rustdoc: fix ICE when delegating functions with impl Trait as argument type#156498
rustdoc: fix ICE when delegating functions with impl Trait as argument type#156498cijiugechu wants to merge 1 commit into
impl Trait as argument type#156498Conversation
rustbot
added
S-waiting-on-review
|
rustbot has assigned @lolbinarycat. Use Why was this reviewer chosen?The reviewer was selected based on:
|
There was a problem hiding this comment.
This is needed because delegated signatures now go through tcx.fn_sig, where the inherited bounds may include internal sizedness bounds.
lolbinarycat
changed the title
impl Trait as argument type
| #![crate_name = "fn_delegation_impl_trait"] | ||
| #![feature(fn_delegation)] | ||
| #![allow(incomplete_features)] | ||
|
|
There was a problem hiding this comment.
I'd like to also see tests of what happens when delegating to a function defined in a separate crate, since rustdoc won't have the HIR in that case. This can be achieved via the aux-crate directive.
I think using build_function will handle this case correctly, but it would be nice to know for sure/
Also, what happens if we #[doc(inline)] a function that was delegated in a separate crate? What about delegating to a function that was delegated in a separate crate?
There's a lot of possible permutations but I think just adding a few more would make me a lot more confident there's no sneaky edge cases that break.
| @@ -0,0 +1,35 @@ | |||
| //! Regression test for <https://github.com/rust-lang/rust/issues/155728> | |||
| #![crate_name = "fn_delegation_impl_trait"] | |||
There was a problem hiding this comment.
| #![crate_name = "fn_delegation_impl_trait"] |
That shouldn't be necessary.
| @@ -0,0 +1,35 @@ | |||
| //! Regression test for <https://github.com/rust-lang/rust/issues/155728> | |||
There was a problem hiding this comment.
| //! Regression test for <https://github.com/rust-lang/rust/issues/155728> | |
| // Regression test for <https://github.com/rust-lang/rust/issues/155728> |
nitpick: Usually regular comments are used for test descriptions, not doc comments. Also it would be nice to have a little blurb describing what the test does without having to follow the link, like "Make sure delegating functions with impl Trait in argument position works works correctly."
rustbot
added
S-waiting-on-author
3 participants
HIR signatures for delegation items contain
InferDelegationplaceholders, clean the inherited fn_sig signature instead.Closes #155728