gecko: changeset 16143:b9032fa50289d5b16663b9b61ec12fd55b51694c

I mistyped one of surkov's review comments for bug 436453. (And the code does finally make sense to me.) This apparently fixes some mochitests that, for some reason, aren't run on tinderbox.

author	L. David Baron <dbaron@dbaron.org>
	Wed, 23 Jul 2008 09:35:38 -0700
changeset 16143	b9032fa50289d5b16663b9b61ec12fd55b51694c
parent 16142	328bf22c060517956ab9ce8928919a41ed862579
child 16144	084567d3ebe69672e9d31d38937748250af1dc7c
push id	1
push user	root
push date	Mon, 20 Oct 2014 17:29:22 +0000
bugs	436453
milestone	1.9.1a2pre

I mistyped one of surkov's review comments for bug 436453. (And the code does finally make sense to me.) This apparently fixes some mochitests that, for some reason, aren't run on tinderbox.

--- a/accessible/src/base/nsAccessibilityUtils.cpp
+++ b/accessible/src/base/nsAccessibilityUtils.cpp
@@ -789,17 +789,17 @@ nsIContent*
 nsAccUtils::FindNeighbourPointingToNode(nsIContent *aForNode, 
                                         nsIAtom **aRelationAttrs,
                                         PRUint32 aAttrNum,
                                         nsIAtom *aTagName,
                                         PRUint32 aAncestorLevelsToSearch)
 {
   nsAutoString controlID;
   if (!nsAccUtils::GetID(aForNode, controlID)) {
-    if (aForNode->IsInAnonymousSubtree())
+    if (!aForNode->IsInAnonymousSubtree())
       return nsnull;
 
     aForNode->GetAttr(kNameSpaceID_None, nsAccessibilityAtoms::anonid, controlID);
     if (controlID.IsEmpty())
       return nsnull;
   }
 
   // Look for label in subtrees of nearby ancestors