Merging gst-examples

Move files from gst-docs into  the "subprojects/gst-docs/" subdir

Merging gst-docs

Move files from gst-devtools into  the "subprojects/gst-devtools/" subdir

Merging gst-devtools

Move files from gst-editing-services into  the "subprojects/gst-editing-services/" subdir

Merging gst-editing-services

Move files from gst-rtsp-server into  the "subprojects/gst-rtsp-server/" subdir

Merging gst-rtsp-server

Move files from gst-ci into  the "subprojects/gst-ci/" subdir

Merging gst-ci

Move files from gst-libav into  the "subprojects/gst-libav/" subdir

Merging gst-libav

Move files from gst-omx into  the "subprojects/gst-omx/" subdir

Merging gst-omx

Move files from gstreamer-vaapi into  the "subprojects/gstreamer-vaapi/" subdir

Merging gstreamer-vaapi

Move files from gst-plugins-ugly into  the "subprojects/gst-plugins-ugly/" subdir

Merging gst-plugins-ugly

Move files from gst-plugins-bad into  the "subprojects/gst-plugins-bad/" subdir

Merging gst-plugins-bad

Move files from gst-plugins-good into  the "subprojects/gst-plugins-good/" subdir

Merging gst-plugins-good

Move files from gst-plugins-base into  the "subprojects/gst-plugins-base/" subdir

Merging gst-plugins-base

Move files from gstreamer into  the "subprojects/gstreamer/" subdir

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

Release 1.19.2

scripts/release: fix license files paths

The previous version only worked when the script was run from
the root source directory

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-docs/-/merge_requests/165>

audiobuffersplit: Remove unneeded buffer_clip wrapper

This is just a small cleanup noticed while reading.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2544>

opencv: cvtracker: add draw property

This property controls the drawing of rectangle around the tracked object.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2454>

opencv: add cvtracker plugin

This adds an object tracker plugin.
Tracker implementations from https://docs.opencv.org/3.4/d0/d0a/classcv_1_1Tracker.html

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2454>

opencv: patch to ensure headers are detected

This patch is used to ensure opencv headers are detected.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2454>

rtph263pdepay: flag keyframes on output buffers

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1091>

clocksync: Add some debug output to the clock waiting code

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/841>

mpegtsmux, mpegpsmux: remove GPL from choice of licenses and add SPDX license identifiers

Some people need to avoid inclusion of GPL code for their use cases and thus
get easily spooked by GPL license headers. This code is actually licensed
under different licenses, only one of which is GPL, and it's already possible
to just upgrade from LGPL to GPL anyway so having the GPL listed explicitly
as one of the choices doesn't really add anything. So remove GPL from the list
and also add SPDX license identifiers while we're at it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2539>

pbutils: codec-utils: fix g-ir-scanner warning

Warning: GstPbutils: gst_codec_utils_h264_get_profile_flags_level:
unknown parameter 'codec_data' in documentation comment, should be 'codecs_data

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1279>

wasapideviceprovider: Add support for dynamic device add/remove

Adding IMMDeviceEnumerator::RegisterEndpointNotificationCallback
in order to support device monitoring.
On OnDeviceAdded(), OnDeviceRemoved(), and OnDefaultDeviceChanged()
callback, wasapi device provider implementation will enumerate
devices again and will notify newly added and removed device
via GstDeviceProvider API.

As a bonus point, this IMMDeviceEnumerator abstraction object
will spawn a dedicated internal COM thread, so various COM thread
related issues of WASAPI plugin can be resolved by this commit.

Fixes: https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/issues/1649
Fixes: https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/issues/1110
Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2484>

x264enc: Strip trailing whitespace from x264's log messages

x264 has linebreaks at the end of its log messages that we should
ignore.

Add G_GNUC_PRINTF to the callback to make sure GCC lets us forward the
format string to another function marked as printf (g_strdup_vprintf)
without triggering -Wformat-nonliteral.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-ugly/-/merge_requests/85>

x264enc: Add the func/line info to the log callback

It looks awkward when it's missing. At least point us at the callback
function.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-ugly/-/merge_requests/85>

x264enc: Reduce log spam about bit depth support

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-ugly/-/merge_requests/85>

alsasink: Allow stop() function to happen during failing writes

In ALSA, there is possible temporary failures that may require a retry,
though in certain situation, this may leak to the write() function
holding on a lock forever preventing the pipeline from going to pause
or stop. Fix this by shortly dropping the lock between retries.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1261>

