Turn on Synced Notification Sending services by default for all users, and save preferences.

Turn on Synced Notification Sending services by default for all users, and save preferences.

This is for the new Synced Notification feature, which allows servers to send notification toasts to the chrome message center (also called the notification center).

As things stand today (before this checkin), Synced Notifications is enabled, but each data type from an individual server needs to be intentionally enabled by the user finding the notification center settings panel and clicking a check box for that service to the "enabled" state.

After lots of rounds of UI discussion we decided the best experience for the user is to default the sending services to enabled instead of the way it defaults to disabled today.

So, the goal of this checkin is just to change the state of the existing synced notification type from disabled to enabled by default.

As it turns out, it wasn't as easy as just changing false to true somewhere.  We opted (after the first round of code review) for a scheme that maintains two preference lists, a list of enabled synced notification services (which we will now add our first sending service to, and in the future we plan to add other services), and a list of preferences that we have "initialized" (or basically turned them on by default once).  We need the second list because we are turning the service on at runtime, and if the user later wants to turn the service off, we need to remember and not just turn it on again.  I intended to describe this part adequately in my changelist, but apparently I didn't describe it well enough, and I'll try to amend the description.

BUG=280250
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=224080

[Pete Williamson, 2013-09-04 22:10:32]

BauerB@ Please review browser_prefs.cc
DewittJ@ Please review everything else

There is a Synced Notifications feature that will present notifications sent
from the server on desktop machines running chrome in the chrome notification
center.  Currently, it is turned off by default.  This change will turn it on by
default for our dogfood users.

The way this change works is to store two preferences, one remembering if we
have ever turned on notifications for G+ before, and one with the current state.

If we have not turned on notifications for G+, then we will turn them on, and
remember that.
If we have ever turned on notifications for G+, we will do nothing, respecting
the state that the user has set.

[Bernhard Bauer, 2013-09-04 22:18:40]

c/b/p LGTM.

[dewittj, 2013-09-05 15:38:44]

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File chrome/browser/notifications/message_center_settings_controller.cc (left):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/message_center_settings_controller.cc:324:
notification_service->SetNotifierEnabled(notifier.notifier_id, enabled);
This affects all types besides webpage, you probably only are trying to remove
it for SYNCED_NOTIFICATION_SERVICE

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc
(right):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:504: //
TODO(reviewers): Is this the right way to save state?  The idea here is to
I prefer a StringListPrefMember like you have in DesktopNotificationService.  In
fact, I'd prefer if you pulled enabled_sync_notifier_id_pref_ into this class as
well.

I think also the particular name you're choosing is not ideal, given that this
is a global, human-readable string that'll be found in the prefs file.

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File
chrome/browser/notifications/sync_notifier/chrome_notifier_service_unittest.cc
(right):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/chrome_notifier_service_unittest.cc:48:
const char kServiceEnabledOnce[] = "ServiceEnabledOnce-";
These should not be copied, they should come from the same string constant that
non-test code is using.

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File chrome/browser/notifications/sync_notifier/synced_notification.cc (right):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/synced_notification.cc:117: //
TODO(petewil): Sometimes on the mac bitmaps can arrive twice, find out why.
If bitmaps arrive twice, this is no longer sufficient, you will need to maintain
a set of pointers or something, I guess.

Below, active_fetcher_count_ is still tested for 0ness before calling Show, but
if there are say 2 fetchers, and the first one calls this twice, then you will
Show before the second one is done.

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/synced_notification.cc:184: LOG(INFO)
<< "****** fetching bitmap " << url;
Remove log statement.

[Pete Williamson, 2013-09-09 17:59:06]

CR fixes per dewittJ

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File chrome/browser/notifications/message_center_settings_controller.cc (left):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/message_center_settings_controller.cc:324:
notification_service->SetNotifierEnabled(notifier.notifier_id, enabled);
On 2013/09/05 15:38:44, dewittj wrote:
> This affects all types besides webpage, you probably only are trying to remove
> it for SYNCED_NOTIFICATION_SERVICE

Done.

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc
(right):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:504: //
TODO(reviewers): Is this the right way to save state?  The idea here is to
On 2013/09/05 15:38:44, dewittj wrote:
> I prefer a StringListPrefMember like you have in DesktopNotificationService. 
In
> fact, I'd prefer if you pulled enabled_sync_notifier_id_pref_ into this class
as
> well.
> 
> I think also the particular name you're choosing is not ideal, given that this
> is a global, human-readable string that'll be found in the prefs file.

Thanks for the feedback!

