hlsdemux2: Fix position tracking

Remove the "pending advance" hack and instead rely on the base stream current
position to track our position (instead of a potentially NULL "current
segment").

Also ensure the media playlists are always refreshed with valid stream time,
even if there is no current segment.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2679>

hlsdemux2: Add a new method to synchronize two media playlist

This allows transfering the stream time of the playlist to an updated
variant/rendition.

This allows updating that information without having a "current segment"

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2679>

adaptivedemux2: Initialize and use stream start/current position

The stream start and current position would be properly set when seeking or
activating a stream after playback started. But it would never be properly
initialized.

Set it to NONE initially to indicate to subclasses that no position has been
tracked yet. This will allow them to detect initial stream usage.

Futhermore, once the initial streams setup is done, make sure that it is set to
a valid initial value:
* The minimum stream time in live
* Or else the period start

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2679>

v4l2videodec: Fix activation of internal pool

If the driver does not support VIDIOC_CREATE_BUFS ioctl, the pool
configuration may get changed, which requires a validation. This would
fail to activate a pool in a case it shouldn't normally fail unless we
are out of memory.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2456>

dv, opusparse: fix duplicate symbols in static build

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/issues/1262

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2670>

d3d11: Enumerate devices via native DXGI API

D3D11CreateDevice() can fail for some reason (e.g., buggy GPU driver)
and then gst_d3d11_device_new() will return nullptr.
Use DXGI API directly instead of relying on the gst_d3d11_device_new()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2663>

examples: Add an example for application texture sharing

This example shows GstD3D11BufferPool usage and a way of
D3D11 texture sharing between application and GStreamer via appsrc.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2646>

examples: Add d3d11decoder zero-copy rendering example

This example code demonstrates D3D11 device sharing between
application and GStreamer. Application can access texture
using appsink and it can be rendered on application's window without
any copy operation.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2646>

d3d11: Add a helper method for GstContext creation

Add gst_d3d11_context_new() method for application to create
GstContext using GstD3D11Device.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2646>

d3d11memory: Add support for wrapping application texture

Add gst_d3d11_allocator_alloc_wrapped() method to support wrapping
application texture using GstD3D11Memory without copy.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2646>

d3d11: Install library headers

Our Direct3D11 abstraction layer has been improved and
it gained good shape from API point of view.
Also, On Windows, GstD3D11 has various advantages over GstGL
in terms of compatibility/stability/feature/performance.
Note that WGL implementation is known to be buggy for some
drivers/vendors/scenario (that's a reason why Google implemented ANGLE).
Moreover, GstGL is not fully optimized for Windows unfortunately.

It's the time to open this interface to application developers
for various optimized processing using our Direct3D11
infrastructure.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2646>

d3d11: Update gst_d3d11_allocation_params_new() interface

Define GST_D3D11_ALLOCATION_FLAG_DEFAULT (0) value instead of
casting enum type everywhere. And pass D3D11_RESOURCE_MISC_FLAG value

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2646>

va: postproc: add scale-method property

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2654>

va: filter/postproc: move unconditional props to vavpp

Only conditional/dynamic properties should be installed/handled
by vafilter.  Thus, move and install the unconditional/static
properties in vavpp.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2654>

va: compositor/filter: use scale method setter

Use the filter's scale method member instead of the sample's member.

Removes the filter_flags member from the composite sample.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2654>

va: filter: add scale method field and setter

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2654>

va: Add vajpegdec element.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1575>

va: allocator: Add hack for no fourcc when surface creation.

This patch adds general mechanism for handling specific hacks. In this
case for jpeg decoder in i965 driver, which cannot create surfaces
with fourcc specified.

From jpeg decoder to the allocator, which creates the surfaces,
there's a non-simple path: basedec pseudo-class adds a hacks guint32
which will be set by actual elements (vajpegdec, in this case) and
basedec will always set the hack to the allocator when the allocator
is instantiated.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1575>

va: Add RGBP format definition.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1575>

va: caps: Add jpeg fields in coded caps.

Given the supported rt formats in a profile/entrypoint config it's
possible to know the supported JPEG colorspace and subsampling. This
patch adds this information in coded caps to a safer autoplugging
after jpegparser.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1575>

va: Add jpegdecoder base class.

This base class is intented for hardware accelerated decoders, but since
only VA uses it, it will be kept internally in va plugin.

