Use a NetLog for test requests sent by about:net-internals

Use a NetLog for test requests seny by the the about:net-internals
connection tester.

BUG=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=148670

[mmenke, 2012-07-26 17:57:15]

Have yet to see a case where this would be useful, but seems better to have it
and not need it that the other way around.

http://codereview.chromium.org/10796112/diff/10001/chrome/browser/net/connect...
File chrome/browser/net/connection_tester.cc (right):

http://codereview.chromium.org/10796112/diff/10001/chrome/browser/net/connect...
chrome/browser/net/connection_tester.cc:121:
session_params.http_server_properties = http_server_properties();
Seemed like the three items above should be set the same order as they are
created.

[eroman, 2012-07-26 18:22:40]

Feels a bit magical for the netlog to come from proxy_request_context, but
LGTM

On Thu, Jul 26, 2012 at 10:57 AM, <mmenke@chromium.org> wrote:

> Reviewers: eroman,
>
> Message:
> Have yet to see a case where this would be useful, but seems better to
> have it
> and not need it that the other way around.
>
>
> http://codereview.chromium.**org/10796112/diff/10001/**
>
chrome/browser/net/connection_**tester.cc<http://codereview.chromium.org/10796112/diff/10001/chrome/browser/net/connection_tester.cc>
> File chrome/browser/net/connection_**tester.cc (right):
>
> http://codereview.chromium.**org/10796112/diff/10001/**
>
chrome/browser/net/connection_**tester.cc#newcode121<http://codereview.chromium.org/10796112/diff/10001/chrome/browser/net/connection_tester.cc#newcode121>
> chrome/browser/net/connection_**tester.cc:121:
> session_params.http_server_**properties = http_server_properties();
> Seemed like the three items above should be set the same order as they
> are created.
>
> Description:
> Use a NetLog for test requests sent by the about:net-internals connection
> tester.
>
> BUG=none
>
> Please review this at
http://codereview.chromium.**org/10796112/<http://codereview.chromium.org/107...
>
> SVN Base:
svn://svn.chromium.org/chrome/**trunk/src/<http://svn.chromium.org/chrome/trunk/src/>
>
> Affected files:
>   M     chrome/browser/net/connection_**tester.cc
>
>
> Index: chrome/browser/net/connection_**tester.cc
> ==============================**==============================**=======
> --- chrome/browser/net/connection_**tester.cc     (revision 147661)
> +++ chrome/browser/net/connection_**tester.cc     (working copy)
> @@ -116,9 +116,10 @@
>      session_params.host_resolver = host_resolver();
>      session_params.cert_verifier = cert_verifier();
>      session_params.proxy_service = proxy_service();
> +    session_params.ssl_config_**service = ssl_config_service();
>      session_params.http_auth_**handler_factory =
> http_auth_handler_factory();
>      session_params.http_server_**properties = http_server_properties();
> -    session_params.ssl_config_**service = ssl_config_service();
> +    session_params.net_log = proxy_request_context_->net_**log();
>      scoped_refptr<net::**HttpNetworkSession> network_session(
>          new net::HttpNetworkSession(**session_params));
>      storage_.set_http_transaction_**factory(new net::HttpCache(
>
>
>

[mmenke, 2012-07-26 18:47:06]

On 2012/07/26 18:22:40, eroman wrote:
> Feels a bit magical for the netlog to come from proxy_request_context, but
> LGTM

I was thinking it made sense for both the proxy request and the URL Request to
log to the same NetLog, since it would be weird if they didn't, but you do have
a point.

I've switched to taking the NetLog in as a parameter.  Simple change, but the
NetLog does have to be passed through 3 layers of ConnectionTester classes.

If I don't hear back from you by tomorrow evening, I'll assume your old LGTM
holds.

[eroman, 2012-07-26 21:33:45]

lgtm

[commit-bot: I haz the power, 2012-07-26 21:35:24]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/mmenke@chromium.org/10796112/16002

[commit-bot: I haz the power, 2012-07-27 00:16:54]

Change committed as 148670