Subj : Re: XML declaration broken going from Mozilla 1.8a6 to 1.8b
To   : netscape.public.mozilla.jseng
From : torntrousers
Date : Sun Mar 13 2005 02:28 am

Brendan Eich <brendan@meer.net> wrote in message news:<42332BBF.9050504@meer.net>...
> Martin Honnen wrote:
> 
> > Maybe Brendan and/or Igor know better whether new XML() in E4X should be 
> > able to consume the complete XML markup from a document including the 
> > XML declaration.
> 
> 
> The specification's authors' intentions are not so clear, but as you 
> note, the spec is: what SpiderMonkey does here is correct per ECMA-357.
> 
> Note that if you make an E4X wrapper around an XML DOM document via
> 
>    var x = new XML(document);
> 
> (a feature not yet implemented in SpiderMonkey and Gecko -- soon!), you 
> shouldn't have any such problems.  So it seems ECMA-357 wants you to be 
> able to parse XML into a DOM and then wrap that with E4X glue, which 
> allows the XML declaration pseudo-PI; or parse a JS string containing 
> XML, which does not.
> 
> I'll raise this issue with ECMA TG1.
> 
> /be

Well this doesn't seem perfect, surely whatever the spec actually says
the intention would have been for this to be possible? It sounds like
I'll have to use a hack along the lines of:

     function getAsXML(s) {
         var sx = s;
         if ("<?xml" == s.substr(0,5)) {
             var l = s.length;
             var i = s.indexOf("?>");
             sx = s.substring(i+2,l);
         }
         var xml = new XML(sx);
         return xml;
     }

Seeing as the spec is unclear and Rhino can deal with this fine
couldn't Spidermonkey be fixed to do a fiddle like this for me?

   ...ant

PS. Thanks for all the help. E4X rocks

.