Sniffing out what different DOMParser-enabled browsers return on invalid XML input. Launch and compare output across browsers.

XML parsererror research

function isValidXML(xmlString) {
    var parser = new DOMParser();
    var parsererrorNS = parser.parseFromString('INVALID', 'text/xml').getElementsByTagName("parsererror")[0].namespaceURI;
    var dom = parser.parseFromString(xmlString, 'text/xml');
    if(dom.getElementsByTagNameNS(parsererrorNS, 'parsererror').length > 0) {
        return false
    }
    return true;
}

function s(x) {var t = new XMLSerializer(); return x ? t.serializeToString(x): undefined;}
var p = new DOMParser();
var arr = [
	"fail",
	"abc",
	"ab",
	"a",
	"",
	null,
	undefined,
	"<ok />",
	"<notok>",
	"<notok><fail></notok>",
	"<notok><a/><b/><c><d/></notok>",
	'<?xml version="1.0" encoding="windows-1251"?><a><b/><c></a>',
	'<?xml version="1.0" encoding="windows-1251"?><invalid xmlns="http://www.w3.org/1999/xhtml"><foo></invalid>',
	'<?xml version="1.0" encoding="windows-1251"?><valid xmlns="http://www.w3.org/1999/xhtml"><foo/></valid>',
	'<?xml version="1.0" encoding="windows-1251"?><valid><foo/><parsererror xmlns="http://www.w3.org/1999/xhtml"/></valid>',  // false detect in Chrome
	'<?xml version="1.0" encoding="windows-1251"?><valid><foo/><parsererror xmlns="http://www.mozilla.org/newlayout/xml/parsererror.xml"/></valid>',  // false detect in Firefox
];
for(var i = 0; i<arr.length; i++) {
	var arg = arr[i];
	var doc = p.parseFromString(arg, "text/xml");
	var parsererror = doc.getElementsByTagName("parsererror")[0];
	var namespace = parsererror ? parsererror.namespaceURI : undefined;
	console.log("\n###xml string:###\n" + arg + "\n###document###\n" + s(doc) + "\n###parsererror node:###\n" + s(parsererror) + "\n###parsererror namespace###\n" +  namespace + "\n###test isValidXML: " + isValidXML(arg) +  "\n##########");
}



// final heuristic function to parse your XML. Beware of falsy alarms (examples in the tests above)
function tryParseXML(xmlString) {
    var parser = new DOMParser();
    var parsererrorNS = parser.parseFromString('INVALID', 'text/xml').getElementsByTagName("parsererror")[0].namespaceURI;
    var dom = parser.parseFromString(xmlString, 'text/xml');
    if(dom.getElementsByTagNameNS(parsererrorNS, 'parsererror').length > 0) {
        throw new Error('Error parsing XML');
    }
    return dom;
}

A disadvantage of this approach is that new DOMParser().parseFromString('INVALID', 'text/xml') reports the following error in Mozilla Firefox's browser console:

XML Parsing Error: syntax error
Location: ....
Line Number 1, Column 1:

The alternative would be to avoid creating an invalid document and use a wildcard for the namespace instead:

dom.getElementsByTagNameNS('*', 'parsererror')