Have browser process shutdown metro viewer on Metro -> Metro restart.

Have browser process shutdown metro viewer on Metro -> Metro restart.

Metro -> Metro restarts can happen as a result of changing about:flags, or a chrome update.  In order for the metro viewer to exit, it is necessary for the browser to notify the metro viewer that it needs to close itself.  This was happening for Metro -> Desktop restarts, but not for Metro -> Metro restarts.

BUG=332063
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=244330

[zturner, 2014-01-09 01:11:57]

cpu: win8\metro_driver\*
jschuh: ui\metro_viewer\metro_messages.h [privileged -> privileged]
sky: chrome\browser\*, ui\aura\*

[zturner, 2014-01-09 01:12:52]

+ananta for non-owners review of chrome\browser\lifetime

[jschuh, 2014-01-09 02:48:46]

ipc security lgtm

[sky, 2014-01-09 16:40:05]

https://codereview.chromium.org/130223002/diff/1/chrome/browser/browser_proce...
File chrome/browser/browser_process_platform_part_aurawin.cc (right):

https://codereview.chromium.org/130223002/diff/1/chrome/browser/browser_proce...
chrome/browser/browser_process_platform_part_aurawin.cc:63: switch (type) {
nit: no point in a switch statement here with only one value, a DCHECK is fine.

https://codereview.chromium.org/130223002/diff/1/chrome/browser/browser_proce...
chrome/browser/browser_process_platform_part_aurawin.cc:65: if
(base::win::GetVersion() >= base::win::VERSION_WIN8) {
If we only care about this for win8, why don't we only register (on line 24) for
win8?

https://codereview.chromium.org/130223002/diff/1/chrome/browser/lifetime/appl...
File chrome/browser/lifetime/application_lifetime.h (right):

https://codereview.chromium.org/130223002/diff/1/chrome/browser/lifetime/appl...
chrome/browser/lifetime/application_lifetime.h:39: void MetroExitHelper();
Description?

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
File ui/aura/remote_root_window_host_win.cc (right):

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
ui/aura/remote_root_window_host_win.cc:150: 
nit: only one newline.

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
File ui/aura/remote_root_window_host_win.h (right):

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
ui/aura/remote_root_window_host_win.h:98: AURA_EXPORT void HandleMetroExit();
Why do we need this instead of having callers do
aura::RemoteRootWindowHostWin::Instance()->HandleMetroExit(); ? I suppose this
question applies to all the Hande functions here. Also, why do these have
'Handle' in them? ExitMetro() would be a lot clearer name for this.

[zturner, 2014-01-10 19:00:25]

https://codereview.chromium.org/130223002/diff/1/chrome/browser/browser_proce...
File chrome/browser/browser_process_platform_part_aurawin.cc (right):

https://codereview.chromium.org/130223002/diff/1/chrome/browser/browser_proce...
chrome/browser/browser_process_platform_part_aurawin.cc:63: switch (type) {
On 2014/01/09 16:40:05, sky wrote:
> nit: no point in a switch statement here with only one value, a DCHECK is
fine.

Done.

https://codereview.chromium.org/130223002/diff/1/chrome/browser/browser_proce...
chrome/browser/browser_process_platform_part_aurawin.cc:65: if
(base::win::GetVersion() >= base::win::VERSION_WIN8) {
On 2014/01/09 16:40:05, sky wrote:
> If we only care about this for win8, why don't we only register (on line 24)
for
> win8?

Done.

https://codereview.chromium.org/130223002/diff/1/chrome/browser/lifetime/appl...
File chrome/browser/lifetime/application_lifetime.h (right):

https://codereview.chromium.org/130223002/diff/1/chrome/browser/lifetime/appl...
chrome/browser/lifetime/application_lifetime.h:39: void MetroExitHelper();
On 2014/01/09 16:40:05, sky wrote:
> Description?

I just removed this function instead.

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
File ui/aura/remote_root_window_host_win.cc (right):

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
ui/aura/remote_root_window_host_win.cc:150: 
On 2014/01/09 16:40:05, sky wrote:
> nit: only one newline.

Done.

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
File ui/aura/remote_root_window_host_win.h (right):

https://codereview.chromium.org/130223002/diff/1/ui/aura/remote_root_window_h...
ui/aura/remote_root_window_host_win.h:98: AURA_EXPORT void HandleMetroExit();
On 2014/01/09 16:40:05, sky wrote:
> Why do we need this instead of having callers do
> aura::RemoteRootWindowHostWin::Instance()->HandleMetroExit(); ? I suppose this
> question applies to all the Hande functions here. Also, why do these have
> 'Handle' in them? ExitMetro() would be a lot clearer name for this.

I believe it's because on component builds we have aura.dll, which hosts
RemoteRootWindowHostWin, so we need to call an exported function.

As for the Handle, I'm not sure.  I left it for consistency, but I can change it
if you feel strongly.  These are basically one-to-one mappings which translate
into "Send an IPC to the Metro Viewer that does X"

[sky, 2014-01-10 19:19:07]

LGTM

[cpu_(ooo_6.6-7.5), 2014-01-10 20:57:55]

lgtm

[commit-bot: I haz the power, 2014-01-11 00:21:27]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/zturner@chromium.org/130223002/230001

[commit-bot: I haz the power, 2014-01-11 03:11:38]

Change committed as 244330