setMayNeedPaintInvalidation in setNeedsOverflowRecalcAfterStyleChange()

third_party/WebKit/Source/core/paint/PaintInvalidator.cpp

 // Copyright 2016 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "core/paint/PaintInvalidator.h"
 
 #include "core/editing/FrameSelection.h"
 #include "core/frame/FrameView.h"
 #include "core/frame/LocalFrame.h"
 #include "core/frame/Settings.h"
@@ skipping 469 matching lines @@
       break;
     default:
       break;
   }
 
   if (object.mayNeedPaintInvalidationSubtree()) {
     context.forcedSubtreeInvalidationFlags |=
         PaintInvalidatorContext::ForcedSubtreeInvalidationChecking;
   }
 
-  if (context.oldLocation != context.newLocation) {
+  if (context.oldLocation != context.newLocation &&
+      !context.paintingLayer->subtreeIsInvisible()) {

[chrishtr, 2017/04/04 17:03:45]

Could you add a test checking that making the tree visible again
will correctly update all these things?

Also, what is the mechanism that does that, especially after
https://codereview.chromium.org/2792063002/?

[Xianzhu, 2017/04/04 18:03:33]

Done.
The mechanism ensuring correct paint invalidation is existing:
- visibility change will cause full paint invalidation through style change
paint invalidation;
- subtreeIsInvisible flag is updated in
PaintLayer::updateDescendantDependentFlags().

     context.forcedSubtreeInvalidationFlags |=
         PaintInvalidatorContext::ForcedSubtreeInvalidationChecking;
   }
 }
 
 void PaintInvalidator::processPendingDelayedPaintInvalidations() {
   for (auto target : m_pendingDelayedPaintInvalidations)
     target->getMutableForPainting().setShouldDoFullPaintInvalidation(
         PaintInvalidationDelayedFull);
 }
 
 }  // namespace blink