[webkit-reviews] review denied: [Bug 26987] Add fast case for Math.floor(a / b) expressions : [Attachment 32298] Enable math.floor(a/b) optimization on interpreter

[bugzilla-daemon at webkit.org]

Oliver Hunt <oliver at apple.com> has denied Zoltan Herczeg
<zherczeg at inf.u-szeged.hu>'s request for review:
Bug 26987: Add fast case for Math.floor(a / b) expressions
https://bugs.webkit.org/show_bug.cgi?id=26987

Attachment 32298: Enable math.floor(a/b) optimization on interpreter
https://bugs.webkit.org/attachment.cgi?id=32298&action=review

------- Additional Comments from Oliver Hunt <oliver at apple.com>
I think the complexity of this patch isn't balanced performance gain.  I
recognise that we need to come up with a solution for the problem of the simple
builtin function, but i don't believe lexical specialisation is the solution. 
It works for call/apply but they have a *huge* overhead on many levels and
aren't cached in js, whereas things like the builtin Math functions (and also a
number of other builtin functions) are cached in ways that defeat lexical
caching, especially in the fast case, and the quantity of calls is likely to be
much larger than that of call and apply.

Just to make it clear, i do believe we want to improve performance of builtins,
i just don't believe lexical caching is the solution.