[petsc-dev] moab nightlybuild failure

[Satish Balay]

On Fri, 28 Jun 2013, Jed Brown wrote:

> Satish Balay <balay at mcs.anl.gov> writes:
> 
> > So we update MOAB.py every week?
> 
> I would update it whenever something new comes out and it has been
> tested to work.  I think it's fine to have gitcommit point at the last
> release up until petsc-dev depends on a more recent feature.
> 
> > Another issue - if we specify both tarballs and gitcommit - both have
> > to point to the equivalent snapshot.
> 
> Yes, snapshots are harder to track that way.
> 
> > [and roll in self.giturls+gitcommit into the the standard url for self.download]
> 
> This is a separate issue, but I have no problem with it.

My point here is - whatever complex scheme we come up with tracking
remote projects tarballs, git, mercurial,svn - all urls should point
to equivalent snapshots - and provide eqivalent functionality

[wrt updates, user over-ride-url or ther features]

Satish