If I instantiate the following object: ```object X...

# getting-started

[Klitos Kyriacou]

If I instantiate the following object:

object X : AbstractMap<Int, Int>() {
    override val entries: Set<Map.Entry<Int, Int>> = object : AbstractSet<Map.Entry<Int, Int>>() {
        override val size = 0
        override fun iterator() = emptySet<Map.Entry<Int, Int>>().iterator()
    }
}

then everything works fine.

However, if I omit the explicit type specification, like this:

object X : AbstractMap<Int, Int>() {
    override val entries = object : AbstractSet<Map.Entry<Int, Int>>() {
        override val size = 0
        override fun iterator() = emptySet<Map.Entry<Int, Int>>().iterator()
    }
}

then everything still compiles just fine, but at runtime I get this error:

java.lang.AbstractMethodError: Receiver class X does not define or inherit an implementation of the resolved method 'abstract java.util.Set getEntries()' of abstract class kotlin.collections.AbstractMap.
	at kotlin.collections.AbstractMap.entrySet(AbstractMap.kt:21)

Playground

I was surprised to see this because this error is documented as:
Normally, this error is caught by the compiler; this error can only occur at run time if the definition of some class has incompatibly changed since the currently executing method was last compiled.
But in my case this error has occurred without any change in the class.

[Daniel Pitts]

I suspect this is a bug in the compiler, since the second example shouldn't have actually compiled, since entries type is wrong, and therefore doesn't override the Map.getEntries() method.

[Daniel Pitts]

If I replace the override val entries with override fun getEntries(), then the compiler gives the expected output. (playground)

[Klitos Kyriacou]

No, that's a totally different thing. If you add the type spec (override fun getEntries(): Set<Map.Entry<Int, Int>>), it still gives you the error. It also says that getEntries overrides nothing. You can't override a val with a getter function.

[Daniel Pitts]

You're right, I didn't test the "positive" case.

[Daniel Pitts]

Either way, definitely a compiler bug.

[Klitos Kyriacou]

Thanks, indeed it is. I found it's already been reported 6 years ago (and still unresolved!)
https://youtrack.jetbrains.com/issue/KT-20070/Overriding-Map.entries-with-a-subtype-to-Set-caused-a-runtime-error