Repository navigation
HDDS-17709. Restore pull_request trigger on master - #11944
adoroszlai wants to merge 1 commit into
Conversation
chihsuan
left a comment
There was a problem hiding this comment.
Thanks for the follow-up! @adoroszlai Restoring full CI for now makes sense.
Just a note: we removed dependabot** from push.branches-ignore in #11871, so dependabot PRs may run CI twice now. Probably minor since this is temporary, so your call. 🙂
| - master | ||
| # - master | ||
| push: | ||
| branches-ignore: |
There was a problem hiding this comment.
Since pull_request runs for master again, should we also bring back 'dependabot**' here? #11871 dropped it from this list because dependabot PRs no longer had a pull_request run. Without it, each dependabot PR to master gets two full CI runs: one for the push to its branch in apache/ozone (concurrency group ci-<sha>) and one for the PR (ci-<PR number>), so neither cancels the other.
branches-ignore:
# only master, builds on schedule
- master
- 'dependabot**'|
Thanks @chihsuan, @smengcl for the review. I'm leaning towards not restoring the trigger. While showing status as a single check is not ideal, it is better than queueing for several hours. I'll keep this open but would prefer to continue with the other PRs. Please take a look at:
|
What changes were proposed in this pull request?
Fork CI Statusmirrorspost-commit.yml, so status appears as a single check. I think it should mirrorci.ymlfor more detailed checks in the PR summary. But the real runs in fork are listed underpost-commit.yml(and I get the same impression by listing runs viaghCLI), so it may not work.Also, it would be nice to get more detailed output from
fork-ciaction in the workflow run. Currently it only shows errors and a summary:Fork CI: 42 processed, 30 checks written, 0 failed.This change partially reverts #11871: re-enable the
pull_requesttrigger onmaster, but keeps theFork CI Statusworkflow so we can continue refining it.https://issues.apache.org/jira/browse/HDDS-17709