000 (01:45 +1100)]
glcontext: add vfunc to retrieve the OpenGL platform version

i.e. the version of EGL, GLX, etc implemented.

https://bugzilla.gnome.org/show_bug.cgi?id=774518

7 years agogl/egl: move get_error_string() into gstegl
Matthew Waters [Wed, 16 Nov 2016 14:38:32 +0000 (01:38 +1100)]
gl/egl: move get_error_string() into gstegl

So others can use it without #include-ing a private header

https://bugzilla.gnome.org/show_bug.cgi?id=774518

7 years agodisplay/egl: implement getting the EGLDisplay of a specific platform
Matthew Waters [Thu, 27 Nov 2014 04:50:04 +0000 (15:50 +1100)]
display/egl: implement getting the EGLDisplay of a specific platform

https://bugzilla.gnome.org/show_bug.cgi?id=774518

7 years agogl/egl: check the feature in the extensions list
Víctor Manuel Jáquez Leal [Wed, 16 Nov 2016 12:05:11 +0000 (13:05 +0100)]
gl/egl: check the feature in the extensions list

https://bugzilla.gnome.org/show_bug.cgi?id=774518

7 years agovmncdec: Sanity-check width/height before using it
Sebastian Dröge [Wed, 16 Nov 2016 18:41:39 +0000 (20:41 +0200)]
vmncdec: Sanity-check width/height before using it

We will allocate a screen area of width*height*bpp bytes, however this
calculation can easily overflow if too high width or height are given
inside the stream. Nonetheless we would just assume that enough memory
was allocated, try to fill it and overwrite as much memory as wanted.

Also allocate the screen area filled with zeroes to ensure that we start
with full-black and not any random (or not so random) data.

https://scarybeastsecurity.blogspot.gr/2016/11/0day-poc-risky-design-decisions-in.html

Ideally we should just remove this plugin in favour of the one in
gst-libav, which generally seems to be of better code quality.

https://bugzilla.gnome.org/show_bug.cgi?id=774533

7 years agottml: Hide behind --enable-experimental
Sebastian Dröge [Wed, 16 Nov 2016 16:50:53 +0000 (18:50 +0200)]
ttml: Hide behind --enable-experimental

The elements are autoplugged and currently assert on invalid streams.

7 years agosmoothwave: Remove empty directory
Sebastian Dröge [Wed, 16 Nov 2016 14:45:27 +0000 (16:45 +0200)]
smoothwave: Remove empty directory

7 years agoreal: Remove empty directory
Sebastian Dröge [Wed, 16 Nov 2016 14:33:41 +0000 (16:33 +0200)]
real: Remove empty directory

7 years agompeg1sys: Remove obsolete, empty directory that only contains a meson.build
Sebastian Dröge [Wed, 16 Nov 2016 14:15:34 +0000 (16:15 +0200)]
mpeg1sys: Remove obsolete, empty directory that only contains a meson.build

7 years agocurl: transfer hangs in case connection was cut between client and server
Imre Eörs [Tue, 15 Nov 2016 15:20:10 +0000 (16:20 +0100)]
curl: transfer hangs in case connection was cut between client and server

https://bugzilla.gnome.org/show_bug.cgi?id=774485

7 years agogl/tests: don't use the default framebuffer
Matthew Waters [Wed, 16 Nov 2016 06:30:54 +0000 (17:30 +1100)]
gl/tests: don't use the default framebuffer

Create our own instead as the default framebuffer may require special
fiddling (like having a visible window) to correctly display/be renderable.

Fixes the remaining GL library tests on OS X

7 years agoglcontext/cocoa: implement empty swap_buffers
Matthew Waters [Wed, 16 Nov 2016 02:39:32 +0000 (13:39 +1100)]
glcontext/cocoa: implement empty swap_buffers

Fixes some GL tests on OS X.

