Remove the last usages of PrefObserver outside of Prefs.

Remove the last usages of PrefObserver outside of Prefs.

This involved:
a) Removing methods that used PrefObserver from PrefChangeRegistrar and PrefMember.
b) Removing implementation of PrefObserver from a few remaining classes.
c) Removing NULL PrefObserver parameter for classes using PrefMember but not wishing for a callback.
d) Updating unit tests and test utility classes.

Also fixed what was most likely a bug (or at least an unintentional change in behavior) that I had introduced in CoreChromeOSOptionsHandler in a previous change.  Before my series of PrefObserver-related changes, there was polymorphism involved and the subclass was only handling pref changes for a couple of preferences (kProxy and kUseSharedProxy, only one of which it registers for itself) if the PrefService was the standard one, not the incognito one, otherwise delegating to the parent class.

[email protected]
BUG=155525
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=171002

[Jói, 2012-12-03 16:44:34]

Finally! Last week was a bit of a wash for me, I was away 3 days due to
sheriffing and a conference, but here is the hopefully last change to remove
PrefObserver in all its splendor.

Cheers,
Jói

[Mattias Nissler (ping if slow), 2012-12-04 10:10:33]

Here's a half-pass from yesterday night, I'll take a look at the updated version
now.

https://codereview.chromium.org/11316163/diff/5045/base/prefs/public/pref_cha...
File base/prefs/public/pref_change_registrar_unittest.cc (right):

https://codereview.chromium.org/11316163/diff/5045/base/prefs/public/pref_cha...
base/prefs/public/pref_change_registrar_unittest.cc:179:
MOCK_METHOD1(OnPreferenceChanged, void(const std::string&));
optional: You can move that declaration to the test fixture class, so you don't
have to declare an extra class for this.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
File chrome/browser/api/prefs/pref_member.h (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
chrome/browser/api/prefs/pref_member.h:11: //    
my_string_.Init(prefs::kHomePage, prefs, NULL /* no observer */);
Update this?

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
chrome/browser/api/prefs/pref_member.h:106: void Init(const char* pref_name,
PrefServiceBase* prefs);
So this wasn't used previously? ;)

Any reason to expose this? Can't we just have the consumer below pass a
NamedChangeCallback() as before?

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/mock_...
File chrome/browser/prefs/mock_pref_change_callback.h (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/mock_...
chrome/browser/prefs/mock_pref_change_callback.h:43:
MOCK_METHOD1(OnPreferenceChanged, void(const std::string&));
As mentioned elsewhere, you could put that declaration on the test fixture
classes that need this.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
File chrome/browser/prefs/pref_notifier_impl_unittest.cc (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
chrome/browser/prefs/pref_notifier_impl_unittest.cc:80:
MOCK_METHOD2(OnPreferenceChanged, void(PrefServiceBase*, const std::string&));
same comment regarding moving the mock declaration.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
File chrome/browser/prefs/pref_service_unittest.cc (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
chrome/browser/prefs/pref_service_unittest.cc:50: MockPrefChangeCallback
obs(&prefs);
Ah, so here you actually don't have a fixture. OK to keep MockPrefChangeCallback
then.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
chrome/browser/prefs/pref_service_unittest.cc:402: ;
nit: remove

[Mattias Nissler (ping if slow), 2012-12-04 10:47:11]

Here are the remaining comments (only nits).

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/profiles/p...
File chrome/browser/profiles/profile_io_data.cc (right):

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/profiles/p...
chrome/browser/profiles/profile_io_data.cc:209:
google_services_username_.Init(prefs::kGoogleServicesUsername, pref_service);
nit: overlong line

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/ui/webui/o...
File chrome/browser/ui/webui/options/core_options_handler.h (right):

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/ui/webui/o...
chrome/browser/ui/webui/options/core_options_handler.h:70: // pref from the
local state or not.
Some people feel strongly about avoiding use of pronouns in comments ;) I think
it's fine, up to you.

