[packman] [opensuse-factory] Fwd: commit libavutil for openSUSE:Factory

[Cristian Morales Vega]

If xine is going to be the only library that uses libavutil a static
version could make sense. But there is no reason to think that's going
to be the case. And then... really, I fail to see the problem here.
Could you please show a specific case where using a dynamic version of
libavutil there is a problem that isn't there when using a static one?


On 24 June 2012 13:47, Dave Plater <davejplater at gmail.com> wrote:
> I have been working on a static build of libavutil included in the hg
> version of xine-lib but was having difficulty getting that version of
> xine-lib to build. Unfortunately I'm computerless atm but this would
> certainly solve the problem.
> Dave
>
> On 6/15/12, Cristian Morales Vega <reddwarf at opensuse.org> wrote:
>> On 15 June 2012 18:25, Dominique Leuenberger a.k.a DimStar
>> <DimStar at opensuse.org> wrote:
>>> Has this be well planned
>>
>> Since xine 1.2.x depends on libavutil unconditionally
>> (https://bugzilla.novell.com/show_bug.cgi?id=762784) it's basically
>> the only solution.
>>
>> But I don't see any real problem here. Can you put an specific
>> combination of circumstances that would be problematic? It would be
>> safer and easier if libavutil versioned symbols in a more detailed way
>> instead of just to avoid using symbols from a library with a different
>> soname, but that problem happens with lots of libraries that don't
>> even version symbols at all.
>> xine is the only package from openSUSE that will link against it. And
>> the most common case is for users to use both xine and libavutil from
>> Packman. So this seems 100% safe to me.
>>
>> _______________________________________________
>> Packman mailing list
>> Packman at links2linux.de
>> http://lists.links2linux.de/cgi-bin/mailman/listinfo/packman
>>