Strip all namespaces from SUSHI responses

[nkuitse]

This is meant to close #485 but has not been tested.

[queryluke]

I was doing some debugging for a Sirsi client and came across this same issue. The problem is that some publishers add namespaces to their SUSHI XML.

Coral removes some namespaces, but I've seen the following in my debugging: ns2, ns3, rr, d, np

Unfortunately, I could not get @nkuitse 's code to work on my development server. Instead I changed their stripNamespaces function to the following:

private function stripNamespaces($string) {
        $sxe = new SimpleXMLElement($string);
        $doc = new DOMDocument();
        $doc->loadXML($string);


        foreach ($sxe->getNamespaces(true) as $name => $uri) {
        	if(!empty($name)){
                $finder = new DOMXPath($doc);
                $nodes = $finder->query("//*[namespace::{$name} and not(../namespace::{$name})]");
                foreach ($nodes as $n) {
                    $ns_uri = $n->lookupNamespaceURI($name);
                    $n->removeAttributeNS($ns_uri, $name);
                }
			}
        }
        return $doc->saveXML(null, LIBXML_NOEMPTYTAG);
}

Or, you can see my branch here.
I've tested my code with the following vendors:

• Brill
• Cambridge University Press
• Elsevier
• Sage
• Springer
• Taylor and Francis

I can either make a new PR or if someone wants to request a change or confirm @nkuitse code works.

[nkuitse]

This fix by @queryluke seems much cleaner than mine in #485. Thanks!

[andyp-uk]

I've tested @queryluke 's fix in PR #611 and this looks good so I've merged this. I think this PR can therefore be closed.