kiosk: Fix kiosk session restart

kiosk: Fix kiosk session restart

Make kiosk apps part of IsKnownUser. Otherwise, a kiosk session
is terminated in the crash-and-restart case.

BUG=599973
TEST=KioskCrashRestoreTest.Basic

Committed: https://crrev.com/d4f0457c76a3bf8bd98a75a220bd5fb9475b51ef
Cr-Commit-Position: refs/heads/master@{#388249}

[xiyuan, 2016-04-06 21:08:17]

xiyuan@chromium.org changed reviewers:
+ bartfab@chromium.org

[bartfab (slow), 2016-04-07 14:55:42]

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
File chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc
(right):

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:45:
CHECK(PathService::Get(chromeos::DIR_DEVICE_EXTENSION_LOCAL_CACHE,
Should we not override this path to avoid littering the actual file system with
test files?

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:59:
~KioskCrashRestoreTest() override {}
Nit: Do you really need to override this?

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:105:
em::DeviceLocalAccountsProto* device_local_accounts =
Nit: const pointer.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:107:
device_local_accounts->clear_account();
Can you not safely assume that the payload starts out empty?

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:109:
em::DeviceLocalAccountInfoProto* account =
Nit: const pointer.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:125: const
std::string local_state_json = base::StringPrintf(
This looks odd. There are many tests that stub out the policy subsystem very
similarly to the way you do but I do not remember any of them having to inject
policies into the local state file.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:136:
NetworkPortalDetectorTestImpl* network_portal_detector =
Nit: const pointer.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:160:
scoped_ptr<SessionManagerClient>(session_manager_client_));
Nit: s/scoped_ptr/std::unique_ptr/

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:168:
scoped_ptr<FakeCWS> fake_cws_;
Nit: s/scoped_ptr/std::unique_ptr/

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:177:
Profile* app_profile = ProfileManager::GetPrimaryUserProfile();
Nit: const pointer.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:179:
extensions::AppWindowRegistry* app_window_registry =
Nit: const pointer.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:181:
extensions::AppWindow* window =
Nit: const pointer.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:183:
EXPECT_TRUE(window);
Nit: s/EXPECT/ASSERT/

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/chr...
File chrome/browser/chromeos/chrome_browser_main_chromeos.cc (right):

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/chr...
chrome/browser/chromeos/chrome_browser_main_chromeos.cc:536:
!user_manager->IsKnownUser(account_id) &&
Why is the kiosk user not a known user? We explicitly add public session users
to the known user list. Is it not just an oversight that we skip kiosk users?

[xiyuan, 2016-04-07 21:43:57]

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
File chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc
(right):

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:45:
CHECK(PathService::Get(chromeos::DIR_DEVICE_EXTENSION_LOCAL_CACHE,
On 2016/04/07 14:55:41, bartfab (slow) wrote:
> Should we not override this path to avoid littering the actual file system
with
> test files?

This is overridden under user data dir when not running on the real device.
[1][2] Browser tests uses a ScopeTempDir for user data dir so we will not
pollute the fs.

[1]:
https://code.google.com/p/chromium/codesearch#chromium/src/chrome/browser/chr...
[2]: 
https://code.google.com/p/chromium/codesearch#chromium/src/chromeos/chromeos_...

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:59:
~KioskCrashRestoreTest() override {}
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> Nit: Do you really need to override this?

Removed.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:105:
em::DeviceLocalAccountsProto* device_local_accounts =
On 2016/04/07 14:55:41, bartfab (slow) wrote:
> Nit: const pointer.

Done.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:107:
device_local_accounts->clear_account();
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> Can you not safely assume that the payload starts out empty?

Removed.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:109:
em::DeviceLocalAccountInfoProto* account =
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> Nit: const pointer.

Done.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:125: const
std::string local_state_json = base::StringPrintf(
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> This looks odd. There are many tests that stub out the policy subsystem very
> similarly to the way you do but I do not remember any of them having to inject
> policies into the local state file.

The code that needs to be tested runs early in PreProfileInit. At this time,
policy system has not yet fully loaded. Device policy cache should be populated
in real system and KioskAppManager is initialized with the cached values. The
code here simulates that.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:136:
NetworkPortalDetectorTestImpl* network_portal_detector =
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> Nit: const pointer.

