Re: DBD::Oracle test failure under DBI 1.49

On Fri, Dec 09, 2005 at 11:45:44AM +0000, Michael Gray wrote:
> DBD::Oracle 1.16 fails a test for me when building under DBI 1.49.
> All was fine with the identical build under DBI 1.48.
>
> This is on SuSE Linux 9.3, 32-bit Athlon.
>
> In a nutshell:
>
> Can't bind a reference (DBI::st=HASH(0x82df350)) at t/50cursor.t line 91.
>
> A full transcript is attached, please let me know if any further info
> would be useful.

I think I know what this is. (DBD::Oracle uses a shortcut to check if a
ref is a DBI handle - that's no longer reliable now handles also have
weak reference magic.)

Although the best fix would be to DBD::Oracle (which I'll do for the
next release), I'll see if I can fix it from the DBI side to avoid a lot
of inconvenience for people.

Tim.

Re: DBD::Oracle test failure under DBI 1.49

On Fri, 9 Dec 2005, Tim Bunce wrote:

> On Fri, Dec 09, 2005 at 11:45:44AM +0000, Michael Gray wrote:
> > DBD::Oracle 1.16 fails a test for me when building under DBI 1.49.
> > All was fine with the identical build under DBI 1.48.
> > This is on SuSE Linux 9.3, 32-bit Athlon.
> > In a nutshell:
> >
> > Can't bind a reference (DBI::st=HASH(0x82df350)) at t/50cursor.t line 91.
>
> Although the best fix would be to DBD::Oracle (which I'll do for the
> next release), I'll see if I can fix it from the DBI side to avoid a lot
> of inconvenience for people.

Just to confirm that DBD::Oracle 1.16 now builds and tests cleanly
under DBI 1.50.

Many thanks

--
Michael