I've discovered two issues with 1.2.2b with the ready event:

For once, the event handler doesn't get a reference to jQuery anymore, instead only a (useless for me) event object is passed.

Test:

test("jQuery(function($) {})", function() {
	stop();
	jQuery(function($) {
		equals(jQuery, $, "ready doesn't provide an event object, instead it provides a reference to the jQuery function, see http://docs.jquery.com/Events/ready#fn");
		start();
	});
});

The above test fails due to another issue: Binding a ready event after the event was issued doesn't do anything anymore. In 1.2.1 the handler would be triggered immediately. Though this behaviour isn't documented anywhere, but changing it definitely breaks stuff, so we should be the docs, too. Eg. adding this to ​http://docs.jquery.com/Events/ready#fn:

<p>If the handler is bound after the event already fired, it is triggered immediately.</p>

No idea how to fix it due to the new special event handling. An introduction to that would be great.

not() fails to filter out a single DOM element. Test:

Index: test/unit/core.js
===================================================================
--- test/unit/core.js	(revision 4206)
+++ test/unit/core.js	(working copy)
@@ -1058,7 +1058,7 @@
 });
 
 test("not()", function() {
-	expect(7);
+	expect(8);
 	ok( $("#main > p#ap > a").not("#google").length == 2, "not('selector')" );
 	ok( $("#main > p#ap > a").not(document.getElementById("google")).length == 2, "not(DOMElement)" );
 	isSet( $("p").not(".result").get(), q("firstp", "ap", "sndp", "en", "sap", "first"), "not('.class')" );
@@ -1066,6 +1066,9 @@
 	isSet( $("p").not($("#ap, #sndp, .result")).get(), q("firstp", "en", "sap", "first"), "not(jQuery)" );
 	ok( $("p").not(document.getElementsByTagName("p")).length == 0, "not(Array-like DOM collection)" );
 	isSet( $("#form option").not("option.emptyopt:contains('Nothing'),[selected],[value='1']").get(), q("option1c", "option1d", "option2c", "option3d" ), "not('complex selector')");
+	
+	var selects = $("#form select");
+	isSet( selects.not( selects[1] ), ["select1", "select3"], "filter out DOM element");
 });
 
 test("andSelf()", function() {

Most likely the optimization in not() caused this to regress:

return this.filter(function() {
	// check to see if the selector is array-like otherwise assume it is just a DOM element
	return ( selector.length && selector[selector.length - 1] !== undefined )
		? jQuery.inArray( this, selector ) < 0
		: this != selector;

I get weird errors in IE 6/7 on the accordion demo page, starting with an illegal argument, sometimes also a out of memory error.

Debugger says the illegal argument error is thrown in line 1077 (elem[ name ] = value;), with name="height" and value="NaNpx". Looks like a wrong height is calculated during an animation, resulting in a NaN.

To reproduce: Checkout trunk/plugins/accordion, update lib/jquery.js to latest, load accordion/demo/ in IE and click on top left accordion.

Works fine with 1.2.1.