What's the policy for bugfixes on K1 and K2? Like,...

# javascript

[Edoardo Luppi]

What's the policy for bugfixes on K1 and K2?
Like, if a problem is found on both of them, how likely is it that you're going to fix it in K1, and not only in K2?
I suppose for a dev it's double the work, so a decision must be taken

[Artem Kobzar]

Right now we will fix it in both of them. Could you share the bug?

[Edoardo Luppi]

Hey! Oh no bugs, it was just a general question to see if I have to start thinking in terms of K2 (enable the preview and try to workaround code that's not supported yet), or not

[Edoardo Luppi]

I now have more or less 30k LOC written for KJS, so I want to be proactive

[Artem Kobzar]

So, in general. If the problem appeared on the codegen side - it will be not a double work, because after frontend we translate the frontend representation into middle-end representation and they are similar for both K1 and K2.
But if the problem appeared at the frontend side (checkers, parsing, type inference, etc) in this case it will be a double work

[Artem Kobzar]

Thank you for your activity. Such kind of activity really helps us to improve the tool. Also, if you can, it will be really interesting to know more about your use-case of KJS.

[Edoardo Luppi]

Thanks!
My main issues with K2 are related to expect/actual typealiases (e.g., in K1 you can suppress the error if the generic type variance is different between platforms, but not in K2 yet, the issue is moved to language design).

I'll send you a PM here tomorrow for our use case, if it's not a problem

[Edoardo Luppi]

Sent you a PM, feel free to give it a look anytime