Update *-focus-by-mouse-then-keydown.html tests to be repaint tests

Update *-focus-by-mouse-then-keydown.html tests to be repaint tests

These tests worked previously but no longer work in our layouttest
repaint infrastructure. These tests intended to do the following:
1) Paint the page
2) Click a form control element (no focus ring)
3) Press a key (causing a focus ring)
4) Paint the page again, this time with the focus ring

The tests actually did the following:
1) Click a form control element (no focus ring)
2) Press a key (causing a focus ring)
3) Paint the page for the first time, including focus ring

This patch updates these tests to be repaint tests. I manually verified
these tests fail if HTMLFormControlElement.cpp does not call
renderer()->setShouldDoFullPaintInvalidation(true). The repaint tests
will have platform-specific results so NeedsRebaseline has been used,
but I uploaded my OSX10.9.4/highdpi results to make the review easier.

The checkbox-focus-by-mouse-then-keydown test is already listed in
FlakyTests (and is flaky [1]) so this was removed from TestExpectations.

[1] http://test-results.appspot.com/dashboards/flakiness_dashboard.html#tests=fast%2Fforms%2Fcheckbox%2Fcheckbox-focus-by-mouse-then-keydown.html

BUG=407961
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=180940

[pdr., 2014-08-27 04:31:39]

pdr@chromium.org changed reviewers:
+ tkent@chromium.org

[tkent, 2014-08-27 04:41:47]

https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
File
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt
(right):

https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt:9:
[3, 4, 55, 29]
We can't know existence of a focus ring with this result, right?

[pdr., 2014-08-27 04:46:55]

On 2014/08/27 04:41:47, tkent wrote:
>
https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
> File
>
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt
> (right):
> 
>
https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
>
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt:9:
> [3, 4, 55, 29]
> We can't know existence of a focus ring with this result, right?

That's right, but we already have focus ring tests (e.g.,
button-reset-focus.html). These tests were just testing that the focus ring was
repainted. If I uncomment the setShouldDoFullPaintInvalidation line in
HTMLFormControlElement, we get no repaints instead of the small repaint in the
-expected.txts

[tkent, 2014-08-27 05:02:47]

On 2014/08/27 04:46:55, pdr wrote:
> On 2014/08/27 04:41:47, tkent wrote:
> >
>
https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
> > File
> >
>
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt
> > (right):
> > 
> >
>
https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
> >
>
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt:9:
> > [3, 4, 55, 29]
> > We can't know existence of a focus ring with this result, right?
> 
> That's right, but we already have focus ring tests (e.g.,
> button-reset-focus.html). These tests were just testing that the focus ring
was
> repainted. If I uncomment the setShouldDoFullPaintInvalidation line in
> HTMLFormControlElement, we get no repaints instead of the small repaint in the
> -expected.txts

The purpose of button-reset-focus-by-mouse.html is to check if the resultant
button has NO focus ring. The purpose of
button-reset-focus-by-mouse-then-keydown.html is to check if the resultant
button has a focus ring.   Even if we removed |m_wasFocusedByMouse = false| in
HTMLFormControlElement::willCallDefaultEventHandler, new tests in this CL
wouldn't fail.
Probably we need enablePixelTesting=true in each of tests.

[pdr., 2014-08-27 05:05:53]

On 2014/08/27 05:02:47, tkent wrote:
> On 2014/08/27 04:46:55, pdr wrote:
> > On 2014/08/27 04:41:47, tkent wrote:
> > >
> >
>
https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
> > > File
> > >
> >
>
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt
> > > (right):
> > > 
> > >
> >
>
https://codereview.chromium.org/504343004/diff/1/LayoutTests/fast/forms/butto...
> > >
> >
>
LayoutTests/fast/forms/button/button-reset-focus-by-mouse-then-keydown-expected.txt:9:
> > > [3, 4, 55, 29]
> > > We can't know existence of a focus ring with this result, right?
> > 
> > That's right, but we already have focus ring tests (e.g.,
> > button-reset-focus.html). These tests were just testing that the focus ring
> was
> > repainted. If I uncomment the setShouldDoFullPaintInvalidation line in
> > HTMLFormControlElement, we get no repaints instead of the small repaint in
the
> > -expected.txts
> 
> The purpose of button-reset-focus-by-mouse.html is to check if the resultant
> button has NO focus ring. The purpose of
> button-reset-focus-by-mouse-then-keydown.html is to check if the resultant
> button has a focus ring.   Even if we removed |m_wasFocusedByMouse = false| in
> HTMLFormControlElement::willCallDefaultEventHandler, new tests in this CL
> wouldn't fail.
> Probably we need enablePixelTesting=true in each of tests.

You're absolutely right, I misread what *focus-by-mouse.html were doing. Updated
patch on the way!

[pdr., 2014-08-27 05:23:46]

PTAL

[tkent, 2014-08-27 05:29:13]

The CQ bit was checked by tkent@chromium.org

[tkent, 2014-08-27 05:29:14]

lgtm

[commit-bot: I haz the power, 2014-08-27 05:30:08]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/pdr@chromium.org/504343004/20001

[commit-bot: I haz the power, 2014-08-27 06:36:01]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/24446)

[commit-bot: I haz the power, 2014-08-27 07:08:35]

Committed patchset #2 (20001) as 180940