Max “Goldstein” Siling

> may pose issues on other architectures Could you elaborate on this, please? As I've written before, if cast to a data pointer truncates, cast to usize also does, so...

> wasm is a well-known architecture where function pointers are in a different address space from data pointers (they're table indices) despite being the same size. Is it relevant? Cast...

> Making people use transmute seems reasonable Note that this requires fnptr to integer to fnptr transmutes to be valid, which is not obvious, cf. https://github.com/rust-lang/unsafe-code-guidelines/issues/286, https://github.com/rust-lang/unsafe-code-guidelines/issues/340 > CHERI is...

It can be done more easily if we use not hotkey, but some IPC method for that (writing to file/fd?). Hotkey is also possible if we know how REPL prompt...

It's _not_ achieved by `.popen().iter_lines()`, because it caches all of process stdout and stderr in memory, see #480.

Are there plans on blessing some solutions as official-ish (while keeping them community-maintained)? Right now there’s a naming problem: for example, nixpkgs currently includes `typst-lsp`, which may be not ideal...

Yeah, sure. I pushed a test that shows that alloc calls are properly forwarded and heap stats counting still works. I’d also like to note that mintex 0.1.3 fixed [an...

I’ve tried playing with this a bit. My first idea was to make the `'py` lifetime invariant, so it may serve as a unique identifier of object’s provenance. Unfortunately, this...

@davidhewitt The idea is taken from the [GhostCell](https://plv.mpi-sws.org/rustbelt/ghostcell/paper.pdf) paper. Basically, the signature of `Python::with_gil()` has `F: for) -> R` in it. If the `'py` lifetime is invariant, then the following...

Yes, `Py` would need to have a runtime tag. I think `#[pyfunction]` is probably okay, since it would be generic over `'py`, which is invariant?