|
|
On May 30, 1:22 pm, "thelf...@xxxxxxxxx" <thelf...@xxxxxxxxx> wrote:
> On May 28, 2:53 pm, Jeff Hobbs <jeff.ho...@xxxxxxxxx> wrote:
> > On May 28, 11:23 am, jdrum...@xxxxxxxxxxx wrote:
>
> > > I have a customer who has a large installed code base written using
> > > oratcl3.3 ( oratclsh ) and Oracle 8i on HPUX 11.
> > > There is a mandatory ( internal ) Oracle upgrade to 10g.
>
> > > We have installed Oratcl 4.4. but found that Oracle no longer provides
> > > oratclsh and the existing code has numerous instances of oramsg(rc)
> > > checks and orafetch calls that expect values instead of return codes.
>
> > > Obviously..not the optimum solution, but.....the time and resources
> > > don't exist to re-code all this... so we're looking at running Oratcl
> > > 3.3 against Oracle 10g. I haven't seen any documents newer than a few
> > > years old that talk about this.
> > > Has anyone done it?
>
> > While I'm not aware of any users of the old oratcl 3 on oracle 10g, I
> > suspect that creating a compat level between oratcl 3 and 4 might be
> > the most time-efficient solution. The newer versions are more feature-
> > rich, in which you can "dumb them down" to the previous behavior.
>
> > Jeff
>
> Jeff, its been a while since I did anything like this, but couldn't a
> 'read trace' on oramsg(rc) be done that pulls the value from the
> function [oramsg rc] combined with the oracompat library that remaps
> orafetch to the old way of doing things, this might get them 99% of
> the way.
Yes, most certainly an easy approach and possible.
Jeff
|
|