[Mattias Nissler (ping if slow), 2012-12-04 10:47:12]

Here are the remaining comments (only nits).

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/profiles/p...
File chrome/browser/profiles/profile_io_data.cc (right):

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/profiles/p...
chrome/browser/profiles/profile_io_data.cc:209:
google_services_username_.Init(prefs::kGoogleServicesUsername, pref_service);
nit: overlong line

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/ui/webui/o...
File chrome/browser/ui/webui/options/core_options_handler.h (right):

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/ui/webui/o...
chrome/browser/ui/webui/options/core_options_handler.h:70: // pref from the
local state or not.
Some people feel strongly about avoiding use of pronouns in comments ;) I think
it's fine, up to you.

[Jói, 2012-12-04 10:55:56]

Thanks Mattias,

I've addressed your comments and uploaded a new patch.

Cheers,
Jói

https://codereview.chromium.org/11316163/diff/5045/base/prefs/public/pref_cha...
File base/prefs/public/pref_change_registrar_unittest.cc (right):

https://codereview.chromium.org/11316163/diff/5045/base/prefs/public/pref_cha...
base/prefs/public/pref_change_registrar_unittest.cc:179:
MOCK_METHOD1(OnPreferenceChanged, void(const std::string&));
On 2012/12/04 10:10:34, Mattias Nissler wrote:
> optional: You can move that declaration to the test fixture class, so you
don't
> have to declare an extra class for this.

Done.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
File chrome/browser/api/prefs/pref_member.h (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
chrome/browser/api/prefs/pref_member.h:11: //    
my_string_.Init(prefs::kHomePage, prefs, NULL /* no observer */);
On 2012/12/04 10:10:34, Mattias Nissler wrote:
> Update this?

Done.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
chrome/browser/api/prefs/pref_member.h:106: void Init(const char* pref_name,
PrefServiceBase* prefs);
On 2012/12/04 10:10:34, Mattias Nissler wrote:
> So this wasn't used previously? ;)
> 
> Any reason to expose this? Can't we just have the consumer below pass a
> NamedChangeCallback() as before?

It's a performance optimization; we already create a do-nothing
NamedChangeCallback in the constructor of PrefMemberBase, no need to do it
again.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/mock_...
File chrome/browser/prefs/mock_pref_change_callback.h (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/mock_...
chrome/browser/prefs/mock_pref_change_callback.h:43:
MOCK_METHOD1(OnPreferenceChanged, void(const std::string&));
On 2012/12/04 10:10:34, Mattias Nissler wrote:
> As mentioned elsewhere, you could put that declaration on the test fixture
> classes that need this.

Leaving this as discussed elsewhere.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
File chrome/browser/prefs/pref_notifier_impl_unittest.cc (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
chrome/browser/prefs/pref_notifier_impl_unittest.cc:80:
MOCK_METHOD2(OnPreferenceChanged, void(PrefServiceBase*, const std::string&));
On 2012/12/04 10:10:34, Mattias Nissler wrote:
> same comment regarding moving the mock declaration.

In this case it needs to derive from PrefObserver, so I think it's cleaner to
make it a separate class from the test class.  WDYT?

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
File chrome/browser/prefs/pref_service_unittest.cc (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
chrome/browser/prefs/pref_service_unittest.cc:50: MockPrefChangeCallback
obs(&prefs);
On 2012/12/04 10:10:34, Mattias Nissler wrote:
> Ah, so here you actually don't have a fixture. OK to keep
MockPrefChangeCallback
> then.

Right, and I don't want to define the PrefValueMatches matcher (from
mock_pref_change_callback.h) in multiple places.

I've inlined a couple of things that weren't using that bit though, i.e. where
we just needed the OnPreferenceChanged mock method and nothing else.

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
chrome/browser/prefs/pref_service_unittest.cc:402: ;
On 2012/12/04 10:10:34, Mattias Nissler wrote:
> nit: remove

Done.

