[FFmpeg-devel] [PATCH] restoring binary compatibility with ffmpeg 0.5

[Måns Rullgård]

Reimar D?ffinger <Reimar.Doeffinger at gmx.de> writes:

> On Mon, Jun 07, 2010 at 12:26:22AM +0200, Michael Niedermayer wrote:
>> On Sun, Jun 06, 2010 at 09:01:10PM +0100, M?ns Rullg?rd wrote:
>> > Reinhard Tartler <siretart at tauware.de> writes:
>> > 
>> > > On So, Jun 06, 2010 at 21:51:27 (CEST), M?ns Rullg?rd wrote:
>> > >
>> > >>> In any case, find below the 'best' fix, that admittedly only works on
>> > >>> gnu platforms. Michael, please comment if you prefer the half fix that
>> > >>> fixes the issue on gcc/gas platforms (and doesn't regress on others) or
>> > >>> bumping major of libavformat.
>> > >>
>> > >> We _already_ have a regression, which we unfortunately didn't discover
>> > >> until now.  Leaving it broken on some platforms while fixing others is
>> > >> simply not acceptable.
>> > >
>> > > Actually, a third way to solve the issue would be to move the affected
>> > > symbols back to libavcodec. How about that?
>> > 
>> > s/libavcodec/libavformat/ methinks.
>> > 
>> > Not acceptable.  The reasons for moving them to libavcodec still
>> > stand.  Some libavcodec interfaces use AVPacket, so the related code
>> > must be there too, or we'd end up with a (loose) dependency of
>> > libavcodec on libavformat, something we can't have.
>> 
>> we can duplicate the symbols and code in lavf under #if
>
> Hm, and this can't cause collissions? At least it would have to be
> linked only into the .so, not the .a...

The last part is easy to take care of.  I'm more worried about the
first bit.

-- 
M?ns Rullg?rd
mans at mansr.com