1. I renamed the string, see if you like the new name better.
2. Just curious, why do you think that a StringListPrefMember is better?  I just
need true/false values.  I'm certainly willing to make the change, I just wanted
to understand it a bit better first.
3. Jun added enabled_sync_notifier_id_pref_ to track which synced notification
services were enabled.  While we can move it into my class, it will change and
likely complicate the interface to the desktop notification service class, and
the code may be simpler as is.  Would you still like me to make the change?

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File
chrome/browser/notifications/sync_notifier/chrome_notifier_service_unittest.cc
(right):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/chrome_notifier_service_unittest.cc:48:
const char kServiceEnabledOnce[] = "ServiceEnabledOnce-";
On 2013/09/05 15:38:44, dewittj wrote:
> These should not be copied, they should come from the same string constant
that
> non-test code is using.

Done.

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
File chrome/browser/notifications/sync_notifier/synced_notification.cc (right):

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/synced_notification.cc:117: //
TODO(petewil): Sometimes on the mac bitmaps can arrive twice, find out why.
On 2013/09/05 15:38:44, dewittj wrote:
> If bitmaps arrive twice, this is no longer sufficient, you will need to
maintain
> a set of pointers or something, I guess.
> 
> Below, active_fetcher_count_ is still tested for 0ness before calling Show,
but
> if there are say 2 fetchers, and the first one calls this twice, then you will
> Show before the second one is done.

OK, I now check to make sure we have at least one bitmap for everything.  In the
future, I'll investigate more, and make sure we use the proper DPI bitmap.

https://codereview.chromium.org/23534048/diff/1/chrome/browser/notifications/...
chrome/browser/notifications/sync_notifier/synced_notification.cc:184: LOG(INFO)
<< "****** fetching bitmap " << url;
On 2013/09/05 15:38:44, dewittj wrote:
> Remove log statement.

Done.

[dewittj, 2013-09-09 18:13:16]

> On 2013/09/05 15:38:44, dewittj wrote:
> > I prefer a StringListPrefMember like you have in DesktopNotificationService.

> In
> > fact, I'd prefer if you pulled enabled_sync_notifier_id_pref_ into this
class
> as
> > well.
> > 
> > I think also the particular name you're choosing is not ideal, given that
this
> > is a global, human-readable string that'll be found in the prefs file.
> 
> Thanks for the feedback!
> 
> 1. I renamed the string, see if you like the new name better.
> 2. Just curious, why do you think that a StringListPrefMember is better?  I
just
> need true/false values.  I'm certainly willing to make the change, I just
wanted
> to understand it a bit better first.
> 3. Jun added enabled_sync_notifier_id_pref_ to track which synced notification
> services were enabled.  While we can move it into my class, it will change and
> likely complicate the interface to the desktop notification service class, and
> the code may be simpler as is.  Would you still like me to make the change?
>

re: #2
I think that what you need is actually a set membership problem.  I'm thinking
that the number of prefs is unbounded with your current scheme, and they use
string prefixes as a way to group themselves.  Since prefs are structured as
json, it makes more sense to have one list of enabled services instead of a
bunch of ungrouped string prefs.

re: #3 
Ultimately this is a design issue; I would like sync to contain its own prefs
rather than DesktopNotificationService (which really services W3C web
notifications and has had other things shoehorned in.)  I could support leaving
the pref here (so that dogfood1 happens sooner) if you made another bug that
blocks launch to refactor the sync prefs into the synced notification service
classes.

[Pete Williamson, 2013-09-13 00:54:16]

thakis@ - Please look at prefs.cc and prefs.h (Feel free to look at more if you
like, but somebody else is covering the rest of the changelist). 

dewittj@ - This should address all your comments.

[Nico, 2013-09-13 04:29:03]

I don't understand what this CL does after reading CL description and bug.

[Pete Williamson, 2013-09-14 00:08:43]

On 2013/09/13 04:29:03, Nico wrote:
> I don't understand what this CL does after reading CL description and bug.

Description updated with more context.  Sorry about that.

[Nico, 2013-09-14 00:22:36]

lgtm

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
File chrome/browser/notifications/sync_notifier/synced_notification.cc (right):

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/synced_notification.cc:92: // dpi
bitmap when appropriate.
(the way images work on mac is that an image can have multiple
"representations"/bitmaps. So an image would have an 1x and a 2x bitmap, and
then the system picks the right one at draw time based on the resolution of the
monitor the window is on. this allows dragging windows between hi and lodpi
monitors)

[dewittj, 2013-09-16 18:08:24]

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc
(right):

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:42: const
char kServiceEnabledOnce[] = "NotificationServiceInitialized-";
This is unused now.