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/profiles/p...
File chrome/browser/profiles/profile_io_data.cc (right):

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/profiles/p...
chrome/browser/profiles/profile_io_data.cc:209:
google_services_username_.Init(prefs::kGoogleServicesUsername, pref_service);
On 2012/12/04 10:47:12, Mattias Nissler wrote:
> nit: overlong line

Done.

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/ui/webui/o...
File chrome/browser/ui/webui/options/core_options_handler.h (right):

https://codereview.chromium.org/11316163/diff/10004/chrome/browser/ui/webui/o...
chrome/browser/ui/webui/options/core_options_handler.h:70: // pref from the
local state or not.
On 2012/12/04 10:47:12, Mattias Nissler wrote:
> Some people feel strongly about avoiding use of pronouns in comments ;) I
think
> it's fine, up to you.

The rest of the comments in this file seem to avoid pronouns so I've updated
this to match.

[Mattias Nissler (ping if slow), 2012-12-04 11:18:33]

LGTM w/ one question regarding premature optimization ;)

And, finally: Thanks so much for all the effort you put into this! And thanks
for bearing with me :)

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
File chrome/browser/api/prefs/pref_member.h (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
chrome/browser/api/prefs/pref_member.h:106: void Init(const char* pref_name,
PrefServiceBase* prefs);
On 2012/12/04 10:55:57, Jói wrote:
> On 2012/12/04 10:10:34, Mattias Nissler wrote:
> > So this wasn't used previously? ;)
> > 
> > Any reason to expose this? Can't we just have the consumer below pass a
> > NamedChangeCallback() as before?
> 
> It's a performance optimization; we already create a do-nothing
> NamedChangeCallback in the constructor of PrefMemberBase, no need to do it
> again.

Is the difference measurable? I usually prefer less/simpler code over premature
optimizations :)

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
File chrome/browser/prefs/pref_notifier_impl_unittest.cc (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/prefs/pref_...
chrome/browser/prefs/pref_notifier_impl_unittest.cc:80:
MOCK_METHOD2(OnPreferenceChanged, void(PrefServiceBase*, const std::string&));
On 2012/12/04 10:55:57, Jói wrote:
> On 2012/12/04 10:10:34, Mattias Nissler wrote:
> > same comment regarding moving the mock declaration.
> 
> In this case it needs to derive from PrefObserver, so I think it's cleaner to
> make it a separate class from the test class.  WDYT?

Yes, you are right, I had missed that this version needs to  extend
PrefObserver.

[Jói, 2012-12-04 11:42:26]

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
File chrome/browser/api/prefs/pref_member.h (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
chrome/browser/api/prefs/pref_member.h:106: void Init(const char* pref_name,
PrefServiceBase* prefs);
On 2012/12/04 11:18:33, Mattias Nissler wrote:
> On 2012/12/04 10:55:57, Jói wrote:
> > On 2012/12/04 10:10:34, Mattias Nissler wrote:
> > > So this wasn't used previously? ;)
> > > 
> > > Any reason to expose this? Can't we just have the consumer below pass a
> > > NamedChangeCallback() as before?
> > 
> > It's a performance optimization; we already create a do-nothing
> > NamedChangeCallback in the constructor of PrefMemberBase, no need to do it
> > again.
> 
> Is the difference measurable? I usually prefer less/simpler code over
premature
> optimizations :)

The performance difference is probably not measurable except by a targeted
microbenchmark.  OTOH, I don't really feel this version of the code is more
complex, and it avoids some unnecessary work.  Up to you, let me know if you
prefer I get rid of the two-parameter PrefMemberBase::Init.

[Jói, 2012-12-04 15:11:16]

Mattias, you may have missed this as you already gave your LGTM, but
I'd like your feedback on this before I commit:

> The performance difference is probably not measurable except by a
> targeted microbenchmark.  OTOH, I don't really feel this version of the
> code is more complex, and it avoids some unnecessary work.  Up to you,
> let me know if you prefer I get rid of the two-parameter
> PrefMemberBase::Init.