7 years agogl: remove empty BUGS file
Matthew Waters [Wed, 16 Nov 2016 05:41:59 +0000 (16:41 +1100)]
gl: remove empty BUGS file

We use bugzilla for bug tracking

7 years agomeson: Build waylandsink
Thibault Saunier [Tue, 15 Nov 2016 20:46:17 +0000 (17:46 -0300)]
meson: Build waylandsink

https://bugzilla.gnome.org/show_bug.cgi?id=774493

7 years agoconfigure: fix build if wayland-client is not available
Luis de Bethencourt [Tue, 15 Nov 2016 17:48:15 +0000 (17:48 +0000)]
configure: fix build if wayland-client is not available

Simplify the PKG_CHECK_MODULES related to Wayland to avoid the confusion
of NOT_FOUND cases when there are 3 nested checks. Group those 3 checks
together since there are no conditions specific to each one.

Thanks to https://ci.gstreamer.net/ for alerting of the problem.

https://bugzilla.gnome.org/show_bug.cgi?id=773927

7 years agoconfigure: fix build if libdrm is not available
Luis de Bethencourt [Tue, 15 Nov 2016 11:59:45 +0000 (11:59 +0000)]
configure: fix build if libdrm is not available

If the NOT_FOUND part of the check PKG_CHECK_MODULES is not written, it
defaults to error. Addind the else clause of this check as
HAVE_WAYLAND="no"

https://bugzilla.gnome.org/show_bug.cgi?id=773927

7 years agoglwindow/cocoa: remove unneeded window code in NSWindow
Matthew Waters [Tue, 15 Nov 2016 13:01:00 +0000 (00:01 +1100)]
glwindow/cocoa: remove unneeded window code in NSWindow

This is all taken care of by GstGLWindowCocoa/GstGLNSView now.

7 years agoglwindow/cocoa: remove our view from the parent when closing
Matthew Waters [Tue, 15 Nov 2016 11:56:25 +0000 (22:56 +1100)]
glwindow/cocoa: remove our view from the parent when closing

Otherwise, when the application reuses the same UIView, we were getting
draw notifications on the previous view/layer's which weren't valid anymore
and were referencing pointers that had been freed.

https://bugzilla.gnome.org/show_bug.cgi?id=753003

7 years agogl/caopengllayer: add a debug category
Matthew Waters [Tue, 15 Nov 2016 10:43:59 +0000 (21:43 +1100)]
gl/caopengllayer: add a debug category

7 years agoglbufferpool: introduce check for GLMemory allocators
Matthew Waters [Tue, 15 Nov 2016 03:36:11 +0000 (14:36 +1100)]
glbufferpool: introduce check for GLMemory allocators

The last missing piece of EGLImage support has been pushed.

7 years agoFix some MSVC warnings about const-ness
Scott D Phillips [Tue, 15 Nov 2016 01:02:54 +0000 (17:02 -0800)]
Fix some MSVC warnings about const-ness

Some miscellaneous warnings about const-ness, either casting away const
where we know it's safe or removing const where it's not correct.

https://bugzilla.gnome.org/show_bug.cgi?id=774293

7 years agoCast away const from GstMetaInfo in *_get_meta_info() functions
Scott D Phillips [Tue, 15 Nov 2016 00:59:48 +0000 (16:59 -0800)]
Cast away const from GstMetaInfo in *_get_meta_info() functions

MSVC warns about the const in the implicit argument conversion in the
calls to g_once_init_{enter,leave}. It's OK so explicitly cast it.

https://bugzilla.gnome.org/show_bug.cgi?id=774293

7 years agoadaptivedemux: Fix startup SEGMENT seeking and setting for live
Seungha Yang [Thu, 10 Nov 2016 14:07:50 +0000 (23:07 +0900)]
adaptivedemux: Fix startup SEGMENT seeking and setting for live

