[Debian-med-packaging] Bug#902820: Bug#902820: gmap: FTBFS on stretch/amd64 if CPU has SSE2 and nothing more

Santiago Vila sanvila at unex.es
Sun Sep 9 19:10:14 BST 2018


On Wed, 18 Jul 2018, Alex Mestiashvili wrote:

> > AFAIK on Debian/amd64 only SSE2 is allowed to be assumed by default.
> > 
> > I tried to compare my build logs with official ones, but there is none:
> > 
> > https://buildd.debian.org/status/package.php?p=gmap
> > 
> > (( Could you please consider source-only uploads? i.e. "dpkg-buildpackage -S" )).
> 
> I just uploaded a new version 2018-07-04-1, but forgot to build source
> only upload. I'll do it for the next upload.
> Meanwhile could you please try to build the new version?

Sorry for the late reply.

I lost access to the low-end machines where this used to fail for me,
so I can't test it any longer. I trust that you did the right thing.

Please note that the autopkgtests are failing:

https://ci.debian.net/data/autopkgtest/testing/amd64/g/gmap/962092/log.gz

Note: gmap.sse42 does not exist.  For faster speed, may want to compile package on an SSE4.2 machine
Note: gmap.sse41 does not exist.  For faster speed, may want to compile package on an SSE4.1 machine
Note: gmap.ssse3 does not exist.  For faster speed, may want to compile package on an SSSE3 machine
Note: gmap.sse2 does not exist.  For faster speed, may want to compile package on an SSE2 machine


Also, since this is a baseline violation, I would try to upload for
stable as well, for the next point release, but this is up to you.

Thanks.



More information about the Debian-med-packaging mailing list