Hello, CI on rust-lang/rust#138961 has picked up a rust-analyzer test that fails with the rustc changes applied – specifically ones that assert that r-a gets the same size of closures as rustc, see this comment for details.
Should we comment these tests out for now? What's your take on this?