Seemingly inconsistent Chrome XPath results

I have a very simple XML document that I've retrieved from a larger parent. See 'accountxml' below:

<accounts xmlns=​"https:​/​/​domain.com/path">​
    <customerid>​sometext</customerid>​
    <login>​sometext​</login>​
    <companyname>​sometext​</companyname>​
    <canmanageclients>​sometext</canmanageclients>​
</accounts>​

Simple enough - just one namespace URL (the URL referred to as ns2 in myNS namespace map below). Querying login:

accountxml.evaluate('//ns2:login',accountxml,myNS,XPathResult.FIRST_ORDERED_NODE_TYPE,null).singleNodeValue;

returns OK:

<login>sometext</login>​

But:

accountxml.evaluate('//ns2:customerid',accountxml,myNS,XPathResult.FIRST_ORDERED_NODE_TYPE,null).singleNodeValue;

Returns:

null

This is such a simple query I'm surprised it breaks. But I can reproduce it 100% of the time. Am I doing som开发者_如何学Goething wrong, or is this a bug in Chrome?

Thanks to Gael for encouraging me to look at capitalization again.

The key to this is the document is taken from a larger parent. Although the small document above is printed exactly as the JS console sees it - with all element names in lower case, the larger parent it is taken from uses initialLowerCase for the element names.

Even though I'm querying the smaller fragment, I still need to use the casing of the parent rather than the casing of the child.

accountxml.evaluate('//ns2:customerId',accountxml,myNS,XPathResult.FIRST_ORDERED_NODE_TYPE,null).singleNodeValue;

I'm not sure if this is a Chrome bug or an expected behavior.