Channels
#stdlib
- d
dmitry.petrov
04/18/2016, 2:23 PMYes. But it's in
for a reason. As well as some other "magic".kotlin.internal - k
kmruiz
04/18/2016, 2:23 PMYou are not going to use the stdlib from Java (I suppose) - d
dalexander
04/18/2016, 2:23 PMI have an easy time imagining some years from now when there’s a larger userbase for Kotlin, and there’s less need to call from Java -> Kotlin, especially on new projects. - d
dmitry.petrov
04/18/2016, 2:23 PMIn fact, Kotlin compiler itself uses stdlib in Java code quite a few times.😱 2 - k
kmruiz
04/18/2016, 2:24 PMblack magic - k
kirillrakhman
04/18/2016, 2:24 PMthat's not a problem if it's usual helper methods without lambdas and reifications and stuff - d
dalexander
04/18/2016, 2:24 PMThat makes sense given that kotlinc started off as pure Java and has been migrating to more Kotlin over time? - k
kmruiz
04/18/2016, 2:26 PMProbably it's "legacy" Java that is going to be rewritten when needed - k
kmruiz
04/18/2016, 2:26 PMHowever you can mark this kind of functionality as internal - k
kmruiz
04/18/2016, 2:34 PM - k
kmruiz
04/18/2016, 2:34 PMWould be possible for Kotlin on Java 8+ - k
kirillrakhman
04/18/2016, 2:34 PMisn't this just for reflection? - k
kmruiz
04/18/2016, 2:35 PMAn intersection type can be either implicitly or explicitly declared in a program. For example, the bound of the type parameter <T extends Number & Runnable> is an (implicit) intersection type. As of RELEASE_8, this is represented by an IntersectionType with Number and Runnable as its bounds.
- k
kmruiz
04/18/2016, 2:36 PMWe can map MutableCollection<R>, RandomAcess like <T extends MutableCollection<R> & RandomAccess> - k
kirillrakhman
04/18/2016, 2:36 PMthat's exactly what intersection types in Kotlin are doing - k
kirillrakhman
04/18/2016, 2:37 PMbut only for parameters - v
voddan
04/18/2016, 2:37 PMclass A<T> where T: B, C - k
kirillrakhman
04/18/2016, 2:37 PMwell, not only parameters, but all the stuff we discussed in the last hour or so - k
kmruiz
04/18/2016, 2:38 PMyep, so in the future when Kotlin moves to Java 8 (I don't know expectations but looking how Android works, maybe later than sooner) we could have something like that - k
kirillrakhman
04/18/2016, 2:38 PMwe already do - k
kmruiz
04/18/2016, 2:38 PMfor return types - k
kmruiz
04/18/2016, 2:40 PMNow this won't work:inline fun <R, T> mutableRandomAccess(vararg value: R): T where T : MutableList<R>, T : RandomAccess, R: Any - k
kirillrakhman
04/18/2016, 2:47 PMit does, but T will always be substituted - k
kirillrakhman
04/18/2016, 2:47 PMwhat you want isfun <R> mutableRandomAccess(vararg value: R): MutableList<R>, RandomAccess - k
kmruiz
04/18/2016, 5:14 PMI'll try it :D - t
thomasnield
04/18/2016, 7:08 PMHey, I don't know if there's a feature that does this already. But I created an extension for all types T that will make it null if a given condition about it is true. I was surprised this was not already part of the language unless I missed it. - t
thomasnield
04/18/2016, 7:09 PMfun <T> T.nullIf(condition: (T) -> Boolean): T? = if (condition.invoke(this)) null else this - u
uhe
04/18/2016, 7:11 PMI think @kirillrakhman proposed something similar a while back - u
uhe
04/18/2016, 7:11 PMinline fun <T> T.assuming(condition: (T) -> Boolean) = if (condition(this)) this else null - u
uhe
04/18/2016, 7:11 PMthough it's the other way around