A socket is leaked when the client passes a wrong session identifier.
A remote attacker could craft malicious RTSP requests and...
(cherry picked from commit eea1cddcb651c874e66999a91962da3b3f85bed5)
Signed-off-by: Pierre Ynard <linkfanel@yahoo.fr>

(cherry picked from commit ffd3fce0a407c61e8e5aafdad666424fe5233609)

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

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

This reverts commit 6cd86bc2.

(cherry picked from commit 444c6a067575f5ec5a002a2204c31e734f50aa04)
Signed-off-by: Ilkka Ollakka <ileoo@videolan.org>
(manually removed string-change)

hmm hmm, we would need a generic en.po too.

(cherry picked from commit 24188e89e466d724935a05fe67b393b7fb3958ce)

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

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
(cherry picked from commit 0e1a2635)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

(cherry picked from commit c4fc3095)

Conflicts:

	modules/access/dvdread.c

(cherry picked from commit d626e678)

Conflicts:

	modules/access/dvdnav.c

The original commit description is rather terse
  "ps demux: fix an issue in ps_pkt_read()"
so I don't really know what the "issue" is. But ps_pkt_size() really
made no sense as is.

This reverts commit d761cf6f.

Conflicts:

	modules/demux/ps.h
(cherry picked from commit 9dc74649)

Paths for error message must be UTF-8; they were locale.
Paths for probing must be UTF-8 also, as vlc_fopen() does the
conversion to locale internally.
(cherry picked from commit f2513588)

Conflicts:

	modules/access/dvdnav.c

Paths given to msg_*() function must be UTF-8, not local.
(cherry picked from commit 140a43f3)

Conflicts:

	modules/access/dvdread.c

(cherry picked from commit bf7f847f)
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>

Pointed-Out-By: Kano

this should fix https://bugs.launchpad.net/ubuntu/+source/vlc/+bug/521850
Tested successfully with compiz and metacity

Todo: adapt configure.ac to ensure XPM is available for skins2 (as done with mozilla)
(cherry picked from commit 7e7ead5d)

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

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

(cherry picked from commit 27e2bbff)
Signed-off-by: Ilkka Ollakka <ileoo@videolan.org>

They may be inconsistent on broken multi-processor hardware. But the
system tick count is causing worse problems with power saving features.
(cherry picked from commit ef61df30)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Rémi Duraffort <ivoire@videolan.org>
(cherry picked from commit 37a4eca2)
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>

Too many quotes.
(cherry picked from commit 5925a384)
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>

(cherry picked from commit 28ad9041)
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>

Fixes issue for me when doing transcode{}:gather: --sout-keep and subtitle
overlaying from dvbs. Otherwise keeps last subtitles on video if new channel
doesn't contain dvbs track.

also consistenty as spu_Create is called on transcode_spu_new()
(cherry picked from commit 41b8c2e1)
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>

We must first close the stream before deleting the tracks, because tracks can be used in StreamClose callback.
This crash could only happen in recents version of live555 library.

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

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

Close #4177
(cherry picked from commit eef97c48)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

(cherry picked from commit 911b23af510484805d35525eb99aec6afc3bb957)
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>

Fix out of range memory access when APIC frame type is out of the range defined in the ID3v2 specification (corrupted tag or vendor defined).

Fix #4163.
Signed-off-by: Fargier Sylvain <fargier.sylvain@free.fr>
Signed-off-by: Rémi Duraffort <ivoire@videolan.org>
(cherry picked from commit 639c195bf4a6e89e32042663857165b7eb05b33b)