forked from spack/spack
-
Notifications
You must be signed in to change notification settings - Fork 0
Telcon: 2015 07 30
Todd Gamblin edited this page Jul 30, 2015
·
4 revisions
- Todd Gamblin (LLNL)
- Greg Lee (LLNL)
- Jim Galarowicz (Krell)
- Matthew LeGendre (LLNL)
- Spack SC15 paper is now past shepherding; making final set of small revisions
- Spack talk is 11:30-12, Wed, Nov 18 according to the SC15 technical program
- This is in the State of the Practice session.
- Ongoing talks with Kitware, NERSC on potential collaboration -- still nothing set in stone.
- Bugfixes from Tom Scogland
- Some patches for master from Mike Nolta
- accept these or move them to develop for next release?
- Mark Miller found bug in
--depth
support in theGitFetchStrategy
. Fixed by Tom Scogland. - Jim still having long
RPATH
issues.- Todd will try this with Jim's pull request code to reproduce.
- Mark Miller asks whether we can query name mangling in Fortran compilers
- Actually an interesting ABI issue because compilers can support multiple underscoring schemes.
- Spack should probably figure out and ensure consistency.
- Currently can build an ABI-incompatible version of the same library and Spack won't know b/c it doesn't understand mangling.
- Mark Miller asking about why Spack can't find manually installed packages in
opt/spack
- How could we support this?
- Matt already implementing support for packages outside the spack build space, but installing into Spack's build space violates the rules. Essentially adding a foreign library to a managed area.