Nick Fitzgerald

Ah cool! I hadn't seen that PR. Would you mind rubber stamping https://github.com/WebAssembly/proposals/pull/194 in the meantime?

Hi, just got back from vacation, thanks for your patience. I am planning a new major version release and one of the features I intend to include is being generic...

(This would be gated on a cargo feature until Rust's `Allocator` trait is stabilized, since it would be nightly only)

Probably a month or two at least. Really depends on how many free cycles I'm able to scrap together.

> EDIT: since there is a blanket impl `Allocator` for all `&T` where `T: Allocator` Yep. Although maybe removing the explicit import for `&Bump` and allowing the blanket impl to...

I think it would be fine to have a couple variations of the read/write methods (one variation being async) that we can `cfg`-enable on/off as the appropriate cargo features are...

They should all use the same underlying helpers as much as possible.

SGTM, modulo one thing: > convert the primary loops in signed / unsigned to macros with a) different bound variations, b) async / no async keyword (or maybe just returning...

Something like that, yes, although I suspect we might get better codegen if we passed all the struct's fields as arguments to each helper function (they won't be `pub` anyways...

But maybe with inlining and SROA that won't actually matter. So I guess just do the struct thing first and then we can profile and tweak things as necessary later.