Georgi Krastev
Georgi Krastev
> It's just conceivable that byname implicits in 2.13.x will make it possible to make some progress, but for now I don't think there's a lot that can be done....
This could be mitigated if case classes use constructors and field access instead of the companion's `apply` / `unapply` but it's tricky to do. I remember attempting that at some...
What would be a good template for the error messages? I'm thinking something like this: > Implicit not found: `TypeClass[A, B, C]`. > Could not prove / compute / infer...
That's an issue in the Scala compiler
One could imagine having a `toLabelledProduct` where you have to pass the labels explicitly. But I'm not sure how useful that would be.
It could also work as a macro when passed a reference to a method or a literal function.
`Lazy` can't know what is `F` - it's probably `Nothing`
I meant that the problem is `F
Yeah the reason is that `Lazy` is using a macro and inside we programatically ask the compiler to infer an implicit with that type, but that doesn't correspond exactly to...
We had to revert the fix, because it breaks kittens