Closed junghans closed 3 years ago
Merging #325 (1625eb4) into master (1ce129a) will increase coverage by
3.1%
. The diff coverage isn/a
.
@@ Coverage Diff @@
## master #325 +/- ##
========================================
+ Coverage 93.6% 96.8% +3.1%
========================================
Files 37 36 -1
Lines 2830 3258 +428
========================================
+ Hits 2651 3156 +505
+ Misses 179 102 -77
Flag | Coverage Δ | |
---|---|---|
clang | ? |
Flags with carried forward coverage won't be shown. Click here to find out more.
Impacted Files | Coverage Δ | |
---|---|---|
core/src/Cabana_LinkedCellList.hpp | 96.9% <0.0%> (-3.1%) |
:arrow_down: |
core/src/Cabana_Distributor.hpp | 94.0% <0.0%> (-1.6%) |
:arrow_down: |
cajita/src/Cajita_LocalGrid_impl.hpp | 93.4% <0.0%> (-1.4%) |
:arrow_down: |
core/src/Cabana_Parallel.hpp | 93.2% <0.0%> (-1.0%) |
:arrow_down: |
cajita/src/Cajita_GlobalMesh.hpp | 91.3% <0.0%> (-1.0%) |
:arrow_down: |
cajita/src/Cajita_Halo.hpp | 97.3% <0.0%> (-0.8%) |
:arrow_down: |
core/src/Cabana_VerletList.hpp | 98.9% <0.0%> (-0.6%) |
:arrow_down: |
core/src/Cabana_Halo.hpp | 95.3% <0.0%> (-0.6%) |
:arrow_down: |
cajita/src/Cajita_Array.hpp | 94.7% <0.0%> (-0.5%) |
:arrow_down: |
core/src/Cabana_Tuple.hpp | 100.0% <0.0%> (ø) |
|
... and 23 more |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 8fe8902...aa703d7. Read the comment docs.
@streeve some cajita tests fail on fedora:rawhide
any ideas?
I'm trying to recreate locally, but no luck yet
@streeve some cajita tests fail on
fedora:rawhide
any ideas?
Man IndexConversion
is a weird test to fail. Perhaps there is an uninitialized value issue somewhere I missed because it is dumping out 0
?
I'm trying to recreate locally, but no luck yet
Did you try the docker container?
docker run -it ghcr.io/kokkos/ci-containers/fedora:rawhide /bin/bash
Did you try the docker container?
Turns out that's super helpful.
fedora:rawhide
tests seem fine - only failing because of mpi (passed with -oversubscribe
)
I reproduced the error with fedora:latest
; it only affects periodic cases, but no fix yet
So Cajita_Halo
failure is the issue (causing the periodic Cajita_IndexConversion
failures
I haven't found the bug that causes Halo
to fail on some distributions, but this now at least runs all the tests. I had to reduce printing in those failing tests (not sure why GHA doesn't truncate) and remove ubuntu:devel
for now as it hangs
I see the Travis still showing up in the branch protection even though the
.travis.yml
has been removed - is there something else in the repo settings we need to change?
I don't have those permissions, but yes I think there's a setting to require Travis
I see the Travis still showing up in the branch protection even though the
.travis.yml
has been removed - is there something else in the repo settings we need to change?I don't have those permissions, but yes I think there's a setting to require Travis
OK - I'll add you to the list. We don't wan't to disable anyway until this one is merged. All other open PRs will also need to be rebased to get the new CI
Initial step in migrating from travis to GHA. WIll continue with #345