It follows the same logic as the others video decoders in the library but.
as JPEG are independet images, there's no need to handle a DBP so no need
of a picture object. Instead a scan object is added with all the structures
required to decode the image (huffman and quant tables, mcus, etc.).

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1575>

va: build: Remove duplicated source.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1575>

cudamemorycopy: Make sure writable caps before removing fields

The caps to be modified may not be writable when D3D11/GL/NVMM
are all disabled.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2664>

tests: skip unit tests for dependency-less elements that have been disabled

Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-bad/-/issues/1136

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2660>

examples: don't try and build jack examples if jack was disabled

Fixes meson build ERROR: Unknown variable "libjack_dep".

Fixes #1301

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2656>

va: allocator: Use always lseek to get dmabuf size.

Gallium drivers historically have reported strange dmabuf sizes, from always
zero to the whole frame (multiple fds). The simplest solution is to use lseek
SEEK_END to get the prime descriptor size.

Also the allocator raises a warning if both values differ in order to report
it to driver.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2574>

va: allocator: Fix translation of VADRMPRIMESurfaceDescriptor

VADRMPRIMESurfaceDescriptor structure describes the offsets from the
point of view of the specific handle (DMABuf). While GstVideoInfo
(and the meta) describes offsets from the point of the view of the
GstBuffer, an aggregate of all the GstMemory (1 per handle).

This changes combined with [Mesa Fix](https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/16813)
fixes decoding failure with AMD driver.

Fixes #1223

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2574>

d3d11: Revert staging buffer pool implementation

The staging buffer pool implementation was added to improve
throughput performance since we can avoid per-frame
CPU copy operation via staging texture but it turned out that
we can not make it thread safe. See
https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1182

Reverting the staging texture implementation as it does not show
any visible value.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2651>

qsvdecoder: Fix system memory alignment

Do right padding, not left

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2651>

d3d11decoder: Fix texture download

Stride applied to src/dst was reversed

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2651>

jpegparse: Add YUV440 sampling support.

Found some samples with this sampling, so in order to support it,
JPEG2000 header cannot be used anymore, a this patch also adds the
enums for each supported colorspace and sampling, avoiding hacks.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2641>

gst: add missing define guard

If compiled with -Dgstreamer:gst_debug=false and we have
GST_REMOVE_DISABLED defined we will get the following compiler error:

```
[...]/libgstreamer-1.0.so.0.2100.0.p/gst.c.o: in function `gst_deinit':
[...]/gst/gst.c:1258: undefined reference to `_priv_gst_debug_cleanup'
[...] hidden symbol `_priv_gst_debug_cleanup' isn't defined
```

Add the missing define guard to avoid this.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2648>

va: Use driver implementation macro rather than function.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2642>

bin: Fix race conditions in tests

The latency messages are non-deterministic and can arrive before/after
async-done or during state-changes as they are posted by e.g. sinks from
their streaming thread but bins are finishing asynchronous state changes
from a secondary helper thread.

To solve this, expect latency messages at any time and assert that we
receive one at some point during the test.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2643>

h265parse: Fix uninitialized width and height when update src caps.

The commit b90d0274 introduces uninitialized width and height when we
consider to change the "pixel-aspect-ratio" for some interlaced stream.
We need to check the resolution in the src caps, and if no resolution
info found, there is no need to consider the aspect ratio.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2630>

d3d11: Fix some typos

Fix up spelling and d3d11testsrc class meta.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2644>

element: Fix requesting of pads with string templates

Previously it was only possible to request them with the exact template
name, e.g. 'src_%s', but not with "instantiated" names that would match
this template, e.g.'src_foo_bar'.

This is now possible and a test was added for this, in addition to
fixing a previously invalid test.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2635>

d3d11compositor: Performance optimization

Removing glvideomixer-like nuance (it was initially referenced)
and rewriting element since it's not an optimal design at all
from performance point of view.

* Remove wrapper bin (and internal conversion/upload/download elements)
  which will waste CPU/GPU resources. Conversion/blending can be done by the
  d3d11compositor element at once.
* Add support YUV blending without RGB conversion.
  The RGB <-> YUV conversion is completely unnecessary since YUV textures
  support blending as well.
* Remove complicated blending operation properties since it's hard
  to use from application point of view. Instead, adding "operator" property
  like what compositor element does.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2631>

d3d11converter: Fix alpha factor update

converter should set entire constant buffer values since it's mapped
with write-discard flag

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2631>

gstwaylandsink: Add rotate-method property

Similar to and inspired by glimagesink and gtkglsink.

