v2.0.x: fortran/base: rename strings.h into fortran_base_strings.h

[ggouaillardet]

rename ompi/mpi/fortran/base/strings.h so it does not get pulled
when /usr/include/strings.h is expected.

Refs #3639

Signed-off-by: Gilles Gouaillardet [email protected]

(cherry picked from commit 08526e8)

[ggouaillardet]

@bwbarrett i checked (only) the free bsd platform.
build was successful, but the overall test failed because --get-stack-traces is a mpirun option that is valid only from v2.x
could you please update your jenkins scripts accordingly ?

[bwbarrett]

yeah, there's a check in the script for version to avoid that bug; not sure what's happening. Unfortunately, I won't be able to look at it in the next week due to meetings and vacation.

[jsquyres]

@bwbarrett Looks like the PRB failed because it's trying to use --get-stack-traces, but that isn't understood here on the v2.0.x branch.

[hppritcha]

bot:lanl:retest

[bwbarrett]

@jsquyres yeah, I got that part. I found the bug; a bad regex for the versions that do / do not support --get-stack-traces. We matched the 1.x series, but not the 2.0.x series properly. Fixed now (I think), although Howard's current build is going to fail same as the first one...

[hppritcha]

are we ready to have jenkins get another go at this one?

[bwbarrett]

Yeah, I pushed what I think is the fix on Thursday.

bot:ompi:retest

[hppritcha]

hmm.. looks like we're still picking up the --get-stack-traces

[jsquyres]

bot:ompi:retest

[jsquyres]

(open-mpi/ompi-scripts#2 should have fixed the --get-stack-traces issue on BSD)

[jsquyres]

Try again, with open-mpi/ompi-scripts#3...

bot:ompi:retest