alsasink: Improve logging in write() function

This moves the "written X frames" lower so that we don't trace
confusing negative values on errors and add the error code in the
"Write error" log.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1261>

openjpeg: add unit test

Test various format supported with subframes.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

jpeg2000parse + openjpeg: Switch striped mode to its own caps

It's not compatible with any other element that use the non-striped
mode. In addition, in this mode, we require that every frame have the
same number of stripes or that the MARKER bit be present, which is
different from the other
formats too.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

openjpegdec: Fix crash with AYUV64 in subframe mode

Remove useless generic fill_frame methods to use
the packed one for AYUV and AYUV64.

Fix gst-launch-1.0 -v videotestsrc  !
video/x-raw,width=640,height=480,format=AYUV64  !  openjpegenc
num-stripes=8   !  openjpegdec max-threads=8  ! videoconvert !
autovideosink sync=false

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

openjpegdec: support for a multithreaded decoding.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

openjpegdec: enable sub frame mode

Rebuild output frame from multiple stripes input.

Keep the first frame and fill it with the following stripes to finish
a complete frame only once.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

mpegtsmux: Require frame alignment for JPEG 2000

We have yet to implement stripe alignment with the required descriptor.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

openjpegdec: Reject stripes for now

They're not implemented.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

openjpegenc: Only allow stripe with image/x-jpc format

It's the only format that our MPEG-TS muxer allows and the carriage of
JPEG 2000 stripes is only defined for MPEG-TS as far as I know.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

openjpegenc: support for a multithreaded encoding.

This commit introduces a multithreaded encoder allowing
to encode mulitple stripes or subframes in separated threads.
This feature aims to enhance the overall latency of a codec
pipeline.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/979>

wpe: context thread dispatch fixes

Use dedicated mutex/cond/flag for jobs being dispatched in the context thread.
The previous code was signalling the thread startup condition, which is wrong.
When WPEContextThread::dispatch() is invoked it means the thread has already
correctly been started up.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2533>

wpe: Properly wait on context thread startup condition

Fixes #1661

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2533>

doc: Update kmssink caps cache

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2530>

kmssink: Add RGB16/BGR16 support

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2530>

kmssink: Sort format according to GST_VIDEO_FORMATS_ALL

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2530>

kmssink: Remove big endian format inversion

This has been a bad interpretation of the DRM docuemntation. The formats are
fixed regardless the CPU, but for some formats, they expressed in the opposite
order as GStreamer. Same change was done in waylandsink 2 years ago.

https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/commit/b393b650ab9bfb9654fc116163ab331907216d74

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2530>

kmssink: Add NV61 support

This identically handled to NV16, so no reason not to inclue it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2530>

kmssink: Add NV24 support

This was tested on RK3566 platform, using vendor DRM driver.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2530>

codecs: av1decoder: Use GstFlowReturn everywhere

The same modification as that of VP8 decoder

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2528>

codecs: vp9decoder: Use GstFlowReturn everywhere

The same modification as that of VP8 decoder

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2528>

codecs: vp8decoder: Use GstFlowReturn everywhere

boolean return value is not sufficient for representing the reason
of error in most cases. For instance, any errors around new_sequence()
would mean negotiation error, not just *ERROR*.
And some subclasses will allocate buffer/memory/surface on new_picture()
but it could be failed because of expected error, likely flushing

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2528>

gst: Initialize optional event/message fields when parsing

These might not exist inside the structure and then we would potentially
keep around uninitialized memory from the caller in the out parameter.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/887>

videodecoder: Add properties to automatically request sync points and vfunc to allow subclasses to handle packet loss / missing data

Subclasses could use the new vfunc to activate packet loss concealment,
for example.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1274>

mpegtspacketizer: memcmp potentially seen_before data

Theoretically the version number is incremented every time there's a new
section, but in a world of streaming we can't easily make that
assumption.

An example of a broken use case is when we're cat-ing two mpeg-ts files
together, which is equivalent of capturing a DVB stream while switching
channels. A set-top box would know that we switched the channels and
reset the demuxer, but in practice this might not happen.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2468>

player: Fix/add various annotations

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2538>

meson: va: Make AV1 support always optional

Otherwise meson configure with -Dva=enabled will be failed
when installed libva version is < 1.8

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2532>

va: caps: Don't use image formats for decoded frames.

