Has this been discussed anywhere? I don't see much benefit in this, and closing such PRs is not a big enough burden, that would justify adding any script for this purpose. I'm strongly against it.

Has this been discussed anywhere?

I think so.
cc @sarahboyce @nessita @LilyAcorn

Has this been discussed anywhere?

I think so. cc @sarahboyce @nessita @LilyAcorn

I'm a bit confused. Why you pinged so many folks? If it was discussed and accepted it's enough to point me to this discussion.

https://docs.google.com/document/d/1cifdwoPGxOiSZa4QiibdqAdWJbjF3BL7YUm3uv61otM/edit?tab=t.0#heading=h.a45nsz1907yp
It doesn’t include details about this PR, as the work had already begun following a discussion with them.

I almost thought this was also a "nonsense PR", there's no ticket linked to in the description, no way of verifying legitimacy other than it mentions GSoC and was able to find this: https://summerofcode.withgoogle.com/programs/2025/projects/xF6r182s 😅

Anyway I agree with Felix... The total time spent creating, testing, reviewing and further maintaining this script seems like it would far outweigh the total time taken closing toast PRs.

If anything I'd recommend removing the toast bit from the docs completely and referring people to the official GitHub docs if they want to learn about PRs 🤔

Anyway I agree with Felix... The total time spent creating, testing, reviewing and further maintaining this script seems like it would far outweigh the total time taken closing toast PRs.

Exactly, we're talking about 5-10 (at most, e.g. 86 "toast" PRs in the last 7 years, ~12 per year) invalid PRs per month, that will take the same number of minutes to close. On the other hand we have 600 LOC PR that need to be understood, reviewed, merged and maintained. Furthermore, it will catch half of those cases and will likely mark at least twice as many valid PRs. I appreciate your work but this is an overkill to me.

👋 here’s Saurabh and Lily’s high-level project description: Automate processes within Django contribution workflow

@shangxiao @felixxm I think we need to keep an open mind for stuff like this. There’s lots of room for more automation, and perhaps the ROI isn’t big enough for just this use case, but it seems like a fine one to experiment with? And I’d hope a triage script doesn’t require as much correctness as other 600 LOC PRs.

Re "legitimacy", if in doubt – assume the best.

👋 here’s Saurabh and Lily’s high-level project description: Automate processes within Django contribution workflow

@shangxiao @felixxm I think we need to keep an open mind for stuff like this. There’s lots of room for more automation, and perhaps the ROI isn’t big enough for just this use case, but it seems like a fine one to experiment with? And I’d hope a triage script doesn’t require as much correctness as other 600 LOC PRs.

Re "legitimacy", if in doubt – assume the best.

I still think we're spending too much time trying to automate instead of actually doing something.

This script will help to close ~1 "toast" PR per month, and in the same time mark dozens of PRs with test and tutorial as invalid that are not (we had thousand of such PRs and most of them were valid). It's debatable that this will improve contribution workflow, especially for contributors 🤷

👍 fair enough. Again this is an in-progress GSoC project, so although the value is debatable in your mind, there’s plenty enough interest from others so important to give this a chance.