login_manager: Add dbus API for restarting entd when enterprise extensions installed

login_manager: Add dbus API for restarting entd when enterprise extensions installed

Change-Id: I19b8533803e73ed16a157314ae6574ccd128c72a

BUG=9282
TEST=
Run (as chronos):

dbus-send --system --type=method_call --print-reply --dest=org.chromium.SessionManager /org/chromium/SessionManager 
org.chromium.SessionManagerInterface.RestartEntd

Verify entd start messages show up in ui.LATEST.
Do this with entd running and not running.
Verified entdwife.sh was getting proper username.

Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=508d0fb

[Chris Masone, 2010-11-19 03:39:37]

Also, can a brutha get a unit test?

http://codereview.chromium.org/5184005/diff/1/interface.h
File interface.h (right):

http://codereview.chromium.org/5184005/diff/1/interface.h#newcode87
interface.h:87: gboolean* OUT_restarted,
I know a lot of these have these boolean out parameters, but it's kinda legacy
and doesn't really make sense -- we didn't really know that when we set out, but
adding new stuff like this call, it probably makes sense to just not have this. 
Whatever you would have stuffed in OUT_restarted, just go ahead and return it.

http://codereview.chromium.org/5184005/diff/1/session_manager_service.cc
File session_manager_service.cc (right):

http://codereview.chromium.org/5184005/diff/1/session_manager_service.cc#newc...
session_manager_service.cc:640: // Shutdown entd if it is currently running,
blocking this thread and
because of this, when you implement the client side of this in
chromeos_login.cc, can you do it like StopSession(), using
::dbus_g_proxy_call_no_reply?  Essentially, it'll just send out a call to
session_manager to do this and not worry about the reply.

[kmixter1, 2010-11-20 03:00:05]

i'll see your unit test and raise you one integration test

http://codereview.chromium.org/5216005

http://codereview.chromium.org/5184005/diff/1/interface.h
File interface.h (right):

http://codereview.chromium.org/5184005/diff/1/interface.h#newcode87
interface.h:87: gboolean* OUT_restarted,
On 2010/11/19 03:39:37, Chris Masone wrote:
> I know a lot of these have these boolean out parameters, but it's kinda legacy
> and doesn't really make sense -- we didn't really know that when we set out,
but
> adding new stuff like this call, it probably makes sense to just not have
this. 
> Whatever you would have stuffed in OUT_restarted, just go ahead and return it.

Done.

http://codereview.chromium.org/5184005/diff/1/session_manager_service.cc
File session_manager_service.cc (right):

http://codereview.chromium.org/5184005/diff/1/session_manager_service.cc#newc...
session_manager_service.cc:640: // Shutdown entd if it is currently running,
blocking this thread and
On 2010/11/19 03:39:37, Chris Masone wrote:
> because of this, when you implement the client side of this in
> chromeos_login.cc, can you do it like StopSession(), using
> ::dbus_g_proxy_call_no_reply?  Essentially, it'll just send out a call to
> session_manager to do this and not worry about the reply.

Done.

[Chris Masone, 2010-11-20 03:17:14]

LGTM

thanks for the integration test!

On Fri, Nov 19, 2010 at 7:00 PM, <kmixter@chromium.org> wrote:

> i'll see your unit test and raise you one integration test
>
> http://codereview.chromium.org/5216005
>
>
>
> http://codereview.chromium.org/5184005/diff/1/interface.h
> File interface.h (right):
>
> http://codereview.chromium.org/5184005/diff/1/interface.h#newcode87
> interface.h:87: gboolean* OUT_restarted,
> On 2010/11/19 03:39:37, Chris Masone wrote:
>
>> I know a lot of these have these boolean out parameters, but it's
>>
> kinda legacy
>
>> and doesn't really make sense -- we didn't really know that when we
>>
> set out, but
>
>> adding new stuff like this call, it probably makes sense to just not
>>
> have this.
>
>> Whatever you would have stuffed in OUT_restarted, just go ahead and
>>
> return it.
>
> Done.
>
>
> http://codereview.chromium.org/5184005/diff/1/session_manager_service.cc
> File session_manager_service.cc (right):
>
>
>
http://codereview.chromium.org/5184005/diff/1/session_manager_service.cc#newc...
> session_manager_service.cc:640: // Shutdown entd if it is currently
> running, blocking this thread and
> On 2010/11/19 03:39:37, Chris Masone wrote:
>
>> because of this, when you implement the client side of this in
>> chromeos_login.cc, can you do it like StopSession(), using
>> ::dbus_g_proxy_call_no_reply?  Essentially, it'll just send out a call
>>
> to
>
>> session_manager to do this and not worry about the reply.
>>
>
> Done.
>
>
> http://codereview.chromium.org/5184005/
>