>>> I don't know if the second error is related, but it is logged at
>>> exactly the same second. Did the first error destabilise the JDBC
> driver?
>> No, the second is another problem - it has problem parsing the server's
>> reply. Can you send me a database and a test case?
>>
>> Roman
>
> The database in the above situation is 66MB :( and as I say I don't
> know what causes it. Sometimes it works.
> Does the stack trace not indicate where in the driver the problem is?

It does, but without real reply from from the server you can't trace the
bug - it really depends on bits and bytes coming from the server. The
second error is decyphering the encoded reply from the server - for some
reason we misinterpret the reply. In this case Firebird server itself
can be buggy (we have had already similar case, a fix in server was
needed), but maybe now some fix in Jaybird is needed. After some
thinking, the first error can be a consequence of the second one, when
the application ignores the NegativeArrayIndexException (though
intuitively I would give not more than 1%, even less, that this is the
case).

> Can I get cvs access (read only) to the source?

You have it already:

cvs -d :pserver:anonymous@...:/cvsroot/firebird
co client-java

or simply download the sources published on SF.net - for each release we
publish also the sources that can be used to build the driver locally.

Roman