Description
https://github.com/greenelab/greenelab.com/actions/runs/3819357107
remote: error: At least 1 approving review is required by reviewers with write access.
It makes sense that you'd want an approver for pull requests, but unfortunately that means the GitHub Action that periodically runs the automatic citation process to update e.g. ORCID papers will fail because it doesn't have permissions.
Apparently there is no way to make an exception for GitHub Actions (yet, but it is a heavily requested feature):
community/community#13836
I could maybe make it such that the action opens a PR instead of committing directly to main
, but then I'd imagine them really piling up over time. Maybe I could try to check if a PR for that is already open and update/add to it instead of creating a new one, but that gets a lot more complex.
@cgreene Thoughts? I'm not just thinking about greenelab.com, but the template in general.
I'm thinking we just bet on GitHub implementing the Actions exception within the next year (since it's so heavily requested), and until then, say that people can't have both branch protection rules and scheduled action runs.
Activity
cgreene commentedon Jan 6, 2023
I agree with the current bet:
If the deploy happened off github pages and only the main was protected, that'd avoid this issue (which, likely, is why manubot manuscripts don't get hit by it). However, that's a lot of additional work and setup would be more painful.
cgreene commentedon Jan 6, 2023
We can unprotect this repo - no concerns on that. I'm also ok with opening a PR
cgreene commentedon Jan 7, 2023
I unprotected the branch, but I'm also coming around to the idea of opening a pull request so it becomes obvious what is being added.
vincerubinetti commentedon Jan 7, 2023
Maybe we can try it in greenelab.com as a pilot and see how it goes. I would probably want to reduce the frequency from weekly to monthly if we're opening up PRs. Given how I've seen other people use/maintain their website, I think most people wouldn't keep up with weekly.
Maybe you could do a twitter poll?
Hopefully this action will make it easy, so I could include both methods in the workflow and just comment one out by default (whichever is less popular).
cgreene commentedon Jan 11, 2023
https://twitter.com/GreeneScientist/status/1613211264171536384?s=20&t=FllxhQl5RrcgXQXRPYRy_g
cgreene commentedon Jan 11, 2023
I have used that action - happy to help out if you want it.