Add a check that we don't send a secure referrer to an insecure URL

Add a check that we don't send a secure referrer to an insecure URL

Unless one of the non-default referrer policies is set, this should
never happen.

BUG=none
R=mef@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=241876

[mef, 2013-12-16 16:39:00]

Looks good. FWIW I've run try job with
https://codereview.chromium.org/110893003/ and it didn't seem to trigger any
meaningful failures.

https://codereview.chromium.org/115803003/diff/20001/net/url_request/url_requ...
File net/url_request/url_request.cc (right):

https://codereview.chromium.org/115803003/diff/20001/net/url_request/url_requ...
net/url_request/url_request.cc:692:
CLEAR_REFERRER_ON_TRANSITION_FROM_SECURE_TO_INSECURE);
I'd keep it as DCHECK to avoid non-critical crashes.

[jochen (gone - plz use gerrit), 2013-12-16 19:25:50]

https://codereview.chromium.org/115803003/diff/20001/net/url_request/url_requ...
File net/url_request/url_request.cc (right):

https://codereview.chromium.org/115803003/diff/20001/net/url_request/url_requ...
net/url_request/url_request.cc:692:
CLEAR_REFERRER_ON_TRANSITION_FROM_SECURE_TO_INSECURE);
On 2013/12/16 16:39:01, mef wrote:
> I'd keep it as DCHECK to avoid non-critical crashes.

IMO a DCHECK is almost pointless. And it is a  critical error if this happens in
the wild.

[mef, 2013-12-16 19:35:12]

On 2013/12/16 19:25:50, jochen wrote:
>
https://codereview.chromium.org/115803003/diff/20001/net/url_request/url_requ...
> File net/url_request/url_request.cc (right):
> 
>
https://codereview.chromium.org/115803003/diff/20001/net/url_request/url_requ...
> net/url_request/url_request.cc:692:
> CLEAR_REFERRER_ON_TRANSITION_FROM_SECURE_TO_INSECURE);
> On 2013/12/16 16:39:01, mef wrote:
> > I'd keep it as DCHECK to avoid non-critical crashes.
> 
> IMO a DCHECK is almost pointless. And it is a  critical error if this happens
in
> the wild.

I agree that it is critical error, but crashing the browser doesn't seem like a
good user experience.
Maybe logging it through UMA and auto-correcting is better?

[jochen (gone - plz use gerrit), 2013-12-18 19:10:59]

ptal

I moved DumpWithoutCrashing() to base, so I can use it from here. This will
correct the referrer and generate a crash report (but don't crash the browser)

[mef, 2013-12-18 20:41:41]

On 2013/12/18 19:10:59, jochen wrote:
> ptal
> 
> I moved DumpWithoutCrashing() to base, so I can use it from here. This will
> correct the referrer and generate a crash report (but don't crash the browser)

Sounds good. Is it ok thing to do from privacy perspective? 
Also, I don't think the move has landed yet.

[jochen (gone - plz use gerrit), 2013-12-18 20:46:34]

On 2013/12/18 20:41:41, mef wrote:
> On 2013/12/18 19:10:59, jochen wrote:
> > ptal
> > 
> > I moved DumpWithoutCrashing() to base, so I can use it from here. This will
> > correct the referrer and generate a crash report (but don't crash the
browser)
> 
> Sounds good. Is it ok thing to do from privacy perspective? 
> Also, I don't think the move has landed yet.

ah, lkgr just hasn't updated yet.

We only send crash reports for users that opted into crash reporting.

[jochen (gone - plz use gerrit), 2013-12-19 08:16:34]

PTAL

[mef, 2013-12-19 15:41:43]

lgtm

[commit-bot: I haz the power, 2013-12-19 15:45:58]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jochen@chromium.org/115803003/60001

[commit-bot: I haz the power, 2013-12-19 17:09:24]

Change committed as 241876

[jamesr, 2013-12-19 22:04:24]

This is failing on the blink canary for browser_tests:ReferrerPolicyTest.History
and content_browsertests:BrowserPluginHostTest.ReloadEmbedder.  See

http://build.chromium.org/p/chromium.webkit/builders/Win7%20%28dbg%29/builds/...

[jamesr, 2013-12-19 22:05:32]

A revert of this CL has been created in
https://codereview.chromium.org/106683004/ by jamesr@chromium.org.

The reason for reverting is: fails browser_tests:ReferrerPolicyTest.History
and content_browsertests:BrowserPluginHostTest.ReloadEmbedder on win7 dbg.

[brianderson, 2013-12-19 22:09:25]

On 2013/12/19 22:04:24, jamesr wrote:
> This is failing on the blink canary for
browser_tests:ReferrerPolicyTest.History
> and content_browsertests:BrowserPluginHostTest.ReloadEmbedder.  See
> 
>
http://build.chromium.org/p/chromium.webkit/builders/Win7%2520%2528dbg%2529/b...

There weren't any notifications sent out, but this patch has been reverted.