Patterns destructing functions are colored the same as getters

[FMorschel]

Today the colouring for geters and methods is different. See the following code:

class Class {
  void Function() get getter => () {};
  void method() {}
}

void foo() {
  var instance = Class();
  instance.getter;
  instance.method;
}

This differenciates both in declaration and on usage even tough both of the above work the same*.

---

Today we have patterns that allow us to deconstruct a variable:

class Class {
  int myField = 0;
}

void foo() {
  var instance = Class();
  if (instance case Class(myField: var myFieldValue)) {}
}

We also can get** a method as a FunctionType:

class Class {
  int foo(int x) => x;
}

void bar() {
  var instance = Class();
  if (instance case Class(foo: var fooFn)) {
    fooFn(3);
  }
}

---

I'd like to request the same colouring to be applied here inside pattern destruction - at least when we have the name before : but I'd be fine with it also being applied to the new variable here if it has the same name.

---

*I've opened dart-lang/language#4159 related to this propriety.
**1. I've opened dart-lang/language#4234 related to this functionality.

**2. @lrhn has commented the following:

We also can get a method as a FunctionType:

Not absolutely sure that's intended, but will have to check the future specification to be sure whether it's allowed by that.

[FMorschel]

Here are some screenshots to help:

Here I've highlighted foo to see the different colouring:

I'd like some more consistency in method highlighting.

[DanTup]

Yeah, I feel like both foos here should be highlighted the same (in the yellow colour). I'd have to do some debugging to figure out why it's not - the colours for semantic tokens are mapped from the servers Highlight kinds. Methods are usually handled around here:

sdk/pkg/analysis_server/lib/src/computer/computer_highlights.dart

Line 413 in 1ac77f5

bool _addIdentifierRegion_method(

So my guess is that we're not going in here for the second foo identifier.

[bwilkerson]

I'd like to wait to see whether method tear-offs are allowed in patterns before we fix this issue. But I agree that if tear-offs are allowed it would be good to color them in such a way that it's more obvious when it's happening.

[lrhn]

The spec allows it.
(If the link target doesn't work, search for "normal property extraction rules".)

Can't see a good reason for allowing tear-offs, but it's there.