Is `Any` unsound?

Soni · 2018-11-08T14:08:24.756Z

Continuing the discussion from Pre-RFC: sum-enums:

In cases when ordering of type arguments is required it can be done based on TypeId of each type. Alternatively for tag we could use TypeId directly

Doesn’t TypeId have a non-uniqueness bug? I mean, I get that it’s a hash, but last time I read about it, IIRC some of the compiler/runtime developers warned that it has a way higher than negligible/acceptable chance of collision for realistically complex types. If this is the case, using it as a discriminant or an ordering key can cause unsoundness and memory unsafety bugs, because it would result in accidentally interpreting values as the wrong type.

Does this mean Any is unsound?

withoutboats · 2018-11-08T14:32:11.756Z

You can usually find known soundness bugs on the issue tracker

github.com/rust-lang/rust

Issue: Collisions in type_id

opened by DaGenix on 2013-11-09

The implementation of type_id from #10182 uses SipHash on various parameters depending on the type. The output size of SipHash is...

C-bug I-unsound 💥 P-low T-lang

system · 2019-03-25T08:31:07.228Z

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.