ICE when building compiler plugin with nll

[real-felix]

I tried the example of compiler plugin and added only the nll feature.

The result is an ICE:

error: internal compiler error: broken MIR in NodeId(186) (Terminator { source_info: SourceInfo { span: src/lib.rs:60:5: 60:40, scope: scope[0] }, kind: _2 = const rustc_plugin::Registry::register_macro(move _3, move _4, move _6) -> [return: bb2, unwind: bb1] }): bad arg #2 (for<'cx, 'r, 's> fn(&'cx mut syntax::ext::base::ExtCtxt<'r>, syntax::codemap::Span, &'s [syntax::tokenstream::TokenTree]) -> std::boxed::Box<syntax::ext::base::MacResult + 'cx> <- for<'r, 's, 't0> fn(&'r mut syntax::ext::base::ExtCtxt<'s>, syntax::codemap::Span, &'t0 [syntax::tokenstream::TokenTree]) -> std::boxed::Box<syntax::ext::base::MacResult>): RegionsOverlyPolymorphic(BrNamed(crate11:DefIndex(1:1022), 'cx), '_#17r)
  --> src/lib.rs:60:30
   |
60 |     reg.register_macro("rn", expand_rn);
   |                              ^^^^^^^^^

error: internal compiler error: broken MIR in NodeId(186) (Terminator { source_info: SourceInfo { span: src/lib.rs:60:5: 60:40, scope: scope[0] }, kind: _2 = const rustc_plugin::Registry::register_macro(move _3, move _4, move _6) -> [return: bb2, unwind: bb1] }): bad arg #2 (for<'cx, 'r, 's> fn(&'cx mut syntax::ext::base::ExtCtxt<'r>, syntax::codemap::Span, &'s [syntax::tokenstream::TokenTree]) -> std::boxed::Box<syntax::ext::base::MacResult + 'cx> <- for<'r, 's, 't0> fn(&'r mut syntax::ext::base::ExtCtxt<'s>, syntax::codemap::Span, &'t0 [syntax::tokenstream::TokenTree]) -> std::boxed::Box<syntax::ext::base::MacResult>): RegionsOverlyPolymorphic(BrNamed(crate11:DefIndex(1:1022), 'cx), '_#17r)
  --> src/lib.rs:60:30
   |
60 |     reg.register_macro("rn", expand_rn);
   |                              ^^^^^^^^^

[pietroalbini]

cc @rust-lang/wg-compiler-nll

[matthewjasper]

Ran into this when trying to bootstrap with NLL. Small repro case:

#![feature(nll)]

fn foo<'r>(_: &'r ()) -> &'static () { &() }

fn main() {
    let f: for<'r> fn(&'r ()) -> &'r () = foo;
}

[nikomatsakis]

This might, like #48071, be due to #33684

[nikomatsakis]

I think we have to fix this -- it's blocking bootstrap and it's a serious ICE. But I'm not sure what's the best fix. I have an idea but it's a bit aggressive -- still, I think I might give it a spin. The basic idea is to rerite the subtyping code to use universes, which should solve #48071 and help pave the way for chalkification. I don't think I have the bandwidth to mentor this, though, and will instead just try and do it. Assigning myself.

[nikomatsakis]

In-progress branch nll-issue-48071-universe-and-sub -- basically works but I have to re-enable user-given type annotations and also I'd like to tweak the performance.

[pnkfelix]

visited for triage. this should be fixed via PR #52488 (which is WIP but hopefully will be ready soon).

[nikomatsakis]

OK, #52488 has landed. This should be fixed.

[Mark-Simulacrum]

@rust-lang/wg-compiler-nll Can someone take a look and confirm this is fixed?

[nikomatsakis]

I can confirm this is fixed.