On Mon, Oct 13, 2008 at 12:18:30PM -0400, Sean Davis wrote: > On Mon, Oct 13, 2008 at 12:10 PM, Simone Giannerini > <sgianner...@gmail.com> wrote: > > [...] > > > >> Your mail is basically unreadable / impossible to attribute what is > >> original and what is cited. Please follow convention and indent. > > > > I am sorry, I am copy/pasting the messages because I switched off receiving > > list mails. I forgot to ask to reply also to my personal address. > > > > Hope this one is readable > > > > [...] > > > >> Ok, didn't see that. But please do understand that 0.5-5 is on CRAN > >> and released. (And yes, that is the version for which I also had to > >> make a fix to get it build on Debian as mentioned). > > > > Ok I 'll stick to that one > > > > [...] > > > >> Not it does not as lib64 != lib. You probably it installed if you > >> create a softlink from > > > >> /usr/lib64/mpi/gcc/openmpi/lib64 > >> > >> to > > > > > /usr/lib64/mpi/gcc/openmpi/lib > > > > >> Dirk > > > > Thank you, the following works for me: > > > > # ln -s /usr/lib64/mpi/gcc/openmpi/lib64/ /usr/lib64/mpi/gcc/openmpi/lib > > # CC="mpicc" R CMD INSTALL Rmpi_0.5-5.tar.gz > > --configure-args=--with-mpi=/usr/lib64/mpi/gcc/openmpi/ > > You can drop the --configure-args. Using the correct mpicc should > take care of all the linking, etc.
As I said, I doubt this works for R. In general, this is sane advide. In R's case, _a lot_ of flags and library mangling happens. Dirk -- Three out of two people have difficulties with fractions.
