It looks like the problem that Chris Gardner was having with the JavaSampler (on the jmeter-user list) was that my patch from March 11 (http://www.mail-archive.com/[EMAIL PROTECTED]/msg03646.html) never got applied. Actually, that's only half of the problem...the other half is that the patch was apparently missing one key file (AbstractSampler), so even if it had gotten applied it still wouldn't have worked.
Anyway, JavaSampler was implementing both PerSampleClonable (indirectly from AbstractSampler) and PerThreadClonable, so it was getting cloned every sample, leading to the undesired behavior. I've made a quick fix for Chris (just updating AbstractSampler -- the correct fix needs to update the implementations of AbstractSampler like my patch did) so hopefully his stuff will work. I can make the real fix this weekend, assuming that I can get my CVS access working.
Mike -- you mentioned you are doing some work on reducing the amount of cloning going on. I would assume that these updates won't get in the way of that? Let me know if it will and we'll figure out the right thing to do.
It's past my bedtime...
Jeremy [EMAIL PROTECTED]
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
