Hi! I have an issue where PRs in a stack get stuck on "Waiting on CI". I have to do a dance of locally

gt sync && gt ss

between every merge to get them to run the PR actions in GitHub. How I think I've ended up in this situation: 1. I had Cursor cloud create a bunch of PRs for me. 2. Those needed a bit of clean up, which I did, and then I wanted create a single clean stack out of those PRs as they were very much related things. 3. So I moved all of them to the same stack with

gt move

4. And finally I submitted everything And now this is the end result. The PRs I've created myself on top of this work nicely as you can see in the screenshot. Is this just a bad practice, or am I doing something wrong as I end up with the waiting on CI issue?

(If applicable) To help us more clearly understand what you're trying to do, please provide additional details about: - Screenshots / debug logs / list of commands you've tried to run so far - Your workflow - Current state

Operating-App/monorepo. E.g. PRs 2867, 2865, 2866, 2859. I've already merged some of those, but some are still open and 2859 seems to be at least having the issue right now.

interesting. Looks like this is based on github actions? have you set up your CI as we reccomend (docs ref) specifically around the triggers?

Yeah, it's GH actions and no it seems that I missed those instructions somehow. I updated the config, but that didn't seem to really fix the issue. As far as I can tell, I can't even see anything in GH actions for the branch of PR#2877 which now the only remaining problematic one as I've merged others.

youll need to push to the branches to retrigger probably, github needs something to know to run

Yeah I did, everything else ran

was the bottom one a no op maybe? might need to push a minor change to trigger it

All of them do say updated in the CLI at least. I made the pr.yml change on the bottom most non-merged PR, the one with unresolved comments, so that should trigger everything to run, right?

the thing here is external to graphite. We are waiting on signal from github. If github does not send then theres not much we can do. If you continue to hit this after youve updated the CI settings to match our reccomendation we can definitely investigate further

Yeah I kind of figured its something like that