Because fragment.timestamp is relative value to period_start,
startup SEGMENT seeking should be pointed to "fragment.timestamp + period_start"

https://bugzilla.gnome.org/show_bug.cgi?id=774196

7 years agogdiscreencapsrc: remove empty gst_gdiscreencapsrc_stop vfunc
Philippe Renon [Mon, 14 Nov 2016 11:23:37 +0000 (12:23 +0100)]
gdiscreencapsrc: remove empty gst_gdiscreencapsrc_stop vfunc

https://bugzilla.gnome.org/show_bug.cgi?id=774400

7 years agoopencv: move base opencv filter class and utilities to gst-libs
Philippe Renon [Thu, 10 Nov 2016 17:42:29 +0000 (18:42 +0100)]
opencv: move base opencv filter class and utilities to gst-libs

https://bugzilla.gnome.org/show_bug.cgi?id=774223

7 years agofbdevsink: Add to meson build
Arun Raghavan [Mon, 14 Nov 2016 17:45:05 +0000 (23:15 +0530)]
fbdevsink: Add to meson build

7 years agokmssink: Trivial naming fix in meson for consistency
Arun Raghavan [Mon, 14 Nov 2016 17:41:11 +0000 (23:11 +0530)]
kmssink: Trivial naming fix in meson for consistency

7 years agokmssink: Enable in meson build
Arun Raghavan [Mon, 14 Nov 2016 15:33:20 +0000 (21:03 +0530)]
kmssink: Enable in meson build

7 years agodash: Fix breakage in unit test caused by gst-indent
Sebastian Dröge [Mon, 14 Nov 2016 15:00:21 +0000 (17:00 +0200)]
dash: Fix breakage in unit test caused by gst-indent

7 years agompdparser: Support multiple Period elements in external xml
Seungha Yang [Mon, 14 Nov 2016 14:31:37 +0000 (23:31 +0900)]
mpdparser: Support multiple Period elements in external xml

External xml could have empty, one or multiple top-level "Period" elements.
Because xml parser cannot parse the multiple top-level elements
(i.e., no root element), we need to wrap a xml in order to make root element.
See also ISO/IEC 23009-1:2014 5.3.2.2

https://bugzilla.gnome.org/show_bug.cgi?id=774357

7 years agomsdk: Initial windows build support
Scott D Phillips [Thu, 3 Nov 2016 17:47:00 +0000 (10:47 -0700)]
msdk: Initial windows build support

https://bugzilla.gnome.org/show_bug.cgi?id=770990

7 years agomeson: add build files for msdk
Scott D Phillips [Thu, 13 Oct 2016 21:30:43 +0000 (14:30 -0700)]
meson: add build files for msdk

https://bugzilla.gnome.org/show_bug.cgi?id=770990

7 years agomsdk: load plugins in h265 and vp8 init
Scott D Phillips [Tue, 27 Sep 2016 21:21:34 +0000 (14:21 -0700)]
msdk: load plugins in h265 and vp8 init

https://bugzilla.gnome.org/show_bug.cgi?id=770990

7 years agomsdk: Width and Height need to be integer multiples of 32
Scott D Phillips [Tue, 27 Sep 2016 21:17:49 +0000 (14:17 -0700)]
msdk: Width and Height need to be integer multiples of 32

https://bugzilla.gnome.org/show_bug.cgi?id=770990

7 years agomsdk: add to build
Scott D Phillips [Fri, 9 Sep 2016 22:43:25 +0000 (15:43 -0700)]
msdk: add to build

https://bugzilla.gnome.org/show_bug.cgi?id=770990

7 years agomsdk: Create plugin for Intel's Media SDK
Josep Torra [Fri, 1 Apr 2016 17:25:25 +0000 (19:25 +0200)]
msdk: Create plugin for Intel's Media SDK

