[Webkit-unassigned] [Bug 14112] Cross-port KDE KJS UnaryPlus and UnaryMinus optimizations

[bugzilla-daemon at webkit.org]

http://bugs.webkit.org/show_bug.cgi?id=14112





------- Comment #3 from mikelima at gmail.com  2007-06-15 01:39 PDT -------
Uhm, the nodes.cpp slipped in, as did the kjs_binding.h change, which I needed
at sometime to build things.

The kjs_binding.h part should be removed, the nodes,cpp, I'd like to keep
(should I send it as a different patch?), since it remove noise while comparing
the two branches.

For the rest, I originally intended to submit a patch to split makenodes.h out
of grammar.y, but I went on a bit further than that.

I can split up the patch, however I need input on how to do that.

In kjs, there has been the makenodes.h split and then the optimization.

Would patche in that order be ok?

About the optimization: are you concerned of debugging?
The unaryplus number -> number optimization is probably almost useless, but the
unaryminus number pattern happens often enough.

Harri porten measured something like a 1.7 improvement over md5-1.js with a
version of KJS including this patch and another one to convert number
arithmeticoperator number -> number. But obviously the tree will be different
from the original.


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