Yoan

I've needed something similar in the current project I'm working on; Basically I wrote a trait which provides `as_slice()` but asserts that `width == stride`. This is super useful for...

There's indeed no need for _ptr, but since it's trivial we might as well do it; this would imitate the "look'n feel" of e.g Vec and friends. I'm also not...

I agree with point 2. As for point 1, I don't see what's wrong; The function uses "copy_nonoverlapping" so it should start with "copy_", because that's what it does. Other...

OK, what about : - `copy_rows_from(src, dst_left, dst_top)` (notice how X goes before Y), for T: Copy. - `clone_rows_from(...)` for T: Clone. I've added "rows" in-between for these reasons :...

So how about `copy_sub_image`/`clone_sub_image` ? I kinda liked `rows` because that's what the implementation is necessarily supposed to do; In the existing API there's `rows()` but not `cols()`, and rightfully...

There was a discussion on this in #12 , you may want to search for `The only function I've been able to make const was new() for vectors` in that...

I've given it a shot today and unfortunately `broadcast` was the only one I could make `const`. As you say, it's obvious that casts/conversions/constructors should and could be `const` when...

Thanks for reporting this. I could probably write an essay about how much I hate this change, but what really matters are the actions I take. My knee jerk reaction...

There are unfortunately a lot of counterarguments, and because I know them, I can't go out there and stand my ground in good faith: - The change was approved more...

My workaround wouldn't allow for construction by members or destructuration : ```rust let v = simd::Vec4 { x: ... ... }; let simd::Vec4 { x, y, ... } = v;...