Initially we tried to use the internal color conversion used in i965
and Gallium drivers when decoding. But this approach has showed
limitations and problems.

This patch removes completely the possible color conversion at
decoding, since it show problems with deinterlacing, for example:

gst-launch-1.0 filesrc location=interlaced.mpg2 ! parsebin ! vampeg2dec ! vadeinterlace ! xvimagesink

Allowing only the surface formats when decoding is more stable.

For color conversion is better to do it explicitly with vapostproc.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2531>

test: bitwriter: Add a test for reset_and_get_data when not byte unaligned.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/886>

bitwriter: Fix a memory leak in reset_and_get_buffer.

We should record the ownership of the data before we reset the bitwriter.
Or we will always dup the buffer data and leak the memory.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/886>

bitwriter: Fix the trailing bits lost when getting its data.

In reset_and_get_data and reset_and_get_buffer, it fails to include
the trailing bits less than 8. So, when the bit_size is not byte
aligned, the trailing bits are lost in the return buffer.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/886>

player: Add missing nullable annotations

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2197>

play: Add missing nullable annotations

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2197>

sys: shm: Define shm_enable and shm_deps before escape meson subdir

Fixes meson configure in tests if the shm plugin is disabled

Fixes #1664

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2534>

d3d11videosink: Display title of content if possible

Update title text of window (currently it's always "Direct3D11 renderer")
when we are rendering on internal HWND, not external HWND.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2529>

d3d11videosink: Remove unused enum value

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2529>

msdkenc: Pass color properties to MediaSDK for encoding

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2523>

avcodecmap: Add support for GBRA_10LE/BE

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-libav/-/merge_requests/142>

msdkh265enc: Add profile main10 still picture for hevc

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2524>

waylandsink: Fix double render check

Our code does not support rendering twice the same wl_buffer in a row, so it
tries to skip that case, but for this it relied on the GstBuffer pointer,
while the cache actually works at the GstMemory level now. To avoid this
compare the GstWlBuffer instead.

This fixes crash when use in zero-copy with videorate element.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2526>

codecs: mpeg2decoder: Use tsg framerate for latency.

Latency setting relies on src pad caps, but they aren't set when the
function is called, and latency is never updated.

In order to fix it, this patch uses TSG framerate first, and if it's
not set yet, sinkpad caps are used to get the framerate.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2514>

d3d11decoder: Refactor for more unified decoding flow

... and various code cleanup.

* Move spreaded decoding API calls into one method
Previously, decoding flow of most codecs are
- Call DecoderBeginFrame() on start_picture()
- Call {Get,Release}DecoderBuffer() on decode_slice()
- Call SubmitDecoderBuffers() and DecoderEndFrame() on end_picture()
Such spreaded API calls make it hard to keep track of status
of decoding. Now it will be done at once in a new method.

* Drop a code for non-zero wBadSliceChopping
When bitstream buffer provided by driver is not sufficient
to write compressed bitstream data, host decoder needs to make use
of wBadSliceChopping so that driver can understand there are
multiple bitstream buffer. But it's a bit unrealistic and
not tested. Since FFMpeg's DXVA implemetaion doesn't support it,
we might be able to ignore the case for now.

* Make code more portable
Consider common logic of GstCodecs -> DXVA translation for all D3D APIs
(i,e., D3D9, D3D11, and D3D12).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2525>

d3d11decoder: Remove duplicated class_init and property related code

Move them into the decoder helper code to remove duplication

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2525>

d3d11: Get rid of "extern "C"" wrapping for GST_DEBUG_CATEGORY_EXTERN

Instead, change the file defining debug category to cpp

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/2525>

videodecoder: Fix min-force-key-unit-interval logic and logging

The new keyframe is needed when the deadline of the buffer has exeeded
the waiting time, not while it is within it.

Also, since we look at the deadline of the frame, log that instead of PTS.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-base/-/merge_requests/1278>

tests: skip cc tests if plugin is disabled

Skip the closedcaption element tests if the
closedcaption option is disabled at compile
time (i.e. -Dclosedcaption=disabled).

v2: rename pangocairo_dep to avoid conflict
with later definition in ext/ttml/meson.build
as suggested by @tpm.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/merge_requests/1388>

rtphdrhext-twcc: Return failure on map failure

This feels like exactly like a case that should fail.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1059>

rtphdrext: Update write() API to return a signed value

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-good/-/merge_requests/1059>