impl SuperTrait for T: SubTrait does not work about chalk

[mikeyhew]

It looks like this is an instance of #12

[scalexm]

@mikeyhew Yes, this is somehow a known bug, because of implied bounds not being completely implemented at the moment.

[mikeyhew]

@scalexm what do you mean by "implied bounds"?

[scalexm]

I mean things like "if T implements Sized, then T necessarily implements DynSized". For these kinds of affirmation to be correct, we must add some requirements on impls (called well-formedness requirements). Currently, these requirements are too restrictive, hence your code is not being accepted. That shall be fixed in the future.

[mikeyhew]

@scalexm the above code is accepted by chalk. I guess it just ignores it though?

[mikeyhew]

@scalexm I tried out your implied-bounds branch, but didn't notice any changes with the above code :/

[scalexm]

@mikeyhew Yes by "not accepted" I meant "gives a wrong answer", sorry :)
That's indeed because the remaining work is to be done in my implied-bounds branch (and that's also why the PR's still open).

[mikeyhew]

@scalexm Oh, I see... I saw that @nikomatsakis was reviewing it, and assumed it was otherwise good to go 😄

[mikeyhew]

I'm gonna close this, since it is a duplicate of #12

[scalexm]

@mikeyhew PS: it works now

[mikeyhew]

@scalexm Sweet!

trait DynSized {}
trait Sized where Self: DynSized {}
impl<T> DynSized for T where T: Sized {}
struct i32 {}
impl Sized for i32 {}

// ?- i32: DynSized
// Unique; substitution [], lifetime constraints []