[Webkit-unassigned] [Bug 18991] SquirrelFish: Major codegen issue in a.b=expr, a[b]=expr

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Wed May 14 06:14:59 PDT 2008


http://bugs.webkit.org/show_bug.cgi?id=18991





------- Comment #7 from cwzwarich at uwaterloo.ca  2008-05-14 06:14 PDT -------
The naive patch I posted is about a 1.4% ms performance regression on ToT
SquirrelFish. My slightly better version (that still misses some cases) is
about a 0.8-0.9% regression. The tracking of assignments alone doesn't seem to
affect much, as it is no change on SunSpider tending towards a 0.1% regression.
I'll try fixing the missing case that causes the problem with eBay and see how
it looks.

In theory, we also have to deal with the situation where parameters are
modified with f.arguments, although this is a pretty terrible programming
technique, so in addition to any checks for assignments in the subscript or on
the right hand side, we also need to check whether the identifier involved is a
function parameter.


-- 
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.



More information about the webkit-unassigned mailing list