For the latest Dart news, visit our new blog at https://medium.com/dartlang .
Posted by Bob Nystrom
Sorry for the long delay. This meeting was just skimming over a bunch of issues in the bug tracker, so forgive me if there isn't a lot of context here.
Import syntax
Kasper and Ivan did the original syntax design. If we want to iterate on that, we should get them in a room.
Catch syntax
Looks weird because it's more like a variable declaration than a parameter declaration. Gilad says we'll figure it out soon and should be a minor change.
Statics in interfaces
Kasper: People find default classes hard to follow. We could consider allowing simple redirect code for constructors in interfaces.
Instance field should be allowed in initializer expression when it is initialized via "this.foo" sugar.
Bug 3263. Gilad says: if the sugar doesn't do what you want, don't use the sugar.
Private names
Dart uses isolates for security, so we don't feel pressure to have JS-private-names-style security.
Abstract or static method named the same as enclosing class
Sorry for the long delay. This meeting was just skimming over a bunch of issues in the bug tracker, so forgive me if there isn't a lot of context here.
Import syntax
Kasper and Ivan did the original syntax design. If we want to iterate on that, we should get them in a room.
Catch syntax
Looks weird because it's more like a variable declaration than a parameter declaration. Gilad says we'll figure it out soon and should be a minor change.
Statics in interfaces
Kasper: People find default classes hard to follow. We could consider allowing simple redirect code for constructors in interfaces.
Instance field should be allowed in initializer expression when it is initialized via "this.foo" sugar.
Bug 3263. Gilad says: if the sugar doesn't do what you want, don't use the sugar.
Private names
Dart uses isolates for security, so we don't feel pressure to have JS-private-names-style security.
The default implementation of an interface must not over-specify generic paramaters
Bug 3279. We're OK with the current semantics.
Bug 3279. We're OK with the current semantics.
Abstract or static method named the same as enclosing class
Bug 3284
Kasper: The VM is slightly broken with respect to reporting. We should clarify the spec.
Lars: It is very confusing.
Gilad: That's why shadowing is a warning.
Kasper: Should say that if it's the same name, it must be a constructor. If it has a return type or abstract, that's an error.
Function expression vs statements
Can we get rid of return types and/or names?
Bob: Idiomatic dart doesn't have return types or names for function expressions. The style guide may discourage it soon. [ed: it does now.]
Kasper: Getting rid of return type for function expression gets rid of the nasty infinite lookahead with generic return types.
Gilad: What about contexts where you can't distinguish between expr or stmt?
Kasper: If it has a return type or a name, it's a statement. Done. It must be in statement position.
Gilad: Let me try to come up with some rules for it to see if it hangs together.
Kasper: The VM is slightly broken with respect to reporting. We should clarify the spec.
Lars: It is very confusing.
Gilad: That's why shadowing is a warning.
Kasper: Should say that if it's the same name, it must be a constructor. If it has a return type or abstract, that's an error.
Function expression vs statements
Can we get rid of return types and/or names?
Bob: Idiomatic dart doesn't have return types or names for function expressions. The style guide may discourage it soon. [ed: it does now.]
Kasper: Getting rid of return type for function expression gets rid of the nasty infinite lookahead with generic return types.
Gilad: What about contexts where you can't distinguish between expr or stmt?
Kasper: If it has a return type or a name, it's a statement. Done. It must be in statement position.
Gilad: Let me try to come up with some rules for it to see if it hangs together.