Add use-counters for Client-Hints features

Add use-counters for Client-Hints features

In order to be aware of usage, this CL adds use-counters for the various
Client-Hints features.
Due to ClientHintsPreferences living in fetch/ (Owned by FetchRequest)
and the will to avoid creating dependencies between fetch/ and
the rest of core/, counted features are re-defined in FetchContext
and counting is done through loader/FrameFetchContext.

BUG=458983
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=197208

[Yoav Weiss, 2015-06-16 12:47:42]

yoav@yoav.ws changed reviewers:
+ japhet@chromium.org, mkwst@chromium.org

[Mike West, 2015-06-16 12:52:27]

I'm not sure whether this indirection is the right way to do things, or whether
simply defining three methods to do the counting would be better. I'll leave it
up to japhet@, as he's more responsible for the current structure than I am.

Either way, LGTM once Nate is happy.

[Nate Chapin, 2015-06-16 17:10:28]

I think moving ClientHintsPreferences to core/loader/ is preferable to me, but
I'm not absolutely opposed to this technique.

@mkwst, while this directory reorg may be my pet project, feel free to
sanity-check my opinions here :)

https://codereview.chromium.org/1189523015/diff/20001/Source/core/fetch/Fetch...
File Source/core/fetch/FetchContext.h (right):

https://codereview.chromium.org/1189523015/diff/20001/Source/core/fetch/Fetch...
Source/core/fetch/FetchContext.h:63: enum FetchUseCounterFeature {
It seems unfortunate to have this enum essentially duplicate the
UseCounter::Feature enum. Could we separate UseCounter::Feature from the rest of
UseCounter and only give core/fetch/ access to the enum?

https://codereview.chromium.org/1189523015/diff/20001/Source/core/html/parser...
File Source/core/html/parser/HTMLPreloadScanner.cpp (right):

https://codereview.chromium.org/1189523015/diff/20001/Source/core/html/parser...
Source/core/html/parser/HTMLPreloadScanner.cpp:524:
handleAcceptClientHintsHeader(equivAttributeValue, m_clientHintsPreferences,
nullptr);
It's not obvious to me why the ResourceFetcher* parameter is nullptr in some
places but not in others.

[Yoav Weiss, 2015-06-16 18:01:44]

On 2015/06/16 17:10:28, Nate Chapin wrote:
> I think moving ClientHintsPreferences to core/loader/ is preferable to me, but
> I'm not absolutely opposed to this technique.
> 
> @mkwst, while this directory reorg may be my pet project, feel free to
> sanity-check my opinions here :)
> 
>
https://codereview.chromium.org/1189523015/diff/20001/Source/core/fetch/Fetch...
> File Source/core/fetch/FetchContext.h (right):
> 
>
https://codereview.chromium.org/1189523015/diff/20001/Source/core/fetch/Fetch...
> Source/core/fetch/FetchContext.h:63: enum FetchUseCounterFeature {
> It seems unfortunate to have this enum essentially duplicate the
> UseCounter::Feature enum. Could we separate UseCounter::Feature from the rest
of
> UseCounter and only give core/fetch/ access to the enum?

I wasn't able to do that without depending on UseCounter.h and without changing
the enum's name and namespace,
which would have forced me to touch upon all the UseCounter::count calls.

I now switched to Mike's suggestion, which seems cleaner than the "translation"
switch.

> 
>
https://codereview.chromium.org/1189523015/diff/20001/Source/core/html/parser...
> File Source/core/html/parser/HTMLPreloadScanner.cpp (right):
> 
>
https://codereview.chromium.org/1189523015/diff/20001/Source/core/html/parser...
> Source/core/html/parser/HTMLPreloadScanner.cpp:524:
> handleAcceptClientHintsHeader(equivAttributeValue, m_clientHintsPreferences,
> nullptr);
> It's not obvious to me why the ResourceFetcher* parameter is nullptr in some
> places but not in others.

It's nullptr in places where we cannot count (off main thread - in
HTMLPreloadScanner) or don't really care to (unit tests).
Counting this in the preloader is not significant, since eventually the parser
would reach the <meta> and would count for the page anyway.

[Nate Chapin, 2015-06-16 18:04:52]

lgtm

[Yoav Weiss, 2015-06-16 20:01:59]

The CQ bit was checked by yoav@yoav.ws

[Yoav Weiss, 2015-06-16 20:01:59]

The patchset sent to the CQ was uploaded after l-g-t-m from mkwst@chromium.org
Link to the patchset: https://codereview.chromium.org/1189523015/#ps40001
(title: "Changed to counting functions")

[commit-bot: I haz the power, 2015-06-16 20:02:59]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1189523015/40001

[commit-bot: I haz the power, 2015-06-16 23:55:08]

Committed patchset #3 (id:40001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=197208