TestSwarm hasn't been happy since this commit, and everyone deserves to be happy.

This reverts commit 318d47b7.

Makes sure start is always called with default context and no arguments. Should stop FF from acting up. Hopefully, this ugly patch can be removed in the future.

Makes parseXML act like parseJSON when given an empty or non-string input: now returns null rather than throwing an exception. Incidently fixes #10527. Unit tests added.

Fixes #11264 or rather seriously limits the risk of global ajaxSettings screwing with script loading in domManip. Gotta love globals and sneaky dependencies. Unit test added.

Fixes #11426: getting the responseText of an xhr should be tried/caught because of IE's inability to give access to binary data. Unit test added.

No more php/js logic duplication. Ensures modules are still loaded separately when using "file:" protocol (makes debugging much easier, /kiss @rwaldron). Ensures test iframes will use the same jquery as specified in the main document. Known issue: chrome will cringe at cross-frame access using file: protocol, apparently chrome sees this as cross-domain... go figure.

Supplements #11356 and fixes unit test failures in FF 3.6.

When the browser window isn't visible, setTimeouts have 1-second granularity.

Okay, Jenkins, you were right this time.

Please folks, hold your questions. We don't explain em, we just fix em.

Reverses the previous search order.

May still be broken in Netscape Navigator 4.

This doesn't fix *all* of them (see the ticket for a supposedly complete list) but these were already handy so it was relatively cheap to fix them. If you need others fixed, add a custom step function as was done here. Thanks @scott_gonzalez!

Not really happy with this, but perhaps Jenkins will be.

Fix whitespace mismatch introduced in bf7a4df2

Easing: Only use the state to calculate the value. Fixes #11284 - Simplify easings to only rely on the state.

Fix prefilter comment typo.