search

libMesh / TIMPI

Templated Interface to MPI
GNU Lesser General Public License v2.1
12 stars 11 forks source link

set_union fixes #106

Closed roystgnr closed 1 year ago

roystgnr commented 1 year ago

Closes #105 IMHO - @jwpeterson should chime in on that too.

Important bug fix here, too. If Peter programmer packs a padded pair into packets, how many packets did Peter programmer pack?

I'm just now realizing I haven't yet tested this with --disable-mpi. Hopefully CI will tell me how bad a mistake that was.

jwpeterson commented 1 year ago

I'm just now realizing I haven't yet tested this with --disable-mpi. Hopefully CI will tell me how bad a mistake that was.

Looks like "without MPI" tests are the only failures here.

roystgnr commented 1 year ago

That's a fair worry. It's exactly the same worry as with insert or merge, though; so I fear the fix looks less like "make TIMPI safer" than like "make C++ safer". They've been going in the right direction on that one; fingers crossed for C++79?

I'll merge this once I have the --disable-mpi fixes in ... but I was screwing up my builds for that and only just figured out why I wasn't replicating the problem, so it'll still take a bit.

moosebuild commented 1 year ago

Job Coverage on 699a11c wanted to post the following:

Coverage

Coverage did not change

Full coverage report

This comment will be updated on new commits.