Pending pull requests from new contributors

Hello,

We currently have several pull requests from new contributors pending review. I know everyone is contributing on a volunteer basis, so I am sharing this mainly to surface the pending items and help with coordination. Are there plans to address any of them in the coming days?

PRs that have not received a code review

This PR was updated by the contributor and needs a second review:

PRs with a Copilot review but no human follow-up

I am not sure if the contributors are waiting for direction or if they have lost interest in the PRs.

PRs to close?

Should we close the following PRs?

I monitor GitHub notifications several times per week. Does the Core Dev Group want me to run a Copilot review on new PRs that seem legitimate and have not received a review?

This could give contributors some initial feedback. However, I would not be able to address technical questions or feedback from the contributors, or merge the PRs.

Thank you for your help.

I'm going to approve and review a few of these but I won't touch AI generated code from first-time contributors which most of these seems to be.

I'm not sure how other maintainers feel but from my point of view these contributors are an exploit waiting to happen(someone with a merge get slightly elevated rights), and if they can prompt I can probably prompt it better and faster than I can review "their" code.

1 Like

Thanks for the summary Martin. People can generate their own live list using this query which will show all unapproved non-draft PRs sorted with those open the longest first. There are currently 39 PRs meeting these criteria, including 9 of mine (I know the subject references "new contributors," but I think timely review for existing contributors is important too).

I've been meaning to start a thread about the PR review process, but I'll do that separately, so it doesn't get lost here. I think we may already have a relevant thread.

Regarding AI, we need to clarify our guidelines around the use of AI so that contributors know what's allowed and what is not, as well as update our agent instructions so that coding agents create PRs which conform to our standards. I'm not personally opposed to AI generated/assisted PRs, but I'm not willing to devote scarce review resources (mine or other developers) to PRs, whether AI or human generated, which don't meet a cost/benefit threshold. This includes gratuitously fixing typos in comments, inventing features that no one asked for, etc. Although it seems obvious to experienced developers, it's not something we can expect new contributors (or coding agents) to understand.

Does the Core Dev Group want me to run a Copilot review on new PRs that seem legitimate and have not received a review?

I think we can set up a workflow to do this. Ideally it would also greet new contributors, set expectations, etc. One of the expectations would be that submitters address review feedback from Copilot (or other agents that we designated). We currently require manual approval for CI workflows from new contributors, which I think is a useful thing to keep in place.

Tom

1 Like

Thank you @abbe98 and @tfmorris for addressing the backlog.

I will open a PR to create a GitHub action to draft those automations.

This is an example of what the automated message would look like.

Someone replied to a topic you are Watching.

This is an example of what the automated message would look like.

Thanks. I think the message itself is a fine start for new contributors, but in this case what we probably wanted to say is "Please wait until issues are triaged before asking to be assigned to them."

For PRs which affect the UI, we should probably ask for a screenshot or movie (depending on whether interaction is involved), which is something I don't think we have in our guidelines.

Tom

1 Like