Partial changes to enable net-internals in iOS.

Changes to compile connection_tester on iOS and adding
of net_internal's HTML and JS to ios's grit_whitelist.

Disabled Firefox settings for iOS ala Android.

Set FTP transaction factory if DISABLE_FTP_SUPPORT is not set.

Used ProxyService::CreateUsingSystemProxyResolver instead of
net::CreateProxyServiceUsingV8ProxyResolver because V8 is
used on IPHone.

Temporarily ifdef'ed out webkit_glue::ResourceLoadTimingInfo
for IOS in load_timing_observer. load_timing_observer is going
away per mmenke. Thus change to load_timing_observer.h and 
load_timing_observer.cc will be deleted.

R=droger@chromium.org, mmenke@chromium.org
BUG=151212
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=172789

[ramant (doing other things), 2012-12-07 02:10:18]

Hi David,
  Tested the Start, Send and Stop logging API calls by adding those buttons in
net-internals. Tested that code on Android to send net_log data via email.

  We could change these API calls (or add more). Start Logging, starts
collecting data into a file. SendLog sends all the data that is accumulated into
the data file and starts accumulating data into a new file. When we stop
logging, we stop collecting data and we delete all the temporary files that were
created.

  We could consider adding a new API called Delete Logs which deletes the files.
Stop doesn't do anything other than stop collecting data. What do you think?

thanks
raman

[ramant (doing other things), 2012-12-10 06:18:11]

Hi David,
  This CL contains only ifdef OS_IOS changes. Separated changes to
start/stop/send net_logs are into https://codereview.chromium.org/11467013/.

thanks
raman

[droger, 2012-12-10 10:23:06]

It looks good, but I wonder if there is a way to reduce the amount of
ifdefs(OS_IOS) or to replace them with more specific (feature) flags.

Also,  are these files compiled on iOS? Maybe you want to change the gyp files
to include them.

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
File chrome/browser/net/connection_tester.cc (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:112: #if !defined(OS_IOS)
Use DISABLE_FTP_SUPPORT instead of OS_IOS.
I think iOS is the only platform that defines this variable.

[droger, 2012-12-10 12:01:16]

+stuartmorgan who might know how to reduce the number of ifdefs.

[stuartmorgan, 2012-12-10 12:56:02]

Please improve the CL description so that it's explaining what this CL actually
does, specifically. "upstream changes" doesn't really say anything meaningful.

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
File chrome/browser/net/connection_tester.cc (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:99: #if !defined(OS_IOS)
Is this necessary? We're already doing ifdef's in the implementation of
CreateProxyService, so couldn't we leave this code as-is?

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:105: return rv;  // Failure.
Is this right? If not being able to create the service is a fatal error, then
why is it okay for us not to have one on iOS?

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:269: // Chrome on Android does not
support Firefox settings.
Update comment.

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
File chrome/browser/net/load_timing_observer.h (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
chrome/browser/net/load_timing_observer.h:33:
webkit_glue::ResourceLoadTimingInfo timing;
I'm not familiar with this code, but it seems worrying that you are removing
"timing" from "LoadTimingObserver". Does this class do anything useful without
this?

Why does this class live in webkit_glue; could it move out so that we could use
it?

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/ui/webui/n...
File chrome/browser/ui/webui/net_internals/net_internals_ui.cc (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/ui/webui/n...
chrome/browser/ui/webui/net_internals/net_internals_ui.cc:632: #if
!defined(OS_IOS)
These changes should be made downstream-only for now; sharing prerender support
is something that's still under investigation.

[ramant (doing other things), 2012-12-11 05:34:15]

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
File chrome/browser/net/connection_tester.cc (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:99: #if !defined(OS_IOS)
On 2012/12/10 12:56:03, stuartmorgan wrote:
> Is this necessary? We're already doing ifdef's in the implementation of
> CreateProxyService, so couldn't we leave this code as-is?

Done.

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:105: return rv;  // Failure.
On 2012/12/10 12:56:03, stuartmorgan wrote:
> Is this right? If not being able to create the service is a fatal error, then
> why is it okay for us not to have one on iOS?

Removed the ifdef'ing out of net::CreateProxyServiceUsingV8ProxyResolver. Will
work on a separate CL to enable net/proxy/proxy_resolver_v8.cc and
net/proxy/proxy/proxy_service_v8.cc

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:112: #if !defined(OS_IOS)
On 2012/12/10 10:23:06, droger wrote:
> Use DISABLE_FTP_SUPPORT instead of OS_IOS.
> I think iOS is the only platform that defines this variable.

Done.

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:269: // Chrome on Android does not
support Firefox settings.
On 2012/12/10 12:56:03, stuartmorgan wrote:
> Update comment.

Done.

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
File chrome/browser/net/load_timing_observer.h (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
chrome/browser/net/load_timing_observer.h:33:
webkit_glue::ResourceLoadTimingInfo timing;
On 2012/12/10 12:56:03, stuartmorgan wrote:
> I'm not familiar with this code, but it seems worrying that you are removing
> "timing" from "LoadTimingObserver". Does this class do anything useful without
> this?
> 
> Why does this class live in webkit_glue; could it move out so that we could
use
> it?

Removed this change from this CL. Will investigate and submit a separate CL for
moving webkit_glue::ResourceLoadTimingInfo out of webkit_glue.

