GrClip no longer translates its clips (to better mimic SkClipStack's behavior)

http://codereview.appspot.com/6445052/



git-svn-id: http://skia.googlecode.com/svn/trunk@4848 2bbb7eff-a529-9590-31e7-b0007b416f81
diff --git a/src/gpu/GrClipMaskManager.h b/src/gpu/GrClipMaskManager.h
index 0551b8c..81be5ed 100644
--- a/src/gpu/GrClipMaskManager.h
+++ b/src/gpu/GrClipMaskManager.h
@@ -33,13 +33,7 @@
  */
 class GrClipMaskCache : public GrNoncopyable {
 public:
-    GrClipMaskCache()
-    : fContext(NULL)
-    , fStack(sizeof(GrClipStackFrame)) {
-        // We need an initial frame to capture the clip state prior to 
-        // any pushes
-        SkNEW_PLACEMENT(fStack.push_back(), GrClipStackFrame);
-    }
+    GrClipMaskCache();
 
     ~GrClipMaskCache() {
 
@@ -86,9 +80,7 @@
      * TODO: can we take advantage of the nested nature of the clips to
      * reduce the mask creation cost?
      */
-    void push() {
-        SkNEW_PLACEMENT(fStack.push_back(), GrClipStackFrame);
-    }
+    void push();
 
     void pop() {
         //GrAssert(!fStack.empty());