Handle better cases when we don't have input frame size
same as encoder input frame size. Also gather input samples
untill we get full frame fillet and not just next block.

Set length based on used samples and not frame_size

Also try to make memory calculations cleaner so they don't
overshoot and write stuff all over.

Refs #8265

In the Web documentation, it reads "text automatically starts scrolling" in automatic mode.

This vout uses ES2, not ES1. The right headers are included by
opengl.h anyway, this was just confusing and wrong.
Signed-off-by: Martin Storsjö <martin@martin.st>

This avoids doing the intermediate copy in some cases.
Signed-off-by: Martin Storsjö <martin@martin.st>

This is the actual width used in the upload calls if the check
matched, so this is the width we should check for.
Signed-off-by: Martin Storsjö <martin@martin.st>

This fixes a regression brought by ff74e967

Forcefully resizing only makes sense when layouts were designed to be similar
in size in the first place. Furthermore, it cuts off a larger layout if a
smaller one is first defined when loading a skin.

Ideally, the skins2 dtd should be extended for skins developers to explicitly
express whether two layouts are tightly-coupled or not, i.e should always be
kept the same size whatever the subsequent resizing performed by the user.

In skins2, a window can have multiple layouts with two different goals.

First goal, layouts have the same original size and only differ in the
presentation (different color, ...). If the active layout gets resized,
the user expects the related inactive layouts with the same size to
automatically resize, should they become active.

Second goal, layouts are different in size, e.g a reduced layout and an
expanded layout. In this case, resizing the active layout doesn't mean
resizing the other inactive layouts, since they were not meant to be
similar in size in the first place.

This patch creates a function that will be used to differentiate these
two different use of layouts.

Don't blacklist yuv converters, since those are needed for DVB subtitle display (thanks a bunch, Martin)

Don't blacklist the adjust filter, since it is supported by MobileVLCKit now

Don't build libmad, since libav can decode mp3 just fine

add mooar debug for ARM builds

Use 'input_item_GetTitleFbName', since this is actually the function used
to display title in the qt4 playlist.

Instead of hardcoding what we think the name of the current item is, make it
configurable via the --input-title-format parameter.

This change makes skins2 more consistent with qt4 that already uses this
means to set the title of the main window.

Default display will now be "artist + title" instead of just "title".
Users who would still prefer the previous behaviour can get it with
--input-title-format '$t'

This fixes a request described at http://forum.videolan.org/viewtopic.php?f=15&t=108976

Error messages were reused from time when there was no uncommitted
switch support. Now we have support for this feature, and we can
supply better error messages, because we know in more details what
has happened at times when error occured.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

These sources have some data appended after the CC units.
Example: 0x91 0xe0 0x65 0x6e 0x67 0x7e 0x3f 0xff
                     e    n    g
Probably a language code.

By default VLC uses "vlc" as syslog ident but in case more than one VLC
is logging this may not be enough.

This commit introduces --syslog-ident parameter which allows setting of
the ident.

Based on patch by Georgi Chorbadzhiyski <gf@unixsol.org>.

Change "cycle through program SIDs" to support cycle both ways (prev/next).
With this commit, it is possible e.g. to quickly change between neighbor
program SIDs back and forth in multi-program stream like DVB streams using
default x/Shift+x keys.
Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

Relicencing by author as it is still not altered.

This is mostly useful if the hw accelerator cannot handle the stream.
The software decoder fallback will then use frame multithreading if it
supports it.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Set to the closest to 5% steps

Close #5671
Patch by "rycius" on trac

Default size was wrong.
Text was really hard to real in that tiny box: lots of line breaks and
scrollbars. Text should fit in a single viewport.
Max size has been increased to 500, screen assumed at least 640px wide.

Also do not check that playlist is non-NULL (always true).