Title: [183794] trunk
- Revision
- 183794
- Author
- [email protected]
- Date
- 2015-05-04 21:29:30 -0700 (Mon, 04 May 2015)
Log Message
Fix assertions in WK1 tests after r183777.
Source/WebCore:
Two issues with the original change:
If not in compositing mode, we shouldn't schedule layer flushes, since
Document::documentWillSuspendForPageCache() only only calls cancelCompositingLayerUpdate()
if in compositing mode.
Second, RenderLayerCompositor::rootBackgroundTransparencyChanged() never stored the
new color, resulting in too many triggered compositing updates.
* rendering/RenderLayerCompositor.cpp:
(WebCore::RenderLayerCompositor::rootBackgroundTransparencyChanged):
LayoutTests:
Behavior difference with the change to RenderLayerCompositor::rootBackgroundTransparencyChanged().
* platform/mac-wk2/tiled-drawing/background-transparency-toggle-expected.txt:
Modified Paths
Diff
Modified: trunk/LayoutTests/ChangeLog (183793 => 183794)
--- trunk/LayoutTests/ChangeLog 2015-05-05 04:24:23 UTC (rev 183793)
+++ trunk/LayoutTests/ChangeLog 2015-05-05 04:29:30 UTC (rev 183794)
@@ -1,3 +1,11 @@
+2015-05-04 Simon Fraser <[email protected]>
+
+ Fix assertions in WK1 tests after r183777.
+
+ Behavior difference with the change to RenderLayerCompositor::rootBackgroundTransparencyChanged().
+
+ * platform/mac-wk2/tiled-drawing/background-transparency-toggle-expected.txt:
+
2015-05-04 Doug Russell <[email protected]>
AX: setting focus via accessibility object needs to set isSynchronizing in resulting selection intent
Modified: trunk/LayoutTests/platform/mac-wk2/tiled-drawing/background-transparency-toggle-expected.txt (183793 => 183794)
--- trunk/LayoutTests/platform/mac-wk2/tiled-drawing/background-transparency-toggle-expected.txt 2015-05-05 04:24:23 UTC (rev 183793)
+++ trunk/LayoutTests/platform/mac-wk2/tiled-drawing/background-transparency-toggle-expected.txt 2015-05-05 04:29:30 UTC (rev 183794)
@@ -46,7 +46,6 @@
(GraphicsLayer
(bounds 785.00 648.00)
(contentsOpaque 1)
- (backgroundColor #CCCCCC)
(tile cache coverage 0, 0 785 x 648)
(tile size 512 x 512)
(top left tile 0, 0 tiles grid 2 x 2)
Modified: trunk/Source/WebCore/ChangeLog (183793 => 183794)
--- trunk/Source/WebCore/ChangeLog 2015-05-05 04:24:23 UTC (rev 183793)
+++ trunk/Source/WebCore/ChangeLog 2015-05-05 04:29:30 UTC (rev 183794)
@@ -1,3 +1,19 @@
+2015-05-04 Simon Fraser <[email protected]>
+
+ Fix assertions in WK1 tests after r183777.
+
+ Two issues with the original change:
+
+ If not in compositing mode, we shouldn't schedule layer flushes, since
+ Document::documentWillSuspendForPageCache() only only calls cancelCompositingLayerUpdate()
+ if in compositing mode.
+
+ Second, RenderLayerCompositor::rootBackgroundTransparencyChanged() never stored the
+ new color, resulting in too many triggered compositing updates.
+
+ * rendering/RenderLayerCompositor.cpp:
+ (WebCore::RenderLayerCompositor::rootBackgroundTransparencyChanged):
+
2015-05-04 Zalan Bujtas <[email protected]>
RenderWidget::setWidgetGeometry() can end up destroying *this*.
Modified: trunk/Source/WebCore/rendering/RenderLayerCompositor.cpp (183793 => 183794)
--- trunk/Source/WebCore/rendering/RenderLayerCompositor.cpp 2015-05-05 04:24:23 UTC (rev 183793)
+++ trunk/Source/WebCore/rendering/RenderLayerCompositor.cpp 2015-05-05 04:29:30 UTC (rev 183794)
@@ -3216,10 +3216,15 @@
void RenderLayerCompositor::rootBackgroundTransparencyChanged()
{
+ if (!inCompositingMode())
+ return;
+
Color documentBackgroundColor = m_renderView.frameView().documentBackgroundColor();
if (m_lastDocumentBackgroundColor.isValid() && documentBackgroundColor.hasAlpha() == m_lastDocumentBackgroundColor.hasAlpha())
return;
+ m_lastDocumentBackgroundColor = documentBackgroundColor;
+
// FIXME: We should do something less expensive than a full layer rebuild.
setCompositingLayersNeedRebuild();
scheduleCompositingLayerUpdate();
_______________________________________________
webkit-changes mailing list
[email protected]
https://lists.webkit.org/mailman/listinfo/webkit-changes
