Add support to update_engine to poke Omaha after an update has been applied

Add support to update_engine to poke Omaha after an update has been applied 
successfully and is awaiting reboot to help ensure the number of actives 
remains accurate.

BUG=chromium-os:12026
TEST=Manual test, unit tests

Change-Id: Ie3397264b0b34e8d423fb9748970f7d330122180

Committed: http://chrome-svn/viewvc/chromeos?view=rev&revision=116fda3

[thieule, 2011-04-13 20:59:38]

I've modified update_engine to send an "update complete, success reboot" event
after the update has been applied and is awaiting reboot.  It reuses the current
schedule logic for  update checks.

PTAL

[thieule, 2011-04-15 21:30:01]

Modified the code to use the <ping> element with active=1 instead of sending an
event.

PTAL

[adlr, 2011-04-15 22:04:51]

http://codereview.chromium.org/6836025/diff/4001/omaha_request_action.h
File omaha_request_action.h (right):

http://codereview.chromium.org/6836025/diff/4001/omaha_request_action.h#newco...
omaha_request_action.h:133: bool ping_only = false);
our style says we can't use default values.

http://codereview.chromium.org/6836025/diff/4001/update_attempter.cc
File update_attempter.cc (right):

http://codereview.chromium.org/6836025/diff/4001/update_attempter.cc#newcode619
update_attempter.cc:619: processor_->StartProcessing();
the other thing that starts processing uses g_idle_add(). It's probably a good
idea to match that here

http://codereview.chromium.org/6836025/diff/4001/update_attempter_unittest.cc
File update_attempter_unittest.cc (right):

http://codereview.chromium.org/6836025/diff/4001/update_attempter_unittest.cc...
update_attempter_unittest.cc:302: TEST_F(UpdateAttempterTest, PingOmahaTest) {
you should probably set up a GMainLoop in here and terminate out when the event
processing is complete.

You can also assert that attempter_.status() is correct after all events have
run

[thieule, 2011-04-15 23:32:57]

http://codereview.chromium.org/6836025/diff/4001/omaha_request_action.h
File omaha_request_action.h (right):

http://codereview.chromium.org/6836025/diff/4001/omaha_request_action.h#newco...
omaha_request_action.h:133: bool ping_only = false);
On 2011/04/15 22:04:51, adlr wrote:
> our style says we can't use default values.

Done.

http://codereview.chromium.org/6836025/diff/4001/update_attempter.cc
File update_attempter.cc (right):

http://codereview.chromium.org/6836025/diff/4001/update_attempter.cc#newcode619
update_attempter.cc:619: processor_->StartProcessing();
On 2011/04/15 22:04:51, adlr wrote:
> the other thing that starts processing uses g_idle_add(). It's probably a good
> idea to match that here

Done.

http://codereview.chromium.org/6836025/diff/4001/update_attempter_unittest.cc
File update_attempter_unittest.cc (right):

http://codereview.chromium.org/6836025/diff/4001/update_attempter_unittest.cc...
update_attempter_unittest.cc:302: TEST_F(UpdateAttempterTest, PingOmahaTest) {
On 2011/04/15 22:04:51, adlr wrote:
> you should probably set up a GMainLoop in here and terminate out when the
event
> processing is complete.
> 
> You can also assert that attempter_.status() is correct after all events have
> run

Done.

[thieule, 2011-04-16 00:48:42]

One more set of changes.

This time, it's to fix a bug I found while testing where the last ping day was
not properly updated.

PTAL

[adlr, 2011-04-18 18:05:28]

http://codereview.chromium.org/6836025/diff/5012/update_attempter_unittest.cc
File update_attempter_unittest.cc (right):

http://codereview.chromium.org/6836025/diff/5012/update_attempter_unittest.cc...
update_attempter_unittest.cc:327: g_main_loop_quit(loop_);
this still may run before the action completes. Can you get notified by the
action processor that processing has completed to exit the main loop?

[thieule, 2011-04-18 21:40:15]

http://codereview.chromium.org/6836025/diff/5012/update_attempter_unittest.cc
File update_attempter_unittest.cc (right):

http://codereview.chromium.org/6836025/diff/5012/update_attempter_unittest.cc...
update_attempter_unittest.cc:327: g_main_loop_quit(loop_);
On 2011/04/18 18:05:28, adlr wrote:
> this still may run before the action completes. Can you get notified by the
> action processor that processing has completed to exit the main loop?

It looks like the processor is a mock so it doesn't really run the action.

[adlr, 2011-04-19 02:43:51]

LGTM!

Sorry for the delay