bjoernager/rust
bjoernager/rust
1
Fork 0
Code Issues Pull requests Activity

Rollup merge of #104411 - lcnr:bivariance-nll, r=compiler-errors

Browse source 
nll: correctly deal with bivariance

fixes #104409

when in a bivariant context, relating stuff should always trivially succeed. Also changes the mir validator to correctly deal with higher ranked regions.

r? types cc ``@RalfJung``
This commit is contained in:
Dylan DPC 2022-11-19 11:54:44 +05:30 • committed by GitHub
commit 00876c68c4
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
10 changed files with 130 additions and 85 deletions
Download patch file Download diff file Expand all files Collapse all files

6
compiler/rustc_const_eval/src/interpret/eval_context.rs
View file

@ -23,7 +23,7 @@ use super::{
MemPlaceMeta, Memory, MemoryKind, Operand, Place, PlaceTy, PointerArithmetic, Provenance,
Scalar, StackPopJump,
};
use crate::transform::validate::equal_up_to_regions;
use crate::util;
pub struct InterpCx<'mir, 'tcx, M: Machine<'mir, 'tcx>> {
/// Stores the `Machine` instance.
@ -354,8 +354,8 @@ pub(super) fn mir_assign_valid_types<'tcx>(
// Type-changing assignments can happen when subtyping is used. While
// all normal lifetimes are erased, higher-ranked types with their
// late-bound lifetimes are still around and can lead to type
// differences. So we compare ignoring lifetimes.
if equal_up_to_regions(tcx, param_env, src.ty, dest.ty) {
// differences.
if util::is_subtype(tcx, param_env, src.ty, dest.ty) {
// Make sure the layout is equal, too -- just to be safe. Miri really
// needs layout equality. For performance reason we skip this check when
// the types are equal. Equal types *can* have different layouts when