Using the Wayland buffer transform API allows to offload
rotate operations to the Wayland compositor. This can have
several advantages:
 - The Wayland compositor may be able to use hardware plane
   capabilities to do the rotation.
 - In case of pre-rotated content on rotated outputs the
   rotations may equal out, potentially allowing the
   compositor to use hardware planes even if they don't
   support rotate operations.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2543>

Add GstMemoryMapInfo to be used with g_auto()

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2412>

GstVideoFrame: Add g_auto() support

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2412>

Add GstBufferMapInfo to be used with g_auto()

We need a separate typedef for this feature because GstMapInfo itself
can be initialized by gst_memory_map() in which case info.memory should
not be unreffed.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2412>

gst-inspect: Fix inspection of third-party plugins

Since commit de57657de1d1916503b4ad451ac13a3e191465f8 inspecting a third-party
plugin would trigger a segfault (Address boundary error) due to the missing
sentinel in the list of GStreamer modules.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2636>

d3d11decoder: Check 16K resolution support

16K decoding is supported by some GPUs

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2629>

ges/gstframepositioner: don't create one compositor per frame meta

Instead, cache the looked up operator property

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2623>

d3d11decoder: Add support for non-zero crop-{x,y} position

AVC and HEVC define crop rectangle and the x/y coordinates might
not be zero. This commit will address the non-zero x/y offset coordinates
via GstVideoCropMeta if downstream supports the meta and d3d11 memory.
Otherwise decoder will copy decoded texture into output frame.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2624>

d3d11converter: Implement filling border color

Equivalent to GST_VIDEO_CONVERTER_OPT_BORDER_ARGB and
GST_VIDEO_CONVERTER_OPT_FILL_BORDER options of GstVideoConverter

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2622>

d3d11converter: Objectify converter

... and set conversion options via property

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2622>

d3d11shader: Remove GstD3D11Quad helper object

It's not very generic to be used for various scenario. Use native D3D11
APIs directly instead.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2622>

d3d11compositor: Don't use GstD3D11Quad

The helper object will be removed

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2622>

d3d11overlaycompositor: Fix HLSL compiler warning

warning X3206: 'Sample': implicit truncation of vector type

And don't use GstD3D11Quad since it will result in redundant API calls

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2622>

cccombiner: expose output-padding property

When schedule=true and output-padding=false, cccombiner will not
inject padding in the output closed caption meta stream.

The property has no effect when schedule=false.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1621>

rtpsession: properly initialise favor-new property

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2625>

hlsdemux2: Look for mpegts synchronization point further

Some mpeg-ts streams have extra data at the beginning. While it's not ideal, we
should be able to cope with it.

Therefore increase the initial search window for at least 4 consecutive
synchronization points to 1kB.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2626>

webrtcbin: Limit sink query to sink pads

This allows the reception of streams that don't exactly match
the codec preferences. In particular, the ssrc in the codec preferences
is local sender SSRC, the other side is expected to send a different SSRC.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2615>

avviddec, video.c, h265parse: Workaround for broken field-based interlaced encoders

Some encoders (e.g. Makito) have H265 field-based interlacing, but then
also specify an 1:2 pixel aspect ratio. That makes it kind-of work with
decoders that don't properly support field-based decoding, but makes us
end up with the wrong aspect ratio if we implement everything properly.
As a workaround, detect 1:2 pixel aspect ratio for field-based
interlacing, and check if making that 1:1 would make the new display
aspect ratio common. In that case, we override it with 1:1.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2577>

gst-env.py: drop 'gst-' prefix from branch name in prompt

Probably leftover from the days where we would have a gst-foo.sh
script to set up the devenv for a particular checkout/branch.

Kind of confusing now if you're working on a named branch and
it just adds an extra gst- prefix in the prompt.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2594>

docs: ensure coretracers plugin index page is index.html

And not blank.html

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2592>

docs: make sure rtspclientsink plugin docs index page is called index.html

.. instead of plugin-index.html.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2592>

gst-inspect: print link to documentation for gstreamer elements

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2592>

srtsrc: add "keep-listening" property to avoid EOS on disconnect

The property 'keep-listening' avoids EOS
when the remote client disconnects.

