My guess would be that the speed difference stems from the Sizzle strategy of searching inside-out from the expression. I.e. using the last part of the query first, and then filtering that set using the previous parts. Perhaps other libraries search outside-in?
In the case of "ul .tocline2" the approach taken by Sizzle seems to fare especially poorly (compared to the others). Of course, this issue would only affect cases where we are using parent-child relations. Cheers, /Per - running IE6 inside CrossOver, so that might slow down my results a bit too On Thu, Oct 23, 2008 at 6:12 PM, John Resig <[EMAIL PROTECTED]> wrote: > Oh, I forgot to mention that I put the test suite online here: > http://ejohn.org/apps/sizzle/test/ > > and the performance suite is here: > http://ejohn.org/apps/sizzle/speed/ > > Taking a quick peek at IE 6 I see a lot of areas in which small improvements > could yield large results ("div p", "div + p", ".class"). As it is it's > faster than all the other major libraries. DOMAssistant has some tricks > which could definitely help here so I'll investigate and report back. > > --John > > > On Thu, Oct 23, 2008 at 11:37 AM, John Resig <[EMAIL PROTECTED]> wrote: >> >> Hey All - >> >> Sizzle is now passing the test suite 100% in IE 6, Firefox 3, and Safari >> 3.1. There is one failing test in Firefox 3.1b1 and in Opera 9.6 (both are >> specific browser bugs, and relatively minor, so I'm filing those with the >> vendors). >> >> I've fixed all of the previously-discussed issues in this thread. >> >> With compliance in order I want to look back and tackle two things: >> - Library-specific hook code (it's looking likely that Sizzle might make >> its way in to jQuery, MochiKit, Dojo, and Prototype). >> - Speed (the performance in IE can still use some work so I'm looking in >> to that) >> >> --John >> >> >> On Mon, Oct 20, 2008 at 12:13 PM, Arnar Birgisson <[EMAIL PROTECTED]> >> wrote: >>> >>> On Mon, Oct 20, 2008 at 17:40, John Resig <[EMAIL PROTECTED]> wrote: >>> >> I don't know what other MochiKitters say about including Sizzle.js as >>> >> a seperate file. Per, Bob? >>> > >>> > So it seems like the major difference is that your selector method >>> > (findChildElements) is able to take an array of results, correct? >>> >>> Yes, correct. This is something we inherited from Prototype, from >>> where the current Selector module was ported. Perhaps we should change >>> the API, I don't know.. >>> >>> cheers, >>> Arnar >> > > > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "MochiKit" group. To post to this group, send email to [email protected] To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/mochikit?hl=en -~----------~----~----~----~------~----~------~--~---
