Javascript syntax error Firefox - demo.html

GoogleCodeExporter commented 8 years ago

Hi,

I tracked down a javascript syntax error in my application to x2js.

FIrefox 34.0.5
https://x2js.googlecode.com/hg/demo.html

- syntax error ... demo.html:1 
- the highlighted line in code is line 1:demo.html - <!doctype html>

The error is unhelpful and seems to point generally to a broken <script> tag. 
This is not the case in my application nor apparently on the demo page.

The demo page still works as expected but I feel it is worth taking note of.

I've searched project issues (keywords: syntax error firefox/firebug) but I 
don't see that this error has been previously reported.

Thanks for the project, it has made life easier having such a good library for 
converting xml<->json.

Kind regards,
Darryl

Original issue reported on code.google.com by darryljcousins on 18 Dec 2014 at 2:06

GoogleCodeExporter commented 8 years ago

Hi, I have the same issue.
I was able to track it and found where it comes from. It seems that the firefox 
DOM parser does like this code:

try {
  parsererrorNS = parser.parseFromString("INVALID", "text/xml").childNodes[0].namespaceURI;
}
catch(err) {                    
  parsererrorNS = null;
}

When trying to parse the INVALID string, firefox DOM parser shows this "syntax 
error" on HTML line 1 in the console, event if the conversion error is 
surrounded by try-catch.

Original comment by geoffrey...@gmail.com on 20 Feb 2015 at 7:37

GoogleCodeExporter commented 8 years ago

Same here. Weird thing is that it works in Chrome perfectly. The XML seems also 
to be OK. But it always fails in Firefox within the try-catch block posted 
above.

Original comment by d...@dg-infotec.de on 27 Mar 2015 at 1:45

GoogleCodeExporter commented 8 years ago

I think the code is trying to solve the fact that some browsers include the 
parsererror element in another namespace (eg: 
http://stackoverflow.com/questions/11563554/how-do-i-detect-xml-parsing-errors-w
hen-using-javascripts-domparser-in-a-cross).

It parses an invalid xml to get the possible namespace for the parsererror 
element.

What I think would be a better solution (altough I don't know if it will work 
for all browsers) is to eliminate this check and use a wildcard to get all 
parsererror elements independent of the namespace.

The patch below show the changes I have made and it has worked for me:

@@ -489,21 +489,9 @@
        var xmlDoc;
        if (window.DOMParser) {
            var parser=new window.DOMParser();          
-           var parsererrorNS = null;
-           // IE9+ now is here
-           if(!isIEParser) {
-               try {
-                   parsererrorNS = parser.parseFromString("<", 
"text/xml").childNodes[0].namespaceURI;
-                   console.log(parsererrorNS);
-               }
-               catch(err) {                    
-                   console.log('error:' + err);
-                   parsererrorNS = null;
-               }
-           }
            try {
                xmlDoc = parser.parseFromString( xmlDocStr, "text/xml" );
-               if( parsererrorNS!= null && xmlDoc.getElementsByTagNameNS(parsererrorNS, 
"parsererror").length > 0) {
+               if(xmlDoc.getElementsByTagNameNS("*", "parsererror").length > 0) {
                    //throw new Error('Error parsing XML: '+xmlDocStr);
                    xmlDoc = null;
                }

Original comment by vladlo...@gmail.com on 2 Jun 2015 at 10:27

Enatel / xmltojson

Javascript syntax error Firefox - demo.html #32