reduced_basis_ex2 and ex3 don't run?

[libmesh]

reduced_basis_ex2 and ex3 don't run?

[benkirk]

On Dec 21, 2012, at 10:48 AM, jwpeterson [email protected] wrote:

In my build directory, if I type make run in reduced_basis_ex1, everything is happy, but if I do that in ex2 or ex3, I currently get:

LIBMESH_DIR=/Users/petejw/projects/libmesh_git/build/.. METHODS="opt dbg" ../../../../examples/reduced_basis/reduced_basis_ex2/
/bin/sh: ../../../../examples/reduced_basis/reduced_basis_ex2/: is a directory
make: *** [run] Error 126

ex4-7 also seem OK.

It is because of this in Makefile.am:

conditionally enable this test

if LIBMESH_ENABLE_SLEPC
if LIBMESH_ENABLE_GLPK
check_SCRIPTS = run.sh
endif
endif

The test is skipped when doing 'make check' (please verify!) but our compatibility 'make run' method is not being smart enough. Perhaps 'make run' should instead just to a 'make check'? Hmm….

-Ben

[jwpeterson]

On Fri, Dec 21, 2012 at 9:52 AM, Benjamin S. Kirk
[email protected]:

On Dec 21, 2012, at 10:48 AM, jwpeterson [email protected]
wrote:

In my build directory, if I type make run in reduced_basis_ex1,
everything is happy, but if I do that in ex2 or ex3, I currently get:

LIBMESH_DIR=/Users/petejw/projects/libmesh_git/build/.. METHODS="opt
dbg" ../../../../examples/reduced_basis/reduced_basis_ex2/
/bin/sh: ../../../../examples/reduced_basis/reduced_basis_ex2/: is a
directory
make: *** [run] Error 126

ex4-7 also seem OK.

It is because of this in Makefile.am:

conditionally enable this test

if LIBMESH_ENABLE_SLEPC
if LIBMESH_ENABLE_GLPK
check_SCRIPTS = run.sh
endif
endif

The test is skipped when doing 'make check' (please verify!) but our
compatibility 'make run' method is not being smart enough. Perhaps 'make
run' should instead just to a 'make check'? Hmm….

Ah, I see. Yeah I don't have SLEPC so that makes sense... still figuring
out this automake stuff.

John

[pbauman]

Since we're using the parallel-tests bits of automake, I wonder if it would be better to put #ifdef's in the test and if, in this case, SLEPc or GLPK are not available, the test just returns 77 (http://ansuz.sooke.bc.ca/entry/233). The test will be reported as skipped and then we just have the test always run.

[benkirk]

On Dec 21, 2012, at 10:59 AM, jwpeterson [email protected] wrote:

Ah, I see. Yeah I don't have SLEPC so that makes sense... still figuring
out this automake stuff.

No problem. The general approach we're taking with the examples is that 'make check' will always compile the source, even if the example dependencies are not satisfied, mostly to give us more API coverage. This is accomplished by setting the check_PROGRAMS target, and that is always built.

As implemented, TESTS itself is only set when the prerequisites are satisfied…

As @pbauman just mentioned, though, we could always run the test too, and change e.g.

  libmesh_example_assert(false, "--enable-slepc --enable-glpk");

such that libmesh_example_assert() return's with code 77 - that'll make the test print "SKIPPED", which I guess is precisely what we're doing…. ;-P

-Ben

[benkirk]

BTW @jwpeterson - did you have your gravatar commissioned? You're one Eddie-Bauer-looing mthrfkr...

[jwpeterson]

On Fri, Dec 21, 2012 at 10:10 AM, Benjamin S. Kirk <[email protected]

wrote:

BTW @jwpeterson https://github.com/jwpeterson - did you have your
gravatar commissioned? You're one Eddie-Bauer-looing mthrfkr...

Heh, I wish I still looked like that... it's from years ago and before
kids. A current pic would have dark circles under eyes, wrinkles, grey
hair, etc.

John

[benkirk]

On Dec 21, 2012, at 11:12 AM, jwpeterson [email protected] wrote:

Heh, I wish I still looked like that... it's from years ago and before
kids. A current pic would have dark circles under eyes, wrinkles, grey
hair, etc.

Yeah, my developer photo is a little disingenuous at this point:
http://libmesh.sourceforge.net/developers.php

also, as @roystgnr is no longer a student, I guess it's time to update that page!