V. Jackson

The `Pod` trait from [bytemuck](https://crates.io/crates/bytemuck) crate does this. It's quite handy for reinterpreting data for buffers, so it would be great if `cgmath` could support it.

Yes, cgmath now uses bytemuck optionally. Enable the feature "bytemuck" to use it. This looks to have been added in #541.

Can confirm this is still a problem.  I encountered this problem when I was approximating the derivative of each cell by...

Ah. After looking at it a bit more, it seems the answer is (if you want 2d/3d noise) use [SuperSimplex](https://docs.rs/noise/0.7.0/noise/struct.SuperSimplex.html), which has been implemented in noise-rs now (#168).

There used to be some code to generate the default Dargent layout, which was supposed to be, in principle, C-compatible. I removed it because it was exceptionally annoying to allocate...

@zilinc I've replicated the failure and worked out what's causing it. Here's an even simpler program which reproduces the error. ``` m: () -> () m b = b |...

This should be somewhat mitigated by Amos and my changes to type inference in Sink/Float. The only place this should occur now is when you have variables under bangs on...

Here's a code reference: https://github.com/NICTA/cogent/blob/0528a3da61f5a89ae78a64fe6dad36c5d788af0d/cogent/src/Cogent/TypeCheck/Base.hs#L605

This appears to be a duplicate of #1445 Note there is a wip pull request to address this issue #1492