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:527: const
base::ListValue* pref_list = profile_->GetPrefs()->GetList(pref_name);
You can just access the list pref member variable's GetValue here.

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.h
(right):

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.h:135:
std::set<std::string> services, base::ListValue& list_value);
We only pass const refs - use a pointer instead for object semantics (or return
an object)

[Pete Williamson, 2013-09-16 21:01:08]

Changed to use the pref directly instead of iterating it, several other small
fixes.

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc
(right):

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:42: const
char kServiceEnabledOnce[] = "NotificationServiceInitialized-";
On 2013/09/16 18:08:24, dewittj wrote:
> This is unused now.

Done.

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:527: const
base::ListValue* pref_list = profile_->GetPrefs()->GetList(pref_name);
On 2013/09/16 18:08:24, dewittj wrote:
> You can just access the list pref member variable's GetValue here.

Done.

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.h
(right):

https://codereview.chromium.org/23534048/diff/21001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.h:135:
std::set<std::string> services, base::ListValue& list_value);
On 2013/09/16 18:08:24, dewittj wrote:
> We only pass const refs - use a pointer instead for object semantics (or
return
> an object)

Done.

[dewittj, 2013-09-17 01:05:21]

lgtm with nits

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc
(right):

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:467:
profile_->GetPrefs()->Set(prefs::kEnabledSyncedNotificationSendingServices,
Can we just use the StringListPrefMember here as well?

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:475:
profile_->GetPrefs()->Set(prefs::kEnabledSyncedNotificationSendingServices,
And here?

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:612:
profile_->GetPrefs()->Set(
And here?

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:617:
profile_->GetPrefs()->Set(
And here?

[Pete Williamson, 2013-09-17 18:40:30]

Uses my pref variable in more places to address DewittJ's nits.

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
File chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc
(right):

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:467:
profile_->GetPrefs()->Set(prefs::kEnabledSyncedNotificationSendingServices,
On 2013/09/17 01:05:21, dewittj wrote:
> Can we just use the StringListPrefMember here as well?

Done.

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:475:
profile_->GetPrefs()->Set(prefs::kEnabledSyncedNotificationSendingServices,
On 2013/09/17 01:05:21, dewittj wrote:
> And here?

Done.

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:612:
profile_->GetPrefs()->Set(
On 2013/09/17 01:05:21, dewittj wrote:
> And here?

Done.

https://codereview.chromium.org/23534048/diff/23001/chrome/browser/notificati...
chrome/browser/notifications/sync_notifier/chrome_notifier_service.cc:617:
profile_->GetPrefs()->Set(
On 2013/09/17 01:05:21, dewittj wrote:
> And here?

Done.

[commit-bot: I haz the power, 2013-09-17 18:43:30]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/23534048/39001

[commit-bot: I haz the power, 2013-09-17 18:55:17]

Step "update" is always a major failure.
Look at the try server FAQ for more details.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=android_db...

[commit-bot: I haz the power, 2013-09-17 23:52:46]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/23534048/55001

[commit-bot: I haz the power, 2013-09-18 00:46:02]

Sorry for I got bad news for ya.
Compile failed with a clobber build on android_clang_dbg.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=android_cl...
Your code is likely broken or HEAD is junk. Please ensure your
code is not broken then alert the build sheriffs.
Look at the try server FAQ for more details.

[commit-bot: I haz the power, 2013-09-18 16:33:14]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/23534048/55001

[commit-bot: I haz the power, 2013-09-18 19:09:16]

Sorry for I got bad news for ya.
Compile failed with a clobber build on android_clang_dbg.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=android_cl...
Your code is likely broken or HEAD is junk. Please ensure your
code is not broken then alert the build sheriffs.
Look at the try server FAQ for more details.

[commit-bot: I haz the power, 2013-09-18 20:24:52]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/23534048/75001

[commit-bot: I haz the power, 2013-09-18 20:53:08]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/23534048/89001

[commit-bot: I haz the power, 2013-09-18 22:39:29]

Sorry for I got bad news for ya.
Compile failed with a clobber build on win_x64_rel.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_x64_re...
Your code is likely broken or HEAD is junk. Please ensure your
code is not broken then alert the build sheriffs.
Look at the try server FAQ for more details.

[commit-bot: I haz the power, 2013-09-19 00:51:53]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/23534048/89001

[commit-bot: I haz the power, 2013-09-19 07:41:02]

Change committed as 224080