Lifting the "Self or associated type" constraint on existentials

I hope this pitch will come to a (positive) conclusion, so the case for removing "Self or associated type as existential" limitation could be moved forward to implementation and the proposal.

Would hate to see this idea die.

1 Like

Same here. In fact I'm not quite sure why protocols are treated special anyways. Starting out with Swift I would have expected them to just work fine with generic type syntax. (Actually I also don't understand why you have to do the typealias dance to make them appear encapsulated in another type, but that's another story…

2 Likes

I'm having trouble determining the state of this proposal today. This thread references another thread that references this thread, but there really hasn't been any discussion in several months. Did this ever make it to a formal SE?

See here: SE-0309: Unlock existential types for all protocols

2 Likes