Cheers,
Jói


On Tue, Dec 4, 2012 at 11:42 AM,  <[email protected]> wrote:
>
>
https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
> File chrome/browser/api/prefs/pref_member.h (right):
>
>
https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
> chrome/browser/api/prefs/pref_member.h:106: void Init(const char*
> pref_name, PrefServiceBase* prefs);
> On 2012/12/04 11:18:33, Mattias Nissler wrote:
>>
>> On 2012/12/04 10:55:57, Jói wrote:
>> > On 2012/12/04 10:10:34, Mattias Nissler wrote:
>> > > So this wasn't used previously? ;)
>> > >
>> > > Any reason to expose this? Can't we just have the consumer below
>
> pass a
>>
>> > > NamedChangeCallback() as before?
>> >
>> > It's a performance optimization; we already create a do-nothing
>> > NamedChangeCallback in the constructor of PrefMemberBase, no need to
>
> do it
>>
>> > again.
>
>
>> Is the difference measurable? I usually prefer less/simpler code over
>
> premature
>>
>> optimizations :)
>
>
> The performance difference is probably not measurable except by a
> targeted microbenchmark.  OTOH, I don't really feel this version of the
> code is more complex, and it avoids some unnecessary work.  Up to you,
> let me know if you prefer I get rid of the two-parameter
> PrefMemberBase::Init.
>
> https://codereview.chromium.org/11316163/

[Mattias Nissler (ping if slow), 2012-12-04 15:24:37]

LGTM, I just failed to submit my reply :-/ Sorry for that.

Also, a big thank you for all the effort you put into this!

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
File chrome/browser/api/prefs/pref_member.h (right):

https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
chrome/browser/api/prefs/pref_member.h:106: void Init(const char* pref_name,
PrefServiceBase* prefs);
On 2012/12/04 11:42:26, Jói wrote:
> On 2012/12/04 11:18:33, Mattias Nissler wrote:
> > On 2012/12/04 10:55:57, Jói wrote:
> > > On 2012/12/04 10:10:34, Mattias Nissler wrote:
> > > > So this wasn't used previously? ;)
> > > > 
> > > > Any reason to expose this? Can't we just have the consumer below pass a
> > > > NamedChangeCallback() as before?
> > > 
> > > It's a performance optimization; we already create a do-nothing
> > > NamedChangeCallback in the constructor of PrefMemberBase, no need to do it
> > > again.
> > 
> > Is the difference measurable? I usually prefer less/simpler code over
> premature
> > optimizations :)
> 
> The performance difference is probably not measurable except by a targeted
> microbenchmark.  OTOH, I don't really feel this version of the code is more
> complex, and it avoids some unnecessary work.  Up to you, let me know if you
> prefer I get rid of the two-parameter PrefMemberBase::Init.

I don't feel too strong, feel free to keep as is.

[commit-bot: I haz the power, 2012-12-04 15:30:19]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11316163/12055

[Jói, 2012-12-04 15:30:38]

Thanks :)  The end result is certainly better than what we had.

In retrospect, I wish I had done fewer, larger batches; I was afraid
of regressions which didn't materialize, and fewer batches would have
probably taken a fair bit less time.

