SuperLU breaks the build

[andlaus]

currently, I get the following if I try to compile opm-porsol

lauser@unni:~/src/opm-porsol|master > make
Makefile:467: *** target pattern contains no `%'. Stop.

investigating the Makefile, one finds the following at lin 467:

SUPERLU_CPPFLAGS = -I./configure: eval: line 41318: syntax error near unexpected token (' ./configure: eval: line 41318:cd yes (version 4.3 or newer)'/include/superlu -DENABLE_SUPERLU

I guess this means that something went wrong in the SuperLU M4 test.

BTW: how can I attach files on the github bugtracker? I would like to provide config.log...

[atgeirr]

You can use the gist feature (button on top of page) to share files/logs etc.

About the error in question, I unfortunately do not have any answer.

[rolk]

@andlaus Just a quick question (as something similar happened to me earlier): Have you tried autoreconf -i before make?

[andlaus]

@rolk: I used dunecontrol. I suppose that this is done automatically...

I uploaded the failed config.log:

git://gist.github.com/3852526.git

(I think that gists are quite cumbersome, though.)

[blattms]

On Mon, Oct 08, 2012 at 06:27:17AM -0700, Andreas Lauser wrote:

I uploaded the failed config.log:

git://gist.github.com/3852526.git

Seems like the one from opm-porsol, doesn't it?

[andlaus]

yes, opm-core suceeded, porsol failed. I opened the issue for opm-core, because porsol does not feature a SuperLU test, right?

[blattms]

On Mon, Oct 08, 2012 at 06:49:27AM -0700, Andreas Lauser wrote:

yes, opm-core suceeded, porsol failed. I opened the issue for opm-core, because porsol does not feature a SuperLU test, right?

Would you test whether the latest commit to my superlu-support branch of
https://github.com/blattms/opm-core.git fixes your issue?

[andlaus]

yes, that seems to work. thanks.

One nitpick: I noted that SuperLU does not appear in the DUNE summary anymore:

Found the following Dune-components: 

----------------------------------------

dune-common.............: version 2.2.0
dune-cornerpoint........: version 1.0
dune-geometry...........: version 2.2.0
dune-grid...............: version 2.2.0
dune-istl...............: version 2.2.0
opm-core................: version 0.1
ALBERTA.................: no
ALUGrid.................: version 1.51 (parallel)
AlgLib for DUNE.........: no
AmiraMesh...............: no
BLAS....................: 
BLAS....................: yes
GMP.....................: yes
Grape...................: no
LAPACK..................: yes
METIS...................: no
MPI.....................: yes (OpenMPI)
OpenGL..................: yes (add GL_LIBS to LDADD manually, etc.)
ParMETIS................: no
SuperLU-DIST............: no
UG......................: yes (parallel)
psurface................: no

is this a known side effect?

[blattms]

On Mon, Oct 08, 2012 at 07:06:24AM -0700, Andreas Lauser wrote:

yes, that seems to work. thanks.

One nitpick: I noted that SuperLU does not appear in the DUNE summary anymore:

Just noticed that, too. I will change the macro name in opm-core such

that the real dune-modules still are using the official DUNE one.

Dr. Markus Blatt - HPC-Simulation-Software & Services http://www.dr-blatt.de
Hans-Bunte-Str. 8-10, 69123 Heidelberg, Germany, USt-Id: DE279960836

[alfbr]

As far as I can see, this issue is resolved. Please reopen if I am mistaken.