DummyTree => CannotHaveAttrs

This looks like a much more specific name than a generic "dummy" prefix.
CannotHaveAttrs also doesn't imply that an implementing tree should
satisfy some bigger contract. EmptyTree and emptyValDef are very unlike
each other, so there's no point in trying to unify them.

Also DummyTree.isEmpty is no longer automatically true. The notion of
trees not having positions and types by design (i.e. EmptyTree + empty
TypeTrees + emptyValDef + the new pendingSuperCall) is quite different
from the notion of an empty tree in a sense of a tree being a null object
(i.e. EmptyTree + empty TypeTrees).

1 files changed, 1 insertions, 1 deletions

diff --git a/src/compiler/scala/tools/nsc/ast/Positions.scala b/src/compiler/scala/tools/nsc/ast/Positions.scala
index 0503c5fb10..49569f5e05 100644
--- a/src/compiler/scala/tools/nsc/ast/Positions.scala
+++ b/src/compiler/scala/tools/nsc/ast/Positions.scala
@@ -20,7 +20,7 @@ trait Positions extends scala.reflect.internal.Positions {
         // When we prune due to encountering a position, traverse the
         // pruned children so we can warn about those lacking positions.
         t.children foreach { c =>
-          if (c.isDummy) ()
+          if (!c.canHaveAttrs) ()
           else if (c.pos == NoPosition) {
             reporter.warning(t.pos, " Positioned tree has unpositioned child in phase " + globalPhase)
             inform("parent: " + treeSymStatus(t))