perl.dbi.dev Google Group

Re: Patch to DBD::Oracle 1.23

sco...@pythian.com (John Scoles) — Wed, 02 Dec 2009 12:36:31 UT

Ok that will be the next one to get done
thanks

Patch to DBD::Oracle 1.23

ma...@ders.cz (Jan Mach) — Wed, 02 Dec 2009 08:29:56 UT

We have published len_char_size property of statement handle. Why? That's
because of nvarchar type.
You have column nvarchar(20) and you get length 40 using PRECISION (it's
right because is BYTE length). But we need for our application CHAR length
and it's done usign len_char_size.
Regards,
Jan Mach

Re: DBD::Oracle, Support binding of integers so they are returned as IVs

rlip...@remotelinux.com (Rudy Lippan) — Wed, 02 Dec 2009 08:12:13 UT

FWIW, Many hosting companies do not send out upgrade notes to their
customers when upgrading some random Perl module. Many people do not
read the entire release notes when upgrading their OS. Many people do
not go looking for release notes before doing a 'sudo cpan install
Module' :) So I do not know if just an upgrade warning would suffice

Re: DBD::Oracle, Support binding of integers so they are returned as IVs

martin.ev...@easysoft.com (Martin J. Evans) — Tue, 01 Dec 2009 19:43:52 UT

Apologies if it seems this change is stalled, I've been extremely busy
attempting to keep the wolf from the door. I will be back on it to
finish it in the next few days. I've so far:
o documented change in trunk for bind_col (edited by Tim)
o documented change in trunk for DBI::DBD
o fixed a problem in sql_type_cast

Re: DBD::Oracle, Support binding of integers so they are returned as IVs

tim.bu...@pobox.com (Tim Bunce) — Sat, 28 Nov 2009 13:43:53 UT

Well, if this was done:
then it would take no work in DBD::File at all. Applications that
actually want IV/UV/NV's for some columns could simply use bind_col() to
ask for it. Also, being done in C, it would also be much faster than a
pure-perl addition to DBD::File.
Tim.
p.s. Seems like there's a reasonable case for restoring the old initial

Re: DBD::Oracle, Support binding of integers so they are returned as IVs

h.m.br...@xs4all.nl (H.Merijn Brand) — Sat, 28 Nov 2009 10:05:31 UT

On Sat, 28 Nov 2009 00:43:48 +0000, Tim Bunce <Tim.Bu...@pobox.com>
wrote:
The plan is still in my head to convert PV("1"\0) to IV(1) if the field
type is numeric on fetch. Low priority though. Probably needs some work
in DBD::File

Re: DBD::Oracle, Support binding of integers so they are returned as IVs

tim.bu...@pobox.com (Tim Bunce) — Sat, 28 Nov 2009 00:43:48 UT

Umm, if the database holds Important Things one could opine that perl
should trust the Important Things that are already in the database.
We can argue it either way.
A compromise might be to default to strict (for SQL_INTEGER and
SQL_DOUBLE) but make SQL_NUMERIC immune from strict.
DBD::SQLite returns only strings, along with most of the pure-perl

Re: DBD::Oracle, Support binding of integers so they are returned as IVs

martin.ev...@easysoft.com ("Martin J. Evans") — Fri, 27 Nov 2009 20:20:53 UT

Greg, the issue I started by trying to address is rather specific to our
use of JSON::XS and roughly described at
[link]. The problem we
faced is not specific to DBD::Oracle and applies to many DBDs (and
certainly DBD::ODBC) although I believe it was addressed in Postgres a

Re: DBD::Oracle, Support binding of integers so they are returned as IVs

g...@turnstep.com (Greg Sabino Mullane) — Fri, 27 Nov 2009 18:03:29 UT

Re: DBD-Oracle-1.23 Makefile.PL patch

h.m.br...@xs4all.nl (H.Merijn Brand) — Fri, 27 Nov 2009 15:45:03 UT

On Fri, 27 Nov 2009 10:17:01 -0500, John Scoles <sco...@pythian.com>
wrote:
Would that be just as bad as True64?
I feel your pain however.

Re: DBD-Oracle-1.23 Makefile.PL patch

sco...@pythian.com (John Scoles) — Fri, 27 Nov 2009 15:17:01 UT

Of you could give it a try on 64 bit power solaris it you want to tug
all of your hair out.

Re: DBD-Oracle-1.23 Makefile.PL patch

ralphdoncas...@gmail.com (Ralph Doncaster) — Fri, 27 Nov 2009 14:41:19 UT

We're running Oracle & perl on a few Unix boxes without a
/usr/share/oracle, so it seems it's not required in some cases. Our
install is in /usr/oracle, and there's no oracle link in /usr/share.
Sorry, I only do perl on Unix.

Re: DBD-Oracle-1.23 Makefile.PL patch

h.m.br...@xs4all.nl (H.Merijn Brand) — Fri, 27 Nov 2009 14:56:05 UT

On Fri, 27 Nov 2009 10:41:19 -0400, Ralph Doncaster
There is a (huge) difference between a full Oracle install (server) or
an Instant Client only. I prefer the latter if the box is to be client
only, because it is much smaller.
So want I, but those silly customers ...
That being said, I think Oracle on Windows is still easier than Oracle

Re: DBD-Oracle-1.23 Makefile.PL patch

h.m.br...@xs4all.nl (H.Merijn Brand) — Fri, 27 Nov 2009 14:31:21 UT

On Fri, 27 Nov 2009 10:21:05 -0400, Ralph Doncaster
Looks sane to me, but I'm not the DBD::Oracle maintainer :)
While you're at it, I'd like to see some more changes:
* make /usr/share/oracle not hardcoded but variable
* make Instant Client work on windows too (Windows instant client works
fine with the current Makefile.PL if installed on

Re: DBD-Oracle-1.23 Makefile.PL patch

ralphdoncas...@gmail.com (Ralph Doncaster) — Fri, 27 Nov 2009 14:21:05 UT

It seems I can't post attachments to the list, so here's the patch inline:
--- ../DBD-Oracle-1.23-3oaWaD/Make file.PL 2009-04-29 10:04:24.000000000 -0500
+++ Makefile.PL 2009-11-23 16:18:31.000000000 -0600
@@ -1055,14 +1055,19 @@
'proc/lib/proc.mk',
'proc16/lib/proc16.mk',
);
- my @mk_oci = (