[petsc-dev] moab nightlybuild failure

Satish Balay balay at mcs.anl.gov
Fri Jun 28 17:29:11 CDT 2013


On Fri, 28 Jun 2013, Barry Smith wrote:

> 
> On Jun 28, 2013, at 3:09 PM, Jed Brown <jedbrown at mcs.anl.gov> wrote:
> 
> > Barry Smith <bsmith at mcs.anl.gov> writes:
> > 
> >>   This could be extended a bit by the gitcommit value in moab.py
> >>   which means don't use moab master by default. But I think this
> >>   could be dangerous to ever use since there is only one value, it is
> >>   not per branch of PETSc-dev, 
> > 
> > It would be perverted to try to keep the branch mapping in the file, but
> > petsc-branch:moab.py could name a commit on a MOAB branch, possibly
> > updated on merge.
> > 
> >>   so better to keep the simple model of petsc master uses moab
> >>   master.
> > 
> > This means that MOAB merging an API change would break PETSc 'master',
> > unless it was synchronized with PETSc merging a matching branch.
> 
>    Exactly. 
> 
> >  I
> > think this is undesirable.
> 
>    If we don't do this then you need to figure out and explain to us simpletons how to 
> 
> > petsc-branch:moab.py could name a commit on a MOAB branch, possibly
> > updated on merge.
> 
>    otherwise it is hopeless.  So each PETSc branch would have to have in it somewhere information about which MOAB "branch/commit" it is suppose to use and switching to (checking out) that PETSc branch means it will automatically switch to that moab branch (oh sorry, "checkout" to that moab branch).  In addition when the PETSc branch is merged to another branch the right information about the moab commit/branch that should be used needs to be put in the right places in that merged branch. Can you do this? Surely git must have some unforgivably clumsy syntax somewhere to allow branches in two repositories to always meet their correct partners that the dance?

The development model you advocate is complicated. So there is no
simple solution to it.  Its the same level of complication as we had
with petsc+buildsystem.

For all other packages - we update in lockstep[primarily
tarballs]. And this works fine. [wrt master, maint branches].

Satish



More information about the petsc-dev mailing list