Re: Mathgl upload issue
Thank you Tobias for the comment. I will chase the culprit with the help of
Sebastien.
Bests
Dimitris
Στις Παρ 25 Ιαν 2013 17:40:01 Tobias Hansen γράψατε:
> Sure, the package will probably build on experimental. But the
> Build-Depends are supposed to clearly state a set of packages that
> suffice for building, and adding a versioned build dependency is a good
> idea unless the versions in all supported Debian distributions satisfy
> the dependency. Just find the culprit and add a versioned build
> dependency. Where is the problem?
>
> Cheers,
> Tobias
>
> On 01/23/2013 01:53 PM, Dimitrios Eftaxiopoulos wrote:
> > Hello,
> >
> > I have prepared version 2.1.1-1 of mathgl for upload to experimental.
> > Since it will go to the NEW queue, I asked Sebastien Villemote to check
> > it and possibly upload it.
> >
> > Sebastien's view is that in unstable, the build dependencies are
> > satisfied but the package fails to compile. This is a bug. The package
> > must compile when build dependencies are satisfied. The build
> > dependencies must be tightened in order to pull in the needed stuff from
> > experimental. Build daemons do not pull in stuff from experimental if
> > the build dependencies can be satisfied in unstable, so the package will
> > FTBFS everywhere.
> >
> > My view is that since the package builds successfully in a clean chroot
> > created and updated with pbuilder with the option "--distribution
> > experimental", it will (most probably) build successfully after an upload
> > to experimental as well. The package appears to be lintian clean apart
> > from compilation flag hardening and missing manpages warnings. All
> > issues related to its behavior upon upload to unstable will be dealt
> > with by me when the time comes for the package to be uploaded to
> > unstable.
> >
> > I ask other members of the list to try the package, share their views on
> > it with me and possibly upload it. Note that only amd64 build is
> > possible currently, due to more symbols file editing needed for other
> > arches.
> >
> > Best regards
> > Dimitris
Reply to: