[cros] Initialize OOBE in parallel when boot animation is disabled.

[cros] Initialize OOBE in parallel when boot animation is disabled.

This includes several fixes that prevent white splashes and UI junk during boot on slow machines:
*) OOBE UI is initialized in parallel and displayed when ready.
*) Status area widget is created initially invisible during boot/signout.
*) Initial cursor visibility initialized in RootWindow, otherwise cursor was visible during boot.
*) 'login visible' signal is sent only when login UI is ready for realz (user pods loaded/GAIA extension loaded/initial OOBE animation finished).


BUG=143298
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=154242

[Ivan Korotkov, 2012-08-29 13:20:29]

Please review:
nkostylev: the whole CL
sky: aura code (2 small fixes for initial cursor and status area visibility).

[Nikita (slow), 2012-08-29 14:10:01]

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
File chrome/browser/chromeos/login/webui_login_display_host.cc (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:47: // Whether to
enablwe tnitializing WebUI in hidden state (see
nit: enable initializing

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:77:
!CommandLine::ForCurrentProcess()->HasSwitch(switches::kDisableNewOobe) &&
nit: extract this flag value too

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:82:
waiting_for_user_pods_ =
How is this handled for OOBE when technically we don't have user pods but still
subscribe to this notification.

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
File chrome/browser/chromeos/login/webui_login_display_host.h (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.h:101: bool
initialize_webui_hidden_;
nit: Feels that renaming this variable hides a 2 methods for displaying WebUI:

parallel: initialize_webui_in_parallel_ == true
postpone: initialize_webui_in_parallel_ == false

in both cases it is initialized hidden.

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
File chrome/browser/chromeos/login/webui_login_view.cc (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_view.cc:245: //
TRACE_EVENT_ASYNC_BEGIN("OOBE", "SetStatusAreaVisible");
cleanup

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
File chrome/browser/resources/chromeos/login/user_pod_row.css (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
chrome/browser/resources/chromeos/login/user_pod_row.css:13: /* opacity: 1; */
Cleanup

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
File chrome/browser/resources/chromeos/login/user_pod_row.js (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
chrome/browser/resources/chromeos/login/user_pod_row.js:452:
bootWallpaperLoaded_: false,
nit: comment

[Nikita (slow), 2012-08-29 14:13:04]

Please use instruction from
http://code.google.com/p/chromium/issues/detail?id=140599 to see how
login-prompt-visible signal reflects reality (that user pods are actually seen).

[sky, 2012-08-29 15:17:29]

root_window.cc LGTM

[Ivan Korotkov, 2012-08-29 16:16:03]

Scott, what about ash? Should I pick someone else or you can take a look?

[sky, 2012-08-29 17:02:42]

http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
File ash/shell/shell_delegate_impl.h (right):

http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
ash/shell/shell_delegate_impl.h:29: virtual bool IsStatusAreaInitiallyVisible()
const OVERRIDE;
How does this differ from IsSessionStarted and IsUserLoggedIn? Does it makes
sense to have a single property that returns an enum?

[Ivan Korotkov, 2012-08-30 12:26:06]

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
File chrome/browser/chromeos/login/webui_login_display_host.cc (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:47: // Whether to
enablwe tnitializing WebUI in hidden state (see
On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> nit: enable initializing

Done.

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:77:
!CommandLine::ForCurrentProcess()->HasSwitch(switches::kDisableNewOobe) &&
On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> nit: extract this flag value too

Done.

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:82:
waiting_for_user_pods_ =
On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> How is this handled for OOBE when technically we don't have user pods but
still
> subscribe to this notification.

initialize_webui_hidden_ is set to false below in this case, so we show
postponed UI and don't wait for the notification.

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
File chrome/browser/chromeos/login/webui_login_view.cc (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_view.cc:245: //
TRACE_EVENT_ASYNC_BEGIN("OOBE", "SetStatusAreaVisible");
On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> cleanup

Done.

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
File chrome/browser/resources/chromeos/login/user_pod_row.css (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
chrome/browser/resources/chromeos/login/user_pod_row.css:13: /* opacity: 1; */
On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> Cleanup

Done.

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
File chrome/browser/resources/chromeos/login/user_pod_row.js (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
chrome/browser/resources/chromeos/login/user_pod_row.js:452:
bootWallpaperLoaded_: false,
On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> nit: comment

Done.

http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
File ash/shell/shell_delegate_impl.h (right):

http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
ash/shell/shell_delegate_impl.h:29: virtual bool IsStatusAreaInitiallyVisible()
const OVERRIDE;
On 2012/08/29 17:02:42, sky wrote:
> How does this differ from IsSessionStarted and IsUserLoggedIn? Does it makes
> sense to have a single property that returns an enum?

It's different in a sense that it's not a state (like UserLoggedIn and
SessionStarted). OTOH, it's value is equivalent to IsSessionStarted(). So it
actually makes sense to revert it at all and use IsSessionStarted directly.

[Nikita (slow), 2012-08-30 12:41:37]

lgtm

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
File chrome/browser/chromeos/login/webui_login_display_host.cc (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:82:
waiting_for_user_pods_ =
On 2012/08/30 12:26:06, Ivan Korotkov wrote:
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> > How is this handled for OOBE when technically we don't have user pods but
> still
> > subscribe to this notification.
> 
> initialize_webui_hidden_ is set to false below in this case, so we show
> postponed UI and don't wait for the notification.

Still it feels that would be cleaner not subscribe to it in the first place.

