[web] javascript reading files

Take a look at the recent threads in this forum on &#106avascript HTTP requests and &#106avascript XML requests. You could use that (write a PHP script that outputs the requested data as an XML page, parse and update current page through &#106avascript). Other than that I know of no way to load things from the server without refreshing the page.

If the host doesn't support anything complex, you could always create a second (0 pixel) frame to query the site. How complex things get depend on what your user interaction is. If it's something like periodically updating stock quotes, it's very simple. If you are doing some sort of form submission, it will get harder as you'll have to have a variable to store whether or not the hidden frame is getting information from the server. Once the frame loads, &#106avascript on it will need to report to update the main frame.

Without looking arround:

function gettext()
{
var http;
var url;

try
{
http = new XMLHttpRequest();
}
catch (e)
{
try
{
http = new ActiveXObject("Microsoft.XMLHTTP");
}
catch (e2)
{
http=false;
}
}
if(!http)
{
return;
}
http.open("GET", "http://yourwebsite/yoururl");
http.onreadystatechange = function ()
{
if (http.readyState==4)
alert(http.responseText);
}

http.send(null);
}

Unfortunately, on IE, the XmlHttpRequest and the XML Document objects are considered to be "ActiveX" objects, so if ActiveX (or ActiveX scripting) are disabled, you're not going to be able to use them.

This is a pity, as these are the only obvious ways of reading file with &#106avascript.<br><br>I suppose in a hacky way you could store the file contents in a .js file as a huge literal string or something, but that's nasty.<br><br>Mark

what about an iframe? they are way more simple then regular frames.

Steps I took to get past this:

1. Create SCRIPT DOM object
2. Keep reference to DOM object
3. Set URL attribute to something like, http://server/rpc/motd.php
The server will now recieve this request, and respond as a &#106avascript file:
rpcMotd_Set('Woo it worked');

(rpcMotd_Set being a function loaded from a JS file on the client, for setting the MOTD box on the game login page).


I ended up expanding on this method to allow more than one RPC request at any one time, deleting the DOM objects as and when they were completed. The way I did this was to assume a timeout of 10 seconds for any RPC request(plenty of time). Each active script DOM object has a timeOut attribute which is initially at 2, lowered by 1 every 5 seconds, when it reaches 0, the object is deleted.

I presume Google Suggest must manage it somehow, and without ActiveX since I'm using firefox. Can't see what their code does tho.

Maybe I told a lie. What do you make of this? It's a little over my head.

When I said you couldn't do it without ActiveX, I meant that you couldn't do it in *IE* without ActiveX.

Of course other browsers allow you to use XML DOM objects and XmlHttpRequest without ActiveX - they have their own, non-ActiveX based versions of these objects.

But if you want to do it in IE with ActiveX object scripting disabled, you can forget it.

Mark

Is's sorta working, the fetchDB function fetches the database file for the co-ordinates it's given and just dumps it for now. It is called whenever the map is scrolled or refreshed.

Have a look at it now. It's a wierd bug if you check the script souce, the second time fetchDB is called there's an exception, but not the first. What's going on? My only guess is that the HTTP request is asynchrynous and the second one starts before the first finishes. If this is the case, how can I make the code wait without hanging the browser?

Also, it doesn't dump the results until the second time either map is scrolled, which it believes is the first. Maybe the actual first attempt was abandoned because of the exception. ????????