Newer Windows issue this warning if we send metadata with each picture:

DXGI WARNING: IDXGISwapChain4::SetHDRMetaData: Redundant invocation on unchanged
metadata could result in presentation performance inefficiency. [ MISCELLANEOUS WARNING #295: ]

Similar code as done in 26954081 for VLC 4.0.

On 3.0 the pool video format doesn't have to match the decoder actual
padding. The quad handling to place the picture will take care of this
with the actual texture size.

We still need to fix the semiplanar dimensions although it's unlikely
the quad will use anything else than an RGB output.

Fixes green line seen in fullscreen.

nurupo authored 1 year ago and Steve Lhomme committed 1 year ago

Follow-up to ed22580d, which did this on
Linux.

(cherry picked from commit f7f346e6)
Signed-off-by: Thomas Guillem <thomas@gllm.fr>

Steve Lhomme authored 1 year ago and Jean-Baptiste Kempf committed 1 year ago

If InitVideoDecCommon() fails, it's already cleaning the decoder,
release the context and freeing p_sys.

We must not do anything and just return the error.


(cherry picked from commit fecebe15)

Pierre Lamot authored 1 year ago and Jean-Baptiste Kempf committed 1 year ago

It looks like windows doesn't consider which window is active the same way X11 does

fix: #28399

KO Myung-Hun authored 2 years ago and Steve Lhomme committed 1 year ago

LIBCn is a fork of kLIBC, and it has socklen_t in sys/socket.h.

(cherry picked from commit 4baa00c9)

(cherry picked from commit d18fd89b)
Signed-off-by: Steve Lhomme <robux4@ycbcr.xyz>

nurupo authored 1 year ago and Steve Lhomme committed 1 year ago

This fixes #28377 -- the issue when double-clicking on
MOUSE_BUTTON_CENTER (mouse wheel) would result in plugins receiving
mouse events only for the first click, as VLC was filtering out all
non-left-mouse-button double-clicks.

The non-left-mouse-button double-click events, instead of being entirely
filtered out, are now being passed as regular mouse button presses
without any indication that they are double-clicks. While it would be
more proper to pass them as double-clicks with the corresponding mouse
button being pressed, that might break some 3rd party plugins that rely
on (vlc_mouse_t.b_double_click == true) to mean that the left mouse
button was double-clicked, without checking if the left mouse button was
actually pressed. Still, even with b_double_click not being set on
non-left-mouse-button double-clicks, passing such double-clicks as
regular press+release events does fix the issue of VLC "eating up" the
second click of a MOUSE_BUTTON_CENTER, so this is a good change.

Steve Lhomme authored 1 year ago and Jean-Baptiste Kempf committed 1 year ago

It's used in Direct3D11CreateFormatResources(). We can update it every time
Direct3D11Open() is called.

Steve Lhomme authored 1 year ago and Jean-Baptiste Kempf committed 1 year ago

There might be holes in the index list but we need to check each of them
to find our driver.

Tristan Matthews authored 1 year ago and Steve Lhomme committed 1 year ago

> 1.3.0 is a medium release reducing memory footprint, increasing again
> the speed on x86 and ARM and extending the API functions.

https://code.videolan.org/videolan/dav1d/-/tags/1.3.0


(cherry picked from commit cf283cb7)
Signed-off-by: Steve Lhomme <robux4@ycbcr.xyz>

CVE-2023-5217
https://crbug.com/1486441

 (CVE-2023-5217)

(cherry picked from commit ab520205)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

We should not use whatever is in the PATH. Especially as the current directory
is set the a temporary directory during uninstallation.

(cherry picked from commit d13608f8)
Signed-off-by: Steve Lhomme <robux4@ycbcr.xyz>

Thomas Guillem authored 1 year ago and Steve Lhomme committed 1 year ago

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

Thomas Guillem authored 1 year ago and Steve Lhomme committed 1 year ago

This will allow to fix a leak by only checking the return code instead
of fetching the media (and forgetting to release it...)

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

Thomas Guillem authored 1 year ago and Steve Lhomme committed 1 year ago

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

François Cartegnie authored 2 years ago and Jean-Baptiste Kempf committed 1 year ago

(cherry picked from commit 2ce189cc)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

We should not subtract width and height values.

Fixes #27976

(cherry picked from commit 4cd819e2) (edited)

edited:
- SetupQuadFlat is in d3d11_quad on 3.0

François Cartegnie authored 2 years ago and Steve Lhomme committed 1 year ago

refs #27771

(cherry picked from commit 11d7708b) (edited)
edited:
- fmt_in is not a pointer in 3.0

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

François Cartegnie authored 2 years ago and Jean-Baptiste Kempf committed 1 year ago

refs #27771

(cherry picked from commit 00e4996d)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Johannes Kauffmann authored 2 years ago and Jean-Baptiste Kempf committed 1 year ago

(cherry picked from commit 2dcd7516)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Steve Lhomme authored 2 years ago and Jean-Baptiste Kempf committed 1 year ago

And fix fribidi dependencies

(cherry picked from commit 7ebe6644)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Steve Lhomme authored 2 years ago and Jean-Baptiste Kempf committed 1 year ago

This shouldn't be enforced. We already use -O2 on release builds.

Introduced in 77f2dac1

(cherry picked from commit 4dce5d8e)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Alexandre Janniaux authored 1 year ago and Jean-Baptiste Kempf committed 1 year ago

VLC 4.0 doesn't need that but 3.0 still has no windowing module, leading
to display size not being reported to the subtitle text renderer, and
leading to blurry subtitles and OSD.

Because the size is coming from the display, it means that the first
time the subtitle is displayed, it will always be blurry depending on
the real display size and original media size.

Regression from 8ff56952.

Fixes #27793