Thread: IE hangs with (1 item remaining)

This is not exactly the right forum for this question, but close enough I reckin'

We have had several reports from our users that IE is 'hanging' with a status of "1 item remaining" when using our internal web application. This is sporadic, but probably happens at least once/day to most users. In some case, it prevent the page from 'working', other times it doesn't appear to affect the page at all.

After doing some research, I've found reference to this as a possible IE bug attempting to load a javascript file (foo.js for example). This would seem to make some sense if the script file happened to have logic needed to support the page. (I.E. submit button not working, AJAX broken, etc).

This has only been mildly annoying, but the natives are growing restless. So

1. has anybody seen this bug before -- it seems to only be in IE
2. how can I determine what the '1 item remaining' is (easily)
3. any line on a possible fix
?

This could be a missing file that the web site is requesting to be loaded, (i.e. CSS, image, JavaScript, etc.). You could try using a program like "Fiddler" to determine who the culprit is.

So, I managed to install a copy of 'Fiddler' and capture a case where the browsers 'hangs' with '1 item remaining' (in fact, it's been sitting there now for over a hour). I also captured a request that completed without a problem.

The request trace is exactly the same between the two (save for the line numbers in Fiddler). Same number of requests, same number of byes, same files, same HTTP return codes.

So, now I'm at a loss as to where to look further. Some googling on the subject is now making me think that a javascript is hanging someplace, but I would have no idea where to start looking for that.

Well I have at least some circumstantial evidence that it's a script getting hung up. When I disable scripts from Internet Options, I was unable to reproduce the problem after about 15 minutes of trying. Yeah, it's an intermittent problem, so this is not conclusive.

When I re-enabled scripts again, I managed to reproduce the problem within 20 attempts. Then, when I disabled scripts using the IE8 debugger, the page finished loading. I then attempted to reproduce the problem for another 100 attempts or so, as was not able to.

After re-enabling scripts in the debugger, I reproduced the problem again in about 2 minutes.

I'm therefore reasonably confident that it's a problem in a script getting hung up. But, what script is it and is it the same script each time? How to tell??

Do you have access to Firefox? You could try this Javascript Debugger add-on: https://addons.mozilla.org/en-US/firefox/addon/216/

Sure, I work exclusively in firefox except when I'm forced to use IE. Alas, this problem does not appear in Firefox. (And yes, IE8 is the corporate standard supported browser).

How about popping it in Visual Studio 2008 and using the JavaScript debugger (opening in IE, of course)?

Is it calling any external JS files from a different domain?

Can you make changes to the JS files for debugging purposes? I'm thinking make it write something to the browser at the beginning and end of each file to narrow down to which one is causing the issue (assuming it's the same one each time).

Once you've located the file, change the position of the debugging element to try and locate the specific element causing the issue.

How about popping it in Visual Studio 2008 and using the JavaScript debugger (opening in IE, of course)?

Yeah, thought of that -- however I don't have VS yet as I have no 'business case' to get it (until we start converting this monster to .NET anyway, sometime in 2011). Maybe this problem will help get it (tho I think the corporate standard is 2005 still. Heck, we still have Word 2000 as the standard).

Is it calling any external JS files from a different domain?

Nope all the js files are in the same directory structure as the application.

Can you make changes to the JS files for debugging purposes?

Yes I can. However, we are talking about 50+ js files here. I'm probably going to have to do this, but first I better confirm that the problem can be reproduced on the development server.