[petsc-dev] Introducing new "test harness" to PETSc

[Barry Smith]

> On Jan 18, 2017, at 1:56 PM, Matthew Knepley <knepley at gmail.com> wrote:
> 
> On Wed, Jan 18, 2017 at 1:54 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> 
> > On Jan 18, 2017, at 1:46 PM, Matthew Knepley <knepley at gmail.com> wrote:
> >
> > On Tue, Jan 17, 2017 at 8:25 PM, Matthew Knepley <knepley at gmail.com> wrote:
> > On Tue, Jan 17, 2017 at 8:19 PM, Barry Smith <bsmith at mcs.anl.gov> wrote:
> >
> > > On Jan 17, 2017, at 8:13 PM, Matthew Knepley <knepley at gmail.com> wrote:
> > >
> > > On Tue, Jan 17, 2017 at 7:59 PM, Jed Brown <jed at jedbrown.org> wrote:
> > >
> > >
> > > Use ${DATAFILESPATH} like we do in all the legacy tests.
> > >
> > > Not working
> >
> >   Why do you even listen to Jed, when you can listen to me. In this case the data files is in the PETSc share directory so you should use the ${PETSC_DIR} to point to it. Only files in the DATAFILESPATH directory should use $DATAFILESPATH
> >
> > Okay, that worked. I have one test converted.
> >
> > Gripes:
> >
> > From the answers to these, it sounds like I will end up writing a piece of Python around this that
> > does largely what my existing Python script does. I wanted to avoid this.
> 
>    What exactly would this "piece of Python around this" consist of? The intention is that default test harness has what you need so anything you think you need to put into a ""piece of Python around this" should be part of the test harness.
> 
> Right now, I am running stuff through 'make'. The % matching does not work, but searchin does. Moreover, in order to replace
> output, I need to type DIFF=<crazy script>, so I would just do that by using Python with a short option, dumping that script,
> calling it, and then deleting. Also, I want to be able to toss in arguments, since this makes it much simpler to debug tests. It
> seems I will need a wrapper to the make.

   Despite that fact that Jed thinks we all have direct access to his ~/bin all of these things need to be added to the test harness. 

    We need to get the % matching to work, we need a REPLACE=1 option, and a OPTIONS= option. These just haven't been done yet.

  Barry

> 
>   Matt
>  
> 
>    Barry
> 
> >
> >   Matt
> >
> >   1) How do I run an individual test run?
> >
> > I use
> >
> >   make -f gmakefile test searchin=plex
> >
> > to run plex tests, and I think
> >
> >   make -f gmakefile test searchin=plex*ex1
> >
> > will work. However, how do I select ex1_2?
> >
> >   2) How do I replace the output which exists with the output of the current test run?
> >
> > This is amazingly useful after changes.
> >
> >   3) How do I give extra arguments with the make interface, rather than test_harness?
> >
> > Once I have all the functionality of my Python stuff,  I will throw it away.
> >
> >   Thanks,
> >
> >      Matt
> >
> > >
> > > arch-c-exodus-master/tests/src/dm/impls/plex/examples/tests/runex1_gmsh_3.sh -v
> > > not ok dm_impls_plex_tests-ex1_gmsh_3 /PETSc3/petsc/petsc-dev/arch-c-exodus-master/bin/mpiexec -n 3 ../ex1 -filename /meshes/square.msh -interpolate 1 -dm_view > ex1_gmsh_3.tmp 2> runex1_gmsh_3.err
> > >
> > > when I have
> > >
> > >   test:
> > >     suffix: gmsh_3
> > >     nsize: 3
> > >     args: -filename ${DATAFILESPATH}/meshes/square.msh -interpolate 1 -dm_view
> > >
> > >    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
> >
> >
> >
> >
> > --
> > 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
> >
> >
> >
> > --
> > 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
> 
> 
> 
> 
> -- 
> 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