[scott.d.phillips@intel.com: gst-indent *.c *.h]
[scott.d.phillips@intel.com: link with libva at build time]
[scott.d.phillips@intel.com: remove unused includes]
[scott.d.phillips@intel.com: Update README]

https://bugzilla.gnome.org/show_bug.cgi?id=770990

7 years agompdparser: wrap unmodified data in protection event
Philippe Normand [Mon, 14 Nov 2016 10:08:24 +0000 (11:08 +0100)]
mpdparser: wrap unmodified data in protection event

PlayReady being the one of the few DRM formats encoding its data with
base64 it was not consistent to have a special case for this. So the
base64 decoding operation now needs to be done by the protection event
consumer, if needed.

https://bugzilla.gnome.org/show_bug.cgi?id=774112

7 years agomssdemux: wrap unmodified data in protection event
Philippe Normand [Mon, 14 Nov 2016 10:06:00 +0000 (11:06 +0100)]
mssdemux: wrap unmodified data in protection event

PlayReady being the one of the few DRM formats encoding its data with
base64 it was not consistent to have a special case for this. So the
base64 decoding operation now needs to be done by the protection event
consumer, if needed.

https://bugzilla.gnome.org/show_bug.cgi?id=774112

7 years agodtlscertificate: Fix error checking in RSA_generate_key_ex() usage
Sebastian Dröge [Mon, 14 Nov 2016 09:32:17 +0000 (11:32 +0200)]
dtlscertificate: Fix error checking in RSA_generate_key_ex() usage

Was broken during the port for OpenSSL 1.1.

https://bugzilla.gnome.org/show_bug.cgi?id=774328

8 years agowinks: undef windows SDK macros before redefining them
Scott D Phillips [Fri, 11 Nov 2016 18:55:57 +0000 (10:55 -0800)]
winks: undef windows SDK macros before redefining them

Redefining an already defined macro causes a warning from the msvc
preprocessor

https://bugzilla.gnome.org/show_bug.cgi?id=774293

8 years agoRemove 'return' from `void` functions
Scott D Phillips [Fri, 11 Nov 2016 18:54:52 +0000 (10:54 -0800)]
Remove 'return' from `void` functions

https://bugzilla.gnome.org/show_bug.cgi?id=774293

8 years agompegtsdemux: fix operator precedence in SAFE_FOURCC_ARGS
Scott D Phillips [Fri, 11 Nov 2016 18:51:49 +0000 (10:51 -0800)]
mpegtsdemux: fix operator precedence in SAFE_FOURCC_ARGS

Type cast has higher precedence than bitwise shift, so the third
argument will truncate to 8 bits and then shift right by 8 bits
resulting in constant zero.

https://bugzilla.gnome.org/show_bug.cgi?id=774293

8 years agohlsdemux: Fix NULL pointer dereference when checking if there is a next fragment
Sebastian Dröge [Sat, 12 Nov 2016 08:38:04 +0000 (10:38 +0200)]
hlsdemux: Fix NULL pointer dereference when checking if there is a next fragment

Thanks to Aleksandr <tumaleksandr@yandex.ua> for reporting and
suggesting the fix.

https://bugzilla.gnome.org/show_bug.cgi?id=774287

8 years agomeson: don't add_global_arguments when being built as a subproject
Scott D Phillips [Thu, 27 Oct 2016 05:36:17 +0000 (22:36 -0700)]
meson: don't add_global_arguments when being built as a subproject

https://bugzilla.gnome.org/show_bug.cgi?id=773568

8 years agogl: x11-xcb is not a required dependency
Tim-Philipp Müller [Fri, 11 Nov 2016 01:05:55 +0000 (01:05 +0000)]
gl: x11-xcb is not a required dependency

Don't fail if it's not found..

8 years agokmssink: do not get kms bo pitch for planar formats
Philipp Zabel [Wed, 19 Oct 2016 12:54:44 +0000 (14:54 +0200)]
kmssink: do not get kms bo pitch for planar formats

The API is ill defined for planar formats in case of dumb BOs,
do not use the reported stride in those cases.

https://bugzilla.gnome.org/show_bug.cgi?id=773473

8 years agokmssink: remove custom gst_kms_sink_get_times
Michael Tretter [Tue, 18 Oct 2016 07:19:09 +0000 (09:19 +0200)]
kmssink: remove custom gst_kms_sink_get_times

The default get_times() function of the base sink is just fine.

