Adrian Seyboldt

> Not "including the world" is a project we've been talking about in the math library since I joined the project (and probably before). Ah, those projects are always great...

I guess if I had checked the libloading issues a bit more thoroughly I might have saved myself quite a bit of time... https://github.com/nagisa/rust_libloading/issues/41 https://github.com/nagisa/rust_libloading/issues/5

This is still an issue. The missing info: CPU: AMD Ryzen Threadripper 1920X OS: arch linux libm version: 2.39 Happens with and without SVML installed

This is still an issue.

I haven't tried with HEAD, but with the most recent build on the conda numba/dev channel: ``` Name Version Build Channel ────────────────────────────────────────────────────────────────────────── numba 0.61.0dev0 np2.0py3.11hb11b5cc_gfca98287a_137 numba/label/dev ``` I guess this...

Some notes on the implementation: - For some reason I get slightly different final llvm modules in the code with and without `no_wrapper`, and the performance is a bit better...

I realized there doesn't really seem a good reason for a `no_wrapper` argument, we can just just the existing `no_cpython_wrapper`...

I was running it in jupyter, and it seems that might be setting a higher recursion limit? I get the same error if I run it with plain python. (both...

Sorry everyone, I had a bit too much fun making a few animations about the linking process :-) Each square represents a llvm ir module of a function, and each...

Sounds good. :-) Feel free to ping me if there is something I can help with.