Lasse R.H. Nielsen

This is more like #8 than dot-shorthands. It's omitting not just a target type, but a target object *and* a parameter taking that object. Sounds similar to Ruby's static tear-off...

That's a result of flow analysis being single-pass. The `do` loop checks first whether the variable is assigned inside the loop. If it is, then it cannot assume that the...

"Overwriting" the type of a variable at some prior point has no effect when the following code had already been analyzed using the old type. All errors have been reported,...

When the loop is encountered, the analysis knows the analysis state before the loop. To analyse the loop condition, it needs to know the state before the condition, which of...

> can the condition be looked at ahead of time in case particular state doesn't matter? In general, no. ```dart // Copies C-style zero-terminated string. List copyStringZ(List chars) { assert(chars.isNotEmpty...

The return type of a setter doesn't matter, because the type isn't used anywhere in the specification. Any returned value is lost, so it's type doesn't matter. If we allow...

> I was thinking `operator []=` was already a setter I think it is, just wanted to be certain that it was included since "setter" sometimes means "something declared using...

Has been asked for before, and rejected. Not saying not to try, you may have better arguments and ideas, but it's worth checking what have been tried so far (it's...

A quick look at the article doesn't show anything particularly wrong. The `JsonCodec` part mentions the `toEncodable` parameter, which I think was the most recent change, so it's probably up-to-date....

+1 An extra unnecessary `implements` is harmless. It's originally an error because it's _probably_ an author mistake to write `implements I, I` *so close that the author is expected to...