[Ivan Korotkov, 2012-08-30 14:54:51]

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
File chrome/browser/chromeos/login/webui_login_display_host.cc (right):

http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
chrome/browser/chromeos/login/webui_login_display_host.cc:82:
waiting_for_user_pods_ =
On 2012/08/30 12:41:37, Nikita Kostylev wrote:
> On 2012/08/30 12:26:06, Ivan Korotkov wrote:
> > On 2012/08/29 14:10:01, Nikita Kostylev wrote:
> > > How is this handled for OOBE when technically we don't have user pods but
> > still
> > > subscribe to this notification.
> > 
> > initialize_webui_hidden_ is set to false below in this case, so we show
> > postponed UI and don't wait for the notification.
> 
> Still it feels that would be cleaner not subscribe to it in the first place.

Ok, makes sense.

[commit-bot: I haz the power, 2012-08-30 15:30:12]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/ivankr@chromium.org/10889024/22001

[sky, 2012-08-30 16:08:52]

On Thu, Aug 30, 2012 at 5:26 AM,  <ivankr@chromium.org> wrote:
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> File chrome/browser/chromeos/login/webui_login_display_host.cc (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_display_host.cc:47: // Whether
> to enablwe tnitializing WebUI in hidden state (see
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> nit: enable initializing
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_display_host.cc:77:
> !CommandLine::ForCurrentProcess()->HasSwitch(switches::kDisableNewOobe)
> &&
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> nit: extract this flag value too
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_display_host.cc:82:
> waiting_for_user_pods_ =
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> How is this handled for OOBE when technically we don't have user pods
>
> but still
>>
>> subscribe to this notification.
>
>
> initialize_webui_hidden_ is set to false below in this case, so we show
> postponed UI and don't wait for the notification.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> File chrome/browser/chromeos/login/webui_login_view.cc (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_view.cc:245: //
> TRACE_EVENT_ASYNC_BEGIN("OOBE", "SetStatusAreaVisible");
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> cleanup
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> File chrome/browser/resources/chromeos/login/user_pod_row.css (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> chrome/browser/resources/chromeos/login/user_pod_row.css:13: /* opacity:
> 1; */
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> Cleanup
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> File chrome/browser/resources/chromeos/login/user_pod_row.js (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> chrome/browser/resources/chromeos/login/user_pod_row.js:452:
> bootWallpaperLoaded_: false,
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> nit: comment
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
> File ash/shell/shell_delegate_impl.h (right):
>
>
http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
> ash/shell/shell_delegate_impl.h:29: virtual bool
> IsStatusAreaInitiallyVisible() const OVERRIDE;
> On 2012/08/29 17:02:42, sky wrote:
>>
>> How does this differ from IsSessionStarted and IsUserLoggedIn? Does it
>
> makes
>>
>> sense to have a single property that returns an enum?
>
>
> It's different in a sense that it's not a state (like UserLoggedIn and
> SessionStarted). OTOH, it's value is equivalent to IsSessionStarted().
> So it actually makes sense to revert it at all and use IsSessionStarted
> directly.
>
> http://codereview.chromium.org/10889024/

Awesome, thanks.

[sky, 2012-08-30 16:08:52]

On Thu, Aug 30, 2012 at 5:26 AM,  <ivankr@chromium.org> wrote:
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> File chrome/browser/chromeos/login/webui_login_display_host.cc (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_display_host.cc:47: // Whether
> to enablwe tnitializing WebUI in hidden state (see
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> nit: enable initializing
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_display_host.cc:77:
> !CommandLine::ForCurrentProcess()->HasSwitch(switches::kDisableNewOobe)
> &&
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> nit: extract this flag value too
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_display_host.cc:82:
> waiting_for_user_pods_ =
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> How is this handled for OOBE when technically we don't have user pods
>
> but still
>>
>> subscribe to this notification.
>
>
> initialize_webui_hidden_ is set to false below in this case, so we show
> postponed UI and don't wait for the notification.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> File chrome/browser/chromeos/login/webui_login_view.cc (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/chromeos/login/...
> chrome/browser/chromeos/login/webui_login_view.cc:245: //
> TRACE_EVENT_ASYNC_BEGIN("OOBE", "SetStatusAreaVisible");
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> cleanup
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> File chrome/browser/resources/chromeos/login/user_pod_row.css (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> chrome/browser/resources/chromeos/login/user_pod_row.css:13: /* opacity:
> 1; */
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> Cleanup
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> File chrome/browser/resources/chromeos/login/user_pod_row.js (right):
>
>
http://codereview.chromium.org/10889024/diff/1/chrome/browser/resources/chrom...
> chrome/browser/resources/chromeos/login/user_pod_row.js:452:
> bootWallpaperLoaded_: false,
> On 2012/08/29 14:10:01, Nikita Kostylev wrote:
>>
>> nit: comment
>
>
> Done.
>
>
>
http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
> File ash/shell/shell_delegate_impl.h (right):
>
>
http://codereview.chromium.org/10889024/diff/2007/ash/shell/shell_delegate_im...
> ash/shell/shell_delegate_impl.h:29: virtual bool
> IsStatusAreaInitiallyVisible() const OVERRIDE;
> On 2012/08/29 17:02:42, sky wrote:
>>
>> How does this differ from IsSessionStarted and IsUserLoggedIn? Does it
>
> makes
>>
>> sense to have a single property that returns an enum?
>
>
> It's different in a sense that it's not a state (like UserLoggedIn and
> SessionStarted). OTOH, it's value is equivalent to IsSessionStarted().
> So it actually makes sense to revert it at all and use IsSessionStarted
> directly.
>
> http://codereview.chromium.org/10889024/

Awesome, thanks.

[commit-bot: I haz the power, 2012-08-30 20:09:39]

Change committed as 154242