On Mon, 2005-08-08 at 10:23 +0200, Roman Rokytskyy wrote:

> Hi,
>
> >> No, not yet. I will do this on the weekend.
>
> Just to confirm your suspections - this is reproducable with classic server
> (both FB 1.5 and FB 2.0) and works ok with super server architecture. The
> statement preparation/usage pattern in your application is somewhat strange,
> but not erroneous.
>
> I will try to create a separate test case to reproduce the issue in classic,
> so maybe it is fixed before the FB 2.0 release.
>
> Roman

Thanks for the confirmation.

The usage pattern in _this_ example is unusual because it is a testcase
that has not got past create and destroy objects. This pattern is not
typical of the application in operation. Once I had an error that I
couldn't trace to my application code, I deliberately exposed the error
in the second test of the testcase.

I can adapt the testcase to work around this limitation, and/or switch
to superserver for the time being.

Thanks again for all you do.

David Johnson