Fix mouse wheel over-scrolls when display is scaled and scroll is paginated

third_party/WebKit/LayoutTests/TestExpectations

 # Intentional failures to test the layout test system.
 Bug(intentional) [ Android Linux Win Mac10.10 Mac10.11 Retina ] harness-tests/crash.html [ Crash ]
 crbug.com/644303 [ Mac10.9 ] harness-tests/crash.html [ Crash Timeout ]
 Bug(intentional) harness-tests/timeout.html [ Timeout ]
 
 # ====== Oilpan-only failures from here ======
 # Most of these actually cause the tests to report success, rather than
 # failure. Expected outputs will be adjusted for the better once Oilpan
 # has been well and truly enabled always.
 
@@ skipping 840 matching lines @@
 
 # Mac does not support fractional scale factor.
 crbug.com/567837 [ Mac ] virtual/scalefactor150/fast/hidpi/static [ Skip ]
 
 # TODO(oshima): Move the event scaling code to eventSender and remove this.
 crbug.com/567837 virtual/scalefactor200/fast/hidpi/static/mousewheel-scroll-amount.html [ Skip ]
 crbug.com/567837 virtual/scalefactor200/fast/hidpi/static/gesture-scroll-amount.html [ Skip ]
 crbug.com/567837 virtual/scalefactor150/fast/hidpi/static/mousewheel-scroll-amount.html [ Skip ]
 crbug.com/567837 virtual/scalefactor150/fast/hidpi/static/gesture-scroll-amount.html [ Skip ]
 
+crbug.com/666133 virtual/scalefactor200withzoom/fast/hidpi/static/gesture-scroll-amount.html [ NeedsRebaseline ]

[aelias_OOO_until_Jul13, 2017/02/16 22:11:13]

It doesn't make sense that this test result would have changed.  It's clearly
intended to scroll by pixel amounts.  It must be a preexisting problem with the
test that deltaHintUnits is incorrect set to Page.  Instead of changing the
expectations to be wrong, please change the event sender framework to set
deltaHintUnits by pixel.

[chengx, 2017/02/17 08:26:02]

Acknowledged.

+crbug.com/666133 virtual/scalefactor200withzoom/fast/hidpi/static/mousewheel-scroll-amount.html [ NeedsRebaseline ]

[aelias_OOO_until_Jul13, 2017/02/16 22:11:13]

Since these don't have any image expectations, please directly include the
rebaselined result in this patch instead of relying on rebaseline bot.  This
makes it easier to examine what changed and also simplifies
cherry-picks/reverts.  (Although, I suspect there shouldn't be any result
changes needed at all if the test framework is changed to configure the events
properly by pixel.O)

[chengx, 2017/02/17 08:26:02]

Acknowledged.

+
 # TODO(ojan): These tests aren't flaky. See crbug.com/517144.
 # Release trybots run asserts, but the main waterfall ones don't. So, even
 # though this is a non-flaky assert failure, we need to mark it [ Pass Crash ].
 crbug.com/389648 crbug.com/517123 crbug.com/410145 fast/text-autosizing/table-inflation-crash.html [ Pass Crash Timeout ]
 
 crbug.com/636207 [ Win Debug ] fast/dom/HTMLImageElement/image-srcset-w-onerror.html [ Failure Pass ]
 
 crbug.com/569139 fast/js/string-replace-2.html [ Failure ]
 crbug.com/569139 fast/js/regexp-caching.html [ Failure ]
 crbug.com/597221 fast/dom/Window/window-postmessage-clone-deep-array.html [ Failure ]
@@ skipping 1486 matching lines @@
 
 # Flaky on trybots
 crbug.com/688486 external/wpt/service-workers/service-worker/fetch-request-redirect.https.html [ Failure Pass ]
 crbug.com/688486 external/wpt/service-workers/service-worker/fetch-request-resources.https.html [ Failure Pass ]
 
 # Sheriff failures Feb 06 2017
 crbug.com/688515 [ Linux ] http/tests/media/video-play-stall.html [ Pass Failure ]
 
 # Sheriff failures Feb 13 2017
 crbug.com/691423 [ Mac ] fast/frames/location-change-no-file-access.html [ Failure Pass ]