Done.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:160:
scoped_ptr<SessionManagerClient>(session_manager_client_));
On 2016/04/07 14:55:41, bartfab (slow) wrote:
> Nit: s/scoped_ptr/std::unique_ptr/

Done.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:168:
scoped_ptr<FakeCWS> fake_cws_;
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> Nit: s/scoped_ptr/std::unique_ptr/

Done.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/app...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:183:
EXPECT_TRUE(window);
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> Nit: s/EXPECT/ASSERT/

Done.

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/chr...
File chrome/browser/chromeos/chrome_browser_main_chromeos.cc (right):

https://codereview.chromium.org/1865133002/diff/1/chrome/browser/chromeos/chr...
chrome/browser/chromeos/chrome_browser_main_chromeos.cc:536:
!user_manager->IsKnownUser(account_id) &&
On 2016/04/07 14:55:42, bartfab (slow) wrote:
> Why is the kiosk user not a known user? We explicitly add public session users
> to the known user list. Is it not just an oversight that we skip kiosk users?

Makes sense to make kiosk apps part of known users. I will do it as the fix.

[bartfab (slow), 2016-04-08 09:44:16]

https://codereview.chromium.org/1865133002/diff/20001/chrome/browser/chromeos...
File chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc (right):

https://codereview.chromium.org/1865133002/diff/20001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:566: const
std::vector<policy::DeviceLocalAccount> device_local_accounts =
We need to persist the current list of public sessions in a pref so that
UpdateAndCleanUpPublicAccounts() can work: Whenever the list of public sessions
configured through policy changes, we need to know what the old list was so that
we can clean up removed sessions.

[xiyuan, 2016-04-08 15:52:14]

https://codereview.chromium.org/1865133002/diff/20001/chrome/browser/chromeos...
File chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc (right):

https://codereview.chromium.org/1865133002/diff/20001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:566: const
std::vector<policy::DeviceLocalAccount> device_local_accounts =
On 2016/04/08 09:44:15, bartfab (slow) wrote:
> We need to persist the current list of public sessions in a pref so that
> UpdateAndCleanUpPublicAccounts() can work: Whenever the list of public
sessions
> configured through policy changes, we need to know what the old list was so
that
> we can clean up removed sessions.

We don't really need kPublicAccounts because device settings cache has a copy of
that. If we don't enforce trusted state, we can get the cache data here via
policy::GetDeviceLocalAccounts. And UpdateAndCleanUpPublicAccounts() would just
work since it gets the old list from |users_|, which we populate here.

[bartfab (slow), 2016-04-11 10:30:23]

On 2016/04/08 15:52:14, xiyuan wrote:
>
https://codereview.chromium.org/1865133002/diff/20001/chrome/browser/chromeos...
> File chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc (right):
> 
>
https://codereview.chromium.org/1865133002/diff/20001/chrome/browser/chromeos...
> chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:566: const
> std::vector<policy::DeviceLocalAccount> device_local_accounts =
> On 2016/04/08 09:44:15, bartfab (slow) wrote:
> > We need to persist the current list of public sessions in a pref so that
> > UpdateAndCleanUpPublicAccounts() can work: Whenever the list of public
> sessions
> > configured through policy changes, we need to know what the old list was so
> that
> > we can clean up removed sessions.
> 
> We don't really need kPublicAccounts because device settings cache has a copy
of
> that. If we don't enforce trusted state, we can get the cache data here via
> policy::GetDeviceLocalAccounts. And UpdateAndCleanUpPublicAccounts() would
just
> work since it gets the old list from |users_|, which we populate here.

That would be racey. What if by the time we call
policy::GetDeviceLocalAccounts(), we actually do get updated policy, not the
cached previous state? Only if we explicitly store the last-known-processed
state outside the policy stack can we be sure that we know what we processed.

[xiyuan, 2016-04-11 17:01:17]

On 2016/04/11 10:30:23, bartfab (slow) wrote:
> That would be racey. What if by the time we call
> policy::GetDeviceLocalAccounts(), we actually do get updated policy, not the
> cached previous state? Only if we explicitly store the last-known-processed
> state outside the policy stack can we be sure that we know what we processed.

Makes sense. Added back the "PublicAccounts" local state.

