AFAIK, whether to allow dismissing the stale approval is configured in the
`.asf.yaml` file:
```
# Require new reviews when new commits are pushed
dismiss_stale_reviews: true
```
I'm open to this. If allowed not to dismiss the stale approval, the
committers who merge these commits in the main branch need to see the
changes again (if needed) to make sure everything goes well. It would be
better to have one CI running to send a note on the changes for the
committers' information.
Best,
Dianjin Wang
On Thu, Jan 16, 2025 at 6:57 PM Max Yang <maxyang...@gmail.com> wrote:
> Sounds reasonable in some cases. Is it possible to add a special CI skip
> cancel instruction if author don't want to cancel APPROVE?
>
> On Thu, Jan 16, 2025 at 5:42 PM jiaqi.zhou <jiaqi...@163.com> wrote:
>
> > Hi all,
> >
> >
> > I personally have an idea: we should not cancel approve after PR has
> been
> > updated.
> >
> >
> > 1. After the reviewer approves the PR, the current PR may still need to
> be
> > rebased (do git rebase main or git merge main)
> > 2. The approved PR will still run CI once to ensure that the current PR
> is
> > compliant
> > 3. If the PR is updated, the reviewer needs to approve it again, which
> > wastes the time of both the committer and the reviewer.
> >
> >
> > For example:the reviewer made some minor comments (such as suggestions
> for
> > changing some typo, or the comments that can be rejected) and approved
> > current PR. The commiter accepted these suggestions and update the PR.
> > After that, reviewer still need click the approve button again (or may
> need
> > to review the PR again). I think this is unreasonable.
> >
> >
> > Thanks
> > Jiaqi
> >
> >
> >
> >
> >
> >
>
