Revert r207560 and r207566 to reland r207511.

Revert r207560 and r207566 to reland r207511.

The win_aura builder appeared to have trouble running
animations in
DesktopBackgroundControllerTest.SetDefaultWallpaper; this
version of the change splits that test into several smaller
tests so that animations can be avoided.

Original change:

ash: Add flags for specifying default wallpaper.

This adds new command-line flags for specifying image files
that will be used as default wallpaper:

  --ash-default-guest-wallpaper-large
  --ash-default-guest-wallpaper-small
  --ash-default-wallpaper-large
  --ash-default-wallpaper-small

This makes it easier for Chrome OS to have per-device custom
wallpaper. (We previously loaded the default wallpaper from
a resource file, but that prevents the same build of Chrome
from being used on different devices.)

BUG=248764
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=207800

[bshe, 2013-06-20 23:19:32]

On 2013/06/20 23:15:14, Daniel Erat wrote:

lgtm

[Daniel Erat, 2013-06-20 23:24:56]

jamescook@ did some Windows builds of ash_unittests on my behalf. We verified
that we could repro the earlier crash in
DesktopBackgroundControllerTest.SetDefaultWallpaper. With this change applied,
DesktopBackgroundControllerTest.* passes.

[commit-bot: I haz the power, 2013-06-20 23:26:44]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/derat@chromium.org/17502005/3001

[bshe, 2013-06-21 00:43:14]

https://codereview.chromium.org/17502005/diff/1/ash/desktop_background/deskto...
File ash/desktop_background/desktop_background_controller_unittest.cc (right):

https://codereview.chromium.org/17502005/diff/1/ash/desktop_background/deskto...
ash/desktop_background/desktop_background_controller_unittest.cc:416:
observer.WaitForWallpaperDataChanged();
Your new test if still good.
For the old test, I just notice this, should we add a
RunDesktopControllerAnimation() here? This maybe the reason for the crash? But I
am not sure why it didn't fail on linux_chromeos bots.

Also, my try job on win8_aura returned this:
"[2080:5844:0620/172717:5211931:FATAL:display_manager.cc(420)] Check failed: 1u
== updated_display_info_list.size() (1 vs. 2)Multiple display test does not work
on Win8 bots. Please skip (don't disable) the test using
|SupportMultipleDisplay()|
[2080:5844:0620/172717:5211931:FATAL:display_manager.cc(420)] Check failed: 1u
== updated_display_info_list.size() (1 vs. 2)Multiple display test does not work
on Win8 bots. Please skip (don't disable) the test using
|SupportMultipleDisplay()|"
Not sure if you have seen it. But this seems may cause new issue even for the
new tests (we can probably wait for the result from win8_aura try bot)?

For reference:
http://build.chromium.org/p/tryserver.chromium/builders/win8_aura/builds/928

[Daniel Erat, 2013-06-21 00:53:31]

https://codereview.chromium.org/17502005/diff/1/ash/desktop_background/deskto...
File ash/desktop_background/desktop_background_controller_unittest.cc (right):

https://codereview.chromium.org/17502005/diff/1/ash/desktop_background/deskto...
ash/desktop_background/desktop_background_controller_unittest.cc:416:
observer.WaitForWallpaperDataChanged();
On 2013/06/21 00:43:14, bshe wrote:
> Your new test if still good.
> For the old test, I just notice this, should we add a
> RunDesktopControllerAnimation() here? This maybe the reason for the crash? But
I
> am not sure why it didn't fail on linux_chromeos bots.

Yeah, I noticed this too when I was updating the tests. The Windows crash looked
different from the failed assert that I saw when I initially wasn't running the
animations there, though, so I think it might've been a separate problem. I
don't know why this omission didn't cause failures anywhere else either. :-(

> Also, my try job on win8_aura returned this:
> "[2080:5844:0620/172717:5211931:FATAL:display_manager.cc(420)] Check failed:
1u
> == updated_display_info_list.size() (1 vs. 2)Multiple display test does not
work
> on Win8 bots. Please skip (don't disable) the test using
> |SupportMultipleDisplay()|
> [2080:5844:0620/172717:5211931:FATAL:display_manager.cc(420)] Check failed: 1u
> == updated_display_info_list.size() (1 vs. 2)Multiple display test does not
work
> on Win8 bots. Please skip (don't disable) the test using
> |SupportMultipleDisplay()|"
> Not sure if you have seen it. But this seems may cause new issue even for the
> new tests (we can probably wait for the result from win8_aura try bot)?
> 
> For reference:
> http://build.chromium.org/p/tryserver.chromium/builders/win8_aura/builds/928

Thanks for catching this! I agree that it probably would've caused more
failures. I've updated the test to use this.

