Graham Markall

The code for recent experiments / investigations during the triage meeting yesterday (2025-06-03) is in https://github.com/gmarkall/numba/tree/issue-10090

> since the leak is triggered by `fn(0)`, i go through the pipeline `fn(0)` and find the first breakpoint yield the leak is a call to `insert_user_function()` of `core/base.py` with...

Just to confirm, because this never got a response from a maintainer before, this is still an issue with numba `main`. Following the OP's steps above (which were clear and...

I'd guess we're at least 6 months out from 0.57. Workarounds from this issue seem to be: - Add `zip_safe=False` to the `setup.py` of the packages using Numba caching: https://github.com/numba/numba/issues/4908#issuecomment-571537473...

Thanks for the request - adding a new target to Numba is quite a lot of work for the initial implementation, and there is a continued maintenance workload for each...

> Just want to bring some fresh air here since there hasn't been any update for a while. Pasting nonsense from ChatGPT simply wastes the time of maintainers and anyone...

There's none that I'm aware of. It would be a significant amount of work so I don't expect this will be likely to change unless someone shows up to do...

It appears that it converts DXIL, which is a subset of LLVM 3.7 IR. I would guess that this exists in support of the Game Porting Toolkit, so I'm not...

When @stuartarchibald reviewed the original PR that added the comment, he also had to go and read the `weakref.finalize` source and outline his understanding to me, which should have been...

From @isVoid in https://github.com/NVIDIA/numba-cuda/pull/145#discussion_r2004347569: > It even took me 30min to understand this refresh PR. I think this is super subtle. Especially the point "the first finalizer registered also calls...