Andreas Borgen Longva

Agree that breaking changes to serialization formats should not happen in patch releases, unfortunate that this slipped through. I can't personally speak to the necessity of the change. All that...

Actually, one small counterpoint to the point about squared distances being unanimous: Some parameters are actually not squared, such as `dhat` and `dmin`, I believe.

Thanks, matrix concatenation/stacking is sorely lacking in `nalgebra`. There's already an ongoing PR for concatenating matrices (#1080), but I think unfortunately that the author of the PR @birktj seems to...

This isn't articulated anywhere, but I have been thinking about stacking/concatenation lately, and I was anyway kind of starting to prefer `stack!` over `bcat!` (which was never a good name...

A small note to my compile-time measurements: Looking at the results again, it's clear that there's quite a bit of parallel compilation of small crates going on, which is why...

@aweinstock314: thanks for clarifying, indeed I hadn't realized that `vstack` and `hstack` are functions, not macros. I'm impressed by what you're proposing here. At the same time, I feel that...

I just completed a new round of reviews for #1080. My personal position is still that we should only have one way of stacking/concatenating matrices in `nalgebra`, and I think...

Huh, I remember noticing `resize_mut` showing up in my profiling at one point. This puzzled me, as I expected it to be mostly a no-op for my case. Would be...

To fix this, maybe we need a new `ResizableAllocator` trait? The cleanest way I can think of immediately, but there may be better ways.

> I'll have to check if it actually works, but I was going to try a trick using `std::mem::swap` and writing things in terms of the non-mut versions. I think...