It kinda suck to have to duplicate this though. Qt4 embedded X video
exhibits the same problem. One ugly "fix" could involve the video
output listening for the parent window event - not sure if this would
work. Another option would be to now create a window in the video
output at all, but then we loose reparenting (useless?). In either
cases, handling of hotkeys in fullscreen mode would seem problematic,
unless we handle fullscreen via the window provider too
(control request?).

This reverts commit 0bed6fc3.

It is broken (at least for ps).

Check the content of the file to validate the extension selection.

 Some checks are light but either the file format does not provide enough
identifier data or too much code would be needed. But it is better than
nothing.

Among them, it will now refuse to load file larger than 500Mbytes.

We do poll character devices, as it works for TTYs and some platform
specific devices.

Block special files are always seekable. File size is undefined for
non-regular files.

That particular key is supposed to work that way.

Note that it will try to use the X11 vout right now, so you won't get any visible video output

Windows-specific keys are missing (found in XFree86?).
Mouse wheel is implemented with mouse events (like X11); afterall it is
not on the KEYboard.

Also libvlccore is not backward compatible, whereas libvlc is.

POSIX does not document the == construction for test; Even if GNU test and bash builtin allow it, configure is run by sh and FreeBSD test does not allow it so that this test is always false there.
Signed-off-by: Christophe Mutricy <xtophe@videolan.org>

macosx: Renamed 'Quick Open File' to 'Open File' and 'Open File' to 'Advanced Open File' and switched the menu order accordingly to provide the same names as the Qt4 intf

Improve spectrum visualization. Change options : the number of bands for the spectrum and spectrometer visualization is now set with a boolean. Ban some actions to avoid segfaults in visualizations.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Pointed-By: Eric Lassauge

The original idea and debug come from basos g patch with the following
comment:

"A dead lock was created when adding an spu filter (like marq)
because the same mutex was used for spu->p stuff and sub filter
chain (p_spu->p->p_chain)."

 I implemented the creation/destruction of the chain outside the sub-filter
callback.