[webkit-reviews] review granted: [Bug 70032] Enable ParallelJobs by default : [Attachment 112859] Patch

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Mon Nov 21 03:58:03 PST 2011


Zoltan Herczeg <zherczeg at webkit.org> has granted Balazs Kelemen
<kbalazs at webkit.org>'s request for review:
Bug 70032: Enable ParallelJobs by default
https://bugs.webkit.org/show_bug.cgi?id=70032

Attachment 112859: Patch
https://bugs.webkit.org/attachment.cgi?id=112859&action=review

------- Additional Comments from Zoltan Herczeg <zherczeg at webkit.org>
There was no objection. r=me with the following minor comments:

(Please check that the patch still apply and works)

View in context: https://bugs.webkit.org/attachment.cgi?id=112859&action=review


> Source/JavaScriptCore/ChangeLog:9
> +	   great speedup for SVG on multicore.

great -> considerable

> Source/WebCore/ChangeLog:11
> +	   great speedup for SVG on multicore.

ditto

> Source/WebCore/ChangeLog:15
> +	   by qualifying ParallelJobs with the WTF namespace (otherwise
> +	   MSVC believes it belongs to WebCore which is likely a compiler bug).


Are you sure? How do other WTF classes work?

> Source/JavaScriptCore/wtf/ParallelJobsGeneric.cpp:87
> +    // The work for the main thread

Comment should be terminated by a point.

> Source/JavaScriptCore/wtf/ParallelJobsGeneric.h:50
> +    void execute(unsigned char* parameters);

unsigned char* ? The thread function expects void*, please change this to
void*.

> Source/WebCore/platform/graphics/filters/FEConvolveMatrix.cpp:459
> +	       // Fallback to single thread mode.

threaded

> Source/WebCore/platform/graphics/filters/FEGaussianBlur.cpp:216
> +	   // Fallback to single thread mode.

ditto.

> Source/WebCore/platform/graphics/filters/FELighting.cpp:277
> +	   // Fallback to single thread mode.

ditto

> Source/WebCore/platform/graphics/filters/FETurbulence.h:74
>      static const int s_minimalRectDimension = (100 * 100); // Empirical data
limit for parallel jobs

Comment should be terminated by a point.


More information about the webkit-reviews mailing list