That is exactly right.
There was a thread by Karsten Strobel on the IB ng's where he actually was
pointing at a place in the source tripping things up.

I've found that you have to have a combination of statements prepared and in
use to trip it up. Like one with three outputs and then one with two, for
example. using one and then the other garbles some things in a global area
where there should be proper distinction made between the statements.

In short, somewhere in the code garbage information is being left behind and
the next guy to come along assumes the information it needs is already in
place and BOOM, problem.

HTH,
Jason Wharton
CPS - Mesa AZ
http://www.ibobjects.com

----- Original Message -----
From: "Ann W. Harrison" <aharrison@...>
To: <ib-support@yahoogroups.com>; <ib-support@yahoogroups.com>
Cc: <Colin.Fraser@...>; "Jason Wharton" <jwharton@...>
Sent: Friday, November 23, 2001 1:41 PM
Subject: Re: [ib-support] Error Message 335544358


> At 05:43 PM 11/15/2001 +1300, C Fraser wrote:
>
> >I have a problem and am not to sure why it is occurring...
> >
> >The error message is :
> >ISC ERROR CODE:335544358
> >ISC ERROR MESSAGE:message length error (encountered 0, expected 8)
> >
> >It happens on the second time I try and save an object from code (using
> >IBO)...
> >
> >Just wondering if anyone could explain the message a bit better or point
> >me in the direction to a source for this information.
>
> That message indicates an error in one of three places: data movement from
> the DSQL processor to the engine, data movement from the engine to the
DSQL
> processor, or engine side of the execution of a stored procedure. In all
of
> those cases, the call parameters have been set up by the DSQL module so
they
> should be correct and there should be no way that your code, or IBO could
> pass bad values.
>
> My questions, both to Colin and to Jason are these: are you saving more
than
> one type of object? Does IBO save and reuse the prepared statement? If
so,
> I think I know what the bug is and the work around is to reprepare the
> statement.
>
> Regards,
>
> Ann
>
>