Cloning XMLDoc's with HTML5 nodeName's breaks on IE

[adebree]

For example:

var xmlStr = "<root><meter /></root>";
var xmlDoc = jQuery.parseXML( xmlStr );
var $xmlDoc = jQuery( xmlDoc );

var $meter = $xmlDoc.find( "meter" ).clone();

Due to "meter" being in the list of HTML5 nodeNames jQuery incorrectly uses the shimCloneNode() function for cloning. This results in $meter being a textNode (nodeType 3) and having the data 'undefined'.

And since the regex to detect the node names is only checking the start of the node name this also causes problems for elements in a HTMLDoc like <metering />

Patch is already done, will place a pull request.

[Rick Waldron]

This is actually a good argument for the removal of shimCloneNode and createSafeFragment

[adebree]

​https://github.com/jquery/jquery/pull/673

[Timmy Willison]

This is valid, but please provide a test case on ​http://jsfiddle.net to demonstrate the issue in this ticket.

@rwaldron: I think jQuery should be able to work without html5shim.

[adebree]

jsfiddle for this bug:

​http://jsfiddle.net/bMBgZ/

Run in IE8 or lower and the nodeType will be 3 instead of the expected 1 showing the clone() does not properly take place.

[Arne de Bree]

Fix #11291. Always clone XML docs with a genuine .cloneNode().

Changeset: bf7a4df22abcea10cf8f97eb31b5fa989fb4daf5