Bug#701710: libav: FTBFS on powerpcspe

[Reinhard Tartler]

On Tue, Feb 26, 2013 at 5:56 PM, Roland Stigge <stigge at antcom.de> wrote:
> Hi,
>
> On 02/26/2013 05:42 PM, Reinhard Tartler wrote:
>> Thank you for the patch. I really think that upstream would be very
>> happy to include changes to configure that would make those switches
>> unnecessary. Would you be willing to send a patch against configure
>> upstream?
>
> Yes, that would be possible. Detection would be similar to the patch for
> the sfst package (see #701510).
>
> I decided to use the switches provided by libav upstream. Autodetection
> of the various powerpc (and other CPU's) variants becomes highly
> sophisticated and when they are added, it's easy to break things in
> unusual environments. In this case, it's better to provide switches to
> the (configure) user that leave her with some obscure failure.

Please have a look here:
http://git.libav.org/?p=libav.git;a=blob;f=configure;h=1aae93b3d243ca192cc12eb46d8607ede8407ebe;hb=HEAD#l2726

TBH, I do not feel comfortable with maintaining this patch outside of
libav, escp. since upstream already does take care about unusual
environments. The libav packaging is already complicated enough.
Therefore, I would like to ask you to reconsider your decision to
implement this with configure switches instead of implementing the
logic in the configuration script.


-- 
regards,
    Reinhard