Received: from punt2.demon.co.uk by lion with SMTP (PP) id <26977-15@lion>;

Received: from punt2.demon.co.uk by lion with SMTP (PP) id <26977-15@lion>;
Wed, 2 Aug 1995 07:02:23 +0100
Received: from punt2.demon.co.uk via puntmail for [email protected];
Wed, 02 Aug 95 04:33:03 GMT
Received: from gw.home.vix.com by punt2.demon.co.uk id aa15485;
2 Aug 95 5:32 +0100
Received: by gw.home.vix.com id AA20960; Tue, 1 Aug 95 16:16:03 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA20956; Tue, 1 Aug 95 16:16:01 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id ab24534;
1 Aug 95 17:24 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa23463;
1 Aug 95 17:24 +0100
Received: from ig.co.uk by lion id <22046-0@lion>;
Tue, 1 Aug 1995 17:24:11 +0100
To: [email protected]
Subject: Announce DBD-Oracle-0.20.tar.Z
Date: Tue, 1 Aug 1995 17:24:11 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

This is just a minor release to consolidate the recent small patches
I've posted.

Changes in DBD::Oracle 0.20, 1 Aug 1995:

Fixed core dump when binding an undef (treated as a NULL)
Binding a string longer that 2000 bytes will use LONG type.
Workaround OSF makefile and oratype.h problems
PL/SQL := construct will no longer confuse dbd_preparse()

Available from ftp.demon.co.uk:/pub/perl/db/DBD/Oracle/ as usual.

I'd appreciate it if you guys could test it since I've just posted
a reply to a message in comp.databases.oracle that may stimualate
more interest in this work.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12152-4@lion>;
Thu, 3 Aug 1995 07:01:19 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 02 Aug 95 21:24:21 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa05662;
2 Aug 95 22:24 +0100
Received: by gw.home.vix.com id AA14307; Wed, 2 Aug 95 06:45:06 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA14303; Wed, 2 Aug 95 06:45:05 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id IAA165142;
Wed, 2 Aug 1995 08:45:04 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA22594;
Wed, 2 Aug 1995 08:44:58 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: Building DBI....what's going on?
Date: Wed, 02 Aug 95 08:44:57 -0500

I thought I had perl5.001m installed, along with MakeMaker v4.16.
However, I get following error message when attempting to build
DBI, which implies that something else is going on. Any ideas?

Mike
[email protected]

$ make all
mkdir ./blib
mkdir ./blib/auto
mkdir ./blib/auto/DBI
Perl lib version (5.000) doesn't match executable version (5.001)
BEGIN failed--compilation aborted at
/usr/local/lib/perl5/ExtUtils/MakeMaker.pm line 10.
BEGIN failed--compilation aborted at -e line 1.
make: The error code from the last command is 22.

Stop.

------------------------------------------------------------------------------
Michael H. Moran | Standard Disclaimer: The content of
[email protected] | this posting is independent of
Commercial Performance | official IBM position.
IBM Corporation, Austin, Texas |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12318-0@lion>;
Thu, 3 Aug 1995 07:01:49 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 02 Aug 95 22:12:24 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26807;
2 Aug 95 23:12 +0100
Received: by gw.home.vix.com id AA17419; Wed, 2 Aug 95 07:31:05 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA17411; Wed, 2 Aug 95 07:30:55 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa26760;
2 Aug 95 15:26 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa16729;
2 Aug 95 15:26 +0100
Received: from ig.co.uk by lion id <02948-0@lion>;
Wed, 2 Aug 1995 15:18:22 +0100
To: [email protected]
Subject: Re: Building DBI....what's going on?
Cc: [email protected]
Date: Wed, 2 Aug 1995 15:18:22 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> I thought I had perl5.001m installed, along with MakeMaker v4.16.
> However, I get following error message when attempting to build
> DBI, which implies that something else is going on. Any ideas?
>
> Mike
> [email protected]
>
> $ make all
> mkdir ./blib
> mkdir ./blib/auto
> mkdir ./blib/auto/DBI
> Perl lib version (5.000) doesn't match executable version (5.001)
> BEGIN failed--compilation aborted at
> /usr/local/lib/perl5/ExtUtils/MakeMaker.pm line 10.