Remove the custom get_times() function, because the default function
already reads the timestamps from the buffers.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
https://bugzilla.gnome.org/show_bug.cgi?id=773473

8 years agosmoothstreaming: use the duration from the list of fragments if not present in the...
Matthew Waters [Thu, 10 Nov 2016 06:20:27 +0000 (17:20 +1100)]
smoothstreaming: use the duration from the list of fragments if not present in the manifest

Provides a more accurate duration for live streams that may be minutes
or hours in front of the earliest fragment.

https://bugzilla.gnome.org/show_bug.cgi?id=774178

8 years agosmoothstreaming: implement adaptivedemux's get_live_seek_range()
Matthew Waters [Thu, 10 Nov 2016 06:18:36 +0000 (17:18 +1100)]
smoothstreaming: implement adaptivedemux's get_live_seek_range()

Allows seeking through the available fragments that are still available
on the server as specified by the DVRWindowLength attribute in the
manifest.

https://bugzilla.gnome.org/show_bug.cgi?id=774178

8 years agoadaptivedemux: reset the manifest failure count after a successful retrieval
Matthew Waters [Wed, 9 Nov 2016 10:20:21 +0000 (21:20 +1100)]
adaptivedemux: reset the manifest failure count after a successful retrieval

This changes the failure case to require a consecutive number of
failures rather than being spread out over the entire stream.

Fixes the case where fetching the manifest was intermittent.

https://bugzilla.gnome.org/show_bug.cgi?id=774177

8 years agoadaptivedemux: call the subclass for duration queries for live sources
Matthew Waters [Wed, 9 Nov 2016 10:18:06 +0000 (21:18 +1100)]
adaptivedemux: call the subclass for duration queries for live sources

Otherwise, the duration query was failing for live sources

https://bugzilla.gnome.org/show_bug.cgi?id=774177

8 years agoadaptivedemux: restart the manifest update task after a seek
Matthew Waters [Wed, 9 Nov 2016 10:15:29 +0000 (21:15 +1100)]
adaptivedemux: restart the manifest update task after a seek

For formats that need to update the manifest to know about new
fragments as they're being written by the server would never receive an
updated fragment list after a seek event

https://bugzilla.gnome.org/show_bug.cgi?id=774177

8 years agompegts: fix typo in GstMpegtsDescriptor
Dongil Park [Wed, 9 Nov 2016 08:10:21 +0000 (17:10 +0900)]
mpegts: fix typo in GstMpegtsDescriptor

https://bugzilla.gnome.org/show_bug.cgi?id=774132

8 years agogl/build: add missing build files
Matthew Waters [Thu, 10 Nov 2016 09:34:53 +0000 (20:34 +1100)]
gl/build: add missing build files

3f7b54988186de17ec4e5b5566074cfaa392819a was incomplete :(

8 years agogl/utils: move gen_shader() to the plugin and remove del_shader()
Matthew Waters [Thu, 10 Nov 2016 08:14:24 +0000 (19:14 +1100)]
gl/utils: move gen_shader() to the plugin and remove del_shader()

gst_gl_context_del_shader() can be replaced by a g_object_unref().
gst_gl_context_gen_shader() should be replaced by using GstGLSLStage.

8 years agoglwindow: remove unused field in public struct
Matthew Waters [Wed, 26 Oct 2016 05:53:27 +0000 (16:53 +1100)]
glwindow: remove unused field in public struct

8 years agoglwindow: remove is_running() function
Matthew Waters [Wed, 26 Oct 2016 05:30:43 +0000 (16:30 +1100)]
glwindow: remove is_running() function

It isn't necessary in correctly written programs.