"Unable to locate class file for inner class"

[elfprince13]

I have two source files, one Java, one Scala.
The Java class definition is as follows:

public class HeapNode<Elem> extends Node<HeapNode<Elem>> implements Cloneable {
	volatile LazyBraunHeapLF.NodeEvalState<Elem> state;
	
	HeapNode(LazyBraunHeapLF.NodeEvalState<Elem> initState){
		state = initState;
	}
	
	@SuppressWarnings("rawtypes")
	public static final AtomicReferenceFieldUpdater<HeapNode, LazyBraunHeapLF.NodeEvalState> updateState 
					= AtomicReferenceFieldUpdater.newUpdater(HeapNode.class, LazyBraunHeapLF.NodeEvalState.class, "state");

	
}

The Scala source contains the following:

object LazyBraunHeapLF {
	
	class EmptyNode[E] extends HeapNode[E](Empty[E]()){
		marked = true
	}
	
	sealed trait Immediate[E]{ this: NodeEvalState[E] =>
		def force:NodeEvalState[E] with Immediate[E] = this
	}
	
	sealed abstract class NodeEvalState[E] {
		def force: NodeEvalState[E] with Immediate[E]
	}
	
	case class Empty[E]() extends NodeEvalState[E] with Immediate[E]
	// The remainder is commented out for a M(N)WE
}

Everything works fine with a clean build, or if the definition of class EmptyNode[E] is commented out; however, after roughly ~1 edit cycle, I get the following error message:

[error] Unable to locate class corresponding to inner class entry for NodeEvalState in owner LazyBraunHeapLF
[error] one error found
[error] (compile:compileIncremental) Compilation failed
[error] Total time: 2 s, completed Aug 8, 2018 11:30:33 AM

I'm using Scala 2.12.6 on Java 8. Notably, the class file in question exists on my file system.

[ghik]

I've been fighting with this issue for a long time. It almost makes the incremental compiler unusable in some situations, forcing me to rebuild the project over and over again.

[SethTisue]

do we have reason to believe this is a scalac bug and not a zinc bug?

[elfprince13]

I have no recollection of how I triggered this bug, and ended up switching that whole project over to dotty anyway.

[SethTisue]

@ghik do you think you might be able to isolate and reproduce this?

[ghik]

@SethTisue since it is related to incremental compilation, it happens seemingly randomly and I'm afraid it won't be easy to reproduce.

[SethTisue]

okay, I'm going to close this because it's almost certainly a zinc bug. unfortunately GitHub doesn't let me transfer issues from org to org, only within the same org.

[dwijnand]

Hand-transferred as sbt/zinc#797.