Mattias Jiderhamn wrote:

> Actually, I haven't since I reasoned this wasn't necessary.

Logical proof a bug can't happen is great, but I still like to perform
actual testing. :) I've just seen too many bugs that couldn't possibly
happen but yet do.

> The only case when the current code would produce the expected output
> while the patched code would not, is if attQName contains both prefix
> and name while attLocalName is the empty string. I figured that if such
> a parser exists, it is erroneous and it shouldn't be JDOMs
> responsibility to handle this. But maybe I'm wrong here?
>
> If I'm wrong, is there a particular set of parser that should be tested
> ("the parsers")?

I'd suggest testing under your Resin 2 and Resin 3 environments where
you found the bug, to make absolutely sure the fix for 3 doesn't hurt 2.
What parser/transformer is Resin 3 using that behaves differently?

-jh-

_______________________________________________
To control your jdom-interest membership:
http://www.jdom.org/mailman/options/jdom-interest/youraddr@(protected)