That message comes from Config.pm. It checks that the version that wrote
and installed Config.pm is the same one that's trying to run it.

You appear to have either not installed perl5.001m (via installperl)
or perhaps you have PERL5LIB or PERLLIB set in your environment.

Just saying perl -e 'use Config' will produce the error.

Umm, then again, you should have got the error when you ran
'perl Makefile.PL' perhaps you're using an old Makefile.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12318-10@lion>;
Thu, 3 Aug 1995 07:04:06 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 02 Aug 95 23:39:25 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa02574;
3 Aug 95 0:39 +0100
Received: by gw.home.vix.com id AA20706; Wed, 2 Aug 95 08:20:11 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA20656; Wed, 2 Aug 95 08:20:06 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id KAA154387;
Wed, 2 Aug 1995 10:20:05 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA21176;
Wed, 2 Aug 1995 10:18:48 -0500
From: [email protected]
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected], [email protected]
Subject: Re: Building DBI....what's going on?
In-Reply-To: Your message of Wed, 02 Aug 95 15:18:22 +0100. <[email protected]>
Date: Wed, 02 Aug 95 10:18:47 -0500

Tim Bunce <[email protected]> writes:
|>
|> > From: [email protected]
|> >
|> > I thought I had perl5.001m installed, along with MakeMaker v4.16.
|> > However, I get following error message when attempting to build
|> > DBI, which implies that something else is going on. Any ideas?
|> >
|> > Mike
|> > [email protected]
|> >
|> > $ make all
|> > mkdir ./blib
|> > mkdir ./blib/auto
|> > mkdir ./blib/auto/DBI
|> > Perl lib version (5.000) doesn't match executable version (5.001)
|> > BEGIN failed--compilation aborted at
|> > /usr/local/lib/perl5/ExtUtils/MakeMaker.pm line 10.
|>
|> That message comes from Config.pm. It checks that the version that wrote
|> and installed Config.pm is the same one that's trying to run it.
|>
|> You appear to have either not installed perl5.001m (via installperl)
|> or perhaps you have PERL5LIB or PERLLIB set in your environment.
|>
|> Just saying perl -e 'use Config' will produce the error.
|>
|> Umm, then again, you should have got the error when you ran
|> 'perl Makefile.PL' perhaps you're using an old Makefile.
|>

Well, none of the above is true. In the last hour, I have completely
rebuilt and installed perl5.001m, doubled checked the environment
(neither PERL5LIB not PERLLIB present), run "perl -e 'use Config'"
(no error message), and run "perl Makefile.PL" (again, no error).

All of this has been done against a copy of DBI v0.58, using
MakeMaker v4.16. Yet, I still can't build DBI. Any other ideas?

Mike
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12318-11@lion>;
Thu, 3 Aug 1995 07:04:14 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 02 Aug 95 23:45:22 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa04585;
3 Aug 95 0:44 +0100
Received: by gw.home.vix.com id AA23727; Wed, 2 Aug 95 09:00:49 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA23723; Wed, 2 Aug 95 09:00:48 -0700
Received: from nando.net.nando.net (parsifal.nando.net)
by merlin.nando.net (4.1/davel-nando/dec93) id AA22119;
Wed, 2 Aug 95 12:00:30 EDT
Received: by nando.net.nando.net (4.1/SMI-4.1) id AA28990;
Wed, 2 Aug 95 12:00:23 EDT
From: Chip Salzenberg <[email protected]>
Message-Id: <[email protected]>
Subject: DBD::Informix?
To: [email protected]
Date: Wed, 2 Aug 1995 12:00:23 -0400 (EDT)
X-Mailer: ELM [version 2.4 PL23]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 256

