This is obsolete according to IRC conversation. But I also think it got
never really executed, code like "case foo in fo*) a ;; foo*) b ;; esac"
would never lead to b in the shells i've tested.
Signed-off-by: Daniel Mierswa <impulze@impulze.org>
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

See 9eb3e68fSigned-off-by: Daniel Mierswa <impulze@impulze.org>
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

LIBS_dvdcss is vague since it's not it's own module but rather just
accidently set on bsdi systems. Just do the check in the modules that
need the hack now.
Signed-off-by: Daniel Mierswa <impulze@impulze.org>
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

Since we used AC_ARG_WITH(groupname) to indicate a group of options we
definitely want the configure output sorted.
Now we use EXTEND_HELP_STRING and do no longer abuse AC_ARG_WITH.
Signed-off-by: Daniel Mierswa <impulze@impulze.org>
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

It remains impossible to enable/disable the filters themselves without
the video output, though.

 * allow setting page all the time,
 * remove dummy exception,
 * fix an object leak,
 * remove impossible error,
 * fix apparently inverted logic.

Use freopen() or the logger... seriously. This was not even thread-safe
(caboom if two messages come at the same time).

Also fix a potential crash in case of stale subscriber
(print debug message to stderr).

The new repositories are now on git.videolan.org/vlc/bindings/
* java => jvlc.git
* phonon => phonon.git
* python-ctypes => python.git

Remove msleep() from vlm. It should work but testing is recommended. I'll take all the blame in case of issues :D.

Based on an idea from Jean-Philippe André