I am late for this discussion, but I think I could tell you some observations. I am working on the next version of Tree Report for Child Nodes add-on. The current public implementation is using Web Browser ActiveX with some freeware treeview JavaScript which is called from HTML page. That script is not optimal for large number of nodes, so I want to replace it. I am currently in benchmark phase with several another free treeview Javascripts. One of them has terrible memory leak which raises with number of nodes. I tried with my Location node and its sub-nodes, approx. 20,000 nodes, and I am getting next values for free RAM after closing the dialog box with the ActiveX control: 303 -> 252 -> 227 -> 199 MB. With another two Javascripts that I am trying I don't have such leaks. Testing within Windows XP SP2.

So, maybe you could try your script example with another Web site. Maybe Google which you tried has some Javascript that has leaks. You could take a look at Microsoft article about IE leaks caused by Javascripts: [url=http://msdn.microsoft.com/en-us/library/Bb250448]Understanding and Solving Internet Explorer Leak Patterns[/url], also this: [url=http://www.javascriptkit.com/javatutors/closuresleak/index.shtml]JavaScript and memory leaks[/url].

By the way, MM versions older then 3.1.0.1205 had memory leaks after closing dialog boxes, especially with ActiveX controls, even without any Javascripts (I already [url=http://www.mediamonkey.com/forum/viewtopic.php?f=19&t=35409]reported that[/url]). To resolve that I am using Script.UnRegisterHandler "wbCtrl_OnChange" and for even older versions which don't have UnRegisterHandler I am using Script.UnRegisterAllEvents, but with it in some cases I need to re-register all needed events again. Unfortunately, UnRegisterHandler and UnRegisterAllEvents cannot resolve problem with the mentioned Javascript that I tried.

[quote="GIL"]wait, 3?
You mean open and close the window 3 times?

Because that's too little, just leave it there literally a couple o minutes[/quote]

Yes, open and close the window 3 times => 3 MB leak on Win 7.

Left it for 5 minutes => further 4 MB leak
i.e. 26 KB per attempt in avarage.
It always oscilates between 41MB and 44MB on each attempt and sometimes there is X KB leak, sometimes there isn't.

It is strange that it is indeed Windows 7 related. So it seems to be a leak in Win 7 IE. We use memory manager (for leak detection) in IDE when developing and it doesn't detect the leak.

wait, 3?
You mean open and close the window 3 times?

Because that's too little, just leave it there literally a couple o minutes

In my case it will fill my memory up,
goes on and on

Rebooted to Windows 7 and true is that there is 3 MB leak and it seems to be a leak in Win 7 IE.
Maybe it isn't a leak, just Win 7 caches a data for next use of IE, because after several attempts (say 3) the memory usage stays at the same level and doesn't grow anymore.

Probably nothing we could fix.

I just fired it up in VMware, Windows XP Pro x86, MM 1294
It doesn't built up either, so it has something to do with Windows 7 or IE8.

Sorry, I meant 3.2.0.1294, of course.

I also thought it could be something related to IE.

I have tested it on Win XP SP2. I am rebooting to Windows 7.

The ActiveX browser is pretty much the same as the installed IE, so could it be something ticked there?
I don't have many extras on my IE, I don't use it much, no bars or anything but are you using Windows 7 x64?

[quote="Ludek"]3.2.1.1294[/quote]
No,
I've got 3.2.[b][i][u]0[/u][/i][/b].1294

GIL,
I tested by using your script and I cannot reproduce.

Everytime the timer hits memory usage grows up, but then (as expected) tails off.

So I don't see a leak there, are you sure you have tested 3.2.1.1294 ?

Thanks, added as http://www.ventismedia.com/mantis/view.php?id=6249

I noticed that you posted this in the bug section as well...

http://www.mediamonkey.com/forum/viewtopic.php?f=7&t=45911

Was afraid of that

Let's hope it's not a really a bug

Thanks anyway, you and this forum are always helpful.
Should I repost somewhere else?

I've not actually used it before, but yes that's certainly how I would have thought it should be implemented. I think this is going to have to be one for the devs to answer.

MMWiki says DestroyControl is only for browsers and references a search script, so maybe it only works on that type of script

This script demonstrates that it does destroy the control, but the memory keeps climbing up.

Did I implement this right?

[size=85](Edit: added DestroyControl to the script and fixed bracket)[/size]