1
Fork 0

Auto merge of #100980 - compiler-errors:normalize-opaque-w-bound-vars, r=lcnr

Normalize opaques w/ bound vars

First, we reenable normalization of opaque types with escaping late bound regions to fix rust-lang/miri#2433. This essentially reverts #89285.

Second, we mitigate the perf regression found in #88862 by simplifying the way that we relate (sub and eq) GeneratorWitness types.

This relies on the fact that we construct these GeneratorWitness types somewhat particularly (with all free regions found in the witness types replaced with late bound regions) -- but those bound regions really should be treated as existential regions, not universal ones. Those two facts leads me to believe that we do not need to use the full `higher_ranked_sub` machinery to relate two generator witnesses. I'm pretty confident that this is correct, but I'm glad to discuss this further.
This commit is contained in:
bors 2022-09-22 12:47:31 +00:00
commit 8ab71ab59f
7 changed files with 43 additions and 24 deletions

View file

@ -218,7 +218,8 @@ pub fn resolve_interior<'a, 'tcx>(
.filter_map(|mut cause| {
// Erase regions and canonicalize late-bound regions to deduplicate as many types as we
// can.
let erased = fcx.tcx.erase_regions(cause.ty);
let ty = fcx.normalize_associated_types_in(cause.span, cause.ty);
let erased = fcx.tcx.erase_regions(ty);
if captured_tys.insert(erased) {
// Replace all regions inside the generator interior with late bound regions.
// Note that each region slot in the types gets a new fresh late bound region,
@ -263,7 +264,7 @@ pub fn resolve_interior<'a, 'tcx>(
// Unify the type variable inside the generator with the new witness
match fcx.at(&fcx.misc(body.value.span), fcx.param_env).eq(interior, witness) {
Ok(ok) => fcx.register_infer_ok_obligations(ok),
_ => bug!(),
_ => bug!("failed to relate {interior} and {witness}"),
}
}