MacOSX/Framework/Examples/test: Don't waste time getting the length of the video while we're doing a drag & drop operation.

shouldn't break anything (tested on mp4/ts files, rtp stream and http/ts
stream). Files play correctly and MP4Box/ffmpeg recognize them
correctly. should fix #1384

MacOSX/Framework: Remove all references to fullscreen mode, it should be the host application's responsibility to transition between fullscreen and window mode.

input/input.c: Define DEMUX_CAN_SEEK and correctly set "seekable" accordingly. (This could introduce regression such as seekable item no more seekable, but I should have greped enough...).

video_output/opengllayer.m: Hopefully fix the tearing issues by correctly updating textures, and using a nice trick to save the current opengl context. (Note, we can't draw to it directly).

MacOSX/VLC_app/VLCMainWindow.m: Set noAutoresizeColumn and MarkerFollowCell for the categories List.

MacOSX/Framework/Examples/test/Controller.m: Updated example to comply with latest framework improvements.

libvlc_video_take_snapshot API change (it would be nice to grep the
VLC source, or at least the bindings directory, for the symbol when
doing such a change).

MacOSX/Framework/VLCMediaPlayer.m: Missing a comment before the detach to match the attach of libvlc_MediaInstancePositionChanged.

Start splitting the ALSA and OSS code. The --v4l2-adev device will be tried first as an ALSA device, and then as OSS upon failure. If no audio device is specified, "default" is used for ALSA, "/dev/dsp" for OSS. You can disable/activate usage of ALSA and OSS using the --v4l2-audio-method option (see vlc -p v4l2 --advanced --help-verbose).
Next step will be to split the module in different source files, the current code is quite a mess.
Also fix compilation of the module if HAVE_ALSA isn't defined.

video_output/opengllayer.m: Don't use AGP texturing until we get a proper frames lifo stack. Set to the layer synchronous, and call setNeedsDisplay when needed.