returning VLC_EGENERIC is no problem, since the vout system is an observer of the change in size and will report the real situation.
returning VLC_SUCCESS is wrong in an asynchronous environmenent, as resizing is only performed later on (or not) depending on GUI configuration.

On the other hand, it allows intf-show to really perform its goal
(react on user interaction via hotkey or mouse)

Revert "Probably fix LoadMessages() on windows. (Untested ... I'm not even sure that the config is initialised when we're called. But that shouldn't matter, we'll just fallback to the default data directory)"

This reverts commit 4c5f93c1.

Probably fix LoadMessages() on windows. (Untested ... I'm not even sure that the config is initialised when we're called. But that shouldn't matter, we'll just fallback to the default data directory)

TODO: Freebox

Maybe we should check the directories do exist first?

Add new common.volume(). Understands use of strings like "+10", "-10" or "10" for relative or absolute volume changes. Fixes volume in http interface.

New video filter : augmented reality. Add a moving ball to the video and handle bounces on the edges thanks to a Sobel gradient computation.
Signed-off-by: Antoine Cellerier <dionoea@videolan.org>

Mention the lua intf move in NEWS. Print an error if lua wasn't found and wasn't explicitly disabled in configure.

Deprecate the C rc, http and telnet interfaces. They are still available as oldrc, oldhttp and oldtelnet.

This actually checks that udev works.

It will allow to resize the video on multiple monitors.

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Avoid nested /* */ comments

Make some room for potential headers, You can now output VLC_CODEC_BMP
We use the same arbitrary value of 200 bytes than ffmpeg.c

Note: if we can ask FFmpeg to output RGBA (32bpp) or RGB48 (48bpp) then
the multiplier should be increased to 4 or 6 (needs to be investigated)

Pointed-out-by: ivoire