[Webkit-unassigned] [Bug 46592] Convert CSSStyleSelector::applyProperty to use function pointers

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=46592





--- Comment #19 from Ryosuke Niwa <rniwa at webkit.org>  2011-01-19 16:01:11 PST ---
(In reply to comment #17)
> Splitting out the style application is a good idea. However, this undebuggable macro/function pointer salad is not an improvement either style or performance wise. I don't see the maintainability or hackability gains either.

I agree.  I don't think using all these macros give us any coding-style improvement over the existing code.

> Any performance testing for this kind of stuff would need to be done on ARM too as indirect branch prediction may be less effective there.

We should probably be testing on both x86 and ARM processors because having lots of small functions may affect L1 & L2 cache hit-rate, and it's really hard to predict what happens. For example, when I implemented recursive and iterative implementations (they're identical otherwise) of merge sort a couple of years ago, recursive version was consistently faster by 10%-30% due to some caching.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.