GitLab

- merge the previous build scripts into one
- allow building outside of vlc source
- add wasm32-emscripten ci

The commit changes the core, and could add complexity & issues

This reverts commit e4b7bce4.

The Webgl context cannot be shared between workers, we need to be able to have :
- shader and program creation and linking
- gl operations
in the same worker.

The module contains a window module that sets
the window type and a gl es2 submodule implementing
webgl function calls.

json_parse is already defined in misc/, the way wasm-emscripten build
works does not support partial linking.

https://trac.videolan.org/vlc/ticket/25156

We need to specify the pthread flag for compilation and linking, otherwise tests will fail.

posix/sort.c won't be added because qsort_r is not supported.

- add vlc_getProxyUrl stub for emscripten
- disable deprecated GL functions for emscripten

The build system assumes OpenGL functions are implemented if the headers are defined.
Which is wrong, so we need to disable the HAVE_GL marker.

we use AM_LINK_IFELSE() because AC_COMPILE will add the "-c" option, and thus the
in wasm-emscripten this function will appear supported even if it is not.

- add thread_id for emscripten

- add weak attribute support for wasm

- handle sout linking when disabled

this commit fixes a mismatch in the prototype of sout_StreamChainNew,
and  adds a stub for sout_instance_ControlsPace to allow linking without
error.

clock_nanosleep() is not supported in emscripten
Implementation from : https://code.woboq.org/userspace/glibc/sysdeps/unix/clock_nanosleep.c.html

Co-Author: Etienne Brateau <etienne.brateau@gmail.com>

- add emscripten target
- set toolchain variables in bootstrap and main.mak
- add meson_system_name for emscripten
- add CMAKE_SYSTEM_NAME

https://blog.chromium.org/2020/08/changes-to-chrome-app-support-timeline.html

sout_instance_ControlsPace is not built nor provided by missing.c when
building with --disable-sout

As pointed out in #25055 the choice of url option delimiter is probably
meant to be that way. std::string.find() returns std::string::npos (aka
-1) in case of no match.
Signed-off-by: Thomas Guillem <thomas@gllm.fr>

For now we don't list other devices

After a mute it seems we need to tell report the new volume otherwise it
assumes it's 0.

We keep the gain so we can compute the proper volume to report on mute.

The requested device is set to the default one in Open and requires the MTA
setup in order for the activation to work. Otherwise it deadlocks.

It must be free'd with CoTaskMemFree and only once.

The aout_dev_t type doesn't need to be shared anymore.

Signed-off-by: Pierre Lamot <pierre@videolabs.io>

Signed-off-by: Pierre Lamot <pierre@videolabs.io>

Signed-off-by: Pierre Lamot <pierre@videolabs.io>

this fixes the animation issue that occurs when Qt 5.15 is used
Signed-off-by: Pierre Lamot <pierre@videolabs.io>

set '_colors' property when forcing colors
Signed-off-by: Pierre Lamot <pierre@videolabs.io>

Signed-off-by: Pierre Lamot <pierre@videolabs.io>

Signed-off-by: Pierre Lamot <pierre@videolabs.io>

Signed-off-by: Pierre Lamot <pierre@videolabs.io>

due to an incident, a preliminary and incomplete version of "enable aligned player controlbar customization" patch was merged. this update ensures that the latest and the most tested version of the patch is available.
Signed-off-by: Pierre Lamot <pierre@videolabs.io>

The has_deint value can never be true at this point.

This code is even dead in 3.0 as DeinterlaceCallback was not called in the init
either.

It's not used by anyone and doesn't match the string of the same core variable.

There doesn't seem to be any display module that transforms these J4xx chromas
into I4xx chromas for their input format. So we don't need 2 filters for this.