Reputation: 1997
I'm writing an application and I'm trying to tie simple AJAX functionality in. It works well in Mozilla Firefox, but there's an interesting bug in Internet Explorer: Each of the links can only be clicked once. The browser must be completely restarted, simply reloading the page won't work. I've written a very simple example application that demonstrates this.
Javascript reproduced below:
var xmlHttp = new XMLHttpRequest();
/*
item: the object clicked on
type: the type of action to perform (one of 'image','text' or 'blurb'
*/
function select(item,type)
{
//Deselect the previously selected 'selected' object
if(document.getElementById('selected')!=null)
{
document.getElementById('selected').id = '';
}
//reselect the new selcted object
item.id = 'selected';
//get the appropriate page
if(type=='image')
xmlHttp.open("GET","image.php");
else if (type=='text')
xmlHttp.open("GET","textbox.php");
else if(type=='blurb')
xmlHttp.open("GET","blurb.php");
xmlHttp.send(null);
xmlHttp.onreadystatechange = catchResponse;
return false;
}
function catchResponse()
{
if(xmlHttp.readyState == 4)
{
document.getElementById("page").innerHTML=xmlHttp.responseText;
}
return false;
}
Any help would be appreciated.
Upvotes: 13
Views: 8254
Reputation: 316
Using Dojo, this can be done using dojo.date.stamp
, just adding the following to the url:
"...&ts=" + dojo.date.stamp.toISOString(new Date())
Upvotes: 1
Reputation: 36745
In jQuery.ajax, you can set the "cache" setting to false:
http://api.jquery.com/jQuery.ajax/
Upvotes: 0
Reputation: 42200
The response header that has worked best for me in the IE AJAX case is Expires: -1
, which is probably not per spec but mentioned in a Microsoft Support Article (How to prevent caching in Internet Explorer). This is used in conjunction with Cache-Control: no-cache
and Pragma: no-cache
.
Upvotes: 1
Reputation:
Read No Problems section in [link text][1] [1]: http://en.wikipedia.org/wiki/XMLHttpRequest
Upvotes: 1
Reputation:
xmlHttp.open("GET","blurb.php?"+Math.random();
I agree with this one.. it works perfectly as a solution to this problem. the problem is that IE7's caching of urls were terrible, ignoring the no-cache header and save the resource to its cache using its url as key index, so the best solution is to add a random parameter to the GET url.
Upvotes: 0
Reputation: 4529
The problem is that IE does wacky things when the response handler is set before open
is called. You aren't doing that for the first xhr request, but since you reuse the xhr object, when the second open is called, the response handler is already set. That may be confusing, but the solution is simple. Create a new xhr object for each request:
move the:
var xmlHttp = new XMLHttpRequest();
inside the select function.
Upvotes: 3
Reputation: 24462
This happens because Internet Explorer ignores the no-cache directive, and caches the results of ajax calls. Then, if the next request is identical, it will just serve up the cached version. There's an easy workaround, and that is to just append random string on the end of your query.
xmlHttp.open("GET","blurb.php?"+Math.random();
Upvotes: 16
Reputation: 37633
It looks like IE is caching the response. If you either change your calls to POST methods, or send the appropriate headers to tell IE not to cache the response, it should work.
The headers I send to be sure it doesn't cache are:
Pragma: no-cache
Cache-Control: no-cache
Expires: Fri, 30 Oct 1998 14:19:41 GMT
Note the expiration date can be any time in the past.
Upvotes: 6