Greg A Watson

Sounds like an approach to counter Cross Site Scripting attacks.

I'm trying to use the response from an Ajax to create a new row in a table element. Code is something like this:

//response parsed into a JS object 'response' var row = document.createElement("tr"); //..following code repeated several times in a function call var cell = document.createElement("td"); var txtNode = document.createTextNode(response.someProperty) cell.appendChild(txtNode); row.appendChild(cell); //....end of repeating code dojo.byId("resultsTable").appendChild(row);

This all works in fine in Firefox. In IE(7), no new row appears - however, if I look at the DOM using the IE Developer Toolbar, the new row & cells have been inserted at the correct position in the DOM.

Any ideas? All help much appreciated...

Thanks

Santosh, I've run your code in Firefox & IE7, and it works perfectly fine - your code also looks good. Could it be that the location of dojo.js doesn't match your file system? When the page is loaded, look in Firebug to see if Dojo has bootstrapped - if so, an object named 'dojo' should have been created.

Greg

Not sure if here or the JSF forum is the best bet, but I'm guessing there's a better chance someone here will understand the problem!

I'm looking at the adoption of Dojo within a fairly large enterprise IT shop where we have a large existing number of applications written in various Java based web frameworks, but primarily JSF.

I like the idea of using the JSF life cycle to service Ajax requests in this model, and we already have a large number of developers skilled in JSF, so it seems a no-brainer. However, I've run into the obvious problem that when I try to add the dojoType attribute to a JSF tag, the validation phase of the markup against the JSF rules throws a hissy fit - it doesn't like the dojoType attribute, so won't parse the page. Has anyone else come up against this problem? I'm wondering if it's feasible / sensible / possible to just extend/replace the existing TLD to allow it as a type (I've spent a bit of time playing around with this and, it may be my ham-fistedness, but it doesn't appear as easy as I'd imagined it would be).

Has anyone else use these approaches together? Any suggestions / pitfalls I should look out for?

Thanks folks

Hi Frank,

I'd be interested in your thoughts on how RIA development with Dojo (and/or the likes of jQuery, Scriptaculous, etc.) sits with the player-based technologies. Although I'm advocating the use of Dojo within my organisation to a degree of richness to existing applications, I'd seen the medium to long term RIA picture being dominated by technologies such as Adobe Flex and Microsoft Silverlight (when it's maturity & market player penetration allow it to be.

However, with the advent of ever quicker & more sophisticated JavaScript VMs such as V8 & Squirrelfish, along with the great work being done by various Open Source toolkits and the proliferation of APIs from the likes of Google, I'm no longer sure that the JavaScript approach is just the "stop gap" I saw it as being. What's your opinion?
[ September 30, 2008: Message edited by: Ulf Dittmer ]

I had a method which opens up http connections to URLs outside our DMZ, and to hide the implementation details of our firewall I make the connection via a proxy. I've previously used the java.net.HttpURLConnection class, and specified a proxy by setting the httpProxy, httpProxy as system properties. The code to open the connection itself was fairly straightforward ('url' is a java.net.URL object);

HttpURLConnection conn = (HttpURLConnection) url.openConnection();

However...another application has deployed into the same JVM where my application lives, and due to co-existence issues (insert long story here) I can no longer set these properties in the JVM. To try to just set the proxy on a connection by connection basis, I've changed my app code to use the Jakarta Commons project; I'm using the HttpClient class to create the connection, and then a HostConfiguration object to set the proxy details;

HttpClient client = new HttpClient(); PostMethod postMethod = new PostMethod(url.toString()); // sets various headers RequestEntity re = new StringRequestEntity(data); postMethod.setRequestEntity(re); HostConfiguration hostConfig = new HostConfiguration(); // the 'proxy' object holds the server name and port hostConfig.setProxy(proxy.getProxyServer(), proxy.getProxyPort()); client.setHostConfiguration(hostConfig); // do the send int status = client.executeMethod(postMethod);

The proxy server/port are identical to the one I used before (when using HttpURLConnection, but I now get the following error from my server code when I tried to do the send;

org.apache.commons.httpclient.HttpMethodDirector I/O exception (java.net.SocketException) caught when processing request: Operation timed out: connect:could be due to invalid address org.apache.commons.httpclient.HttpMethodDirector Retrying request

Any help much appreciated,

Thanks, Greg