is this an evaluation bug, or do I miss something?

# announcements

[Mikael Alfredsson]

is this an evaluation bug, or do I miss something?

[Mikael Alfredsson]

I would expect the first line to resolve to false since all variables are -1 and the last evaluation tells me that it actually should be false

[spand]

Seems odd. This isnt javascript where runningEntryId is actually a string or something?

[Mikael Alfredsson]

kotlin and they are Long.

[spand]

Yeah but is it kotlin/js ?

[Mikael Alfredsson]

sorry, no, its for Android

[Kiryushin Andrey]

Looks like it?.id ?: -1 expression resolves to the value of Int type when used as a part of a bigger expression

[Mikael Alfredsson]

quickly verified.. (it?.id ?: -1L) works.
this must be a compiler bug since if (1L== 1) will return a compiler error.

[spand]

You can always submit it but there are limits to how far the compiler can go. I think what you have is basically: 1L == (-1 as Number) which compiles also.

[Kiryushin Andrey]

I think the question is more about the Elvis operator behavior, it is really misleading here. It would make sense to
• either always infer Any type for the expression like it?.id ?: -1 in case left-hand part is inferred to be other than Int - this would make this pitfall more visible (and btw if you replace -1 with a reference to variable of Int type, the actual type inferred for the expression would really be Any)
• or emit the run-time conversion of the right-hand numeric constant to Long into the generated bytecode.