[petsc-dev] PCGAMG: set default type (AGG); PCSetUp_GAMG crashes if not set

[Matthew Knepley]

On Wed, Oct 30, 2013 at 9:49 AM, Jed Brown <jedbrown at mcs.anl.gov> wrote:

> Peter Brune <prbrune at gmail.com> writes:
>
> > It might be best to just pair things.  Prefix matching sounds like a
> recipe
> > for ripping one's hair out over minutiae.
>
> Let's ask petsc-dev.
>
> All, how should we deal with people forgetting to clear dynamic
> functions in XXDestroy_YY?
>

So the rationale for this is when we change backend types?

Is it worth anything more than a convention? If so, what about just having
a count in PetscObject that gets checked in the teardown process.

   Matt


> >
> > - Peter
> >
> >
> > On Wed, Oct 30, 2013 at 9:20 AM, Jed Brown <jedbrown at mcs.anl.gov> wrote:
> >
> >> Alternatively, we could have a PetscObject function that strips all
> >> functions that match a prefix like PCGAMGAgg* or PCGAMG*.
> >>
> >> Peter Brune <prbrune at gmail.com> writes:
> >>
> >> > As I think I said on bitbucket (they're having a massive outage,
> >> > apparently), I'm going to create a branch where I make sure that each
> >> > composition is paired, because while it looks like this rule was
> followed
> >> > in TS, it certainly isn't followed anywhere else, and therefore didn't
> >> > appear to be a convention.
> >> >
> >> > - Peter
> >> >
> >> >
> >> > On Wed, Oct 30, 2013 at 8:17 AM, Jed Brown <jedbrown at mcs.anl.gov>
> wrote:
> >> >
> >> >> Mark Adams <mfadams at lbl.gov> writes:
> >> >>
> >> >> > I can't read bitbucket anymore.
> >> >> >
> >> >> > How can I undo:
> >> >> >
> >> >> >   ierr =
> >> >> >
> >> >>
> >>
> PetscObjectComposeFunction((PetscObject)pc,"PCSetCoordinates_C",PCSetCoordinates_AGG);CHKERRQ(ierr);
> >> >>
> >> >> ierr =
> >> >>
> >>
> PetscObjectComposeFunction((PetscObject)pc,"PCSetCoordinates_C",NULL);CHKERRQ(ierr);
> >> >>
> >> >> just like PCDestroy_FieldSplit, KSPDestroy_GMRES, and hopefully all
> >> >> subtype destructors.
> >> >>
> >> >> > and I'm thinking we can sort of deprecated the PCSetCoordinates
> stuff.
> >> >> >  Lets get ex54 working with PCSetCoordinates and remove it from all
> >> other
> >> >> > tests.  Doing it the right way, with CreateRigidBodyModes solves
> >> Peter's
> >> >> > desire of getting it out of GAMG.  If a user of PCSetCoordinates
> >> breaks,
> >> >> > when the change something, then we just tell them to update to the
> new
> >> >> > regime.  Old users, like Sanjay G, do not need to be bothered.
> >> >>
> >> >> Sounds fine to me.
> >> >>
> >>
>



-- 
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/20131030/b1f49f14/attachment.html>