Get the library/pathes from the installed PRL files in PREFIX/plugins or
PREFIX/qml and add it to Libs.Private in the .pc files. So the libraries are
kept in those folders rather than copied in the libs directory.

We get exactly the libs and order that qmake would use when linking.

This way we don't need a forced call during each init.

And InitArea becomes the real CommonInit

That's pretty much all it does now and it's more explicit in each module when
we place the picture after some dimension changed.

No functional change

Structures with no field should not be used, it's Undefined Behavior:
https://stackoverflow.com/questions/3849334/sizeof-empty-structure-is-0-in-c-and-1-in-c-why

And the table is using PICTURE_PLANE_MAX not VOUT_MAX_PLANES.

This way, the UI elements are a lot more accessible when working with the visual editor.

Vibhoothi authored 5 years ago and Felix Paul Kühne committed 5 years ago

Before adding this constraint, when we resize the clear-queue and repeat, shuffle
gets overlappeed in the Split View which is not intented and this commeit fixes it

Signed-off-by: Felix Paul Kühne <felix@feepk.net>

Data views are now removed from/added to the superview instead of hidden/unhidden

Additionally, the media lookup was cleaned and generalized beyond video.

From man (2) open:

O_TMPFILE  requires  support by the underlying filesystem; only a subset of
Linux filesystems provide that support.  In the initial implementation, support
was provided in  the  ext2,  ext3,  ext4,  UDF, Minix,  and shmem filesystems.
Support for other filesystems has subsequently been added as follows: XFS
(Linux 3.15); Btrfs (Linux 3.16); F2FS (Linux 3.16); and ubifs (Linux 4.9)

One must check for two different error codes, EISDIR and ENOENT, when trying to
determine whether the kernel supports O_TMPFILE functionality.

Fixes #22250

regression by db8827e5

This gives a much better picture quality.

If the call fails (COLORONCOLOR not supported) this has no effect and the old
ugly stretch mode is used.

Thomas Guillem authored 5 years ago and François Cartegnie committed 5 years ago

This fixes assert triggered from vout_SizeWindow() when the fullscreen state is
changed while the vout is not started.

Thomas Guillem authored 5 years ago and François Cartegnie committed 5 years ago

And move window deletion from vout_Close() to vout_Release(). That way, the
window is always valid during the lifetime of the vout. This simplify window
NULL checks.

Note: vout_IntfDeinit() is called from vout_Close(). So, theoretically, no
functions touching the window can be called after this call. This may change in
the future if we want to control the vout from functions directly.

Romain Vimont authored 5 years ago and Thomas Guillem committed 5 years ago

In Qt, on capability changes, a signal was emitted for all capabilities
regardless of whether they actually changed.

Instead, only emit a signal for capabilities which changed.

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

Romain Vimont authored 5 years ago and Thomas Guillem committed 5 years ago

The capabilities are stored in a bitset. When it changes, it may be
useful to know its old value, to know which capability changed.

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

Romain Vimont authored 5 years ago and Thomas Guillem committed 5 years ago

There were helpers for retrieving the selected title, chapter or track,
but a helper for the selected program was missing.

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

Romain Vimont authored 5 years ago and Thomas Guillem committed 5 years ago

A title list can be held, so vlc_player_GetTitleList() returns a pointer
to a non-const title list.

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