Re: smallint mapping issue

On Fri, Jul 29, 2005 at 12:23:43PM +1200, Oliver Jowett wrote:
> Christian Cryder wrote:
>
> [...]
>
> > Does that help explain where I'm coming from here?
>
> Sure, I understand exactly where you're coming from.
>
> If you want a stable, unchanging, tested-for-your-application driver, I
> suggest you import from whatever spot in CVS serves your needs and
> maintain it yourself. You can selectively backpatch bugfixes from the
> official driver as needed. (this is exactly what Open Cloud does for our
> product, BTW, it's just that because I also hack on the official driver
> the changes aren't usually all that big -- any fixes we need I try to
> push back into the official driver)

> I don't really want to see the official driver become even more of a
> maze of twisty little backwards-compatibility hacks than it already is..

> We have perhaps 1/4 of a
> developer working on the driver between myself, Kris, and Dave, so we
> really don't have tons of development time spare :/ And when a
> complete CVS history is there for the taking.. it's not like we're
> forcing you to use a particular version.

Christian,

You can find the cause of your despair here:

http://gborg.postgresql.org/cgi-bin/cvsweb.cgi/pgjdbc/org/postgresql/jdbc2/AbstractJdbc2ResultSet.java.diff?r1=1.74;r2=1.75;cvsroot=pgjdbc

As you can see there, the "bug/feature" you are asking for seems to be
just the matter of the following line of code:

case Types.SMALLINT:
return new Short(getShort(columnIndex));

You really don't need any seasoned developer to check out a stable
release from CVS (using a tag), rolling back this change above and
compiling a driver customized to your needs. You don't even need to
know Java. Considering what you said about your business, it would be
surprising that you don't have anyone in-house skilled enough for
that. The time your team will spent on this issue, or the fees you
gonna pay to the OpenCloud-like company which will deal with it for
you, will both surely be pale compared to the licensing costs you were
paying before Postgres :-)

Cheers,

Marc.