Benjamin Saunders

Retrying immediately makes sense to me. I think the only reason we haven't done this is because it's fiddlier than just dropping a flag.

That would be great, thanks! Note 1.91 isn't actually out for another few days, so it won't work in stable rust *quite* yet and there may be an MSRV concern...

I expect it's probably beneficial anywhere that it can be used without unreasonable contortions; `insert` was just an example.

I feel like a panic when performing an operation that requires elements on an empty matrix would be less prone to unpleasant surprises. Note that `Iterator::min` returns `Option`.

I think our current interface is safer and more readable than the C++-style one. `mut` on the extension structs is fundamentally reasonable because we do mutate their `p_next`, and the...

> the point is to get rid of mut in this case - but I think our only possibility to do so is by building the p_next chain individually /...

> mut looked weird given the nature of Info being read-only. It seems normal to me to have something that's mutable while you set it up and then passed elsewhere...

> Wdyt about the proposal? 100% on board with incremental. > You mentioned two points against traitifying, that equally apply to https://github.com/ash-rs/ash/pull/953? Yep. Will discuss further in #879.

> There is a workaround however, but it doesn't seem ideal: I don't think this is worse than using a type that has no lifetimes at all. Ideally, of course,...

> Additionally, we could make an ffi.rs module with contents like the following That seems reasonable to me, if there's no more direct way to tell bindgen to translate the...