This was mentioned on the jquery-dev group or somewhere else, I don't remember exactly. It's about Mootools not extending properly the document object I think, I'm sure it has been fixed in the trunk.
I suspect this is it: http://github.com/jeresig/sizzle/commit/0a9df6e0b68a91dd1f59bfd6e7941843cba32a7a On Jan 20, 3:26 am, Neil Craig <neil.big.cr...@gmail.com> wrote: > Hi > > I'm working on a project that used to implement Mootools 1.11. It has > been decided however that we switch over to jQuery since it is much > faster and more light-weight. > > However, because we have created numerous widgets in Mootools, we > still have to use it while the jQuery equivalent widgets are being > developed. We have implemented the noConflict method with various > degrees of success. > > We have noticed that after upgrading from jQuery 1.2.6 to 1.3, that > using the old Mootools 1.11 along with jQuery, renders the class > selector useless. > > Executing something like jQuery(".myClass") or jQuery("div.myClass") > returns an error: "context.getElementsByClassName is not a function". > It occurs in the following section: > > if ( document.documentElement.getElementsByClassName ) { > Expr.order.splice(1, 0, "CLASS"); > Expr.find.CLASS = function(match, context) { > return context.getElementsByClassName(match[1]); > }; > > } > > Strangely enough the expression "if > ( document.documentElement.getElementsByClassName )" yields a result > of true, but context.getElementsByClassName is still undefined. The > "context" in this case is the document object itself. > > I'm suspecting that it has got something to do with Mootools that is > throwing a spanner into the works. Has anyone encountered a similar > problem? Or has anyone got some comments on the subject?