ci: don't cancel in-progress jobs in matrix if others fail - #1213
#1213ci: don't cancel in-progress jobs in matrix if others fail#1213G-Rath wants to merge 1 commit into
Conversation
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #1213 +/- ##
==========================================
- Coverage 94.44% 93.93% -0.52%
==========================================
Files 3 3
Lines 360 412 +52
Branches 132 155 +23
==========================================
+ Hits 340 387 +47
- Misses 18 23 +5
Partials 2 2 ☔ View full report in Codecov by Sentry. |
evenstensberg
left a comment
There was a problem hiding this comment.
I think fail fast is good. When the CI fails, you'll most likely have to rerun the jobs anyway
|
Yes, but you're more likely to have to re-run it more because you might have os-specific failures - for example I won't know if #1211 is going to pass on any OSs because the macOS job failed; now that #1215 is landed that should be unblocked but now I'll just learn if any-but-not-all of the OSs are passing (and repeat until all jobs are passing...) Ultimately up to you though - please do close this PR if you'd still prefer to keep fail-fast enabled :) |
|
I think we will close this, but thank you so much for the contribution, appreciate you! |
This PR contains a:
Motivation / Use-Case
Matrixes "fail fast" by default, which can be annoying as failures in linting and the test suite by the time a pull request are technically expected to be because of one specific dimension (e.g. an os, since most people don't develop on multiple os's at the same time) - so this disables that :)
Breaking Changes
Not a breaking change
Additional Info