Leaf Petersen

> The only muddled point yet, as far as I can see, is that `C(super._x)` where `C` is declared in a different library than the class that declares `_x` uses...

Seems good to me to document this. I'm not sure what is intended by this part though: > It'd also be nice if something about the interplay between `abstract` and...

Closing, we have a design.

> So, if a conditional expression has a context type, let the static type of the conditional expression be that context type if we can't find a better LUB, defined...

Just noting that I expect this issue to become much more prominent with switch expressions. For example, the following simple definition of a map function over a lisp style linked...

This looks good to me (obviously this generalizes to switches). I believe that in the analogous place in function [literal return inference](https://github.com/dart-lang/language/blob/master/resources/type-system/inference.md#function-literal-return-type-inference) we use greatest closure. I think there are...

> Here is an issue that matches 'using bounds information to restrict choices during inference': #3567. That issue isn't about union types, but it's certainly about using bounds information. I...

> Any news guys? I'd say that this request is still open, and seems to have some level of support, but there are a lot of higher priority things occupying...

I was just reading over this section, this seems wrong to me. If I'm understanding this correctly, the result is that all nullable types are assignment compatible with all other...

Hey Patrice - I'm still way lost in the weeds on this section. Concretely to your change, I think something went wrong in the drafting. You're trying to define when...