CI: Fix install-with-conda job #7103
Conversation
| if: >- | ||
| always() && | ||
| (steps.install.outcome == 'failure' || steps.tests.outcome == 'failure') && | ||
| failure() && |
There was a problem hiding this comment.
This is a better way to express what we need: run this step iff any of the previous steps failed.
https://docs.github.com/en/actions/reference/workflows-and-actions/expressions#failure
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #7103 +/- ##
==========================================
+ Coverage 79.61% 79.61% +0.01%
==========================================
Files 566 566
Lines 43546 43546
==========================================
+ Hits 34663 34666 +3
+ Misses 8883 8880 -3 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
danielhollas
changed the title
| channels: conda-forge | ||
|
|
||
| # Use mamba because conda is running out of memory | ||
| # see https://github.com/conda-incubator/setup-miniconda/issues/274 |
There was a problem hiding this comment.
This is no longer necessary since conda now uses libmamba solver by default.
| conda config --set solver libmamba | ||
| # Temporary workaround: https://github.com/mamba-org/mamba/issues/488 | ||
| - run: rm /usr/share/miniconda/pkgs/cache/*.json |
There was a problem hiding this comment.
This step started failing. I haven't investigated too deeply, but simply removing this seems to work.
There was a problem hiding this comment.
from:
6da1333
- The switch to mamba required a workaround to delete any existing JSON
files in/usr/share/miniconda/pkgs/cache. This seems a known problem:
RuntimeError on cache in multi-users case mamba-org/mamba#488
Using conda now thus probably makes this unnecessary.
GeigerJ2
left a comment
GeigerJ2
left a comment
There was a problem hiding this comment.
LGTM! Thanks for the explanations in the comments, very helpful.
Seems like the "workaround" has stopped workarounding and instead started breaking the job so let's remove it! Also added a Slack notification step so that we notice such breakage earlier in the future.