[petsc-dev] transitioning more PETSc tests to new harness before next release

[Matthew Knepley]

On Sat, Oct 7, 2017 at 9:08 PM, Jed Brown <jed at jedbrown.org> wrote:

> Matthew Knepley <knepley at gmail.com> writes:
>
> > On Sat, Oct 7, 2017 at 6:26 AM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> >
> >>
> >>    We'll be preparing another release in the next couple of months. It
> >> would be nice to transition more examples to the new test harness.
> >>
> >>    How can we organize this? My original plan was to try to do a
> >> test/tutorial directory at a time in a branch and move them through
> next.
> >> Can we start to do this or is there still something missing in the test
> >> harness?
> >
> >
> > I want one more thing in the test harness, the ability to specify the
> > number of procs on the fly. I tried to do this myself, but
> > screwed it up. Can someone do it correctly?
>
> Test output very often changes when you change the number of processes
> so the diff tests would often fail.  What about just printing the exact
> command that would be run?  Then you can change number of processes or
> other options without needing extra steps.
>

This seems masochistic. I want to run a test with a different number of
processes.
So my best option is print out the options, cut & paste that soewhere,
alter it to
what I want, and run? Why would we have EXTRA_OPTIONS?

  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

https://www.cse.buffalo.edu/~knepley/ <http://www.caam.rice.edu/~mk51/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/petsc-dev/attachments/20171007/f7fb8ca0/attachment.html>