Cache the boolean value in instance variable instead of calling usingDelegatedRenderer

Cache the boolean value in usingDelegatedRenderer.
As we are going to use this api at many places, caching the value at the first instance, will save the checking every time.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=251665

[sivag, 2014-02-12 11:59:00]

PTAL..

[no sievers, 2014-02-12 17:36:59]

https://codereview.chromium.org/160073005/diff/1/content/browser/renderer_hos...
File content/browser/renderer_host/render_widget_host_view_android.cc (right):

https://codereview.chromium.org/160073005/diff/1/content/browser/renderer_hos...
content/browser/renderer_host/render_widget_host_view_android.cc:108: static
bool g_using_delegated_renderer =
nit: remove 'g_' prefix since the scope of the variable is not global

https://codereview.chromium.org/160073005/diff/1/content/browser/renderer_hos...
content/browser/renderer_host/render_widget_host_view_android.cc:110:
switches::kEnableDelegatedRenderer) &&
nit: indent is wrong in this whole block

[aelias_OOO_until_Jul13, 2014-02-12 19:36:48]

I'm not a fan of this.  statics and globals have a way of coming back to bite us
when the process/threading model or the call order changes, so it's a best
practice to avoid them.  It would be better cache this as an instance variable
in RWHVA if you think it's worth it.  But you also don't present any perftest
data or a count of the number of calls per frame so I'm not sure the cost is
worth making any changes at all.

[aelias_OOO_until_Jul13, 2014-02-12 23:02:34]

Also, we're converging on delegated rendering everywhere, and when that happens
we can just delete the flag.

[sivag, 2014-02-13 16:53:13]

On 2014/02/12 19:36:48, aelias wrote:
> I'm not a fan of this.  statics and globals have a way of coming back to bite
us
> when the process/threading model or the call order changes, so it's a best
> practice to avoid them.  It would be better cache this as an instance variable
> in RWHVA if you think it's worth it.  But you also don't present any perftest
> data or a count of the number of calls per frame so I'm not sure the cost is
> worth making any changes at all.

I kept a log to see the count of the calls, as this api is used twice in
OnSwapCompositorFrame
i can see a increase in count based on the fps, so it varies with pages (with
and without animation, scroll,zoom).
Calling this many times which internally does a switch search is not required.
We can maintain a member variable and cache it, but are we going to clean up the
code in a short span to move completely to delegated rendering then this would
be corrected in that process.

Shall we cache it for now , or cleanup later?
Please advice.

[sivag, 2014-02-14 11:30:20]

On 2014/02/12 23:02:34, aelias wrote:
> Also, we're converging on delegated rendering everywhere, and when that
happens
> we can just delete the flag.

When i tested on google.com site with a simple scroll this got called 90 times.
3 times per frame.

I have cached the switch value in an instance variable , 
in the constructor just once and used it all over the file.
Please take a look.

[sivag, 2014-02-14 11:30:50]

PTAL..

[no sievers, 2014-02-14 18:42:47]

lgtm with 1 nit

https://codereview.chromium.org/160073005/diff/90001/content/browser/renderer...
File content/browser/renderer_host/render_widget_host_view_android.cc (right):

https://codereview.chromium.org/160073005/diff/90001/content/browser/renderer...
content/browser/renderer_host/render_widget_host_view_android.cc:159:
using_delegated_renderer_(UsingDelegatedRenderer()) {
I'd remove the function now and inline the code here, so that it doesn't
accidentally get called elsewhere instead of using the boolean.

[aelias_OOO_until_Jul13, 2014-02-14 19:18:20]

lgtm modulo same comment as sievers@

[sivag, 2014-02-17 07:51:13]

https://codereview.chromium.org/160073005/diff/90001/content/browser/renderer...
File content/browser/renderer_host/render_widget_host_view_android.cc (right):

https://codereview.chromium.org/160073005/diff/90001/content/browser/renderer...
content/browser/renderer_host/render_widget_host_view_android.cc:159:
using_delegated_renderer_(UsingDelegatedRenderer()) {
On 2014/02/14 18:42:48, sievers wrote:
> I'd remove the function now and inline the code here, so that it doesn't
> accidentally get called elsewhere instead of using the boolean.

Done.

[sivag, 2014-02-17 07:51:20]

The CQ bit was checked by siva.gunturi@samsung.com

[commit-bot: I haz the power, 2014-02-17 07:51:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/siva.gunturi@samsung.com/160073005/170001

[commit-bot: I haz the power, 2014-02-17 12:33:52]

Change committed as 251665