ohh.. i see. thanks for clarification. This is exactly how it is treated in our internal repo based on 1.7 and we will push it into trunk soon.
On Sat, Dec 14, 2013 at 5:01 PM, Ralph Castain <r...@open-mpi.org> wrote: > I think what Brian is saying is that we don't generally link IB libraries > into the upper-level library - instead, they are only linked to the > specific components that need them. The core code (i.e., non-component) > should not be calling anything in IB itself. So anything that specifies > "with IB" should be in the .m4 for the component that will use IB. > > The oshmem_configure_options.m4 code should be in the top-level config > directory. > > On Dec 14, 2013, at 6:34 AM, Mike Dubman <mi...@dev.mellanox.co.il> wrote: > > Hi, > In our internal oshmem v1.7 based branch it is > in oshmem/config/oshmem_configure_options.m4 and not in > config/oshmem_configure_options.m4 > > Is that a way to go to have it under oshmem/config? > Will check why it is missing. > > Thanks > M > > > On Fri, Dec 13, 2013 at 11:21 PM, Barrett, Brian W <bwba...@sandia.gov>wrote: > >> Mellanox - >> >> In cleaning up some code, I noticed that the OSHMEM_SETUP_CFLAGS test is >> testing for infiniband libraries and then linking them into the OSHMEM >> library. This is fairly different than what we do for the MPI layer; is >> there a reason those tests are in the top-level configure and not in a >> component's configure test? >> >> Thanks, >> >> Brian >> >> -- >> Brian W. Barrett >> Scalable System Software Group >> Sandia National Laboratories >> >> >> >> _______________________________________________ >> devel mailing list >> de...@open-mpi.org >> http://www.open-mpi.org/mailman/listinfo.cgi/devel >> > > _______________________________________________ > devel mailing list > de...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/devel > > > > _______________________________________________ > devel mailing list > de...@open-mpi.org > http://www.open-mpi.org/mailman/listinfo.cgi/devel >
