Open
Description
I've started a discussion on Zulip (https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/Cross.20crate-hygiene.20roots) - there are some unresolved questions about how we want the general solution ('correct' cross-crate decoding of and ) to work. However, it should be possible to make a hack that just fixes the problems for macros: when we're checking the span of a token for a foreign macro, fall back to the edition of that foreign crate if the is .
Originally posted by @Aaron1011 in #84429 (comment)