The function is a bit ugly but it's difficult to do this using
pure arithmetic.

Fixes #6378
(cherry picked from commit 1e9dc296)
Signed-off-by: Pierre Ynard <linkfanel@yahoo.fr>

This will only be used if requested explicitly,
that is to say '--v4l2-chroma=h264'.

(cherry picked from commit 28db7725)

DBCS countries assign their currency symbol to '\', ASCII 92. This causes
an unexpected behavior when converting a directory separator on OS/2.
In case of GNU libiconv, it does not treat '\', ASCII 92, as a currency
symbol at all. So let's mimic GNU libiconv.
Signed-off-by: Rafaël Carré <funman@videolan.org>
(cherry picked from commit 8acdc1f5)
Signed-off-by: KO Myung-Hun <komh@chollian.net>
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

OS/2 supports UCS-2 only instead of UTF-16.
Signed-off-by: Rafaël Carré <funman@videolan.org>
(cherry picked from commit 2cc70e68)
Signed-off-by: KO Myung-Hun <komh@chollian.net>
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

(cherry picked from commit 5785e63a)

Close #6403

This patch adds isScrollable() to a control, so that we can distinguish
controls that require mouse wheel (playtree, slider) from those that don't.
In the latter case, we can then forward events to vlc core.

this fixes #6457.
(cherry picked from commit f3afffbc)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Without this flag, PulseAudio will tweak the buffer parameters, but
will not account for hardware latencies. As a consequence, the live
caching parameter is not really followed - the latency is longer.

Longer latency becomes a big problem when trying to synch with another
live source (such as a V4L2 video capture device). Then audio is too
late and gets discarded or messed up at output.
(cherry picked from commit d8a96604)

based on patch by Richard Grenville <pyxlcy@gmail.com>
(cherry picked from commit 66da497b)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

(cherry picked from commit cd470819)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Close #6451
(cherry picked from commit ef790e37)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

According POSIX the va_list is defined in stdarg.h
http://pubs.opengroup.org/onlinepubs/009695399/basedefs/stdarg.h.html

On some systems (like uClibc) the detection of vasprintf will fail
unless stdarg.h is included and build will later fail due to a vasprintf
redeclaration.
Signed-off-by: Natanael Copa <ncopa@alpinelinux.org>
Signed-off-by: Rafaël Carré <funman@videolan.org>
(cherry picked from commit 2859c5db)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Fixes undocumented/poorly documented Windows behavior. RemoveFilter function does a silent Release().
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
(cherry picked from commit 3f6ccc0d)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

(cherry picked from commit ea068351)

The slider is not displayed, anyway. But sometimes it flickers through
for a short moment which does not look good.
(cherry picked from commit 2fc5dc5f)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

(cherry picked from commit d612827c)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

Note that the origin values can also be negative when we have our window
on a secondary screen.
(cherry picked from commit 1cee2e3b)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

(cherry picked from commit e6820b61)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

(cherry picked from commit 3646cf43)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

(cherry picked from commit bf916c35)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

Coeffs have been recomputed.

Close #6132
(cherry picked from commit a712e052)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

When increasing speed to 1.50 (command + =), '1.50x' is shown on OSD, but '10x' is shown on menu (Playback -> 'Playback Speed').
Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>
(cherry picked from commit 70a5e26593b671ab5d267b3c527fda4ba83deae1)

This still doesn't work with ordered chapters though.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
(cherry picked from commit 50f3a9b9)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
(cherry picked from commit 0b1b7516)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

(cherry picked from commit 38a4fd74)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>
(cherry picked from commit fbdbd7da)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Close #6423
(cherry picked from commit 36641267)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Close #6424
(cherry picked from commit f7f78e8a)
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

macosx: make sure that the http interface (and other clients) can toggle the fullscreen mode when using VLC's fullscreen implementation (close #6009)
(cherry picked from commit 2a2bb7af)

This fixes an endless loop when piping commands to stdin with echo
for example

In particular this happens when piping commands into stdin

(cherry picked from commit 9dfcc9e548abefb260a10df57f79151790a1bf03)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

(cherry picked from commit 2368d54953fc4e7731d29ae8749c5ffb3ecb03b8)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

The if condition was only true in a very very rare case, and even then the code is
incorrect (e.g. when the detached window is in fullscreen and the user closes
the main window).
(cherry picked from commit 80e730f1e4d67f4a00428220ed170d477c85032a)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

This simplifies the code in several lines.
(cherry picked from commit de742db20a39f116ca2505926fe1a0e9c20de47a)
Signed-off-by: David Fuhrmann <david.fuhrmann@googlemail.com>

glXSwapIntervalEXT() does not return a value, contrary to the
SGI equivalent. In case of error, the X11 error handler is used.

This patch fixes compilation with conforming headers. It should still
be compatible with buggy Mesa headers where 'int' is returned.

For reference, see the specification here:
http://www.opengl.org/registry/specs/EXT/swap_control.txt
http://www.opengl.org/registry/api/glxext.h
(cherry picked from commit 9cf66c2856ab3a209fe15ec5662a8cd927ee20bc)