Ilmar Wilbers wrote:
Hi,
No, I will need to install the MTL4 backend manually on all the build
slaves. The build slaves run on different machines, setting a path for
only the fenics server wouldn't be sufficient.
I was suggesting that the MTL4 path could be relative to the dolfin
source, i.e. ./sandbox/mtl_src/boost/numeric/mtl4. But if you could
install as usual, that would be better.
If you send me the instructions, I will be happy to do so!
It simple:
svn co https://svn.osl.iu.edu/tlc/trunk/mtl4/trunk mtl4
I would put everything from mtl4/boost in a system-wide directory, e.g.
next to uBLAS in /usr/local/include/boost
Again, there is nothing to build and install. Just add -DHAS_MTL4 to the
CFLAGS and add any non-standard include path
/Dag
ilmar
Dag Lindbo wrote:
Hello all!
I would like to use the buildbot for self-testing, but the MTL4 backend
is not activated. One needs to add
-DHAS_MTL4 -I<path to mtl4 svn clone>
at a suitable stage. Can I do this with some custom buildbot flags or
does the self-test script need to be updated? I could put the MTL4
source in the _my_ sandbox and give that as the include path (if custom
buildbot flags exist). Otherwise one would need to "install" MTL4 on the
build slaves (which is not hard since MTL4 is header-only).
Thanks!
/Dag
------------------------------------------------------------------------
_______________________________________________
DOLFIN-dev mailing list
DOLFIN-dev@xxxxxxxxxx
http://www.fenics.org/mailman/listinfo/dolfin-dev
