Add ApplicationDelegate::Quit()

Add ApplicationDelegate::Quit()

This CL adds ApplicationDelegate::Quit(), which ApplicationImpl calls
before shutting down the main run loop. In this function, an
ApplicationDelegate should shut down anything that relies on the main
run loop being active.

R=eseidel@chromium.org, jamesr@chromium.org

Committed: https://chromium.googlesource.com/external/mojo/+/e89197930bc3cdd75c494b48f8ffbdb5b6ea1a0a

[blundell, 2015-02-23 20:14:29]

blundell@chromium.org changed reviewers:
+ eseidel@chromium.org, jamesr@chromium.org

[blundell, 2015-02-23 20:14:30]

jamesr: for the cl
eseidel: for the sky naming change

[jamesr, 2015-02-23 20:27:31]

https://codereview.chromium.org/947253003/diff/1/mojo/public/cpp/application/...
File mojo/public/cpp/application/application_delegate.h (right):

https://codereview.chromium.org/947253003/diff/1/mojo/public/cpp/application/...
mojo/public/cpp/application/application_delegate.h:38: // the delegate can no
longer rely on the main run loop still running.
"the main run loop" is a feature of the chromium application runner, not the
public ApplicationDelegate type which doesn't know about it.

Can you think of a way to describe this in terms of concepts that exist in this
layer of the code?

[blundell, 2015-02-23 20:35:42]

https://codereview.chromium.org/947253003/diff/1/mojo/public/cpp/application/...
File mojo/public/cpp/application/application_delegate.h (right):

https://codereview.chromium.org/947253003/diff/1/mojo/public/cpp/application/...
mojo/public/cpp/application/application_delegate.h:38: // the delegate can no
longer rely on the main run loop still running.
On 2015/02/23 20:27:31, jamesr wrote:
> "the main run loop" is a feature of the chromium application runner, not the
> public ApplicationDelegate type which doesn't know about it.
> 
> Can you think of a way to describe this in terms of concepts that exist in
this
> layer of the code?

I was following the terminology used in the comment for
ApplicationImpl::Terminate():

// Quits the main run loop for this application.

Does that terminology similarly need abstracting?

[jamesr, 2015-02-23 20:42:14]

hmm, well I suppose this is consistent then.  ApplicationImpl::Terminate() is a
bit odd IMO but that's neither here nor there.

lgtm

[eseidel, 2015-02-23 21:07:33]

lgtm2

[eseidel, 2015-02-23 21:07:41]

lgtm too

[blundell, 2015-02-24 13:06:27]

On 2015/02/23 20:42:14, jamesr wrote:
> hmm, well I suppose this is consistent then.  ApplicationImpl::Terminate() is
a
> bit odd IMO but that's neither here nor there.
> 
> lgtm

I also find ApplicationImpl::Terminate() odd.

[blundell, 2015-02-24 13:06:54]

Committed patchset #1 (id:1) manually as
e89197930bc3cdd75c494b48f8ffbdb5b6ea1a0a (presubmit successful).