[petsc-dev] automatically updated git sample repositories

[Matthew Knepley]

On Thu, May 22, 2014 at 9:27 PM, Jed Brown <jed at jedbrown.org> wrote:

> Barry Smith <bsmith at mcs.anl.gov> writes:
>
> >     We want our users to us a revision control system, we’d like them to
> have it at some publicly available place. We should make the process for
> them to do this as trivial as possible. I suggest the following
> >
> > 1) we select 2 linear examples, 2 nonlinear examples and 2 TS examples
> (structured grid and unstructured grid) from our suite as prototypical
> examples for users to start with
> > 2) have a git repository for each (named 3.5-sample-linear-structured
> for example) which contains the source file(s) and the simplest makefile
> > 3) we make the actual source in these examples to be identical to
> examples in petsc-maint (and another set with petsc-master) through some of
> Jed’s git magic.
>
> We can do this with git subtree split.  If we only ever edit these via
> the PETSc repository, it can be completely automated (because there will
> never be a merge conflict).
>
> > We then tell people the first step for them is
> >
> > 1) make a bitbucket account
> > 2) clone the appropriate example repository from petsc into their
> bitbucket account
> > 3) clone from their bitbucket repository to their machine and start work
> >
> > if they do this and have code questions we can trivially pull their
> repository and futze with it ourselves rather than badly emailed source
> code etc.
>
> Good, but they have to set public permissions for us to do that.


Can't they just add the petsc account to a private repo?


> > What suggestions do you all have to improve on this idea?  Unless
> someone proves this is a bad idea let’s do this before the next release.
>
> Those should probably be placed in $PETSC_DIR/examples/.  I think we
> only want one example of each (otherwise it's confusing).  Don't know
> what to do about structured vs. unstructured.
>

I do think one place we lag Deal.II is having fully documented examples.

   Matt

-- 
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
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20140522/de84caa1/attachment.html>