Matthew Lutze
Matthew Lutze
Figure out how sub-effecting works with typematch
I think it's stale wrt the new new typer stuff.
> I am not sure it is worth dealing with restrictable variants... I would only do it if it is easier. The main complexity is in the back end. I...
> > > I am not sure it is worth dealing with restrictable variants... I would only do it if it is easier. > > > > > > The...
> @magnus-madsen @mlutze The type of `None` is still `Option[t]` right (and not `Unit -> Option[t]`)? Right
> Oopsies... that's me. Thought it was clearer than `decls.map(visitDecl)` I guess it's a subjective design question, but its 2 vs 1 at the moment :P
> True but then we also have to refactor stratifier for consistency :)) I can live with that burden.
> The diff is very large, it is just because of a missing merge/rebase? I feel like I still dont understand git. Yeah, the git ghost is about. I'll fix...
> Can you write me a short description of how to best review this? No, but I can write a long one. ### Before this PR Unification of types is...
> In the final version, how will Monomorpher conduct unification? > > Will it use constraints? One constraint? Or will it just call deeper into the solver (i.e. some specific...