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/ui/webui/n...
File chrome/browser/ui/webui/net_internals/net_internals_ui.cc (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/ui/webui/n...
chrome/browser/ui/webui/net_internals/net_internals_ui.cc:632: #if
!defined(OS_IOS)
On 2012/12/10 12:56:03, stuartmorgan wrote:
> These changes should be made downstream-only for now; sharing prerender
support
> is something that's still under investigation.

Done.

[mmenke, 2012-12-11 05:46:49]

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
File chrome/browser/net/load_timing_observer.h (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
chrome/browser/net/load_timing_observer.h:33:
webkit_glue::ResourceLoadTimingInfo timing;
On 2012/12/11 05:34:16, ramant wrote:
> On 2012/12/10 12:56:03, stuartmorgan wrote:
> > I'm not familiar with this code, but it seems worrying that you are removing
> > "timing" from "LoadTimingObserver". Does this class do anything useful
without
> > this?
> > 
> > Why does this class live in webkit_glue; could it move out so that we could
> use
> > it?
> 
> Removed this change from this CL. Will investigate and submit a separate CL
for
> moving webkit_glue::ResourceLoadTimingInfo out of webkit_glue.

Don't think you need to, at least not as a prereq for this CL.  This class is
going away soon (Hopefully - working on the CLs now), and gathering this data
will all be done in the network stack.  Suggest you just #ifdef out
ChromeNetLog's dependencies on load_timing_observer, for now.

[ramant (doing other things), 2012-12-11 19:41:40]

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
File chrome/browser/net/connection_tester.cc (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:105: return rv;  // Failure.
On 2012/12/11 05:34:16, ramant wrote:
> On 2012/12/10 12:56:03, stuartmorgan wrote:
> > Is this right? If not being able to create the service is a fatal error,
then
> > why is it okay for us not to have one on iOS?
> 
> Removed the ifdef'ing out of net::CreateProxyServiceUsingV8ProxyResolver. Will
> work on a separate CL to enable net/proxy/proxy_resolver_v8.cc and
> net/proxy/proxy/proxy_service_v8.cc

CreateProxyServiceUsingV8ProxyResolver is not built on IPhone because there is
no V8. Used  CreateUsingSystemProxyResolver instead for IOS

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
File chrome/browser/net/load_timing_observer.h (right):

https://codereview.chromium.org/11450023/diff/28001/chrome/browser/net/load_t...
chrome/browser/net/load_timing_observer.h:33:
webkit_glue::ResourceLoadTimingInfo timing;
On 2012/12/11 05:46:49, Matt Menke wrote:
> On 2012/12/11 05:34:16, ramant wrote:
> > On 2012/12/10 12:56:03, stuartmorgan wrote:
> > > I'm not familiar with this code, but it seems worrying that you are
removing
> > > "timing" from "LoadTimingObserver". Does this class do anything useful
> without
> > > this?
> > > 
> > > Why does this class live in webkit_glue; could it move out so that we
could
> > use
> > > it?
> > 
> > Removed this change from this CL. Will investigate and submit a separate CL
> for
> > moving webkit_glue::ResourceLoadTimingInfo out of webkit_glue.
> 
> Don't think you need to, at least not as a prereq for this CL.  This class is
> going away soon (Hopefully - working on the CLs now), and gathering this data
> will all be done in the network stack.  Suggest you just #ifdef out
> ChromeNetLog's dependencies on load_timing_observer, for now.

Because this class is going away, added back the #if !defined(OS_IOS). When we
get rid of this class, we don't need to make any changes on IOS side (IMO, it is
the least disruptive).

[droger, 2012-12-12 09:42:29]

LGTM.

You might want to compile these files on iOS (change the gyp).

[stuartmorgan, 2012-12-12 10:36:21]

LGTM with nit

https://codereview.chromium.org/11450023/diff/59001/chrome/browser/net/connec...
File chrome/browser/net/connection_tester.cc (right):

https://codereview.chromium.org/11450023/diff/59001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:271: // Chrome on Android and IOS do not
support Firefox settings.
iOS (IOS is to the iOS platform as ANDROID is to the Android platform: something
you'd never call it except when you specifically need all-caps)

[mmenke, 2012-12-12 17:24:16]

LGTM

[ramant (doing other things), 2012-12-12 18:23:56]

On 2012/12/12 09:42:29, droger wrote:
> LGTM.
> 
> You might want to compile these files on iOS (change the gyp).

Done.

I was making these changes on iOS and submitting the CL from windows.

[ramant (doing other things), 2012-12-12 18:25:50]

https://codereview.chromium.org/11450023/diff/59001/chrome/browser/net/connec...
File chrome/browser/net/connection_tester.cc (right):

https://codereview.chromium.org/11450023/diff/59001/chrome/browser/net/connec...
chrome/browser/net/connection_tester.cc:271: // Chrome on Android and IOS do not
support Firefox settings.
On 2012/12/12 10:36:21, stuartmorgan wrote:
> iOS (IOS is to the iOS platform as ANDROID is to the Android platform:
something
> you'd never call it except when you specifically need all-caps)

Done.

[commit-bot: I haz the power, 2012-12-12 18:26:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rtenneti@chromium.org/11450023/70001

[commit-bot: I haz the power, 2012-12-12 19:20:43]

Retried try job too often on linux_chromeos for step(s) interactive_ui_tests

[commit-bot: I haz the power, 2012-12-12 22:10:41]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/rtenneti@chromium.org/11450023/70001

[commit-bot: I haz the power, 2012-12-13 02:45:00]

Change committed as 172789