Just came across a curious situation. Suppose `A` ...

# getting-started

[Mark]

Just came across a curious situation. Suppose A extends B and we have two extension funs declared in different packages:

fun A.foo()
fun B.foo()

How can we call both of these from the same class/file which is in the same package as A? Because as soon as we include the import for B.foo(), it will be used for both calls.

[Ving]

What about import aliases ?

[Mark]

Nice - thanks!

[Mark]

I feel there should be some kind of lint warning.

[Ving]

Yup, it's a bit misleading.

[Mark]

Took me a while to realise how it was possible that an instance with type A was resolving to use B.foo() instead of A.foo()

[Mark]

I guess at some point, I added a call to B.foo() which added the import, which then meant that a pre-existing a.foo() switched from resolving A.foo() to B.foo()

[phldavies]

you can also (a as B).foo() to force B.foo() to be resolved

[Mark]

Hmm, I thought I tried that

[Mark]

I mean, the type was already A and so the IDE showed the as A as redundant. And anyway, it clicked through to B.foo()

[phldavies]

Ah sorry, missed the multiple packages part...

[phldavies]

You'd need to import the foo from the same package additionally (explicit imports have a higher resolution priority)

[Mark]

It was solved by just doing: import packageB.foo as bfoo instead of import packageB.foo . No need to import current package A.foo . Although I suppose it might be good for clarity to import both.

[phldavies]

Unfortunately I think IntelliJ will consider the import redundant and try to remove it 😄

[Mark]

Oh, and found a bug - Optimize Imports… (in AS) doesn’t sort import aliases

[Mark]

Actually, Android Studio allowed me to add it fine, presumably because import alias

[Mark]

Hmm, perhaps not a bug - it just puts them to the end

[phldavies]

It doesn't remove on optimize but it is highlighted as unused 🤔

[Mark]

Need to be an import alias

[phldavies]

alias is probably clearer though and avoids needing the explicit cast