[wip] ash: Update expectations in dock window test.

[wip] ash: Update expectations in dock window test.

BUG=none

[sadrul, 2014-02-20 04:16:40]

Hi! When investigating https://codereview.chromium.org/173443002, I noticed
these tests were failing, and noticed farther that the test expectations don't
quite make sense. Can you check if this expectation is correct?

[varkha, 2014-02-20 12:42:11]

On 2014/02/20 04:16:40, sadrul wrote:
> Hi! When investigating https://codereview.chromium.org/173443002, I noticed
> these tests were failing, and noticed farther that the test expectations don't
> quite make sense. Can you check if this expectation is correct?

Not sure that I understand. The windows should get docked in this test, not
minimized. Why is the test failing?

[sadrul, 2014-02-20 13:29:39]

On 2014/02/20 12:42:11, varkha wrote:
> On 2014/02/20 04:16:40, sadrul wrote:
> > Hi! When investigating https://codereview.chromium.org/173443002, I noticed
> > these tests were failing, and noticed farther that the test expectations
don't
> > quite make sense. Can you check if this expectation is correct?
> 
> Not sure that I understand. The windows should get docked in this test, not
> minimized. Why is the test failing?

The test passes locally for me with this change applied on ToT (I have sent out
a tryjob to verify), which means the first window is somehow getting minimized
(in ToT), which is what I found confusing, and wanted to check whether it was
expected or not.

[varkha, 2014-02-20 13:41:38]

On 2014/02/20 13:29:39, sadrul wrote:
> On 2014/02/20 12:42:11, varkha wrote:
> > On 2014/02/20 04:16:40, sadrul wrote:
> > > Hi! When investigating https://codereview.chromium.org/173443002, I
noticed
> > > these tests were failing, and noticed farther that the test expectations
> don't
> > > quite make sense. Can you check if this expectation is correct?
> > 
> > Not sure that I understand. The windows should get docked in this test, not
> > minimized. Why is the test failing?
> 
> The test passes locally for me with this change applied on ToT (I have sent
out
> a tryjob to verify), which means the first window is somehow getting minimized
> (in ToT), which is what I found confusing, and wanted to check whether it was
> expected or not.

Not expected. I have just had an unrelated change committed that had this test
pass on linux_chromeos bot.
Are you changing the screen resolution?

[sadrul, 2014-02-20 14:09:52]

On 2014/02/20 13:41:38, varkha wrote:
> On 2014/02/20 13:29:39, sadrul wrote:
> > On 2014/02/20 12:42:11, varkha wrote:
> > > On 2014/02/20 04:16:40, sadrul wrote:
> > > > Hi! When investigating https://codereview.chromium.org/173443002, I
> noticed
> > > > these tests were failing, and noticed farther that the test expectations
> > don't
> > > > quite make sense. Can you check if this expectation is correct?
> > > 
> > > Not sure that I understand. The windows should get docked in this test,
not
> > > minimized. Why is the test failing?
> > 
> > The test passes locally for me with this change applied on ToT (I have sent
> out
> > a tryjob to verify), which means the first window is somehow getting
minimized
> > (in ToT), which is what I found confusing, and wanted to check whether it
was
> > expected or not.
> 
> Not expected. I have just had an unrelated change committed that had this test
> pass on linux_chromeos bot.

The test passes on ToT both with and without the change in this CL. The change
in this CL simply shows that the first window is somehow getting minimized.
However, GetWindowBoundsInScreen() returns the bounds of the window before it
was minimized, which is why the test passes on ToT. But, as the change in this
CL shows, the first window is indeed getting minimized, which since unexpected,
seems to be a bug (perhaps only in the test?) that exists on ToT.

> Are you changing the screen resolution?

Nope. I am not making any additional changes. (if you apply the patch locally,
or simply add the WindowState related lines from this patch, the test should
pass for you too)

[varkha, 2014-02-20 14:17:16]

On 2014/02/20 14:09:52, sadrul wrote:
> On 2014/02/20 13:41:38, varkha wrote:
> > On 2014/02/20 13:29:39, sadrul wrote:
> > > On 2014/02/20 12:42:11, varkha wrote:
> > > > On 2014/02/20 04:16:40, sadrul wrote:
> > > > > Hi! When investigating https://codereview.chromium.org/173443002, I
> > noticed
> > > > > these tests were failing, and noticed farther that the test
expectations
> > > don't
> > > > > quite make sense. Can you check if this expectation is correct?
> > > > 
> > > > Not sure that I understand. The windows should get docked in this test,
> not
> > > > minimized. Why is the test failing?
> > > 
> > > The test passes locally for me with this change applied on ToT (I have
sent
> > out
> > > a tryjob to verify), which means the first window is somehow getting
> minimized
> > > (in ToT), which is what I found confusing, and wanted to check whether it
> was
> > > expected or not.
> > 
> > Not expected. I have just had an unrelated change committed that had this
test
> > pass on linux_chromeos bot.
> 
> The test passes on ToT both with and without the change in this CL. The change
> in this CL simply shows that the first window is somehow getting minimized.
> However, GetWindowBoundsInScreen() returns the bounds of the window before it
> was minimized, which is why the test passes on ToT. But, as the change in this
> CL shows, the first window is indeed getting minimized, which since
unexpected,
> seems to be a bug (perhaps only in the test?) that exists on ToT.
> 
> > Are you changing the screen resolution?
> 
> Nope. I am not making any additional changes. (if you apply the patch locally,
> or simply add the WindowState related lines from this patch, the test should
> pass for you too)

I will investigate, thanks. Most likely the call to UpdateDisplay in line 51
sets the height that is not enough for the two windows and the checking was
inadequate to catch that.

[varkha, 2014-02-20 15:32:46]

I have put up CL https://codereview.chromium.org/173843003/ to address this.