[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#681589: givaro: uses system include syntax for local header



Hi release team,

Could you please remove givaro, fflas-ffpack and linbox from testing
if you don't want to make a new transition from libgivaro0 to
libgivaro1? Thanks.


Regards,
Lifeng

On 11:42 Sat 07/14/12 Jul     , Neil Williams wrote:
> OK, in that case, givaro_check.m4 is irretrievably broken and givaro
> needs to provide pkg-config support which linbox can use instead.
> 
> givaro_check tries to link a test program against libgivaro0 simply by
> including the givinteger.h header and this *was* a valid method for
> older versions of givaro.
> 
> > > 1: linbox and other packages which use givaro_check.m4 will fail to
> > > build from source because gmpxx.h cannot be found by the configure
> > > check.
> > 
> > The FTBFS bug of linbox in sid is likely caused by givaro API
> > incompatibility.
> 
> It definitely is - however, even if givaro migrates API, linbox will
> still not get a valid result from givaro_check.m4. Givaro can help this
> situation by providing a givaro.pc file.
>
> > I would file an RFS for linbox-1.3.2 to fix the bug
> > if Vincent Danjean is no longer interested in this package (I got
> > permission from Tim Abbott to adopt linbox and other SAGE-related
> > packages).
> 
> There is no point thinking about uploading libgivaro1 without talking
> to the release team who are unlikely to want a new transition at this
> point. See #678769 - that's why I blocked this bug by #678769.
> 
> At this stage of the release freeze, the release team may be more
> willing to consider removing givaro, linbox and fflas-ffpack.
>
> This problem should have been noticed before the 3.7.0-1 upload. It is
> difficult to fix now and you should talk to the release team about what
> can be done and whether givaro can be released as part of Wheezy.
> Currently, it looks unlikely.

-- 

Attachment: signature.asc
Description: Digital signature


Reply to: