Merge AcceleratedCompositingForOverflowScroll into PreferCompositingToLCDText.

Merge AcceleratedCompositingForOverflowScroll into PreferCompositingToLCDText.

This removes the AcceleratedCompositingForOverflowScroll setting and
changes code that checked for it to instead check for the
PreferCompositingToLCDText setting.

R=abarth@chromium.org, vollick@chromium.org, abarth
BUG=365851
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180492

[danakj, 2014-08-17 21:37:04]

There are 5 test failures after this change. 3 are straightforward rebase
because the text in the test changed and it is part of the resulting png.

2 are not straightforward...

1.
compositing/layer-creation/fixed-position-nonscrollable-body-mismatch-containers.html

This test used to trigger composited fixed position. And test that the fixed pos
element would be composited because an ancestor scrolls. Now it triggers
composited fixed pos and composited overflow scroll. The resulting output looks
very different. In fact, it seems like the test was failing before? I'm not
sure. The diff looks like this though:

---
/tmp/results/compositing/layer-creation/fixed-position-nonscrollable-body-mismatch-containers-expected.txt
+++
/tmp/results/compositing/layer-creation/fixed-position-nonscrollable-body-mismatch-containers-actual.txt
@@ -6,7 +6,58 @@
     {
       "bounds": [800, 600],
       "contentsOpaque": true,
-      "drawsContent": true
+      "drawsContent": true,
+      "children": [
+        {
+          "position": [8, 72],
+          "bounds": [302, 302],
+          "drawsContent": true,
+          "children": [
+            {
+              "position": [1, 1],
+              "bounds": [285, 285],
+              "children": [
+                {
+                  "bounds": [285, 800],
+                  "shouldFlattenTransform": false,
+                  "drawsContent": true,
+                  "children": [
+                    {
+                      "shouldFlattenTransform": false
+                    }
+                  ]
+                }
+              ]
+            },
+            {
+              "children": [
+                {
+                  "position": [1, 286],
+                  "bounds": [285, 15],
+                  "drawsContent": true
+                },
+                {
+                  "position": [286, 1],
+                  "bounds": [15, 285],
+                  "drawsContent": true
+                },
+                {
+                  "position": [286, 286],
+                  "bounds": [15, 15],
+                  "drawsContent": true
+                }
+              ]
+            }
+          ]
+        },
+        {
+          "position": [10, 100],
+          "bounds": [100, 100],
+          "contentsOpaque": true,
+          "drawsContent": true,
+          "backgroundColor": "#00FF00"
+        }
+      ]
     }
   ]
 }

2. compositing/overflow/universal-accelerated-overflow-scroll.html

This test used to enable composited overflow scroll. Now it enabled composited
fixed position too. Since the test uses a fixed position element, its tree dump
changed because the fixed thing gets promoted. I'm not sure if this still tests
what it should be testing or if the test should even exist. The diff looks like
this:

---
/tmp/results/compositing/overflow/universal-accelerated-overflow-scroll-expected.txt
+++
/tmp/results/compositing/overflow/universal-accelerated-overflow-scroll-actual.txt
@@ -761,20 +761,46 @@
           ]
         },
         {
-          "children": [
-            {
-              "position": [375, 20],
-              "bounds": [50, 100],
-              "contentsOpaque": true,
-              "drawsContent": true,
-              "backgroundColor": "#008000"
-            },
-            {
-              "position": [15, 140],
-              "bounds": [290, 220],
-              "drawsContent": true
-            }
-          ]
+          "position": [375, 20],
+          "bounds": [50, 100],
+          "contentsOpaque": true,
+          "drawsContent": true,
+          "backgroundColor": "#008000"
+        },
+        {
+          "position": [15, 140],
+          "bounds": [50, 100],
+          "contentsOpaque": true,
+          "drawsContent": true,
+          "backgroundColor": "#008000"
+        },
+        {
+          "position": [135, 140],
+          "bounds": [50, 100],
+          "contentsOpaque": true,
+          "drawsContent": true,
+          "backgroundColor": "#008000"
+        },
+        {
+          "position": [255, 140],
+          "bounds": [50, 100],
+          "contentsOpaque": true,
+          "drawsContent": true,
+          "backgroundColor": "#008000"
+        },
+        {
+          "position": [15, 260],
+          "bounds": [50, 100],
+          "contentsOpaque": true,
+          "drawsContent": true,
+          "backgroundColor": "#008000"
+        },
+        {
+          "position": [135, 260],
+          "bounds": [50, 100],
+          "contentsOpaque": true,
+          "drawsContent": true,
+          "backgroundColor": "#008000"
         },
         {
           "children": [

[abarth-chromium, 2014-08-18 00:32:56]

I'd be inclined to just land this CL.  The configuration tested by these tests
no longer exists.  If we've lost some test coverage, that's ok.  We'll get it
back in the future.  So, LGTM.  (Please wait for feedback from vollick, who
might have a different perspective.)

[Ian Vollick, 2014-08-18 00:44:18]

On 2014/08/18 00:32:56, abarth wrote:
> I'd be inclined to just land this CL.  The configuration tested by these tests
> no longer exists.  If we've lost some test coverage, that's ok.  We'll get it
> back in the future.  So, LGTM.  (Please wait for feedback from vollick, who
> might have a different perspective.)

Agreed. Lgtm2

[danakj, 2014-08-18 15:05:51]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2014-08-18 15:06:28]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/danakj@chromium.org/486493002/80001

[danakj, 2014-08-18 16:02:59]

Hm, the
virtual/gpu/compositedscrolling/scrollbars/custom-scrollbars-paint-outside-iframe.html
test fails with this change, the green box becomes red.

[danakj, 2014-08-18 18:30:42]

On 2014/08/18 16:02:59, danakj wrote:
> Hm, the
>
virtual/gpu/compositedscrolling/scrollbars/custom-scrollbars-paint-outside-iframe.html
> test fails with this change, the green box becomes red.

This test fails because CompositedScrollingForFrames is now enabled with the
PreferCompositing flag, and that behaviour breaks this test. I'm going to file a
bug about it because it seems like a bug.

[danakj, 2014-08-18 18:34:24]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2014-08-18 18:35:20]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/danakj@chromium.org/486493002/140001

[commit-bot: I haz the power, 2014-08-18 19:34:13]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  linux_blink_dbg on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_dbg/builds/2...)
  linux_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_rel/builds/2...)
  mac_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/mac_blink_rel/builds/20252)
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/23284)

[danakj, 2014-08-18 19:35:01]

The CQ bit was checked by danakj@chromium.org

[commit-bot: I haz the power, 2014-08-18 19:36:13]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/danakj@chromium.org/486493002/160001

[danakj, 2014-08-18 20:39:54]

Committed patchset #6 manually as 180492 (presubmit successful).