scottmcm

Jack as written up a new summary of the proposal to have it available in one place: https://hackmd.io/@jswrenn/SJmHwgdov

Thinks like "gee, this should have a way to do more than success/failure" are exactly the kind of design work that the type needs 🙂 For example, people have suggested...

I'm not a good reviewer for anything `Pin`-related, sorry. r? libs

Ok, with `UnOp::PtrMetadata` landed I picked this up again and added some consumption cases too. Hopefully this is about right... @rustbot ready

(Rebased to make sure there's no conflicts after #125893 )

> then we also need to address the top issue of making the FF64/FF32 constructors unsafe - unless we find a good reason not to. What do you think? With...

I don't know if it'd help at all, but cc https://github.com/bytecodealliance/wasmtime/pull/7719/files#diff-2041f67049d5ac3d8f62ea91d3cb45cdb8608d5f5cdab988731ae2addf90ef01 which will convert larger `mul`s back to smaller `mul`s and `mulhi`s. So I think if you can recognize the...

Since there's no 128-bit `iconst` it's `imul.i128 x (uextend i128 (iconst 2))`, but with https://github.com/bytecodealliance/wasmtime/pull/7693/files#diff-2041f67049d5ac3d8f62ea91d3cb45cdb8608d5f5cdab988731ae2addf90ef01R92 and https://github.com/bytecodealliance/wasmtime/pull/7670/files#diff-bf1f1ea991da38588a1785fc7012072e7a685411ea6a262c1fab18379411ef50R94 hopefully it actually already works 🤞

For extend we don't need to wait on `or` patterns, because it could use https://github.com/bytecodealliance/wasmtime/blob/f1fe2afffa9430b4a7765459a462cb32a4d81de8/cranelift/codegen/src/prelude_opt.isle#L113-L116 which is a special extractor added in #7710 for exactly the "there might be an...

If you're willing to use more nightly things, as this PR proposes, I'd recommend [`as_chunks(_mut)`](https://doc.rust-lang.org/nightly/std/primitive.slice.html#method.as_chunks) instead of any of the chunking iterators. Basically, rather than using `foo.chunks_exact(8)`, use `let (foo_chunks,...