Add support to auto-shutdown on lid close if no one is logged in

Add support to auto-shutdown on lid close if no one is logged in

Since there should be no background processing going on and boot is fast, we
shouldn't waste battery sleeping when no one is logged in.  This is implemented
using simple state files and upstart events. It is simplistic but minimizes
suspend dependencies at the cost of some minor login+lid-close raciness.

As a side benefit, this ensures that if a user types in a passphrase, then
shuts the lid before pressing enter, the password won't stay resident in the
password box for the next person that opens it.

BUG=1780
TEST=none

[Chris Masone, 2010-03-08 22:32:20]

http://codereview.chromium.org/671023/diff/1/4
File src/platform/init/logout.conf (right):

http://codereview.chromium.org/671023/diff/1/4#newcode8
src/platform/init/logout.conf:8: # TODO(wad) should this be triggered on stopped
ui instead?
that's very possible.  I think this would happen _before_ we try to take down
the UI.  
(see
https://groups.google.com/a/chromium.org/group/chromium-os-dev/browse_thread/...)
Which would mean that if you trigger a log out and then close the lid, it could
shut down before chrome finishes writing state to disk, which would lead to
"chrome shut down uncleanly" infobars.

[Will Drewry, 2010-03-08 22:56:14]

http://codereview.chromium.org/671023/diff/1/4
File src/platform/init/logout.conf (right):

http://codereview.chromium.org/671023/diff/1/4#newcode8
src/platform/init/logout.conf:8: # TODO(wad) should this be triggered on stopped
ui instead?
On 2010/03/08 22:32:21, cmasone wrote:
> that's very possible.  I think this would happen _before_ we try to take down
> the UI.  
> (see
>
https://groups.google.com/a/chromium.org/group/chromium-os-dev/browse_thread/...)
> Which would mean that if you trigger a log out and then close the lid, it
could
> shut down before chrome finishes writing state to disk, which would lead to
> "chrome shut down uncleanly" infobars.

Hmm well it just calls shutdown -h now which is all the power button does now --
that's supposed to avoid that case already.

However, no reason to worry about it.  I'll change it to stopped and if we find
some reason to bump it up, we can do it later

[Chris Masone, 2010-03-08 23:00:20]

lgtm