Auto merge of #37129 - arielb1:erased-normal, r=eddyb

normalize types every time HR regions are erased

Associated type normalization is inhibited by higher-ranked regions.
Therefore, every time we erase them, we must re-normalize.

I was meaning to introduce this change some time ago, but we used
to erase regions in generic context, which broke this terribly (because
you can't always normalize in a generic context). That seems to be gone
now.

Ensure this by having a `erase_late_bound_regions_and_normalize`
function.

Fixes #37109 (the missing call was in mir::block).

r? @eddyb
This commit is contained in:
bors 2016-10-16 04:22:21 -07:00 committed by GitHub
commit 6572a46311
13 changed files with 38 additions and 38 deletions

View file

@ -583,7 +583,8 @@ impl_trans_normalize!('gcx,
ty::FnSig<'gcx>,
&'gcx ty::BareFnTy<'gcx>,
ty::ClosureSubsts<'gcx>,
ty::PolyTraitRef<'gcx>
ty::PolyTraitRef<'gcx>,
ty::ExistentialTraitRef<'gcx>
);
impl<'gcx> TransNormalize<'gcx> for LvalueTy<'gcx> {
@ -603,6 +604,18 @@ impl<'gcx> TransNormalize<'gcx> for LvalueTy<'gcx> {
// NOTE: Callable from trans only!
impl<'a, 'tcx> TyCtxt<'a, 'tcx, 'tcx> {
/// Currently, higher-ranked type bounds inhibit normalization. Therefore,
/// each time we erase them in translation, we need to normalize
/// the contents.
pub fn erase_late_bound_regions_and_normalize<T>(self, value: &ty::Binder<T>)
-> T
where T: TransNormalize<'tcx>
{
assert!(!value.needs_subst());
let value = self.erase_late_bound_regions(value);
self.normalize_associated_type(&value)
}
pub fn normalize_associated_type<T>(self, value: &T) -> T
where T: TransNormalize<'tcx>
{