That horrible hack caused all subtitles to be parsed as Unicode if one
(but not necessarily all) opened subtitles started with a UTF-8 or
UTF-16 Byte Order Mark. If any other subtitle was neither in UTF-8 nor
in UTF-16 with a BOM, that hack failed.

This should not be needed (BOM is a zero-width character). If any
stream reader barks on a BOM, it really needs fixing on its own.
The subtitle demuxer was fixed in the previous commit.

In the case of transparent conversion from UTF-16 to UTF-8, the BOM
will be converted to UTF-8 as any other character. Thus the stream
reader can (correctly) detect a UTF-8 text stream.

If an UTF-8 BOM is found at the beginning of the text file, skip it
(it can confuse some parsers). Also mark the subtitle track explicitly
as UTF-8 encoded.

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

10L

Now, the sidebar will not  will not resized while resizing the window.
Also, this commit sets a min width for the the sidebar. But it can
still be hidden, just drag the slider to the left.

When calling a modal dialog, Qt switches to a new event loop.
The call may still have some event to process in its own queue,
and deleting it when doing the Vout update will have those events
refering to a no longer existing widget.

Since 42a7f46d it is required to re-run gettextize because
we run autoreconf.
Unfortunately gettextize doesn't run in non interactive mode (this
is an advertised "feature").

This gpg-error patch will appear in the next release so it doesn't
need maintenance:
http://git.gnupg.org/cgi-bin/gitweb.cgi?p=libgpg-error.git;a=commit;h=c35dbfdaa2e30e08185e4a468278a29e2c48b500

Tested on HP Touchpad up to 1280x720

1920x1080 gives:
E/OMX-VDEC-1080P: ERROR: IP Requirements(#1: 2097152) Requested(#2: 4177920)
Which probably means that provided buffers are too small for the decoder
to operate

This is required on HTC Sensation. That device uses a special
tiled pixel format, which probably contains a bit of padding, too.
The size that the omx codec suggests for the output port is bigger
than what VLC calculates, and we thus should keep the larger value
instead of overwriting it with what we calculated ourselves.
Signed-off-by: Rafaël Carré <funman@videolan.org>

close #6926

As Qt >= 4.6.0 provides AbstractAnimation we no longer need QTimer
based animators.
PixmapAnimator allows to be used by Q*Painter.
Unused still frame support is removed, but can easily be implemented.

macosx: fixed HiDPI fspanel background artwork, which accidently included a shadow which lead to drawing artifacts since the 2 variants didn't match their sizes

no user-visible change.

This is a broken concept, since according to the HIGL, one point always equals one pixel on non-HiDPI screens regardless of the actual DPIs or the actual screen resolution.

Additionally, this interferes with our support for HiDPI

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

Signed-off-by: Felix Paul Kühne <fkuehne@videolan.org>

changes are not read anyway

Previously, the open panel always converted the mrl to an (not available)
mount point when you want to setup a playback without menus

QPixmapCache can't compute using an undefined maxsize.
Introduced by a2314241