Isaac Garzón

@kupiakos, that's a very interesting idea! I might be missing something obvious, but I think your code can be simplified to this: ```rust use core::mem::{ManuallyDrop, MaybeUninit, size_of}; use core::ptr::addr_of; const...

Would getting the pointer after the assignment make it valid? i.e.: ```rust // Set 8 byte chunks sequentially until we bump into the field pointer let mut buf_offset = 0;...

@vespakoen @phkahler ping.

> [@phkahler](https://github.com/phkahler) shall we start building/releasing x64 version(s) on Windows in addition to the regular 32 bit x86? Given the limitations of the 32-bit version, and the fact that it...

Fair. Then don't drop the 32-bit version. Still, on 64-bit hardware the 64-bit version is able to take advantage of the available resources, so building it would definitely be useful...

I have a branch where I also added code to avoid creating the zero length line segment, but that's not what this PR is about. Here's a video demonstrating the...

@ruevs @phkahler ping. Unlike the issue that #1592 is fixing, which only creates redundant segments, this PR fixes a bug in the splitting logic that introduces a usability annoyance, as...

> ...so is the conclusion simply "too hard, fuggedaboutit in any kind of future that could be considered near"...? Sigh. I'm not a maintainer, and I sympathise with your frustration,...

Adding some debugging prints, this is what I get when loading the model: ``` 96 params 98020 entities 12 groups g001: 0 entities g002: 0 entities g003: 1390 entities g004:...

Can you describe the exact steps that you're taking so I could try to reproduce it and diagnose what's happening there? (I assume it's still the same model, and you're...