Remove some useless cast changes. Apply the patch always as it is
correct on all architecture. Drop the errno code part which is wrong
(live needs the Winsock error codes, not the <io.h> ones).

fffmpeg -> ffmpeg

Ref #3682 and #3666
Now, we have
[022d6ffc] main tls client debug: TLS client session initialized
instead of
[004e6ffc] gnutls tls client error: Certificate could not be verified

However, this still doesn't work fine, since we got a:
"access_http acccess error: failed to read answer"
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

This patch contains various updates to the new HTTP UI

Mobile views have only been tested on Android.

Update - Added EQ functionality
Update - Corrected rounding mistakes and EQ units
Update - Added Mobile EQ
Update - Added Mobile Flash View
Update - Minor Display Adjustments
Update - Revert Playlist.xml for backward compatability
Update - Added playlist_jstree.xml for UI usability
Update - Added Batch VLM Command interface
Update - Added VLM command error reporting
Update - Added functionality for future EQ merge into status.xml
Update - Modified Makefile.am to include new files
Update - Corrected main EQ button
Update - Modified author comments
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

THIS IS TOTALLY UNSUPPORTED.
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

And use autoreconf instead of autogen for tremor.

VLC will fallback to lavc

The whole module_config_t is loaded/saved anyway, which is sufficient
for immediate values.

this also prepares a mainwindow object and includes some clean-up

Close #4988

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

SSE function used in PackedYUV plane was bad, removed
copying data using Memcpy8BMMX was slower than vlc_memset -> changed
Sepia8ySSE4_1 changed to Sepia8ySSE2, due to some changes I made to speed it up and clean
Minor changes in indentation and comments
Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Signed-off-by: Jean-Baptiste Kempf <jb@videolan.org>

Use the high order bits for the intrinsic value type (hint, integer, float
or string) and the low order ones for the more specific sub-types.

Values are 8-bits.

Variable types and configuration item types are separate. Fortunately
they use the same bits so it worked fine.

If the high nibble is set, the type is item type and the low nibble is
necessarily zero.

The FFmpeg configure script autodetects those correctly. Of course,
this assumes that the compiler defaults or flags are adequate.