It can be useful to a keep a pipeline alive
when the srt connection drops remotely.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/967>

srtsrc: remove dead code

Remove code useless since
132e3a1af9deb1d2cdf84d22b954967207bcc03a

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/967>

qsvh264dec: Use newly added avcC data parsing API

Use gst_h264_parser_parse_decoder_config_record() method to parse
codec_data.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2449>

h264parse: Use newly added avcC data parsing API

Use gst_h264_parser_parse_decoder_config_record() method to parse
codec_data.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2449>

h264decoder: Use newly added avcC data parsing API

Use gst_h264_parser_parse_decoder_config_record() method to parse
codec_data.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2449>

h264parser: Add an API for AVCDecoderConfigurationRecord parsing

Add a method for AVC configuration date parsing

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2449>

matroskademux: Avoid integer-overflow resulting in heap corruption in WavPack header handling code

blocksize + WAVPACK4_HEADER_SIZE might overflow gsize, which then
results in allocating a very small buffer. Into that buffer blocksize
data is memcpy'd later which then causes out of bound writes and can
potentially lead to anything from crashes to remote code execution.

Thanks to Adam Doupe for analyzing and reporting the issue.

CVE: CVE-2022-1920

https://gstreamer.freedesktop.org/security/sa-2022-0004.html

Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1226

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2612>

qtdemux: Fix integer overflows in zlib decompression code

Various variables were of smaller types than needed and there were no
checks for any overflows when doing additions on the sizes. This is all
checked now.

In addition the size of the decompressed data is limited to 200MB now as
any larger sizes are likely pathological and we can avoid out of memory
situations in many cases like this.

Also fix a bug where the available output size on the next iteration in
the zlib decompression code was provided too large and could
potentially lead to out of bound writes.

Thanks to Adam Doupe for analyzing and reporting the issue.

CVE: tbd

https://gstreamer.freedesktop.org/security/sa-2022-0003.html

Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1225

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2610>

matroskademux: Fix integer overflows in zlib/bz2/etc decompression code

Various variables were of smaller types than needed and there were no
checks for any overflows when doing additions on the sizes. This is all
checked now.

In addition the size of the decompressed data is limited to 120MB now as
any larger sizes are likely pathological and we can avoid out of memory
situations in many cases like this.

Also fix a bug where the available output size on the next iteration in
the zlib/bz2 decompression code was provided too large and could
potentially lead to out of bound writes.

Thanks to Adam Doupe for analyzing and reporting the issue.

CVE: CVE-2022-1922, CVE-2022-1923, CVE-2022-1924, CVE-2022-1925

https://gstreamer.freedesktop.org/security/sa-2022-0002.html

Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1225

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2610>

avidemux: Fix integer overflow resulting in heap corruption in DIB buffer inversion code

Check that width*bpp/8 doesn't overflow a guint and also that
height*stride fits into the provided buffer without overflowing.

Thanks to Adam Doupe for analyzing and reporting the issue.

CVE: CVE-2022-1921

See https://gstreamer.freedesktop.org/security/sa-2022-0001.html

Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1224

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2608>

queuearray: Fix potential heap overflow when expanding GstQueueArray

Check that elt_size*newsize doesn't overflow when expanding a
GstQueueArray, which has the potential for a heap overwrite.

Co-authored-by: Sebastian Dröge <sebastian@centricular.com>
Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1232

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2606>

smpte: Fix integer overflow with possible heap corruption in GstMask creation.

Check that width*height*sizeof(guint32) doesn't overflow when
allocated user_data for mask, potential for heap overwrite when
inverting.

Fixes https://gitlab.freedesktop.org/gstreamer/gstreamer/-/issues/1231

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2603>

decklinkvideosink: Fix selection of > PAL widescreen modes

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2601>

docs: harness: Fix example

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2602>

decodebin3: fix EOS event sequence

See docs:
https://gstreamer.freedesktop.org/documentation/additional/design/seqnums.html?gi-language=c#seqnums-sequence-numbers

Per docs:
When a sink element receives an EOS event and creates a new EOS
message to post, it should copy the seqnum from the event to the
message because the EOS message is a consequence of the EOS event
being received.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2588>

vaapi: Do not disable the whole vpp when some va operations not available.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2578>

adaptivedemux2: Prevent duplicate symbols on static builds

Uses prelude header files with #defines to rename DASH and MSS
symbols duplicated in their old standalone versions.
Also redefines soup-related functions when building it for
adaptivedemux2 to prevent symbol conflicts there.

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

va: basedec: Select output format with the used chroma in decoder.

Currently, video format is decided with downstream caps intersection,
but that's not correct since chroma is not considered. The video
decoders have to decide the output format given the used chroma, not
by the downstream caps negotiation.

