Try TI_1_Einfuehrung.avi to trigger the bug.

Thanks a lot to Ludovic Fauvet who made the hard debugging work and found the idea.

(cherry picked from commit 8796e33a)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

If main() is not found in pthread 10 line above it won't be find now.
And the patch submitter is looking for a better test

This reverts commit bb1d1eb5.
(cherry picked from commit c065c55c)

This reverts commit 31cef7fc.

This is fixed upstream in mplayer now. Our libass is up to date with r29367 of Mplayer.
(cherry picked from commit efa45eaf)
Signed-off-by: Derk-Jan Hartman <hartman@videolan.org>

(cherry picked from commit a6fdca45)
Signed-off-by: Derk-Jan Hartman <hartman@videolan.org>

(cherry picked from commit 6520cc40)
Signed-off-by: Derk-Jan Hartman <hartman@videolan.org>

kO -> ko, kO/s -> kb/s
(cherry picked from commit ca24a7a3)

demux/dirac: don't send a chained set of blocks to packetizer
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
(cherry picked from commit 758096d5)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Yes it's invalid, but there's little point in rejecting it
(cherry picked from commit 4046e4c0)

file:// URI syntax on Windows features a leading slash before the drive
letter (file:///C:/Foobar/...). This is irrelevant on WinCE, that only
has a root directory.
(cherry picked from commit c4402537)

(cherry picked from commit d997857b)

(cherry picked from commit f239f6fb)

reset the current status item.

May JVC go to h**l
(cherry picked from commit bcd95bbd)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Fix #2720 *only* if the ML is empty. This is a real *HACK* !! (so please let this commit only in the branch).

The ticket #2720 does still exist if the ML isn't empty.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

(cherry picked from commit dfb106d9)

(cherry picked from commit ce6c6a96)

Signed-off-by: Christophe Mutricy <xtophe@videolan.org>

This segfault happend because :
 * the variable "audio-device" wasn't created
 * hotkeys module (and others) expect the variable to exist.
(cherry picked from commit 94ca05ff)
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>

Support for the older API has dropped as of now.
Patch by Francois Cartegnie <fcvlcdev at free.fr>
Signed-off-by: Jean-Paul Saman <jpsaman@videolan.org>
(cherry picked from commit f77a9a62)
Signed-off-by: Jean-Paul Saman <jean-paul.saman@m2x.nl>

This reverts commit 503c860a.

contrib: prevent some rare linking errors on OSX by providing the minimal OS version also at link-time

On Windows, it hopefully solves crashes at termination for skins2

Removed references to the old HTTP interface removed by commit [ad16fc37]
Signed-off-by: Derk-Jan Hartman <hartman@videolan.org>
(cherry picked from commit 4ef9c9fd)
Signed-off-by: Derk-Jan Hartman <hartman@videolan.org>

Add -lws2 for access_output_rtmp

(cherry picked from commit 875d8873)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

(cherry picked from commit d65ad9ba)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

What is probably a libtool version mismatch prevents us from
easily rebootstrapping. To avoid that, we patch Makefile.in and
that's it.

Seems that NSIS does not like it:
http://buildbot.videolan.org/builders/NB-I386-Win32Cross-Branch/builds/125/steps/Create%20package/logs/stdio

(cherry picked from commit 7935d5a8)
Signed-off-by: Derk-Jan Hartman <hartman@videolan.org>

I'll look at this again tomorrow.

This reverts commit be40e542.