False positive moving methods within sealed abstract class hierarchy

[adriaanm]

trafficstars

Needs clarification: https://github.com/typelevel/cats/pull/2335#discussion_r205950655

"Only sealed abstract classes that provide implicit instances to companion objects are allowed here, since they don't affect usage outside of the file."

[kailuowang]

When concrete methods are being moved between abstract sealed classes whose usage is in a single inheritance hierarchy, the change should be deemed binary compatible. E.g.

sealed abstract class A {def a = 1 }
sealed abstract class B extends A
class C extends B //only usage of A and B

If you move def a = 1 to B it should be deemed as binary compatible.

[adriaanm]

Thanks!

[dwijnand]

Another piece of the "Scala working mode" epic.

https://github.com/lightbend/migration-manager/issues/34.