Matt Keeter

I'll look into it! It seems like Zig doesn't support asan easily, but it's definitely possible that my own code is breaking things 😆

For what it's worth, I'm using Inkwell on an ARM Macbook (M1 Max) + LLVM 14 without any issues, so I don't think it's fundamentally broken! (Inkwell commit 40d7ba0d387819140ca85c9bbf14ccdbd199ceee, according...

Ah, I see what you mean! I'm seeing the same error that you do when running the test suite (on 40d7ba0d387819140ca85c9bbf14ccdbd199ceee) ``` cargo test --all --target=aarch64-apple-darwin --features=llvm14-0 ``` ``` LLVM...

I'm running into a similar issue: I'm compiling **many** tiny functions, and the overhead of allocation and `mprotect` is a big deal. The idea of implementing a custom `DynasmApi` is...

Good catch, the screenshot was using the stock `Assembler`; my comment was more of a +1 to the notion of reusing structures in general. I've now switched to a `VecAssembler`...

Cool, I'll take a look in the next day or two and post some benchmarks. In case you're curious, my hilarious failures turned out to be `dcache` and `icache` getting...

The changes on `dev` make a meaningful difference! Here's what I'm seeing on one test case, logging time spent in the JIT: - Version 1.2.3: 19.2 ms - `dev`: 13.7...

This is just swapping `dynasm = 1.2.3` to `dynasm = { git = ..., branch = "dev" }` in `Cargo.toml`, tested on an older branch that was using the stock...

Yup, that looks like self-intersection, which is a known issues with the whole Dual Contouring family of algorithms. At this point, I'd expect meshing to be - Manifold - Watertight...

> 1. Is anyone working on a Rust binding of this already? Not that I know of! > 2. Would you start from the C API or the C++ API...