Leaf Petersen

@stereotype441 Thoughts on this? I think one possible way to enable this would be to add "intermediate" types to the promotion stack when promoting. That is, when promoting from `T?`...

I guess the question of whether the declaration itself is an error is only relevant for the variant in which we use a separate declaration form for static extensions, i.e.:...

This is not planned.

@scheglov points out that allowing these causes obvious termination problems with the subtyping algorithm. Specifically, given (for example) `T extends FutureOr` and `Future`, an attempt to prove `FutureOr

> About phrasing, I always gets worried when we "erase a union" and choose one of the sides as the canonical one. The other side is equally part of the...

> We want to move away from owning all these `is*Error` matchers, so I think I'll avoid adding a new one for `isTypeError`. That's fine, I just observed the presence...

SGTM. I landed the CastError deprecation today, I'd guess it will make it into 2.8 but not guaranteed.

@jakemac53 raised the issue of stack traces. We need some way of uniquely reporting locations in files. It feels very unfortunate if tools end up surfacing stack traces (and other...

The language leaves it deliberately implementation defined what constants are defined for use in conditional imports, IIRC. The limiting factor is largely around the ability of the build systems to...

@lrhn > ```dart > static extension NumList on List { > ``` I don't know what the `T extends num` is doing above, do you mean to have that in...