On Tue, Feb 10, 2004 at 11:20:42PM -0000, Andy Hassall wrote:
> Just to pre-empt the probably soon to be frequently-asked question, here's
> one data point: DBD::Oracle 1.15 appears to work OK on Oracle 10.1.0.2 on
> Linux.
Great, thanks.
> t/ph_type....ok 11/19 Placeholder behaviour for ora_type=1 (the default)
> varies with Oracle version.
> Oracle 7 didn't strip trailing spaces, Oracle 8 did, until 9.2.x
> Your system doesn't. If that seems odd, let us know.
As you can guess, that's how DBD::Oracle 1.15 now addresses the change in
placeholder behaviour.
> Makefile.PL had some trouble with detecting the version; patch attached
> addresses one aspect of that. However, it fell through the first two version
> detection options and ended up with the 'scary but handy' part where it
> parses the Oracle home directory for numbers. My Oracle home's called
> Ora101; had to alter the check of $inspdver{rdbms_ver} to allow a version
> >10, and tweak the regex to add a zero on the end.
>
> SQL*Plus can report the version of the Oracle home in a readily parseable
> way:
>
> $ echo DEFINE _SQLPLUS_RELEASE | sqlplus -S /nolog
> DEFINE _SQLPLUS_RELEASE = "1001000200" (CHAR)
>
> Any objections to me writing a patch to run sqlplus as above to get this
> value? It seems to me that if you can't run sqlplus, you're unlikely to be
> able to compile DBD::Oracle anyway - and besides, there's the other 4
> existing methods to fall back on. Also appears neater than parsing through
> obscure files? (I can't find unix.rgs or inspdver in the docs)
That would be great. Thanks.
Tim.