There is a bug that I reported that causes transformer threads to never
exit when an error is encountered while parsing the XML. I think it was
reported as fixed, but I have not tested it yet. We are still using
XalanJ1 on production servers for now because there are still some
pretty serious bugs in XalanJ2. I think there is a lot of good work put
into XalanJ2, so hopefully we will be able to upgrade soon.
Chris
[EMAIL PROTECTED] wrote:
>We have a multi-threaded RMI server that uses Xalan to transform XML docs.
>When we have a moderate load on the server (5-10 users), we can quite often
>cause the server to deadlock. This is using Xalan 2.1.0 on Windows/NT SP6
>/ Sun JVM 1.3.0. We have had similar results using Win2K/IBM JVM 1.2.2
>
>Examination of the thread dump shows threads stopped in Locale.hashcode().
>We have seen the fix applied to Xalan 2 regarding Locale and have looked at
>the source where it now clones the Locale object before use.
>
>We have been using Xalan since it was Lotus/XSL. We are considering other
>tools because of this problem. Are there any known issues remaining with
>Locale and/or thread deadlocks when multiple threads are transforming XML
>concurrently?
>
>Any information would be greatly appreciated...
>
>Steve Horne
>Systems Architect
>Computer Sciences Corp.
>