This patch changes that. Still, caps feature is selected by caps
negotiation, then, with the preferred caps feature, the output format
is search within that caps feature.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2569>

multiqueue: fix potential crash on shutdown

The mq we get out of the weak ref might be NULL if we're
shutting down, which could cause assertion failures or
crashes.

It might also cause miscompilations where the compiler just
optimises away the NULL check because it jumps to a code path
that then dereferences the pointer which clearly isn't going
to work. Seems like something like this happens with gcc 11.

Fixes #1262

Co-authored-by: Doug Nazar <nazard@nazar.ca>
Co-authored-by: Sebastian Dröge <sebastian@centricular.com>
Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2599>

.editorconfig: Add some indentation settings

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/1812>

multiqueue: fix warning: ‘is_query’ may be used uninitialized in this function

Fixes:
../plugins/elements/gstmultiqueue.c: In function ‘gst_multi_queue_loop’:
../plugins/elements/gstmultiqueue.c:2394:19: warning: ‘is_query’ may be used uninitialized in this function [-Wmaybe-uninitialized]
 2394 |     if (object && !is_query)
      |                   ^~~~~~~~~

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2593>

adaptivedemux2: Improve reverse playback buffering.

In reverse playback, store the lowest running time in each GOP
as the input_time for buffering purposes. That means we end up
storing at least a complete GOP before declaring buffering
100%

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2510>

adaptivedemux2: reverse playback running times

Account for running time moving non-monotonically in
reverse playback by tracking the highest running time
seen at each point.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2510>

vulkan: add vulkan overlay compositor element

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2470>

vkfullscreenquad: add support for disabling clearing

e.g. if drawing over the top of an existing image, we don'w want to
clear.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2470>

vkfullscreenquad: support setting blend operations

Allows e.g. blending with an existing image.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2470>

vkfullscreenaud: create the vulkan command pool upfront

Allows outside to reuse the same command buffer pool much more easily.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2470>

vulkan: add some missing GAutoPtr definitions

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2470>

vulkan: remove unneeded #ifdef for GAutoPtr

Not necessary anymore as we depend on a new enough Glib version

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2470>

vulkan: move element register definition to relevant element headers

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2470>

d3d11converter: Add support for colorimetry conversion

Handle color primaries and gamma functions.
HDR <-> SDR conversion (tone mapping) should be implemented as well
but not a part of this patch.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2581>

d3d11convert: Don't passthrough if colorimetry conversion is required

Different input/output colorimetry requires conversion

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2581>

d3d11window: Use SDR colorspace whenever possible

PQ uses completely different light level scale
and it should not be used for SDR bt2020.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2581>

d3d11converter: Rewrite conversion object

Rewriting GstD3D11Converter (equivalent to GstVideoConverter)
to optimize some conversion path and clean up.

* Extract YUV <-> RGB conversion matrix building method to
  utils. It will be used by other implementation
* Use calculated offset values for YCbCr <-> YPbPr conversion
  instead of hardcoded values
* Handle color range adjustment
* Move transform matrix building helper function to utils.
  The method will be used by other elements
* Use single constant buffer. Multiple constatne buffer for
  conversion pipeline is almost pointless
* Remove lots of duplicated HLSL code and split pixel shader
  code path into sampling -> colorspace conversion ->
  shader output packing
* Avoid floating point precision error around UV coordinates
* Optimize RGB -> YUV conversion path

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2581>

v4l2codecs: Add P010 pixel format

Copy V4L2_PIX_FMT_P010 define from linux header.
V4L2_PIX_FMT_P010 is the little endian definition of P010 so map
it GST_VIDEO_FORMAT_P010_10LE.
Add it v4l2 default video formats to allows v4l2 decoders to
enumerate and use it.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2590>

docs: update technical howto in moving-plugins

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2583>

ci: Disable werror for documentation build

The documentation build is running on an ancient Fedora 30 image that
cannot be updated at present, and it's now triggering deprecation
warnings in system headers:

https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2568#note_1418956

Disable werror here, it's not actually useful.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2568>

clock: Use g_atomic_rc_box for refcounting entry clocks

g_atomic_rc_box was added in GLib 2.58, and we require 2.62 now, so we
can fix the FIXME and use this.

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2568>

Bump GLib requirement to >= 2.62

Can't require 2.64 yet because of
https://gitlab.freedesktop.org/gstreamer/cerbero/-/issues/323

Part-of: <https://gitlab.freedesktop.org/gstreamer/gstreamer/-/merge_requests/2568>