Hey all, some of our devs have recently starting reporting Graphite review dismissals, seemingly caused by restacking a PR after a partial merge. The dismissal message mentions that "The base branch was changed.", but it should only have changed from the previous parent PR to the

main

branch. See

/pr/Didask/odc/23770

for example, thank you 👍

(If applicable) To help us repro, please send the following if you haven't already: - PR links, screenshots, and/or debug output if applicable - Repro steps - Does this happen consistently? Just to you or others on your team also? - Is this a regression? Is the problem new? - If you're having an issue with the CLI, please send us debug logs using

gt feedback --with-debug-context <any message>

- PR links, screenshots, and/or debug output if applicable

•

/pr/Didask/odc/23770

•

/pr/Didask/odc/23655

•

/pr/Didask/odc/24597

- Repro steps

Not certain about the steps taken, I haven't reproduced it yet, but I'm guessing they were: • Open a PR stack with two PRs • Review both, merge the first one •

gt get

and

gt submit

the second to restack it on

main

• The second PR review is dismissed

- Does this happen consistently? Just to you or others on your team also?

Several team members have reported this

- Is this a regression? Is the problem new?

New as far as I'm aware

- If you're having an issue with the CLI, please send us debug logs using

gt feedback --with-debug-context <any message>

N/A

Hey Jimmy, my first thought here is this is caused by a github branch protection rule (BPR)

dismiss stale approvals

. We have some documentation on how to handle this without using the BPR setting here

As far as I can tell we don't have that setting enabled. These aren't getting dismissed by GitHub, but by Graphite

interesting. Thanks for the quick response. And are you using our action then?

gt get

and

gt submit

the second to restack it on

main

I guess this is needed for merge conflict cases versus relying on our auto restack?

And are you using our action then?

Are you talking about the meregeability check? If so, yes

no, the github action we provide in lieu of dismiss stale approvals https://github.com/marketplace/actions/dismiss-stale-approvals

As far as I can tell we don't have that action either, I'm not sure where the dismissal is coming from

this one maybe

I don't think we had multiple contributors for these branches, and I'm not sure this rule is related to the Graphite app dismissing the review directly? It seems more like a GitHub mergeability check, not a reason to remove an existing review

Looks like we do require this one to be turned off

Fair enough, but can you confirm that this is leading to review dismissal? I'm not convinced the two issues are linked

i cannot confirm that for certain, the only other thing i would double check is that you dont have that github action. Could you share the actions on this repo?

Sure, here they are

awesome thanks for that. Will let you know what else i can come up with

Okay that makes more sense, thank you for investigating. I'll disable "Require approval of the most recent reviewable push" and see if that fixes the issue 👍