Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

also maintain uniform case "CSeq" throughout.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

The order of aliased entries was dependent on the build machine. This
is not a problem for good input, but affects duplicate detection.

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

Multiple identical FourCC in the same ES category have never worked and
never will work. So far the linear search loop would always pick the
first match.

--enable-update-check enables update support in core, which is not needed
for the mac platform. To avoid linking libgcrypt into the core, instead a
new configure switch is introduced to disable sparkle support directly.

Now, for disabling sparkle support, one needs to pass --disable-sparkle.
It completely removed sparkle from the GUI, and does not copy the
framework to the final bundle as well.

tracks aren't init yet, so we can't compute duration
when falling back from missing mehd

Could have been at least some leaks, and some nasty things
due to smooth box handling.

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

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>