Re: [Q] Does Oraperl run with Perl5 ?

From: Tim Bunce <Tim.Bunce_at_ig.co.uk>
Date: 1995/08/01
Message-ID: <DCMxIq.F38_at_ig.co.uk>#1/1


In article <3vatoj$e42_at_pong.lasc.lockheed.com>, Terry Greenlaw <z50816_at_mip.lasc.lockheed.com> wrote:
> In article <DCC3Iv.H4v_at_ig.co.uk>, Tim Bunce <Tim.Bunce_at_ig.co.uk> writes:
> > In article <3v58t0$e7a_at_pong.lasc.lockheed.com>,
> > Terry Greenlaw <z50816_at_mip.lasc.lockheed.com> wrote:
> > > In article <3v45p0$2i6_at_ixnews3.ix.netcom.com>, rsingh_at_ix.netcom.com (Rashpal Singh) writes:
> > > >
> > > > Does anyone know if Oraperl will run
> > > > under Perl5, or should I stick with Perl 4.036 ?
> > > >
> > > > Thanks in advance.
> > >
> > > Use DBI and DBD/Oracle with Perl5 for Oracle7 access.
> > > It comes with an Oraperl emulation package, and has quite
> > > a few improvments over the original.
> > >
> > Er, reality check! It's still an unstable alpha release which is
> > currently significantly slower than Oraperl 2.4. Whilst I greatly
> > appreciate your enthusiasm it's not really ready for production use.
> >
> > Tim.
>
> I guess after dealing with Oracle's bug-ridden application code for so
> long (IMHO, Forms 4.x will NEVER work right on X/Motif, e-mail for gory
> details), using code like DBI/DBD that works as advertised must have gone
> to my head ;-}
>

Yes, battling with Oracle's code quality and interface stability is not fun.

Somebody told me that until recently the rdbms/demo/*.h OCI files were not tested before release since they thought that no one used them! I was worried when I first saw that they were in a /demo/ directory.

> Actually, Tim, on our environment (HP/UX 9.x), the combination of Perl5,
> Oracle 7, and DBI/DBD, is significantly faster and more stable than using
> Perl4 and Oraperl. It may be related to Perl5 enhancements. We used to
> have quite a few OraPerl processes hang under Perl4.
>
> I just converted a 300 line SQL*Plus HTML generation script into a 100 line
> Perl5/DBI program with a 5x speedup so I'm sure not gonna complain about speed.
>

Well that's great. Right now it doesn't use a row cache (oraperl 2.4 does). Once that's in and I've added some other assorted enhancements it should run considerably quicker than oraperl 2.4 and have more functionailty (in/out named vars, possibly including arrays, automatically fetched and stored by reference to perl variables etc).

> I would say that DBI and DBD were closer to late beta than alpha WRT Oracle,
> but you're the one who knows what's really going on behind the curtains, so
> I'll take your word for it. I wish the CDE group at Oracle was as modest about
> their code quality!
>

[blush]

The reason I still call it an Alpha is that it's built on the DBI and the DBI interface is still alpha (read 'incomplete'). The Oraperl emulation layer could be called beta.

I'm also a little worried about having time to 'support' a large number of users. I really hope that existing users will help out new users and leave me the time I need to continue developing this (and my real work).

For those who would like to know about portability I know it's being built/tested/used on:

	Solaris 1, Solaris 2, HP-UX 9, Unisys U6000/300, SCO 3,
	Sequent DYNIX/ptx 2, Motorola M88100 (SVR3.2), DEC UNIX 3.2,
	DEC Alpha, OSF/1 v2, SGI Iris Indigo R4000 Irix 5.3.

For more information and to keep informed about progress people can join the perldb mailing list by sending a message to perldb-interest-request_at_vix.com (it may take a few days to be processed). Discussion related to this work should take place on the mailing list.

Releases of the software are available from ftp.demon.co.uk in /pub/perl/db/ in the DBI and DBD/Oracle directories.

> Anyway, thanks again for the goodies. My goal with all this is to develop a
> freeware replacement for Oracle Forms (using TkPerl for GUI'ness). My guess is
> you'll have DBI/DBD production ready before I am...
>

:-)  

> Terry Greenlaw (on-site _at_ Lockheed-Martin) Encompass Technologies
> z50816_at_mip.lasc.lockheed.com greenlaw_at_entech.is.net
>

Tim. Received on Tue Aug 01 1995 - 00:00:00 CEST

Original text of this message