This is already how the UWP vout works but by not compiling some code.

It matches the API calls better and is more readable.

We already know it's d3d9

Zhao Zhili authored 6 years ago and Thomas Guillem committed 6 years ago

Since p_sys is created by vlc_obj_calloc.

Signed-off-by: Thomas Guillem <thomas@gllm.fr>

Zhao Zhili authored 6 years ago and Thomas Guillem committed 6 years ago

Mutexes haven't been initialized yet and Clean will call vlc_mutex_destroy.

Signed-off-by: Thomas Guillem <thomas@gllm.fr>

Zhao Zhili authored 6 years ago and Thomas Guillem committed 6 years ago

Since p_sys is created by vlc_obj_calloc.

Signed-off-by: Thomas Guillem <thomas@gllm.fr>

...before destroying them. Otherwise we can no longer keep track of
which picture buffer the display server still references, leading to
potential corruption.

To force the server to release all buffers, we must first detach the
any outstanding buffer from the display surface.

Alexandre Janniaux authored 6 years ago and Rémi Denis-Courmont committed 6 years ago

Picture is released by the core, but we can only release it for real after
receiving a wl_buffer::release event.

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

unused

video stream resync/vout can't be reused fast enough.
partial revert of 0160bec1

Code relying on AVX-2 (like dav1d) cannot work without an alignment of 32 bytes
at least. Even avcodec has a requirement of 64 (likely to get at least 32 on all
planes). So it's probably time to upgrade.

picture_Setup ensures a horizontal alignment to 32 bytes and a vertical to 16
bytes. So for all buffers allocated by the core, we have a size multiple of 512
bytes.

It is more consistent with the way of using/defining enums in the rest of the libvlc API.

Hamza Parnica authored 6 years ago and Hugo Beauzée-Luyssen committed 6 years ago

Signed-off-by: Hugo Beauzée-Luyssen <hugo@beauzee.fr>

Fix #17368

ref #17368

This API is mostly meant to simplify thumbnailing management, by
exposing a simple picture type, that can be probed for its buffer or
saved to a file

dav1d may return the same buffer/picture_t multiple times. In that case we
can't output the same picture_t pointer multiple time, so we clone it
every time.

Fixes decoding of https://code.videolan.org/videolan/dav1d/uploads/86ea12faad3cd7bd6ab8e6cca34a5b01/wanderers_aq_1.webm