Use ChromeBrowserMainPartsAndroid.

Use ChromeBrowserMainPartsAndroid.

Includes several changes:
1) Stubs out jankometer
2) Makes some background-related profile code guarded by
ENABLE_BACKGROUND
3) Remove SingletonProcess because it's guaranteed by the OS.
4) If-def out first-run code and process startup code as it's necessary
handled in the Java layer on Android. In order to improve startup
performance, Chrome for Android runs startup code in the java layer
asynchronously with loading the native shared library. Re-using this
code would result in us blocking until the entire library is loaded into
memory.
5) Removes UMA code. UMA sessions are handled differently on Android as
the browser isn't ever killed and is instead paused/resumed.
6) Also removes things like ManagedMode/PageCycler which are unused on Android.

BUG=138184
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=154029

[Yaron, 2012-08-29 01:57:27]

erg: chrome/browser/profiles
thestig: the rest

So this does include a bunch of if-defs... I'm really not sure there's a good
abstraction to make it clearer though. There's just many things (e.g. session
service, process singleton, translate infobar, managed mode, first run) which
aren't used. We can go case-by-case and maybe pull out helper functions in the
case of some of the first_run code but I think start-up is necessarily different
on Android. I'm open to suggestions though.

As for exercising this code, it now links into unit_tests_apk although remains
largely untested. dtrainor is working on a test harness for Android that will
use this for some java instrumentation tests.

[Elliot Glaysher, 2012-08-29 17:06:25]

trivial one statement addition in profiles/ lgtm

[Lei Zhang, 2012-08-29 19:23:57]

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/browser_proc...
File chrome/browser/browser_process_impl.cc (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/browser_proc...
chrome/browser/browser_process_impl.cc:483: if (!remote_debugging_server_.get())
{
Since |remote_debugging_server_| isn't used anywhere else, care to just #ifdef
it out on Android altogether?

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
File chrome/browser/chrome_browser_main.cc (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
chrome/browser/chrome_browser_main.cc:944: void
ChromeBrowserMainParts::RunPageCycler() {
#ifdef out the method definition as well? Then you can take out the code on
lines 933-934 as well.

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
File chrome/browser/chrome_browser_main.h (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
chrome/browser/chrome_browser_main.h:166: scoped_ptr<StartupBrowserCreator>
browser_creator_;
Are the StartupBrowserCreator class and ProcessSingleton class built on Android?
If yes, you may want to exclude them.

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/jankometer_a...
File chrome/browser/jankometer_android.cc (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/jankometer_a...
chrome/browser/jankometer_android.cc:5: #include "chrome/browser/jankometer.h"
Why is this file an 'A+' an not just an 'A' ? Is git confused? I think it will
end up showing the history for panel_strip.cc even though it is suppose to be a
new file.

[Yaron, 2012-08-29 20:48:33]

comments addressed, includes a rebase although only affects gyp. Sorry, though.
Lei: ptal

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/browser_proc...
File chrome/browser/browser_process_impl.cc (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/browser_proc...
chrome/browser/browser_process_impl.cc:483: if (!remote_debugging_server_.get())
{
On 2012/08/29 19:23:57, Lei Zhang wrote:
> Since |remote_debugging_server_| isn't used anywhere else, care to just #ifdef
> it out on Android altogether?

Done.

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
File chrome/browser/chrome_browser_main.cc (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
chrome/browser/chrome_browser_main.cc:944: void
ChromeBrowserMainParts::RunPageCycler() {
On 2012/08/29 19:23:57, Lei Zhang wrote:
> #ifdef out the method definition as well? Then you can take out the code on
> lines 933-934 as well.

Was trying to minimize if-defs but sure it's more complete that way.

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
File chrome/browser/chrome_browser_main.h (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/chrome_brows...
chrome/browser/chrome_browser_main.h:166: scoped_ptr<StartupBrowserCreator>
browser_creator_;
On 2012/08/29 19:23:57, Lei Zhang wrote:
> Are the StartupBrowserCreator class and ProcessSingleton class built on
Android?
> If yes, you may want to exclude them.

StartupBrowserCreator was already excluded via gyp exclude. Removed
ProcessSingleton as well.

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/jankometer_a...
File chrome/browser/jankometer_android.cc (right):

http://codereview.chromium.org/10885023/diff/1009/chrome/browser/jankometer_a...
chrome/browser/jankometer_android.cc:5: #include "chrome/browser/jankometer.h"
On 2012/08/29 19:23:57, Lei Zhang wrote:
> Why is this file an 'A+' an not just an 'A' ? Is git confused? I think it will
> end up showing the history for panel_strip.cc even though it is suppose to be
a
> new file.

I think it's just git being confused. I did a local copy from jankometer_mac.cc
and then a git add of the file.

[Lei Zhang, 2012-08-29 20:53:35]

lgtm

[commit-bot: I haz the power, 2012-08-29 22:43:11]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/yfriedman@chromium.org/10885023/8001