[petsc-dev] petsc rewind of next branch?

[Matthew Knepley]

On Fri, Jun 16, 2017 at 9:54 AM, Jed Brown <jed at jedbrown.org> wrote:

> smithc11 <smithc11 at rpi.edu> writes:
>
> > I suppose this email should/could go to a petsc mailing list.  If you
> > agree, please let me know which one and I'll gladly send it there.
>
> Cc'd petsc-next.
>
> > I was just reading through the petsc developer's guide
> > https://bitbucket.org/petsc/petsc/wiki/developer-instructions-git
> > and an older presentation of yours
> > https://jedbrown.org/files/20141113-Software.pdf
> > on git workflows involving a master, next, and topic branches.
> >
> > In petsc, is the next branch rewound (reset to point at the head of
> > master) periodically (i.e., after a release)?
>
> After a feature release.
>
> > If so, what git commands are used for the rewind?  In the gitworkflows
> > docs: https://git-scm.com/docs/gitworkflows a 'reset --hard' is used
> > followed by an 'announcement'. I'm guessing the announcement tells
> > folks who cloned the repo that they should expect a history change in
> > next?
>
> Yup.  Satish has been doing this lately, but I expect the procedure
> looks like
>

Satish, is this already scripted? If not, maybe we should work on that for
this
release. I don't know if anyone else could currently release, so we should
rachet
up the automation. I am fine starting it.

  Matt


>   git checkout master
>   ... set release version strings
>   git commit -am'PETSc 3.8 feature release'
>   git tag -s
>   git checkout maint
>   git merge --ff-only master
>   git checkout master
>   ... set PETSC_VERSION_RELEASE=0 (a flag meaning that this version is not
> a release)
>   git checkout next
>   git reset --hard master
>   git push origin maint master +next
>
> Then send an email announcing the release and that 'next' has been
> rewound.  The easiest thing for users is to run
>
>   git branch -D next
>
> in which case any subsequent
>
>   git checkout next
>
> will create it anew.
>



-- 
What most experimenters take for granted before they begin their
experiments is infinitely more interesting than any results to which their
experiments lead.
-- Norbert Wiener

http://www.caam.rice.edu/~mk51/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20170618/951ea397/attachment.html>