Re: [Sqlrelay-discussion] Bug report, empty cursor/blob, with code to reproduce problem

[David M. <dav...@fi...>]

Ok, I think I took care of it.

It looks like it had to do with not re-initializing the null indicator,
which oracle doesn't re-initialize for cursor binds.  So, since the
first bind variable came back as NULL in the first query and the null
indicator for the first bind variable wasn't reset, even though a valid
cursor came back in the second query, the null indicator was still set
to true and that overrode the cursor and just sent a NULL back for the
cursor.

It should be fixed in 0.39.4 which was just released.

Give it a try, let me know what you find.

Dave
dav...@fi...

On Wed, 2007-09-12 at 12:47 +0200, Ingmar Brouns wrote:
> Hi David,
> 
> Thanks for your reply.
> 
> I was curious whether there has already been made some progress in
> tracing
> the cause of this bug. 
> The strange thing about  this bug is that seems to be related with the
> freeing of 
> resources. When you comment out the first query after you excecuted
> the first 
> query once in the bug situation,  then the second query will
> repeatedly give 
> faulty results, even with the first query commented out. In more
> complex situations 
> I have seen that sometimes the client connections are kept while they
> should 
> have been freed at the end of a php script. That behavior always
> occured 
> together with an instance of this bug.   
> If can provide you with more  information, please let me know.
> 
> Friendly greetings, 
> 
> Ingmar 
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2005.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________ Sqlrelay-discussion mailing list Sql...@li... https://lists.sourceforge.net/lists/listinfo/sqlrelay-discussion