Update TestURLRequest constructor interface in unit tests.

Update TestURLRequest constructor interface in unit tests.

The webkit and android_webview unit tests that create TestURLRequest objects
are updated in this patch to explicitly specify which network delegate they
use, rather than extracting it from the context.  This is usually but not
always NULL.

This change is motivated by a desire to remove external accesses to a request
context's delegate. (If a block of code needs to use a NetworkDelegate, it
should not query the URLRequestContext for it, but instead should already have
access to whatever delegate the context uses.)

BUG=146587
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=188920

[tedv, 2013-02-25 03:43:13]

michaeln: Please review webkit/*
mnaganov: Please review android_webview/*

These changes depend on https://codereview.chromium.org/12328072/ which adds the
new TestURLRequest constructor interface.

[mnaganov (inactive), 2013-02-25 10:32:47]

LGTM for android_webview

[michaeln, 2013-02-27 00:55:02]

lgtm2

[tedv, 2013-03-10 20:32:57]

mmenke: Can you review net/url_request/url_request.h?

I added a protected network_delegate() accessor to URLRequest, which derived
classes in unit tests like having.

michaeln: Can you re-review appcache_request_handler_unittest.cc?

I reworked this a bit to make the use of a NULL NetworkDelegate* more explicit. 
It's a bit more pedantic, but given that several unit test objects track the
network delegate object, it seemed better to specify how exactly the NULL value
propagates through the system.

[mmenke, 2013-03-10 20:40:56]

https://codereview.chromium.org/12321106/diff/5001/net/url_request/url_request.h
File net/url_request/url_request.h (right):

https://codereview.chromium.org/12321106/diff/5001/net/url_request/url_reques...
net/url_request/url_request.h:652: NetworkDelegate* network_delegate() { return
network_delegate_; }
I don't think we need this.  The one class that uses it can just cache the value
from its constructor.

As the main entry point to the network stack, I prefer not to expand the
interface except when needed.

These functions are mostly for URLRequestJobs, which have their own copy of the
delegate pointer.  I don't want to introduce ambiguities as to how they should
access the delegate.

[tedv, 2013-03-10 22:35:08]

On 2013/03/10 20:40:56, mmenke wrote:
> I don't think we need this.  The one class that uses it can just cache the
value
> from its constructor.
> 
> As the main entry point to the network stack, I prefer not to expand the
> interface except when needed.
> 
> These functions are mostly for URLRequestJobs, which have their own copy of
the
> delegate pointer.  I don't want to introduce ambiguities as to how they should
> access the delegate.

I understand-- I removed that accessor and reworked as you suggested, so that
shouldn't change any code in net/*.

[mmenke, 2013-03-11 01:01:31]

Not an owner of any files in this CL any more, of course, but with that change,
LGTM.

https://codereview.chromium.org/12321106/diff/10001/webkit/appcache/appcache_...
File webkit/appcache/appcache_request_handler_unittest.cc (right):

https://codereview.chromium.org/12321106/diff/10001/webkit/appcache/appcache_...
webkit/appcache/appcache_request_handler_unittest.cc:120: explicit
MockURLRequest(
nit:  While you're here, explicit isn't needed any longer.

[commit-bot: I haz the power, 2013-03-17 18:00:39]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/tedv@chromium.org/12321106/10001

[commit-bot: I haz the power, 2013-03-17 18:32:53]

Retried try job too often on linux_aura for step(s) content_unittests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_aura...

[commit-bot: I haz the power, 2013-03-18 20:40:27]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/tedv@chromium.org/12321106/21002

[commit-bot: I haz the power, 2013-03-19 02:53:31]

Change committed as 188920