Hi Stamatis,

Thank you for raising this topic.

Currently, we have three options for merging a PR that contains multiple
commits. Below, I’ve outlined these options in order of preference below.
Please let me know what everyone thinks about it.

1. Request Authors to Squash Merge:
In this approach, we ask the PR authors to squash their commits into a
single coherent change. IMO, this is the safest option as the author is
involved and no manual steps are required.

2. Maintainers Squash-Merge Locally and Push to GitBox:
The maintainers perform a squash merge locally on the PR. Afterward, they
push the resulting change to GitBox. This method was successfully used for
merging PR-359, and it serves as a reliable backup option.

3. Maintainers Squash-Merge Locally and Force Update the Author’s Branch,
Then Use GitHub to Merge:
While all maintainers are allowed to update the PR, and is clearly
communicated when a PR is created, but we should be mindful that some
contributors might perceive it as aggressive if they are unaware of this
behavior.

Action Items:
I'd like to document these options in the PULL_REQUEST_TEMPLATE.md file
and clarify them in the contribution guidelines section of the README to
prevent any confusion.

Feel free to share your thoughts or additional suggestions!
Best,
Ashvin


On Fri, Mar 15, 2024 at 2:44 AM Stamatis Zampetakis <zabe...@apache.org>
wrote:

> Hey everyone,
>
> It seems that branch protection rules that are setup in GitHub via the
> .asf.yaml file [1] can be bypassed if someone pushes directly the
> changes to the GitBox repository. I tested this while merging PR-359
> [2].
>
> If we don't want to allow pushing directly to GitBox maybe we should
> raise an infra ticket to block this (if possible).
>
> On the other hand, instead of asking the contributors to manually
> squash and rebase their PRs we could leave this window open for this
> or other similar use-cases.
>
> I don't really mind either way but just wanted to bring this up. In
> case we want to act upon it this thread would be a good place to
> start.
>
> Best,
> Stamatis
>
> [1]
> https://github.com/apache/incubator-xtable/blob/7c7b54b3870791578b4689b29cf8794866eefa52/.asf.yaml#L36
> [2]
> https://github.com/apache/incubator-xtable/pull/359#issuecomment-1999265918
>

Reply via email to