62b272d25c
move `considering_regions` to the infcx it seems weird to prove some obligations which constrain inference vars while ignoring regions in a context which considers regions. This is especially weird because even for a fulfillment context with ignored regions, we still added region outlives bounds when directly relating regions. tbh our handling of regions is still very weird, but at least this is a step in the right direction imo. r? rust-lang/types |
||
|---|---|---|
| .. | ||
| traits | ||
| autoderef.rs | ||
| infer.rs | ||
| lib.rs | ||