Now, on to PrefService :D

Cheers,
Jói


On Tue, Dec 4, 2012 at 3:24 PM,  <[email protected]> wrote:
> LGTM, I just failed to submit my reply :-/ Sorry for that.
>
> Also, a big thank you for all the effort you put into this!
>
>
>
>
https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
> File chrome/browser/api/prefs/pref_member.h (right):
>
>
https://codereview.chromium.org/11316163/diff/5045/chrome/browser/api/prefs/p...
> chrome/browser/api/prefs/pref_member.h:106: void Init(const char*
> pref_name, PrefServiceBase* prefs);
> On 2012/12/04 11:42:26, Jói wrote:
>>
>> On 2012/12/04 11:18:33, Mattias Nissler wrote:
>> > On 2012/12/04 10:55:57, Jói wrote:
>> > > On 2012/12/04 10:10:34, Mattias Nissler wrote:
>> > > > So this wasn't used previously? ;)
>> > > >
>> > > > Any reason to expose this? Can't we just have the consumer below
>
> pass a
>>
>> > > > NamedChangeCallback() as before?
>> > >
>> > > It's a performance optimization; we already create a do-nothing
>> > > NamedChangeCallback in the constructor of PrefMemberBase, no need
>
> to do it
>>
>> > > again.
>> >
>> > Is the difference measurable? I usually prefer less/simpler code
>
> over
>>
>> premature
>> > optimizations :)
>
>
>> The performance difference is probably not measurable except by a
>
> targeted
>>
>> microbenchmark.  OTOH, I don't really feel this version of the code is
>
> more
>>
>> complex, and it avoids some unnecessary work.  Up to you, let me know
>
> if you
>>
>> prefer I get rid of the two-parameter PrefMemberBase::Init.
>
>
> I don't feel too strong, feel free to keep as is.
>
> https://codereview.chromium.org/11316163/

[commit-bot: I haz the power, 2012-12-04 15:30:51]

Presubmit check for 11316163-12055 failed and returned exit status 1.


Running presubmit commit checks ...

** Presubmit Messages **
You added one or more #includes of files that are temporarily
allowed but being removed. Can you avoid introducing the
#include? See relevant DEPS file(s) for details and contacts.
  chrome/browser/extensions/extension_prefs_unittest.cc
      Illegal include: "chrome/browser/prefs/mock_pref_change_callback.h"
    Because of "!chrome/browser/prefs" from chrome/browser/api/prefs's
specific_include_rules. \
  chrome/browser/prefs/pref_notifier_impl_unittest.cc
      Illegal include: "chrome/browser/prefs/mock_pref_change_callback.h"
    Because of "!chrome/browser/prefs" from chrome/browser/api/prefs's
specific_include_rules. \
  chrome/browser/prefs/pref_service_unittest.cc
      Illegal include: "chrome/browser/prefs/mock_pref_change_callback.h"
    Because of "!chrome/browser/prefs" from chrome/browser/api/prefs's
specific_include_rules. \
  chrome/browser/prefs/scoped_user_pref_update_unittest.cc
      Illegal include: "chrome/browser/prefs/mock_pref_change_callback.h"
    Because of "!chrome/browser/prefs" from chrome/browser/api/prefs's
specific_include_rules.

** Presubmit ERRORS **
Missing LGTM from an OWNER for files in these directories:
    chrome/browser/ui/ash
    chrome/browser/ui/webui/options/chromeos
    chrome/browser/ui/webui/options
    chrome/browser/chromeos/login
    chrome/browser/ui/sync
    chrome/browser/plugins
    chrome/browser/profiles
    chrome/browser/ui
    chrome/browser/chromeos
    chrome/browser/ui/cocoa
    chrome
    chrome/browser/ui/webui
    chrome/browser/download
    chrome/browser/spellchecker
    chrome/browser/net
    chrome/browser/extensions
    chrome/browser/password_manager
    chrome/browser/tab_contents
    chrome/browser/ui/views
    chrome/browser/chromeos/input_method

Presubmit checks took 3.9s to calculate.

[Jói, 2012-12-04 15:32:58]

[email protected] for usage updates (mostly removing NULL parameter from
PrefMember::Init, plus switching over a couple of final classes inheriting from
PrefObserver).  Already reviewed by mnissler@ who is an OWNER for the API being
updated.

Cheers,
Jói

[commit-bot: I haz the power, 2012-12-04 15:34:33]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/11316163/12055

[Ben Goodger (Google), 2012-12-04 16:24:07]

lgtm

[commit-bot: I haz the power, 2012-12-04 18:14:09]

Change committed as 171002