Closed SMoraisAnsys closed 2 days ago
Thanks for opening a Pull Request. If you want to perform a review write a comment saying:
@ansys-reviewer-bot review
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 83.16%. Comparing base (
0e38674
) to head (4ecf96c
). Report is 1 commits behind head on main.
Temporary passing to draft due to CI typo (deleted breaking lines)
Changes:
vale
related folder to use version>=3
ofvale
;warning
intoerror
.The reason for the last change is related to #4880. The problem this PR tackles is the fact that a PR from a fork, dependabot, ... ends up failing the
doc-style
job of our CICD.From my understanding of the problem, it is related to the permissions associated to GitHub token in the event pull_request_target. Since this permissions can differ from one context to another, some PR might add changes passing
doc-style
but doing so while addingvale
warnings. Those are pending bombs as they will impact future PR(s) where the permissions are not enough (for example, cases where GitHub token doesn't have write permission of Review API). Indeed, in such cases, the report results are performed via logging command and creating annotations similar togithub-pr-check
reporter as a fallback. However, the behavior is different as in this case warnings cause reviewdog to exit with status code 1 and make thedocumentation-style
fail.Proof that author access has no impact: PR #4882 comes from my fork of pyaedt and has the same permission issue.
This seems to be related to https://github.com/reviewdog/reviewdog/issues/1408 and the workaround we are applying here is to update every
warning
severity toerror
.Closes #4880