Some progress...
by using a garbage collection bin (a dummy div to appendChild and then empty
the innerHTML), the DOM usage is stable now. Code:
jQuery.fn.discard = function(){
var garbageBin = document.getElementById('IELeakGarbageBin');
if (!garbageBin) {
garbageBin = document.createElement('DIV');
garbageBin.id = 'IELeakGarbageBin';
//garbageBin.style.display = 'none';
document.body.appendChild(garbageBin);
}
this
.unbind() //unbind all handlers
.each(function(){ // move the element to the garbage bin
garbageBin.appendChild(this);
garbageBin.innerHTML = '';
});
garbageBin = null;
}
Usage: e.g. $("test>tbody>*").discard();
But there are stills some memory keep increasing in IE when those addRows
and deleteRows repeat again and again. Not knowing why.
On 5/22/07, Jacky <[EMAIL PROTECTED]> wrote:
I have some problem in creating the ticket. It always timeout when I
submitting... I will try later.
Anyway, here is an example about it.
Load it in IESieve , cilck refresh rows or set interval and watch the
memory and dom usage.
<html>
<head>
<title>Test Mem Leak in IE6</title>
<script type="text/javascript" src="jquery-1.1.2.js"></script>
<script type="text/javascript">
var datas = [
[1,2,3,4,5],
[1,2,3,4,5],
[1,2,3,4,5],
[1,2,3,4,5],
[1,2,3,4,5],
[1,2,3,4,5]
];
$(document).ready(function(){
$("#remove").click(removeAllRows);
$("#refresh").click(refreshRows);
$("#interval").click(startRefresh);
});
function startRefresh(){
var interval = setInterval(refreshRows,5*1000);
}
function refreshRows(){
removeAllRows();
for(var i=0; i<datas.length; i++){
var tr = document.createElement("tr");
for(var j=0; j<datas[i].length; j++){
var td = document.createElement("td");
var d = datas[i][j]*Math.floor(Math.random
()*100+1);
td.innerHTML =
"<a href='#' onclick='return
false;'>"+d+"</a>" +
"<input type='button'
onclick='buttonClick(this)' value='"+datas[i][j]+"'/>";
$(tr).append(td);
td = null;
d = null;
}
var tb = $("#test>tbody").append(tr);
}
}
function removeAllRows(){
var tb = $("#test>tbody")
.find("*").unbind().end()
.html("");
}
function buttonClick(obj){
obj.value = 'btn';
}
</script>
</head>
<body>
<table id="test" border="1">
<thead>
<tr>
<th>th1</th>
<th>th2</th>
<th>th3</th>
<th>th4</th>
<th>th5</th>
</tr>
</thead>
<tbody>
</tbody>
</table>
<input type="button" value="remove all rows" id="remove"/>
<input type="button" value="refresh all rows" id="refresh"/>
<input type="button" value="start interval" id="interval"/>
</body>
</html>
--
Best Regards,
Jacky
網絡暴民 http://jacky.seezone.net
On 5/21/07, Brandon Aaron <[EMAIL PROTECTED]> wrote:
>
>
> If you could create a ticket and a test case for this it would help a
> lot! I believe I could find some time to investigate it in detail
> soon.
>
> --
> Brandon Aaron
>
> On 5/21/07, Jacky < [EMAIL PROTECTED]> wrote:
> > Hi all,
> >
> > I'm dealing with the IE memory problem.
> > There is an IE leak called 'pseudo leak', where removeChild() is not
> really
> > removing child from memory.
> > It is pseudo because it would be cleared after refresh.
> > (some detail:
> >
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/IETechCol/dnwebgen/ie_leak_patterns.asp
>
> > )
> >
> > However, my app is not page-driven, it would stay on same page and
> > refreshing ajax call, refreshing table rows, etc. So the memory would
> just
> > keep raising by the pseudo leak.
> >
> > It is said in some site that by setting innerHTML = "" would prevent
> the
> > leak.
> > So I try to do something like:
> >
> > $("tbody")
> > .find("*").unbind().end() //unbinding all handler to prevent
> another
> > closure leak
> > .find("td,tr").html("").remove().end() //setting html
> > ="" and remove itself
> > .html(""); //setting tbody html to empty
> >
> > I use IESieve ( http://home.wanadoo.nl/jsrosman/ ) to check
> > for the the dom object creation but it just keep increasing.
> > Is there anyone try to tackle this 'leak' before?
> > --
> > Best Regards,
> > Jacky
> > 網絡暴民 http://jacky.seezone.net
> >
>
--
Best Regards,
Jacky
網絡暴民 http://jacky.seezone.net
