Imported From: https://issues.scala-lang.org/browse/SI-8295?orig=1
Reporter: @xeno-by
Affected Versions: 2.10.0, 2.11.0
See #8063

@gkossakowski said:
Depends how you define isCaseClass, if you define it as "has case modifier and is a class" then object's module class satisfies that definition.

@xeno-by said (edited on Feb 16, 2014 9:10:44 PM UTC):
I don't think this definition is a good fit for a public API, because it doesn't quite map on the surface of the language we're reflecting upon.

@xeno-by said:
This brings up a general problem of module classes being a very confusing part of the API (I remember having a lot of troubles grokking them when I was just starting with scalac), but that's a story for a different ticket.

@xeno-by said:
Unfortunately, I don't think we can fix this in 2.11.0.

First, the change is too risky to be applied in the last minute (even though in the compiler noone uses isCaseClass in its weird sense).

Second, immediate application of the required change (having separate isCaseClass and isCaseObject) leads to significant uglification of typical sealed case analysis (e.g. as in shapeless), because knownDirectSubclasses returns module classes for objects, which makes what currently is a simple isCaseClass check non-trivial.

stability wins here, I think