Oli Scherer

> some changes are accompanied with a blog post explaining the rationale for the change with instructions to users. This part was done: https://blog.rust-lang.org/2025/03/02/Rustup-1.28.0.html

These attributes affect how the serialization code is generated, but isn't available at runtime. I don't think it makes sense to single out the attributes here, as any manual impl...

> as there is no way to know that a field is missing or not Right, but with a schema like thing made available to the serialized and deserializer this...

Ah I didn't realize the Schema was unnecessary. Yea in that case the question is how far you want to take the protection. Users can already write arbitrary serialization impls...

We don't have the capacity for reviewing an entire crate for feasability. If you want to make a case, I recommend: 1. creating benchmarks, ideally by swapping out the quote...

Unfortunately I don't use changelogs myself, so my intrinsic motivation for maintaining it is not very high. I do understand the benefit, but I'd need folk to maintain it who...

I rebased the PR, reordered some commits and squashed the clippy/rustdoc/rustfmt commits into the main commit. Each commit now builds on its own and does the minimal work necessary to...

@bors delegate+ r=me,petrochenkov once CI is happy

Miri doesn't support the linker hacks that linkme is using. Few platforms do. We'd either have to have some sort of impl for all the platforms linkme supports or we'd...

Maybe the shim we use to read permission is unfinished and doesn't work at all even though it could work on Unix?