Tom Lane wrote:
> I would say that that is a very bad decision in the JDBC driver and
> should be reverted ... especially if the driver is not bright enough
> to notice the context in which the parameter is being used. Consider
> for example
...
>
> You are trying to mask a server problem in the driver. This is not a
> good idea. The server problem is short-term (yes, we've finally agreed
> how to fix it, and it will happen in 7.4), but a client-library hack to
> mask it will cause problems indefinitely.
>
> regards, tom lane
>
Tom,
Thanks for the quick reply. I will back out the jdbc change. It was
one of those changes I did reluctantly. I have been pushing back for a
couple of releases now saying that this is a server bug and needs to be
fixed there. But it didn't seem like that was ever going to happen so I
finally gave in. For some users this bug of not using indexes for
int2/int8 makes it impossible for them to use postgres. This happens
for users who are using a database abstraction layer that doesn't allow
the user to actually touch the sql being sent to the server. Therefore
they have no opportunity to work around the underlying bug and can't use
postgres as their database because of the performance problems.
I am glad to here this is finally getting resolved for 7.4.
thanks,
--Barry