What !? :smile:

# language-proposals

[addamsson]

What !? 😄

[Ruckus]

😆

[Ruckus]

I only mean it half seriously, as it's way too late to change now, but I wish it would have been that way from the start.

[addamsson]

Yep, too bad 😞

[adam-mcneilly]

If it's too late to change, what about adding a compiler warning if you don't use a safe operator on a platform type? 🤔

I think you could start a civil war with that one, though. Some would be grateful, someone would probably be very annoyed.

[karelpeeters]

The language would be completely useless if you weren't allowed to do that, a warning would be okay.

[Ruckus]

@adam-mcneilly Maybe add an inspection for it, but don't have it set to warning by default, but then people could go change it to a warning (or error) if they want.

[adam-mcneilly]

+1

[adam-mcneilly]

I like that because it allows us to introduce a capability a lot of people would benefit from (I think?) but without forcing any new behavior on people that didn't care previously.

[littlelightcz]

I think there is already a "weak warning" for this in IDEA, which asks you to specify the type explicitly.

[karelpeeters]

That's if you leak platform types though.

[benleggiero]

I am 1000% for a compiler warning/error for not acknowledging anything possibly-null, platform or not

[dmitry.petrov]

Platform types are non-denotable so far, it's never too late to change their rendering in IDEA

[Ruckus]

🤔

[benleggiero]

Wait, that's a thing I can do, @dmitry.petrov

[dmitry.petrov]

That would be quite a big change, actually (because such types are frequently mentioned in the test data). Also, at least discuss it with someone else from the Kotlin team, at least to get a second opinion. I'm kinda indifferent to it, but, well, it might be more important than I think.

[Ruckus]

I'm with @dmitry.petrov on this one. I don't really think it's worth changing at this point.