search

JuliaTime / TimeZones.jl

IANA time zone database access for the Julia programming language
Other
86 stars 52 forks source link

disable fail-fast on GHA CI #328

Closed oxinabox closed 2 years ago

oxinabox commented 3 years ago

Fail-fast on GHA doesn't work great. The benchmark job erroring shouldn't prevent the actual testing jobs from running. Since the benchmarking job can error if you change how any types are defined https://github.com/JuliaTime/TimeZones.jl/pull/327#issuecomment-818623379

codecov-io commented 3 years ago

Codecov Report

Merging #328 (5808d45) into master (906cea0) will decrease coverage by 0.20%. The diff coverage is n/a.

Impacted file tree graph

@@            Coverage Diff             @@
##           master     #328      +/-   ##
==========================================
- Coverage   93.80%   93.59%   -0.21%     
==========================================
  Files          32       32              
  Lines        1534     1531       -3     
==========================================
- Hits         1439     1433       -6     
- Misses         95       98       +3
Impacted Files Coverage Δ
src/compat.jl 0.00% <0.00%> (-100.00%) :arrow_down:
src/types/zoneddatetime.jl 96.00% <0.00%> (-0.06%) :arrow_down:
src/parse.jl 94.44% <0.00%> (-0.02%) :arrow_down:
src/tzdata/compile.jl 95.50% <0.00%> (+0.32%) :arrow_up:

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 906cea0...5808d45. Read the comment docs.

omus commented 3 years ago

Are you sure fast-fail works this way? Seems strange that it looks at jobs that aren't part of it's on matrix

oxinabox commented 3 years ago

No, I am not sure.

omus commented 3 years ago

When set totrue, GitHub cancels all in-progress jobs if any matrix job fails. Default: true https://docs.github.com/en/actions/reference/workflow-syntax-for-github-actions#jobsjob_idstrategyfail-fast

Seems like this documentation is wrong then. Would be good to open an issue.

omus commented 2 years ago

So the solution to this is to actually set fast-fail: false on just the benchmark job. With this then a failure from the benchmark job will not impact any other running job but jobs with fast-fail: true will cause other jobs to be cancelled. I'll make a PR for this.