Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
Description
If we don't run source release test on pull request, it's difficult that we detect source release test failure on pull request.
For example,
https://github.com/apache/arrow/pull/5852 introduced source release test failure but we can't detect it on pull request.
Then
https://github.com/apache/arrow/pull/5889 tried to fix the source release test failure and merged but it didn't fix the failure. We couldn't detect it because we didn't run source release test on pull request.
Then
https://github.com/apache/arrow/pull/5893 will really fix the source release test failure.
Why did we not run source release test?
https://github.com/apache/arrow/pull/5889#issuecomment-558096239
> The reason why I've removed them from running on each PR to reduce the number of the triggered builds. See the mailing list thread about the queueing concerns. The docs build takes fairly long time, in contrary of the source release test.
The mailing list discussion:
[CI] Docker-compose refactor and GitHub Actions
https://lists.apache.org/thread.html/8fefdebd76742fe4ffb2ee7d0fa88ddf2b3229c55935c146b0d97e52@%3Cdev.arrow.apache.org%3E
Attachments
Issue Links
- links to