Bug 100736 – [EFL][WK2][AC] Avoid storing dirty rects in a Vector inside EwkViewImpl

• Component: WebKit EFL
• Status: RESOLVED FIXED

[Chris Dumez]

http://trac.webkit.org/changeset/132483 is storing dirty rectangles in a Vector data member inside EwkView and then creates a Region from them in EwkViewImpl::displayTimerFired(). It would therefore be more optimal to keep the dirty rects as a Region instead of Vector to avoid iterating over the Vector to construct a Region from it.

[Chris Dumez]

Created attachment 171402 [details]
Patch

[Chris Dumez]

I have tried running in MiniBrowser:
http://trac.webkit.org/export/132877/trunk/Websites/webkit.org/blog-files/3d-transforms/poster-circle.html

I could not reproduce infinite loop problem and the animation seems to work properly.

I enabled AC using:
--- a/Tools/Scripts/webkitperl/FeatureList.pm
+++ b/Tools/Scripts/webkitperl/FeatureList.pm
@@ -392,7 +392,7 @@ my @features = (
       define => "ENABLE_TEXT_AUTOSIZING", default => 0, value => \$textAutosizingSupport },
 
     { option => "tiled-backing-store", desc => "Toggle Tiled Backing Store support",
-      define => "WTF_USE_TILED_BACKING_STORE", default => isQt(), value => \$tiledBackingStoreSupport },
+      define => "WTF_USE_TILED_BACKING_STORE", default => (isQt() || isEfl()), value => \$tiledBackingStoreSupport },
 
     { option => "touch-events", desc => "Toggle Touch Events support",
       define => "ENABLE_TOUCH_EVENTS", default => (isQt() || isBlackBerry() || isEfl()), value => \$touchEventsSupport },

[Kenneth Rohde Christiansen]

Comment on attachment 171402 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=171402&action=review

> Source/WebKit2/UIProcess/API/efl/EwkViewImpl.h:275
> +    OwnPtr<WebCore::Region> m_dirtyRegion;

Ah I didn't even know we had a region, but of course we do.

[Chris Dumez]

Yael, would you please take a look before we commit this?

[Yael]

(In reply to comment #4)
> Yael, would you please take a look before we commit this?

LGTM,

I am not sure why you did not observe a crash. I got the crash I posted in https://bugs.webkit.org/show_bug.cgi?id=100288#c0 very consistently.

[Chris Dumez]

(In reply to comment #5)
> (In reply to comment #4)
> > Yael, would you please take a look before we commit this?
> 
> LGTM,
> 
> I am not sure why you did not observe a crash. I got the crash I posted in https://bugs.webkit.org/show_bug.cgi?id=100288#c0 very consistently.

To be clear, I did not experience the crash with my patch (and yours) applied so it is not surprising that I couldn't reproduce the patch. I merely wanted to make sure that my patch did not cause a regression (i.e. reintroduce the crash you had fixed).

Could someone please cq+ then?

[WebKit Review Bot]

Comment on attachment 171402 [details]
Patch

Clearing flags on attachment: 171402

Committed r133046: <http://trac.webkit.org/changeset/133046>

[WebKit Review Bot]

All reviewed patches have been landed.  Closing bug.