Make RegistryControlledDomainService a static-function container class, rather than a Singleton

Make RegistryControlledDomainService a static-function container class, rather than a Singleton

BUG=none
TEST=none


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=107539

[Ryan Sleevi, 2011-10-25 23:44:03]

vandebo: Tripped up one of my CLs due to trying to access the RCDS on a
non-joinable worker thread. While making the Singleton leaky is one option,
there doesn't seem to be any need any longer for it to be a singleton.

It hasn't needed to be a Singleton since
http://src.chromium.org/viewvc/chrome?view=rev&revision=35196 landed

[vandebo (ex-Chrome), 2011-10-26 00:05:48]

LGTM

[commit-bot: I haz the power, 2011-10-26 04:52:47]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rsleevi@chromium.org/8395025/3001

[commit-bot: I haz the power, 2011-10-26 04:52:49]

Presubmit check for 8395025-3001 failed and returned exit status 1.

Running presubmit commit checks ...

** Presubmit Warnings **
License must match:
.*? Copyright \(c\) 2011 The Chromium Authors\. All rights reserved\.\n.*? Use
of this source code is governed by a BSD-style license that can be\n.*? found in
the LICENSE file\.\n
Found a bad license header in these files:
  net/base/registry_controlled_domain.cc \
  net/base/registry_controlled_domain.h

[Ryan Sleevi, 2011-10-26 22:52:15]

Mark, Pawel - Could you provide guidance on what the "correct" thing to do here
for the .h and .cc?

They both bear clear markings of the Mozilla Tri-License GPL/LGPL/MPL, with the
comment that the implementation was inspired by (but mostly re-written?) from
original Mozilla sources.

Is it correct to add a Chromium-specific header like net/base/cookie_monster.*
or net/http/http_chunked_decoder.*? Is it better to leave it alone, like
net/http/md4.*?

[Mark Larson, 2011-10-27 00:41:43]

Please use the Chromium-specific header like cookie_monster. 

This is not a blanket recommendation for MPL/tri-license code that we modify,
but in this case it's OK (and the simplest path forward).

[commit-bot: I haz the power, 2011-10-27 06:50:42]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rsleevi@chromium.org/8395025/7001

[commit-bot: I haz the power, 2011-10-27 08:00:22]

Change committed as 107539

[wtc, 2011-10-27 22:44:09]

LGTM.  Good cleanup!

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
File net/base/registry_controlled_domain.cc (left):

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
net/base/registry_controlled_domain.cc:60: RegistryControlledDomainService*
test_instance_;

Is test_instance_ not used?

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
File net/base/registry_controlled_domain.h (right):

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
net/base/registry_controlled_domain.h:203: typedef const struct DomainRule*
(*FindDomainPtr)(const char *, unsigned int);

Nit: list this typedef at the beginning of the private section.

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
File net/base/registry_controlled_domain_unittest.cc (left):

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
net/base/registry_controlled_domain_unittest.cc:255: }  // namespace

Why did you need to reduce the unnamed namespace?  Just
curious.

[Ryan Sleevi, 2011-10-27 22:48:03]

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
File net/base/registry_controlled_domain.cc (left):

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
net/base/registry_controlled_domain.cc:60: RegistryControlledDomainService*
test_instance_;
On 2011/10/27 22:44:09, wtc wrote:
> 
> Is test_instance_ not used?

Not in the refactor, no. Previously it was used by GetInstance() - further
indicating that the Singleton<>-ness was simply related to testing.

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
File net/base/registry_controlled_domain.h (right):

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
net/base/registry_controlled_domain.h:203: typedef const struct DomainRule*
(*FindDomainPtr)(const char *, unsigned int);
On 2011/10/27 22:44:09, wtc wrote:
> 
> Nit: list this typedef at the beginning of the private section.

Will send a quick cleanup CL.

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
File net/base/registry_controlled_domain_unittest.cc (left):

http://codereview.chromium.org/8395025/diff/7001/net/base/registry_controlled...
net/base/registry_controlled_domain_unittest.cc:255: }  // namespace
On 2011/10/27 22:44:09, wtc wrote:
> 
> Why did you need to reduce the unnamed namespace?  Just
> curious.

friend scope. Because RegistryControlledDomainService is in net::, the implicit
forward declare from "friend class RegistryControlledDomainTest" declares a
net::RegistryControlledDomainTest.

The same pattern is used throughout the GTest friending, as far as I could tell.