It's done in the decoder device we keep during the whole lifetime of the
filter (and the pictures that will be pushed).

Make sure we're using a MMAL pipeline.

And test the input video context.

Otherwise initialize the hardware locally.

And store it in the video context that we push.

John Cox authored 5 years ago and Steve Lhomme committed 5 years ago

Signed-off-by: Steve Lhomme <robux4@ycbcr.xyz>

Regression from c515dc46

The transformation must be done after the fmt_out is fully configured.

Pierre Lamot authored 5 years ago and Thomas Guillem committed 5 years ago

  QUrl::None doesn't encode space, thus generate illformed URL

  this patch follows up "qt: fix url encoding of medialibrary endpoints"

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

Pierre Lamot authored 5 years ago and Thomas Guillem committed 5 years ago

  this fixes issues when we get the callback from a node after we changes view
  and end up displaying the content of a folder within the device view

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

Pierre Lamot authored 5 years ago and Thomas Guillem committed 5 years ago

  this fixes issues when we get the callback from a node after we changes view
  and end up displaying the content of a folder within the device view

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

Pierre Lamot authored 5 years ago and Thomas Guillem committed 5 years ago

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

Pierre Lamot authored 5 years ago and Thomas Guillem committed 5 years ago

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

Pierre Lamot authored 5 years ago and Thomas Guillem committed 5 years ago

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

It's not handled by the core.

It was used to fix decoder device creation from the core but it's not needed
since 2210c192

We don't use the video_format from it anymore. Only the vout pointer was used.

So we can create the vout thread and the decoder device without knowing the
actual display dimensions that will be used.

The vout always has a usable window, even if it's not enabled.

The signature descrambling call changed to using several steps with
intermediate variables. We know that the signature field must be
URL-decoded before descrambling, so we now rely on that.

Bernhard M. Wiedemann authored 5 years ago and Thomas Guillem committed 5 years ago

if available (in GNU tar >= 1.28)
and use ustar format to not have pax headers add nondetermistic bits
to make skins2/default.vlt reproducible.

See https://reproducible-builds.org/

 for why this is good.

This patch was done while working on reproducible builds for openSUSE.

Signed-off-by: Bernhard M. Wiedemann <bwiedemann@suse.de>
Signed-off-by: Thomas Guillem <thomas@gllm.fr>

The headers are often incompatible and create compilation issues when
rebuilding.

The headers are often incompatible and create compilation issues when
rebuilding.