[petsc-dev] Scotch 6.0.8
Smith, Barry F.
bsmith at mcs.anl.gov
Wed Aug 28 11:11:13 CDT 2019
> On Aug 28, 2019, at 3:30 AM, Lisandro Dalcin <dalcinl at gmail.com> wrote:
>
>
>
> On Wed, 28 Aug 2019 at 06:26, Smith, Barry F. <bsmith at mcs.anl.gov> wrote:
>
>
> > On Aug 27, 2019, at 3:22 AM, Lisandro Dalcin via petsc-dev <petsc-dev at mcs.anl.gov> wrote:
> >
> > There is a new scotch release. We should update configure to this version. I'm a bit confused about the current status PTscotch.py in BuildSystem, it seems to be using the scotch/master from the official repo.
>
> It is using
>
> self.version = '6.0.7'
> self.versionname = 'SCOTCH_VERSION.SCOTCH_RELEASE.SCOTCH_PATCHLEVEL'
> self.gitcommit = 'v'+self.version
> self.download = ['git:https://gitlab.inria.fr/scotch/scotch',
> 'https://gitlab.inria.fr/scotch/scotch/-/archive/'+self.gitcommit+'/scotch-'+self.gitcommit+'.tar.gz',
> 'http://ftp.mcs.anl.gov/pub/petsc/externalpackages/scotch-'+self.gitcommit+'.tar.gz']
>
> which means it is using the tag v6.0.7 from the repo which in a sane world corresponds to release 6.0.7 exactly.
>
>
> Well, yes, but I'm still wondering why we default to git rather than an official release tarball.
Why not? It makes it trivial to switch to other branches/packages for testing.
Is there some inherent reason that using "official release tarball" for any package instead of the the appropriate git branch is better? We don't have an official policy on which to use (maybe we should) but my feeling is to use the git repository unless for a particular package there is a reason not too.
Personally I'd like to see the idea of "official release tarballs" get lost in the mists of time and everyone just make tags in their repository for releases. There is the question of built documentation in the tarball but I won't even put docs in the releases, just put them on the web. It is 2019.
Barry
>
> To switch to the new release just change the 7 to an 8; test and make a merge request. We are not official accepting merge requests yet but despite this people are shoving them in like crazy, you might as well join the crowd.
>
>
> No pressure, I'll do it once merge request acceptance is back to normal.
>
> --
> Lisandro Dalcin
> ============
> Research Scientist
> Extreme Computing Research Center (ECRC)
> King Abdullah University of Science and Technology (KAUST)
> http://ecrc.kaust.edu.sa/
More information about the petsc-dev
mailing list