John you should start with Sizzle then, where the basic build will be something like:
function Sizzle(selector, context){ return Array.prototype.slice.call((context || document).quertySelectorAll(selector)); }; Makes sense for every Sizzle based library, famous or "home made" one, no? Regards On Sat, Aug 22, 2009 at 7:32 PM, John Resig <jere...@gmail.com> wrote: > > Additionally, if there is an appreciable difference, it's likely that > we would just change the logic in jQuery itself to only initialize the > specific code branches once (which is something that we should be > doing anyway). > > --John > > > > On Sat, Aug 22, 2009 at 2:01 PM, Kevin Dalman<kevin.dal...@gmail.com> > wrote: > > > > For this idea to be taken seriously, I think you would need to create > > such a version and benchmark it VS the all-in-one jQuery.js. IF you > > can show significantl performance differences, then perpahs your idea > > could gain some traction (though I'm not in favor of it). But if we > > are only talking about *a few milliseconds* difference in normal use, > > then this would not justify the increased maintenance and code > > required at every level - development, testing, documentation, library > > linking, etc. IMO, the performance improvemetns would need to be > > significant and *noticeable* to justify all this extra work and > > complication. Making users add branching logic to every page is not a > > small matter. > > > > My 2-cents. > > > > /Kevin > > > > On Aug 21, 4:01 am, ludovic <ludothebe...@gmail.com> wrote: > >> @DBj > >> > >> Not exactly. > >> > >> Before use, we generate a few browser-specific jquery files, like > >> jquery.ie7.js, jquery.ie6.js. Theses files are generated server-side > >> from the current jquery.js file from a support detection database. For > >> example, for IE6, we already know that the box model isn't supported. > >> So we replace all occurrences of if( jQuery.support.boxModel ) > >> { instruction a } else { instruction b } by instruction b. > >> > >> We still have a jQuery file like we have currently. It can work on a > >> generic browser as the feature detection is made dynamically. It is a > >> bit slower as there are more conditions. The file name is jquery.js. > >> This file is downloaded only if there is no browser-specific jquery > >> file. > >> > >> Now, when the page is loading, we can have two policices : > >> - One based on user agent : if it is a user agent corresponding to one > >> of our precomputed files, we download the precomputed one, and not > >> jquery.js. Else, we download the generic file, I mean jquery.js > >> > >> - Another one based on client-side detection, with for example > >> <!--[if lt IE 7]> > >> <!-- Internet Explorer before IE7 --> > >> <script src="jquery.2.ie_before_7.js" ></script> > >> <![endif]--> > >> > >> So, what I propose is what you called "the ultimate solution" and my > >> precedent message explained how I would generate the different files. > >> > >> Regards, > >> Ludovic > > > > > > > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "jQuery Development" group. To post to this group, send email to jquery-dev@googlegroups.com To unsubscribe from this group, send email to jquery-dev+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/jquery-dev?hl=en -~----------~----~----~----~------~----~------~--~---