[bartfab (slow), 2016-04-12 15:08:13]

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
File chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc
(right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:121: //
Prepare the policy data store in device policy cache.
Nit: s/store/to store/

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
File chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc (right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:95: // A string
pref that gets set when a public account is removed but a user is
Nit: s/public/device local/

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:127: bool
IsDeviceLocalAccountUser(const user_manager::User* user) {
This should be a getter on User, where it can be implemented in
DeviceLocalAccountUserBase.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:926: // If the
user is logged into a public account that has been removed from the
Nit: s/public/device local/

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:928: if
(IsLoggedInAsPublicAccount()) {
This will need changing to catch all device-local accounts.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:950: // Try to
remove any public account data marked as pending removal.
Nit: s/public/device local/

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:994:
IsLoggedInAsPublicAccount() || IsLoggedInAsKioskApp();
It would be more future-proof to either create an
IsLoggedInAsDeviceLocalAccount() getter or use
IsDeviceLocalAccountUser(GetActiveUser()).

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:1294:
user_manager::User* user = nullptr;
Nit: Use a unique_ptr.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
File chrome/browser/chromeos/login/users/chrome_user_manager_impl.h (right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.h:204: const
std::vector<std::string>& old_public_accounts);
Nit: s/old_public_accounts/old_device_local_accounts/

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.h:244:
user_manager::User* CreateUserFromDeviceLocalAccount(
Nit: Return an std::unique_ptr to indicate ownership transfer.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/extensio...
File chrome/browser/extensions/api/file_system/file_system_api_unittest.cc
(right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/extensio...
chrome/browser/extensions/api/file_system/file_system_api_unittest.cc:401:
user_manager::User* manual_kiosk_app_user = user_manager_->AddKioskAppUser(
Nit 1: const pointer.
Nit 2: #include "components/user_manager/user.h"

[xiyuan, 2016-04-12 16:22:51]

Sorry for the rebasing. I want to pull in the std::unique_ptr changes.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
File chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc
(right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/app_mode/kiosk_crash_restore_browsertest.cc:121: //
Prepare the policy data store in device policy cache.
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit: s/store/to store/

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
File chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc (right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:95: // A string
pref that gets set when a public account is removed but a user is
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit: s/public/device local/

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:127: bool
IsDeviceLocalAccountUser(const user_manager::User* user) {
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> This should be a getter on User, where it can be implemented in
> DeviceLocalAccountUserBase.

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:926: // If the
user is logged into a public account that has been removed from the
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit: s/public/device local/

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:928: if
(IsLoggedInAsPublicAccount()) {
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> This will need changing to catch all device-local accounts.

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:950: // Try to
remove any public account data marked as pending removal.
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit: s/public/device local/

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:994:
IsLoggedInAsPublicAccount() || IsLoggedInAsKioskApp();
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> It would be more future-proof to either create an
> IsLoggedInAsDeviceLocalAccount() getter or use
> IsDeviceLocalAccountUser(GetActiveUser()).

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.cc:1294:
user_manager::User* user = nullptr;
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit: Use a unique_ptr.

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
File chrome/browser/chromeos/login/users/chrome_user_manager_impl.h (right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.h:204: const
std::vector<std::string>& old_public_accounts);
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit: s/old_public_accounts/old_device_local_accounts/

Done.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/chromeos...
chrome/browser/chromeos/login/users/chrome_user_manager_impl.h:244:
user_manager::User* CreateUserFromDeviceLocalAccount(
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit: Return an std::unique_ptr to indicate ownership transfer.

Done. Also moved User's ctor/dtor to public to work with std::unique_ptr.

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/extensio...
File chrome/browser/extensions/api/file_system/file_system_api_unittest.cc
(right):

https://codereview.chromium.org/1865133002/diff/60001/chrome/browser/extensio...
chrome/browser/extensions/api/file_system/file_system_api_unittest.cc:401:
user_manager::User* manual_kiosk_app_user = user_manager_->AddKioskAppUser(
On 2016/04/12 15:08:13, bartfab (slow) wrote:
> Nit 1: const pointer.
> Nit 2: #include "components/user_manager/user.h"

Done.

[xiyuan, 2016-04-12 16:24:47]

Description was changed from

==========
kiosk: Fix kiosk session restart

Kiosk apps are part of device local accounts but not part of IsKnownUser.
The end session condition for no-longer existing device local accounts
should include kiosk apps as well. Otherwise, the session is terminated
when restarting a kiosk session.

