Didn't notice this one there... Might help to read all mail when out for
a day ;D
He does have a good point though, in that you should look through any
attached code to see if any unexpected or odd bits are there.. The hard
part, is knowing what will cause a problem and what won't.
What I usually do in this case (warning: this is not the easiest way for
most, just easier for me); is to start by blanket disable of all script
in the page, to trace the error down to the closest spot. You can
disable chunks of jscript by using the /* comment inside here */ method.
It would be useful if you can put down all the modifications you've made
to the Xwiki (even if just for yourself), as this gives you an
indication of your own changes as well as a starting point for
debugging.
Do you have the script debugger installed by any chance? Sometimes this
can help give an indication of where the error to start looking. It can
be disabled if it becomes a nuisance (point of interest: a fair chunk of
Microsoft's web pages will kick this debugger into overdrive).
http://www.microsoft.com/downloads/details.aspx?familyid=2f465be0-94fd-4
569-b3c4-dffdf19ccd99&displaylang=en
-----Original Message-----
From: THOMAS, BRIAN M (SBCSI) [mailto:bt0008@att.com]
Sent: 02 November 2006 23:29
To: xwiki-users(a)objectweb.org
Subject: [xwiki-users] FW: Don't click on the blue 'e'...
Well, here's an interesting bit of news in response to my plea for help
locally from a coworker who is a web developer.
Any clues on where to go to remove this offending bit of code as a test,
or reasons it should be there? Brandon: does this ring any bells?
brain[sic]
-----Original Message-----
...
I know of no clever IE debugging techniques.
In my experience, however, IE's crashes normally occur when loading
embedded objects like applets, Flash, SVG, fonts, etc, or when people do
some unorthodox trickery in CSS or scripting. So when troubleshooting
problems like this, my instinct is to dig through the source to see if
there's anything odd going on.
(Having validating markup and CSS helps this process.)
My first observation is that the page is written in XHTML, and IE
doesn't support XHTML[1]. This isn't normally a problem, necessarily,
if the page is delivered as text/html (as these pages are), because IE
will look at it and treat it as HTML "tag soup", and ignore the non-HTML
XHTML stuff as bad markup to work around.
However, the site is also emitting an <?xml ... ?> prolog at the start
of the page. This has traditionally caused problems with IE in my
experience, because IE isn't sure if it should look at it as a raw XML
document or HTML. This issue is compounded by the fact that the page
isn't well-formed XML at all. (There are HTML-style <BR> tags in there,
for example.)