Its likely that someone force pushed into delivery. I chatted to Laszlo too via slack and he mentioned he had to rebase his local delivery branch again for a PR because he saw commits on gh of another PR in his PR - this is another indication that delivery got reverted back to an older state.

It would also explain why two (latest) PRs (hopefully nothing more) disappeared after they were integrated using the github merge button.

you can check this by opening a PR, e.g:
https://github.com/apache/netbeans/pull/4007

scrolling down, then clicking on the hash of:
"ebarboni merged commit 74e1e5f into apache:delivery 5 days ago"

the merge commit is gone.

If we can confirm that its only the 2 PRs which are missing - and everything else is fine, we could simply apply them again. Luckily the remote branches are still there since nobody removed them yet - so no work is lost. Unfortunately I don't think its possible to reopen merged PRs (since this would not make sense workflow wise). All we would have to do is to reopen+merge.

I don't know who can force push into delivery directly, maybe there is a way to restrict that if it isn't.

I didn't have a lot of time to investigate this so far unfortunately, we should compare delivery with master, maybe someone even has a pristine delivery branch locally, from before the incident.

best regards,
michael

On 02.05.22 11:32, Eric Barboni wrote:

Hi folks and hi to cc’ed PR writer.

It seems that PR #4004 and #4007 did not apply correctly to delivery branch and by consequence to master branch. Somehow should be my fault

Michael founds the issue on Friday and found the other PR missing.

I try to explore the milestone finding non present PR too. I found the same.

Could PR owner of PR belonging to milestone 14 double check ?

Would be nice if Martin and Michael could repropose their respective PR to be on RC3.

Best Regards

Eric

Reply via email to