David Seddon

Well, I just released 3.5 and I'm afraid it still isn't including the Python 3.13 wheels, except for Windows 😞 . I'm guessing it's because the setup-python Github action doesn't...

I tried setting up [this PR](https://github.com/seddonym/grimp/pull/163) just for the purposes of publishing the additional wheels, but no joy. Here's [the release workflow](https://github.com/seddonym/grimp/actions/runs/11258259337). The setup-python step does seem to install Python...

I think the issue might be with Maturin's discovery of Python interpreters. The good news is, I tried explicitly setting the interpreter and the wheels have now built and published!...

Thanks for looking at this - just FYI I am on the fence about adding this to Grimp - when I've looked at this before, the results you get on...

Oh, I've just realised what this is related to - https://github.com/seddonym/import-linter/pull/250. Sorry if I'm giving you mixed messages 😄 I've added a link in the description. As I mentioned in...

Thanks @K4liber for your continued thinking about this! > When it comes to the point nr 1, I would really encourage you guys to welcome such functionality in the Grimp...

Thanks for the comments everyone. My current thinking about our end goal would be for Grimp to have a high level function something like this: ```python @dataclass class CyclicDependency: parent:...

> I am not sure what is a justification on CyclicDependency.sibilings attribute existence. Maybe it can just be a property calculated over the cycles? Its a detail actually, but I...

> How about next to the max_cycles_per_dependency parameter we can introduce max_cycles_overall? It will allow for ASAP feedback if one executes the contract check with max_cycles_overall=1. Not a bad idea,...

> It seems like we still need to "find it all" and then reduce the results based on the max_cycles_per_dependency value. Could you say a little bit more about why...