Matthew Lutze

Results 310 comments of Matthew Lutze

That's weird, it's part of the standard library

Can't reproduce. Anyway, this is part of a known class of errors where we cannot handle error types. Should be resolved with the new solver.

- Introduce new kind `Region` - Introduce `RegionUnification` -- presumably just identity - Copy or generalize the treatment of associated effects in EffUnificationX to work for heaps. They're mostly treated...

> Meanwhile, we can start refactoring with: > > ``` > type alias Heap[r: Eff] = r > ``` I bet this will shine a light on many bugs

What is the motivation here? This seems to be increasing the complexity of using what I see as a fundamental effect.

> > What is the motivation here? > > This seems to be increasing the complexity of using what I see as a fundamental effect. > > We use RichString...

This code no longer exists. However, the loop replacing it is still a little smelly wrt the documentation.

Yeah, in principle we should handle them properly.

Default for tests or default for everything?