Closed weaverba137 closed 3 years ago
FYI #547 documents the pixsim test failure and likely cause (a change in specter, but requiring a non-negative input flux cut in desisim). We could proceed with this PR even if that test is still failing.
The "suspiciously low XYZ flux" problem appears occasionally in the nightly tests and previously travis tests; I usually ignore and rerun.
OK, the tests were successful after a restart, but I still think the low flux problem should actually be fixed, not ignored. In particular, experts know to restart the tests (and know how to restart), but newcomers would have to be aware that this is a known issue.
PS, I still have a few minor changes to make to this PR, like removing the .travis.yml
file, so please don't merge immediately.
@sbailey, thank you merging very soon.
Test failures are related to #507, ignoring.
This PR migrates desisim to use GitHub Actions instead of Travis CI for unit tests.
While testing, a version of #507 reappeared, although this time it is complaining of suspiciously low ELG or LRG flux.
See this run for full details.
Oddly though, #534 did not reappear. Perhaps that was a red-herring and desimodel/branches/test-0.12 does have the fix installed.