Introduce NavigatorID IDL interface

Introduce NavigatorID IDL interface

Introduce NavigatorID IDL interface to avoid duplication and match the
specification:
http://www.whatwg.org/specs/web-apps/current-work/multipage/timers.html#navigatorid
http://www.whatwg.org/specs/web-apps/current-work/multipage/timers.html#navigator
http://www.whatwg.org/specs/web-apps/current-work/multipage/workers.html#workernavigator

R=arv@chromium.org, haraken@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=156396

[do-not-use, 2013-08-20 11:05:05]

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorBase.h
File Source/core/page/NavigatorBase.h (right):

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorBas...
Source/core/page/NavigatorBase.h:37: // FIXME: The following should be moved to
NavigatorID and
Planning to address this in a follow up patch. Should not be controversial.

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorBas...
Source/core/page/NavigatorBase.h:42: // FIXME: The following is not part of the
specification and is only
Planning to address this in a follow-up patch. My personal preference would be
to move this to Navigator and keep exposing this on Navigator.idl only since
this does not seem to be standard. However, advice is welcome.

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorID.h
File Source/core/page/NavigatorID.h (right):

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorID....
Source/core/page/NavigatorID.h:44: static String userAgent(const
NavigatorBase*);
Note that in theory, we could have 2 versions of this userAgent(), one taking a
Navigator* and another taking a WorkerNavigator*. As a result, we the
userAgent() virtual method in WorkerBase would become unnecessary and the
WorkerBase base class would eventually disappear.

I was not sure which way to go:
- Removing WorkerBase and a virtual function call is a plus
- However, the duplication in NavigatorID would not look great

[haraken, 2013-08-20 12:52:52]

LGTM

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorBase.h
File Source/core/page/NavigatorBase.h (right):

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorBas...
Source/core/page/NavigatorBase.h:42: // FIXME: The following is not part of the
specification and is only
On 2013/08/20 11:05:06, Christophe Dumez wrote:
> Planning to address this in a follow-up patch. My personal preference would be
> to move this to Navigator and keep exposing this on Navigator.idl only since
> this does not seem to be standard. However, advice is welcome.

+1 to moving the methods to Navigator.idl only.

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorID.h
File Source/core/page/NavigatorID.h (right):

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorID....
Source/core/page/NavigatorID.h:44: static String userAgent(const
NavigatorBase*);
On 2013/08/20 11:05:06, Christophe Dumez wrote:
> Note that in theory, we could have 2 versions of this userAgent(), one taking
a
> Navigator* and another taking a WorkerNavigator*. As a result, we the
> userAgent() virtual method in WorkerBase would become unnecessary and the
> WorkerBase base class would eventually disappear.
> 
> I was not sure which way to go:
> - Removing WorkerBase and a virtual function call is a plus
> - However, the duplication in NavigatorID would not look great

I'd prefer sharing code between the main thread and workers as much as possible.
In that sense, the current CL looks better to me.

[arv (Not doing code reviews), 2013-08-20 13:40:44]

LGTM

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorBase.h
File Source/core/page/NavigatorBase.h (right):

https://codereview.chromium.org/22927015/diff/1/Source/core/page/NavigatorBas...
Source/core/page/NavigatorBase.h:42: // FIXME: The following is not part of the
specification and is only
On 2013/08/20 12:52:52, haraken wrote:
> On 2013/08/20 11:05:06, Christophe Dumez wrote:
> > Planning to address this in a follow-up patch. My personal preference would
be
> > to move this to Navigator and keep exposing this on Navigator.idl only since
> > this does not seem to be standard. However, advice is welcome.
> 
> +1 to moving the methods to Navigator.idl only.

+1 and add deprecation warnings on these as well. IE11 did not add them so these
should be safe to remove on the open web.

https://codereview.chromium.org/22927015/diff/2/Source/core/page/NavigatorID.cpp
File Source/core/page/NavigatorID.cpp (right):

https://codereview.chromium.org/22927015/diff/2/Source/core/page/NavigatorID....
Source/core/page/NavigatorID.cpp:2: * Copyright (C) 2008 Apple Inc. All Rights
Reserved.
This came up before... are you using git mv and it still picks the wrong base
file for comparison? If that is the case you should be able to use "git cl
upload --no-find-copies" or something like that.

[do-not-use, 2013-08-20 13:42:15]

Committed patchset #3 manually as r156396 (presubmit successful).