Implement return position impl trait / opaque type support #4689
Conversation
Makes sense to me. More generally, after we do ItemTree work, it might makes sense to only store resolved things in |
| TypeCtor::OpaqueType(opaque_ty_id) => { | ||
| let bounds = match opaque_ty_id { | ||
| crate::OpaqueTyId::ReturnTypeImplTrait(func, idx) => { | ||
| let datas = |
There was a problem hiding this comment.
I wonder if we should switch to data/datum terminlology...
crates/ra_hir_ty/src/lower.rs
Outdated
| let func = match ctx.resolver.generic_def() { | ||
| Some(GenericDefId::FunctionId(f)) => f, | ||
| _ => { | ||
| // this shouldn't happen |
There was a problem hiding this comment.
This needs an assert or an explanation why it isn't an assert
There was a problem hiding this comment.
Ok, added a panic instead. I'm sometimes not sure whether to assert or just ignore things like this in cases where there's an easy way to bail out without panicking.
There was a problem hiding this comment.
Good question...
I think for compilery things, it's better to panic, but for IDE things, it's probably better to try to soldier on.
I guess, at least debug_assert always makes sense.
This is working, but I'm not that happy with how the lowering works. We might need an additional representation between `TypeRef` and `Ty` where names are resolved and `impl Trait` bounds are separated out, but things like inference variables don't exist and `impl Trait` is always represented the same way. Also note that this doesn't implement correct handling of RPIT *inside* the function (which involves turning the `impl Trait`s into variables and creating obligations for them). That intermediate representation might help there as well.
flodiebold
force-pushed
the
chalk-impl-trait-changes
branch
from
b2339e3
to
0d2328f
Compare
|
@matklad, I addressed your comments. |
|
bors r+ |
🎉 |
This is working, but I'm not that happy with how the lowering works. We might need an additional representation between
TypeRefandTywhere names are resolved andimpl Traitbounds are separated out, but things like inference variables don't exist andimpl Traitis always represented the same way.Also note that this doesn't implement correct handling of RPIT inside the function (which involves turning the
impl Traits into variables and creating obligations for them). That intermediate representation might help there as well.