Removed old login screen from source.

Removed old login screen from source.
Deleted kEnableLoginImages flag.

BUG=chromium-os:7214
TEST=Old login screen should not show up even under automation tests for login.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=71313

[whywhat, 2010-12-14 10:24:58]

PTAL, I fixed browsertest.

[Dmitry Polukhin, 2010-12-14 14:06:18]

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/existing_user_controller.h (right):

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/existing_user_controller.h:166: static
ExistingUserController* current_controller_;
Yeah one more static pointer to current controller :(

It looks like delete_scheduled_instance_ can be removed and current_controller_
should be enough.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/login_screen_browsertest.cc (left):

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/login_screen_browsertest.cc:76:
IN_PROC_BROWSER_TEST_F(LoginScreenTest, TestBasic) {
It would be good to add in-process tests for new login - we don't have them at
all.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/wizard_controller.cc (right):

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/wizard_controller.cc:562: if (controller)
It should be DCHECK now.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/wizard_controller.cc:908: if
(first_screen_name.empty() &&
Also we need to check that first screen is login. Otherwise we will create
wizard controller that will try to show login screen and create login existing
user controller. Let's optimize this path here.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/wizard_controller.h (right):

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/wizard_controller.h:255: bool is_test_mode_;
It looks like we don't need it anymore.

http://codereview.chromium.org/5809001/diff/6001/chrome/common/chrome_switche...
File chrome/common/chrome_switches.cc (left):

http://codereview.chromium.org/5809001/diff/6001/chrome/common/chrome_switche...
chrome/common/chrome_switches.cc:1216: const char kEnableLoginImages[]          
  = "enable-login-images";
You will also need to remove it from /sbin/session_manager_setup.sh

[whywhat, 2010-12-16 10:47:52]

Working on the test.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/existing_user_controller.h (right):

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/existing_user_controller.h:166: static
ExistingUserController* current_controller_;
On 2010/12/14 14:06:18, Dmitry Polukhin wrote:
> Yeah one more static pointer to current controller :(
> 
> It looks like delete_scheduled_instance_ can be removed and
current_controller_
> should be enough.

Done.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/wizard_controller.cc (right):

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/wizard_controller.cc:562: if (controller)
On 2010/12/14 14:06:18, Dmitry Polukhin wrote:
> It should be DCHECK now.

Done.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/wizard_controller.cc:908: if
(first_screen_name.empty() &&
On 2010/12/14 14:06:18, Dmitry Polukhin wrote:
> Also we need to check that first screen is login. Otherwise we will create
> wizard controller that will try to show login screen and create login existing
> user controller. Let's optimize this path here.

Done.

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
File chrome/browser/chromeos/login/wizard_controller.h (right):

http://codereview.chromium.org/5809001/diff/6001/chrome/browser/chromeos/logi...
chrome/browser/chromeos/login/wizard_controller.h:255: bool is_test_mode_;
On 2010/12/14 14:06:18, Dmitry Polukhin wrote:
> It looks like we don't need it anymore.

Done.

[whywhat, 2011-01-12 14:05:56]

Added a new user login testcase. Please, take another look.

[Dmitry Polukhin, 2011-01-13 09:17:28]

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
File chrome/browser/chromeos/login/existing_user_controller_browsertest.cc
(right):

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
chrome/browser/chromeos/login/existing_user_controller_browsertest.cc:32:
MockLoginPerformerDelegate(ExistingUserController* controller)
Please add explicit.

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
chrome/browser/chromeos/login/existing_user_controller_browsertest.cc:65:
virtual ~ExistingUserControllerTest() {}
There is no point in such destructor - it will be generated by compiler
automatically.

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
chrome/browser/chromeos/login/existing_user_controller_browsertest.cc:71:
MockCryptohomeLibrary* mock_cryptohome_library_;
Could you please don't mix data-members and function-members. Data-members are
visible inside functions even if they textually declared behind in source.
Parsing of member-function bodies is delayed until full class definition is
parsed (it's compiler magic).

http://codereview.chromium.org/5809001/diff/22001/chrome/common/chrome_switch...
File chrome/common/chrome_switches.cc (left):

http://codereview.chromium.org/5809001/diff/22001/chrome/common/chrome_switch...
chrome/common/chrome_switches.cc:1224: const char kEnableLoginImages[]          
  = "enable-login-images";
You will also need to remove the switches from /sbin/session_manager_setup.sh
file.

[whywhat, 2011-01-13 10:21:56]

Will send a change for review that removes the flag from
session_manager_setup.sh shortly.

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
File chrome/browser/chromeos/login/existing_user_controller_browsertest.cc
(right):

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
chrome/browser/chromeos/login/existing_user_controller_browsertest.cc:32:
MockLoginPerformerDelegate(ExistingUserController* controller)
On 2011/01/13 09:17:28, Dmitry Polukhin wrote:
> Please add explicit.

Done.

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
chrome/browser/chromeos/login/existing_user_controller_browsertest.cc:65:
virtual ~ExistingUserControllerTest() {}
On 2011/01/13 09:17:28, Dmitry Polukhin wrote:
> There is no point in such destructor - it will be generated by compiler
> automatically.

Done.

http://codereview.chromium.org/5809001/diff/22001/chrome/browser/chromeos/log...
chrome/browser/chromeos/login/existing_user_controller_browsertest.cc:71:
MockCryptohomeLibrary* mock_cryptohome_library_;
On 2011/01/13 09:17:28, Dmitry Polukhin wrote:
> Could you please don't mix data-members and function-members. Data-members are
> visible inside functions even if they textually declared behind in source.
> Parsing of member-function bodies is delayed until full class definition is
> parsed (it's compiler magic).

Done.

[Dmitry Polukhin, 2011-01-13 10:23:51]

LGTM