Fix number of expected select tests from commit 879799fe. Thanks to @jitter for the catch.

Make sure that the contents of the element is still in place when html(Function) is called. Fixes #6733.

Make .attr(name, null) equivalent to removeAttr(name). (Was roughly this before - but is now consistent across platforms). Fixes #6341.

Removing unique check on closest(Array) (it wasn't appropriate there and broke functionality elsewhere) - follow-up to commit 8c41325a. Also added another test for closest(Array) after a discussion with @jitter in 26b0e913.

Make sure that opacity is being reset properly on a show animation. Additionally expose jQuery.isNaN from the data module.

Add a feature test for options inside a disabled select. Follow-up to 157a383d.

Only do the hasAttribute check if the method exists - IE falls back to the === null check below which allows this to still work. Follow-up to cb40495b - thanks to @jitter for the catch.

Bug fix for commit 2c4b2080 - technique wasn't working as expected in WebKit browsers. Thanks to @jitter for the bug fix.

Ensure .val(null) sets an empty string on IE6/7; fixes #5163.

Make sure that requests without a body don't set contentType, and a zero-length body is sent rather than null. Possible fix for #6811 and #6674.

Don't have .val() return selected-but-disabled options, or selected options inside a disabled optgroup. Doesn't change the .val() returned for a disabled select. Fixes #3240, adapted from Nathan Hammond's patch there.

Use origSettings.context unless it's undefined; for that case use merged settings as context for backcompat with jQuery <1.3. Fixes #6727.

For JSON and script requests, set the scriptCharset before the url so IE won't interpret it the wrong way; fixes #4855.

Allow data to be bound to Flash objects (but still stopping short of attaching to applets. Fixes #6121.

Make sure that a clone of the animate speed properties are used to avoid recursion problems. Fixes #3583.