[petsc-dev] Building petc4py inside PETSc source archive

[Satish Balay]

On Tue, 8 Dec 2020, Satish Balay via petsc-dev wrote:

> On Tue, 8 Dec 2020, Lisandro Dalcin wrote:
> 
> > On Tue, 8 Dec 2020 at 18:48, Satish Balay via petsc-dev <
> > petsc-dev at mcs.anl.gov> wrote:
> > 
> > > On Tue, 8 Dec 2020, Antonio T. sagitter wrote:
> > >
> > > >
> > > > Is the project https://gitlab.com/petsc/petsc4py/ yours or it's another
> > > > different one?
> > > >
> > >
> > > We'd like to get this going - but due to "git subtree" errors - we are
> > > unable to make progress here [i.e keep this in sync with changes in petsc
> > > repo]
> > >
> > >
> > Satish, I think at this point we should kill the petsc/petsc4py repo. This
> > will keep confusing people.

I guess we can kill the repo - if we can generate the petsc4py tarball in the same process that creates petsc tarball.

[then the utility of the separate repo is minimal]

Satish

> 
> I'm retrying this now with 2.17.1
> 
> It doesn't crash - but it adds in the complete (old) petsc history to the petsc4py repo.
> 
> Perhaps we should just ignore this issue - and use this mode?
> 
> [and when subtree gets fixed - we can recreate a more pristine petsc4py repo]
> 
> current:
> 
> balay at sb /home/balay/git-repo/petsc4py (master=)
> $ du -sh .git
> 2.5M	.git
> 
> after "git subtree push" from petsc repo:
> 
> balay at sb /home/balay/test/petsc4py (master>)
> $ du -sh .git
> 219M	.git
> 
> Satish
>