This is a migrated thread and some comments may be shown as answers.

Officially supported platforms?

12 Answers 216 Views

General Discussions

This is a migrated thread and some comments may be shown as answers.

12 Answers, 1 is accepted

answered on 31 Mar 2008, 11:50 PM

We have concluded testing under Windows 2000. We found issues with your RadScriptManager, as well as ComponentArt's script combiner. Both script combiners cause IE6 in Windows 2000 to issue lots of erroneous JavaScript errors. My best guess is that some underlying component in the operating system is failing to handle the HTTP request/response properly, or that the JavaScript engine itself can't handle it. We needed to remove both your script manager, as well as ComponentArt's script manager (and yes, we tried it with just one or the other). This cleared up all Windows 2000 issues. I am going to assume that something in the nature of combining scripts into one request causes issues in Win2k and leave it at that. Once all our customers have migrated to XP or above, we'll reinstate the combiners. If you are able to reproduce and solve the problem, however, please do let me know so that I can bring the RadScriptManager back. (Our 3 JS requests went up to 21 or more without it!) Thanks.

answered on 01 Apr 2008, 03:38 PM

Telerik.Web.UI_2008_1_401_trial_hotfix.zip

Hello Jeremy,

We must admit we have introduced an issue in the "Futures" build concerning RadScriptManager behavior in IE6. We believe the problems are already addressed and you can find a trial version of the hotfix attached to my response. Please, give it a go and let us know if that helps.

Sincerely yours,
Konstantin Petkov
the Telerik team

Instantly find answers to your questions at the new Telerik Support Center

answered on 03 Apr 2008, 11:01 AM

ScriptManager.JPG

AJAXEnabledWebSite1.zip

Hello Jeremy,

We tested the behavior locally with the "Futures" build of "Prometheus" controls but couldn't reproduce the problem. Please find our test project attached to this support thread.
Also find attached a screenshot as part of our local tests we made on Win 2000 and IE6. What might be the difference in your case? Are we missing something?

Kind regards,
Helen
the Telerik team

Instantly find answers to your questions at the new Telerik Support Center

answered on 03 Apr 2008, 08:55 PM

One difference I see right away is that in our tests, the application was being served from IIS6 and IIS7 from external servers, and the Windows 2000 box was simply consuming the content. You're actually running Cassini on the Win2k box. That might be one source of the issue (I still highly suspect a flaw in the HTTP subsystem).

If you still can't repro anything, then you probably don't have a project of sufficient complexity. Our web application is over 300 MB and has over 5000 files, so it's pretty difficult to set you up with a test case that adequately covers it. Our JavaScript load consisted of your library, some scripts from ComponentArt, some scripts from the AJAX Control Toolkit, and a few scripts from our own assemblies. The total JavaScript load in debug mode including the AJAX Extensions was over 250K, spread out over about 3-4 requests. Your ScriptManager was serving two requests, one for 88K and one for 24K. That should give you an idea of the level of complexity involved here.

The primary and most visible error that we received was "Telerik is undefined", but we also got several other errors. Errors occurred inside script files from you, ComponentArt, and us, and there were even some syntax errors directly in the HTML document. None of these errors were "real", however (the scripts are actually perfect) -- which is why I suspect some sort of bug/overflow in the HTTP or JavaScript system. Removing the ScriptManager cleared up 100% of the errors.

Again, we had the exact same problem with the ComponentArt script handler. The effect was similar -- random JavaScript errors that don't really exist -- but they occurred in different places. It could be that some invalid character is being transmitted during the combine, which modern systems correctly interpret (or ignore), but which causes Win2k to stop parsing or error out.

If I get some free time, I'll attempt to isolate this in a test project of my own, but I'm pretty swamped right now.

Thanks!