[thelist] JavaScript variable scope wrong?

Wed Aug 3 08:40:23 CDT 2005

---- Original Message -----
From: "Paul Waring" <paul at xk7.net>
To: <thelist at lists.evolt.org>
Sent: Wednesday, August 03, 2005 2:26 PM
Subject: [thelist] JavaScript variable scope wrong?

> I'm having a bit of trouble with what I think is down to variable scope
> in JavaScript (though I'm not entirely sure if that's the case, I
> haven't written this much client-side scripting for several years). At
> the moment I've got code along the lines of the following:
>
> <script language="JavaScript" type="text/javascript">
> <!--
>
> var req;
> var response;
>
> function loadXMLDoc(url)
> {
> req = new XMLHttpRequest();
> req.onreadystatechange = processReqChange;
> req.open("GET", url, true);

This is an async request (that's what the true is doing). This may or may
not be related to the problem - it's possible that if you do it
synchronously, the rest of the code will wait until the resource is
returned. It might be worth reading up on this.

> req.send(null);
> }
>
> function processReqChange()
> {
> if ( req.readyState == 4 )
> {
> if ( req.status == 200 )
> {
> response = req.responseXML.documentElement;
> }
> }
> }

Why don't you make the function return a variable instead of assigning it to
response, which already exists? It seems that reusing response might be
confusing the issue.

>
> loadXMLDoc('urlToXML');
> alert("1. " + response); // prints "1. undefined"
>
> function init()
> {
> loadXMLDoc('urlToXML');
> alert("2. " + response); // prints "2. [object element]
> }
>
> -->
> </script>
> </head>
> <body onLoad="init();">
>
> I've snipped the code to do things like use ActiveXObject for IE (I'm
> only testing in Mozilla at the moment and the browser detection code
> works anyway, don't see much point in posting huge chunks of it) and the
> error checking for the processReqChange() function.
>
> Anyway, the problem is the two loadXMLDoc() calls (both are reading in
> the same URL). At the moment, the first one sets response to undefined,
> and the second one sets it to [object element] (which is correct, I've
> checked what's actually in that object and it's what I want). Here's
> where the weird behaviour that I don't understand starts though:
>
> 1. If I comment out the first loadXMLDoc call, the response variable in
the
> main <script> section *and* in the init() function is set to undefined,
> which is obviously a bit useless as I can't get at the XML I've tried to
> read in.
>
> 2. If I comment out the second loadXMLDoc call but leave the first one
> running, then within the init() function the response variable is set as
> I want it to be, and all is well. However, if I then get rid of the
> alert("1. " + response); code (because I don't want stuff like that on
> pages that non-developers are going to see) then the response within the
> init() function is undefined (this also happens if I comment out the
> alert("1.") part and leave both loadXMLDoc calls in).
>
> The best guess I can make is that either the response variable isn't
> being set *or* the script is somehow waiting for the loadXMLDoc function
> to finish before calling the alert("1.") code.
>
> Sorry if this seems a bit of a convoluted question, but I have very
> little idea as to why the script is playing up like this. The JavaScript
> console in Mozilla doesn't throw up any errors, except if I try to
> access the response variable anywhere when it is still undefined (e.g.
> response.getElementsByTagName()).
>
> Thanks in advance for any help/pointers,
>
> Paul
>
> --
> Rogue Tory
> http://www.roguetory.org.uk
> --
>
> * * Please support the community that supports you.  * *
> http://evolt.org/help_support_evolt/
>
> For unsubscribe and other options, including the Tip Harvester
> and archives of thelist go to: http://lists.evolt.org
> Workers of the Web, evolt !
>