Internet Explorer dies : "Operation aborted" with RichFaces

[gabrielk]

Hi everyOne ,

We are currently using richfaces 3.2.1.GA
On several pages we have an issue with Internet Explorer IE6 and IE7 .
Randomly the browser gives an alert popup saying "can not open this site XXX. Operation aborted" . Then IE just ... dies !

It seems to be the same problem as described here:
http://weblogs.asp.net/infinitiesloop/archive/2006/11/02/Dealing-with-IE-_2600_quot_3B00_Operation-Aborted_2600_quot_3B002E00_-Or_2C00_-how-to-Crash-IE.aspx

To me it is the calendar element who is the No1 suspect.
Do you also have these kind of problems?
I can not find any message in the Forum or any bug track talking about this issue. Strange...

How do you deal with it?

We use :
- JSF Sun 1.2
- richfaces 3.2.1.GA

[ilya_shaikovsky]

could you please check this also with the beta which released today(http://repository.jboss.org/maven2/org/richfaces/ui/richfaces-ui/3.3.0.BETA1/).. or at least with 3.2.2SR1?

[nbelaevski]

Hi,

We encountered the issue while developing framework code and modal panel component but then solved the problem. Usually happens when script tries to modify children collection when children not yet created (e.g. append something as its next sibling).

Can you please post the simplest page code the problem is reproduced at?

[gabrielk]

Thanks both of you for your replies

- I tried to use 3.3.0Beta and the 3.2.2.SR1. Does not work either
- Tried to remove the <rich:calendar tag> in the page - IE does not abort the Operation.
- But! Tried the calendar without any other tag, no problem...
It is the interaction between the Calendar and some other tag that make IE crash. Now we have to find what.

(...later...)
we stripped the JSF code to find what tuple of tags make crash.
It appears that it is the combination of Spring-Faces validators + RichFaces calendar.
We use Sprign Faces 2.0.2. We are going to try with the newest 2.0.5

So the "suspect" is now (as usual...) the <sf:... /> tags. Or, to be more precise, the interaction between rich:calendar and the <sf:clientTextValidator />
We're going to see if the upgrade solves the problems.

Thank you both once again!
More to follow...

[nbelaevski]

Good, please keep us posted

[gabrielk]

Hi,

We found the guilty component. It was some <rich:calendar>... We changed it and put some normal "inputText" . The generated code is simple but i did not find where it is generated. So, i did not send a patch.

[nbelaevski]

"GabrielK" wrote:
Hi,

We found the guilty component. It was some <rich:calendar>... We changed it and put some normal "inputText" . The generated code is simple but i did not find where it is generated. So, i did not send a patch.

I guess the problem was not with the calendar, but with Spring tag? Can you please post full page code so we can check?