RE: [Jquery-dev] Massive Speed Improvements

RE: [Jquery-dev] Massive Speed Improvements

Woo hoo, this sounds great! After doing some recent exploration in the area
of find/filter I was going to suggest some work in that area. The new
benchmarking framework makes it a lot easier to focus on the bottlenecks.
I will have some time next week because I'm off work, so if you need help
squashing bugs on anything let me know. I'm also very good at creating bugs
if you need that. :)
-----Original Message-----
From: jquery-dev-bounces@bassistance.de
[mailto:jquery-dev-bounces@bassistance.de] On Behalf Of John Resig
Sent: Wednesday, December 20, 2006 5:49 PM
To: jquery-dev@bassistance.de
Subject: [Jquery-dev] Massive Speed Improvements
Hey Everyone -
I've got a nice surprise to go with the upcoming 1.1 release (I'm currently
shooting for a Jan 14th release with an alpha release early Jan - let me
know how that sounds for everyone).
These past two days I've been working on drastically increasing the speed of
the selector capabilities of jQuery. I've achieved speed increases across
the board - ranging from 400-1000%. The result is a library that is
completely faster than Prototype, and mostly faster than cssQuery.
Yehuda has been helping me in my work, he's built a complete speed-test
suite that's serving to be incredibly useful. You can test out all the
different libraries (jQuery new, jQuery "old", Prototype SVN, and cssQuery)
here:
http://24.47.149.250:3000/jq_test.html
I plan on pushing this into SVN once I can be completely certain that
nothing has regressed. I also managed to fix a rather serious bug with the
CSS selector engine (all nth-child selections started at 0, instead of 1,
like the spec outlines). For this reason, I can't push this live until we
know, for certain, that we're going to be starting work on 1.1.
Let me know what you think of the improvements, they're not completely there
yet (I think I can still squeeze some ms out of the code). But as it stands,
I'm very very pleased. We're going to be able to do stuff like $(".foo") and
$("div#bar") without any significant decrease in speed - and other things
like $("tr:nth-child(even)") and $("table