ReSharper even flags them as useless in C#

here's the actual method:

inline fun <reified T : Any> HttpException.bodyAs(gson: Gson): T? =
        gson.fromJson<T>(response().errorBody().string(), object : TypeToken<T>() {}.type)

I see now.

for reference http://stackoverflow.com/a/260779/615306

That's a bug indeed. You do not expose it. The object that your function creates has all the right to use this protected constructor.

that's how I see it, too

while we're on the topic of exposing, do you have an opinion on https://youtrack.jetbrains.com/issue/KT-15347?

I don't. I have to see use-cases. At the first sight it seems to that Kotlin correctly protects you from bugs here. If you can call it in Java then what can you do with the result if the result type is not accessible to you?

In Kotlin, such code can cause IAE at run-time.

(unfortunately)

Scala has the same problem, although it just silently accepts such code, and fails with IAE.

Probably it should be downgraded to a warning...

Problem is, it can get rather messy with big hairy class hierarchies mixing public and package-private classes. Something like:

public interface Base {}
/*package-private*/ abstract class Impl implements Base {}
public class Derived1 extends Impl {}
public class Derived2 extends Impl {}

Now, when you use

Derived1

and

Derived2

in expression,

Impl

can be implicitly inferred, BE will emit CHECKCAST, boom. Actually you can write similar code in Java (that would implicitly infer

Impl

), it'd just take a bit more of "creativity" to do so.

So yeah, possibly it should be a warning that would somehow suggest declaring types explicitly.

An intention to add types explicitly would be great.

Unfortunately, that's not always possible without additional refactorings. E.g.,

foo(bar(if (flag) Derived1() else Derived2())) // should extract variable here

is there a way to suppress the "destructure" inspection for specific types?

It would be nice if I could annotate a data class with a non-small number of members so that it wouldn't get constantly flagged for "use destructuring declaration" in lambdas

(other than that, I like the inspection)

made an issue since I couldn't find anything for it https://youtrack.jetbrains.com/issue/KT-15422

(on a side note,

Destructure

shouldn't be a spelling error by default in IntelliJ 😉 )

welp, I updated one of my projects to 1.1-M04 and it causes javac to crash 👀

@User uploaded a file: Untitled and commented: should I report this to YouTrack or the Java bug reporting page? I've no idea yet what exactly causes it... it's a big project unfortunately

never attached a debugger to javac before, so that was interesting https://youtrack.jetbrains.com/issue/KT-15424

In 1.1-M04, the following code fails when executed:

fun test() {
    val action = mock<() -> Unit>()
    MyJavaClass().doSomething(action)
}

public class MyJavaClass {

    void doSomething(Runnable r) {
        r.run();
    }
}

Now

mock<() -> Unit>()

uses Mockito to create a mock object. This mock object will by default return

null

when invoked:

action.invoke()

will return

null

. Since M04, the following ISE is thrown:

java.lang.IllegalStateException: invoke(...) must not be null

	at test.SAMTestKt$sam$Runnable$3ede1671.run(SAMTest.kt)
	at test.MyJavaClass.doSomething(MyJavaClass.java:6)
	at test.SAMTest.test(SAMTest.kt:12)

Is this intended?

i have created https://youtrack.jetbrains.com/issue/KT-15431 for this

Just saw that a new version of Kotlin plugin was available in AS. Updated it and now I cannot open my project anymore. I assume that’s because the latest version is v1.0.6-eap-123-IJ2016.2-1. Unfortunately, that’s the only version AS offers me to install and I need to download it manually from the plugin repository.

And 1.0.6-eap-123-Studio2.3-1 is unavailable on https://plugins.jetbrains.com: https://plugins.jetbrains.com/plugin/download?pr=androidstudio&updateId=31264

After updating to

1.1-M04

I'm getting a lot of

org.jetbrains.kotlin.codegen.CompilationException: Back-end (JVM) Internal error: Couldn't inline method call ...

and it always caused at

MethodInliner.java:752

I'm unable to reproduce it on a small project. Somehow it related to bound callable reference inside inlined lambda. Maybe it is a known issue? Would it be enough to file an issue just with compiler error message?