Change to handle changed x264_encoder_headers and nal-stuff

it's checked to match from r19973 on libavcodec

Isn't fork() supposed to be slow on Windows?

This will be needed to replace aout_buffer_t with block_t.
We also need to replace end_date with i_length, which is a bit more
involved.

In case they are known to be unsupported at compilation time, they are now
defined to be zero to be nice with the compiler.

When allocation is false, aout_BufferFree is simply not called.

The info panel is flushed and the current epg event is undefined when
stopping the associated input.

This reverts commit 6fc6b16d.

I need it for input_item_SetEpgOffline (next commit).

As a side effect, the EPG informations are no more lost when stopping.

Also segregate flags per architecture from now on.

Signed-off-by: Rémi Denis-Courmont <remi@remlab.net>

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

This makes SSE3/SSSE3 in the same switch. Maybe this should be fusionned with SSE/SSE2 ?

This should be useful to record TNT HD streams
Thanks to Thannoy and Laurent

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

It's so slow that it's not usable on large datasets.

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

This is ugly, but can save a few megabytes at download-, install- and
run-time when avcodec is linked in statically.