BUG=599973
TEST=KioskCrashRestoreTest.Basic
==========

to

==========
kiosk: Fix kiosk session restart

Make kiosk apps part of IsKnownUser. Otherwise, a kiosk session
is terminated in the crash-and-restart case.

BUG=599973
TEST=KioskCrashRestoreTest.Basic
==========

[xiyuan, 2016-04-15 17:50:15]

ping?

[bartfab (slow), 2016-04-18 13:09:25]

lgtm

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
File components/user_manager/user.h (right):

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
components/user_manager/user.h:118: virtual bool IsDeviceLocalAccountUser()
const;
Nit: The "User" is redundant.

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
File components/user_manager/user_manager_base.cc (right):

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
components/user_manager/user_manager_base.cc:599: // public accounts whose data
has not been removed yet is not ephemeral.
Nit: s/public/device local/

[xiyuan, 2016-04-18 16:53:27]

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
File components/user_manager/user.h (right):

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
components/user_manager/user.h:118: virtual bool IsDeviceLocalAccountUser()
const;
On 2016/04/18 13:09:25, bartfab (slow) wrote:
> Nit: The "User" is redundant.

Done.

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
File components/user_manager/user_manager_base.cc (right):

https://codereview.chromium.org/1865133002/diff/80001/components/user_manager...
components/user_manager/user_manager_base.cc:599: // public accounts whose data
has not been removed yet is not ephemeral.
On 2016/04/18 13:09:25, bartfab (slow) wrote:
> Nit: s/public/device local/

Done.

[bartfab (slow), 2016-04-18 16:57:48]

lgtm

[xiyuan, 2016-04-18 17:17:38]

xiyuan@chromium.org changed reviewers:
+ achuith@chromium.org, mtomasz@chromium.org

[xiyuan, 2016-04-18 17:17:39]

+achuith for owner's review:
  /components/user_manager/*
+mtomasz for owner's review:
  c/b/extensions/api/file_system/file_system_api_unittest.cc

Thanks.

[achuithb, 2016-04-18 17:22:29]

On 2016/04/18 17:17:39, xiyuan wrote:
> +achuith for owner's review:
>   /components/user_manager/*

lgtm for /components/user_manager

[mtomasz, 2016-04-19 01:50:31]

file_system_api_unittest.cc lgtm

[xiyuan, 2016-04-19 15:48:02]

The CQ bit was checked by xiyuan@chromium.org

[commit-bot: I haz the power, 2016-04-19 15:48:13]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1865133002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1865133002/100001

[commit-bot: I haz the power, 2016-04-19 18:23:04]

Description was changed from

==========
kiosk: Fix kiosk session restart

Make kiosk apps part of IsKnownUser. Otherwise, a kiosk session
is terminated in the crash-and-restart case.

BUG=599973
TEST=KioskCrashRestoreTest.Basic
==========

to

==========
kiosk: Fix kiosk session restart

Make kiosk apps part of IsKnownUser. Otherwise, a kiosk session
is terminated in the crash-and-restart case.

BUG=599973
TEST=KioskCrashRestoreTest.Basic
==========

[commit-bot: I haz the power, 2016-04-19 18:23:05]

Committed patchset #6 (id:100001)

[commit-bot: I haz the power, 2016-04-22 19:13:54]

Description was changed from

==========
kiosk: Fix kiosk session restart

Make kiosk apps part of IsKnownUser. Otherwise, a kiosk session
is terminated in the crash-and-restart case.

BUG=599973
TEST=KioskCrashRestoreTest.Basic
==========

to

==========
kiosk: Fix kiosk session restart

Make kiosk apps part of IsKnownUser. Otherwise, a kiosk session
is terminated in the crash-and-restart case.

BUG=599973
TEST=KioskCrashRestoreTest.Basic

Committed: https://crrev.com/d4f0457c76a3bf8bd98a75a220bd5fb9475b51ef
Cr-Commit-Position: refs/heads/master@{#388249}
==========

[commit-bot: I haz the power, 2016-04-22 19:13:55]

Patchset 6 (id:??) landed as
https://crrev.com/d4f0457c76a3bf8bd98a75a220bd5fb9475b51ef
Cr-Commit-Position: refs/heads/master@{#388249}