oshmem: add pkg-config files #4737
Conversation
|
@jladd-mlnx please review, otherwise this won't make 4.0 |
|
@ggouaillardet Do you foresee progressing this PR? |
Refs open-mpi#4730 Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
ggouaillardet
force-pushed
the
topic/oshmem_pkgconfig
branch
from
d64b592 to
4994090
Compare
|
I restored the original indentation and this PR is ready for merge @jladd-mlnx can you please review it ? |
|
Ping. Do we want this PR? |
|
Probably yes, we want this, but I'm not familiar enough to review. |
|
@jladd-mlnx re-ping. Please review or assign someone to review. Thanks. |
|
@amaslenn can you take a look, please. |
| # static linking (they're pulled in by libopen-rte.so's implicit | ||
| # dependencies), so only list these in Libs.private. | ||
| # | ||
| Libs: -L${libdir} @OMPI_PKG_CONFIG_LDFLAGS@ -loshmem -l@OMPI_LIBMPI_NAME@ |
There was a problem hiding this comment.
(minor) Does it make sense to create a OMPI_LIBOSHMEM_NAME to replace hardcoded oshmem?
There was a problem hiding this comment.
If this is the only issue left on this PR, I'd say that it's ready / done / can be merged.
There was a problem hiding this comment.
Nothing apart from this one from my side.
There was a problem hiding this comment.
@ggouaillardet If you still want this one, I'd say you're clear to merge.
|
Re-ping. 5.0 branching is targeted for April 30th. If you want this in for 5.0, please try to get it in master by then. Thanks! |
|
Can one of the admins verify this patch? |
|
@ggouaillardet @open-mpi/ucx - What's the fate of this PR? |
|
This is from 2018. A lot has happened on master since then; you might want to check that this PR is on-par with the OMPI package files that are on master. |
|
bot:retest |
|
@amaslenn @mellanox-github can you verify this is still a good patch and does not need updating? |
@yosefe could you please take a look or assign someone? |
|
I told you this stuff was from 2018 and it was probably wrong. Did anyone actually test this before it was merged? I just looked at the merge commit email that went by, and I see that it is definitely, positively wrong. This needs to be reverted. If you're going to actually merge something into master -- especially from a PR from 3 years ago -- you need to actually review it and test it. Don't just rubber stamp it and merge it in without thinking. 👎 👎 👎 |
Refs #4730
Signed-off-by: Gilles Gouaillardet gilles@rist.or.jp