duplicates:

• SR-7695 WritableKeyPath mutation with multiple didSets does not propagate change

relates to:

• SR-5442 Calling subscript setter with a WritableKeyPath results in wrong invocation order of embedded setters

Issue Description:

Example code courtesy of Vermeulen (JIRA User):

struct Foo {
    var value: Int { didSet {} }
}

struct Bar {
    var foo: Foo { didSet {} }
}

var bar = Bar(foo: Foo(value: 1))

bar.foo[keyPath: \Foo.value] = 2
print(bar.foo.value) // 1

Value is changed to 2, but reading the value reflects '1'.

Removing the didSet causes the printout to be the correct value, 2.

Similar to SR-5442, this also causes the didSet to evaluate its block out of order.