Bruno Kolenbrander

All the data is here: https://github.com/mejrs/data_rs3 (note that the npc data is rather out of date). The files/folders of interest are the npcids folder, npc_morph_collection and npc_name_collection. You can use...

You'll need to rebase on main to make CI work (because of https://github.com/PyO3/pyo3/pull/2489 ) :)

I asked (someone else) a while ago and they recommended that approach. Personally I think this is very misbehaved and kind of a ridiculous corner case so my preference would...

CI failure is because `impl From for LinkedList` is only available for Rust 1.56+. `vec![1,2.3].into_iter().collect()` is a decent way to make one for a test, I think. > What would...

> I understand this relies on __traverse__ and __clear__, right? These are only relevant if you have cycles in your struct that the garbage collector must break. If you don't...

> However, the `PyErr` type isn't available in `pyo3-macros` which means I can't trivially write a conversion from the custom exception type to `PyErr` nor can a `new_err` return it....

Do we know why this segfaults? I can imagine why, but I glanced through the capi documentation and didn't see it called out anywhere. > While I don't disagree, there...

I'd like something better than slapping unsafe on it and calling it a day. What about making `Python::with_gil` and `Python::acquire_gil` panic (more likely: abort) if called from within `__traverse__`? We...

> I once experimented with a derive macro Could something like this work? With the `__traverse__` impl above as an unsafe escape hatch? ```rust #[pyclass] struct GcIntegration { #[pyo3(gc)] self_ref:...

> I wonder if it would be worth adding a "Performance Suggestions" or similar section to the docs, perhaps the FAQs? Regardless of the outcome of this issue I think...