phanikumv commented on code in PR #75447:
URL: https://github.com/apache/airflow/pull/75447#discussion_r4357390431


##########
.github/workflows/ts-sdk-release.yml:
##########
@@ -83,7 +83,20 @@ jobs:
           test "${DISPATCH_REF}" = "refs/tags/${RELEASE_TAG}"
           test "${DISPATCH_SHA}" = "$(git rev-parse HEAD)"
           test "$(git rev-parse "refs/tags/${RELEASE_TAG}^{commit}")" = "$(git 
rev-parse HEAD)"
-          git merge-base --is-ancestor HEAD "$(git rev-parse 
refs/remotes/origin/main)"
+          # A release is cut from main or from a core release branch. 
Cherry-picks onto a
+          # release branch get new SHAs, so containment in main alone is not 
enough.
+          contained_in=""
+          for ref in $(git for-each-ref --format='%(refname)' \
+              refs/remotes/origin/main \
+              'refs/remotes/origin/v*-test' \
+              'refs/remotes/origin/v*-stable'); do

Review Comment:
     Before this, passing the check meant the commit had gone through main's 
required review. `v*-stable` is under the .asf.yaml ruleset, but `v*-test` 
isn't, so a commit pushed directly to v3-4-test (or to any new branch matching 
`v*-test`; the glob also accepts something like `vx-test`) now passes. Is that 
intended?
   
   Accepting only stable isn't an option for 3.4 since v3-4-stable doesn't 
exist yet. If we keep `-test`, I'd at least anchor the match to real release 
branch names, e.g.
   `[[ "${ref}" =~ ^refs/remotes/origin/v[0-9]+-[0-9]+-(test|stable)$ ]]`, so 
it matches what DEVELOPMENT.md says.



-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to