[commit-bot: I haz the power, 2013-06-21 00:55:36]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/derat@chromium.org/17502005/27001

[Daniel Erat, 2013-06-21 05:11:08]

I'm still seeing failures on win8_aura (while win7_aura is fine). This time,
it's returning the small wallpaper in places where it should be using the large
wallpaper. I suspect that UpdateDisplay() is broken on Windows 8 even in the
single-display case and the displays aren't getting created as requested.

I feel like I've spent too much time trying to work around testing bugs on a
platform that we don't care about with regard to this change, so I've added
SupportsMultipleDisplays() calls to the beginning of all of these tests to make
them not run on platforms where the testing framework seems to be buggy. We'll
still have testing coverage on win7_aura, along with all of the Chrome OS
builders that we actually care about.

[commit-bot: I haz the power, 2013-06-21 05:11:24]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/derat@chromium.org/17502005/42001

[commit-bot: I haz the power, 2013-06-21 10:49:04]

Step "update" is always a major failure.
Look at the try server FAQ for more details.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_aura...

[commit-bot: I haz the power, 2013-06-21 11:48:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/derat@chromium.org/17502005/42001

[commit-bot: I haz the power, 2013-06-21 11:54:01]

Change committed as 207800

[oshima, 2013-06-21 14:20:33]

On win8, most tests that rely on real host window size fails because
you can't change host window size on win8/metro. We probably
should add similar method like SupportsHostWindowResize()
to skip such tests on win8 bots.



On Thu, Jun 20, 2013 at 10:11 PM, <derat@chromium.org> wrote:

> I'm still seeing failures on win8_aura (while win7_aura is fine). This
> time,
> it's returning the small wallpaper in places where it should be using the
> large
> wallpaper. I suspect that UpdateDisplay() is broken on Windows 8 even in
> the
> single-display case and the displays aren't getting created as requested.
>
> I feel like I've spent too much time trying to work around testing bugs on
> a
> platform that we don't care about with regard to this change, so I've added
> SupportsMultipleDisplays() calls to the beginning of all of these tests to
> make
> them not run on platforms where the testing framework seems to be buggy.
> We'll
> still have testing coverage on win7_aura, along with all of the Chrome OS
> builders that we actually care about.
>
>
https://codereview.chromium.**org/17502005/<https://codereview.chromium.org/1...
>

[Daniel Erat, 2013-06-21 14:22:37]

I guess that the reason that this wasn't a problem before was that the
tests that were exercising
DesktopBackgroundController::GetAppropriateResolution() were in
chrome/browser/chromeos/login instead of part of
DesktopBackgroundControllerTest. :-(


On Fri, Jun 21, 2013 at 7:20 AM, oshima <oshima@chromium.org> wrote:

> On win8, most tests that rely on real host window size fails because
> you can't change host window size on win8/metro. We probably
> should add similar method like SupportsHostWindowResize()
> to skip such tests on win8 bots.
>
>
>
> On Thu, Jun 20, 2013 at 10:11 PM, <derat@chromium.org> wrote:
>
>> I'm still seeing failures on win8_aura (while win7_aura is fine). This
>> time,
>> it's returning the small wallpaper in places where it should be using the
>> large
>> wallpaper. I suspect that UpdateDisplay() is broken on Windows 8 even in
>> the
>> single-display case and the displays aren't getting created as requested.
>>
>> I feel like I've spent too much time trying to work around testing bugs
>> on a
>> platform that we don't care about with regard to this change, so I've
>> added
>> SupportsMultipleDisplays() calls to the beginning of all of these tests
>> to make
>> them not run on platforms where the testing framework seems to be buggy.
>> We'll
>> still have testing coverage on win7_aura, along with all of the Chrome OS
>> builders that we actually care about.
>>
>>
https://codereview.chromium.**org/17502005/<https://codereview.chromium.org/1...
>>
>
>