Would anyone know if a DBD::Informix exists or is in progress?
adTHANKvance
--
Chip Salzenberg, aka <[email protected]>
"And remember to worship at the railroad of your choice."
-- Mike Nelson, MST3K: "The Amazing Transparent Man"
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12152-24@lion>;
Thu, 3 Aug 1995 07:04:52 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 03 Aug 95 00:53:32 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26439;
3 Aug 95 1:53 +0100
Received: by gw.home.vix.com id AA23731; Wed, 2 Aug 95 09:00:57 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA23721; Wed, 2 Aug 95 09:00:43 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa03152;
2 Aug 95 16:39 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa26866;
2 Aug 95 16:40 +0100
Received: from ig.co.uk by lion id <05233-0@lion>;
Wed, 2 Aug 1995 16:38:51 +0100
To: [email protected]
Subject: Re: Building DBI....what's going on?
Cc: [email protected]
Date: Wed, 2 Aug 1995 16:38:51 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim Bunce <[email protected]> writes:
> |>
> |> > From: [email protected]
> |> >
> |> > I thought I had perl5.001m installed, along with MakeMaker v4.16.
> |> > However, I get following error message when attempting to build
> |> > DBI, which implies that something else is going on. Any ideas?
> |> >
> |> > Mike
> |> > [email protected]
> |> >
> |> > $ make all
> |> > mkdir ./blib
> |> > mkdir ./blib/auto
> |> > mkdir ./blib/auto/DBI
> |> > Perl lib version (5.000) doesn't match executable version (5.001)
> |> > BEGIN failed--compilation aborted at
> |> > /usr/local/lib/perl5/ExtUtils/MakeMaker.pm line 10.
> |>
> |> That message comes from Config.pm. It checks that the version that wrote
> |> and installed Config.pm is the same one that's trying to run it.
> |>
> |> You appear to have either not installed perl5.001m (via installperl)
> |> or perhaps you have PERL5LIB or PERLLIB set in your environment.
> |>
> |> Just saying perl -e 'use Config' will produce the error.
> |>
> |> Umm, then again, you should have got the error when you ran
> |> 'perl Makefile.PL' perhaps you're using an old Makefile.
>
> Well, none of the above is true. In the last hour, I have completely
> rebuilt and installed perl5.001m, doubled checked the environment
> (neither PERL5LIB not PERLLIB present), run "perl -e 'use Config'"
> (no error message), and run "perl Makefile.PL" (again, no error).
>
> All of this has been done against a copy of DBI v0.58, using
> MakeMaker v4.16. Yet, I still can't build DBI. Any other ideas?
>
I presume 'I still can't build DBI' means you get the same error as above.

Okay, try these things:

perl Makefile.PL verbose

check Makefile definition of $(PERL)

delete all '@' prefixes from commands in Makefile

Check for stray old Config.pm's in the directories listed by
perl -l -e 'print join(":", @INC)'

> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12318-17@lion>;
Thu, 3 Aug 1995 07:05:03 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 03 Aug 95 01:06:59 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa29961;
3 Aug 95 2:06 +0100
Received: by gw.home.vix.com id AA14925; Wed, 2 Aug 95 06:55:08 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA14921; Wed, 2 Aug 95 06:55:06 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id OAA08513;
Wed, 2 Aug 1995 14:51:47 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id OAA21122; Wed, 2 Aug 1995 14:56:42 +0100
Message-Id: <[email protected]>
Subject: DBD::mSQL site
To: [email protected], [email protected]
Date: Wed, 2 Aug 1995 14:56:42 +0100 (BST)
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 173

DBD::mSQL is available from:

ftp://ftp.mcqueen.com/pub/languages/perl5/DBPerl

The latest release version is 0.59.

A.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12152-26@lion>;
Thu, 3 Aug 1995 07:05:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 03 Aug 95 01:49:17 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa10532;
3 Aug 95 2:49 +0100
Received: by gw.home.vix.com id AA25989; Wed, 2 Aug 95 09:41:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA25984; Wed, 2 Aug 95 09:41:35 -0700
Received: from morpheus.cis.yale.edu (morpheus) by minerva.cis.yale.edu
with SMTP id AA05731 (5.67a/IDA-1.5 for <[email protected]>);
Wed, 2 Aug 1995 12:41:23 -0400
Received: by morpheus.cis.yale.edu id AA02504 (5.67a/IDA-1.5
for interest perl <[email protected]>);
Wed, 2 Aug 1995 12:41:23 -0400
Date: Wed, 2 Aug 1995 12:41:22 -0400 (EDT)
From: Royal Hansen <[email protected]>
X-Sender: rivory@morpheus
To: interest perl <[email protected]>
Subject: ora_do
Message-Id: <Pine.SOL.3.91.950802123905.2383A-100000@morpheus>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

Will someone post a syntactically correct ora_do command. Even better,
will someone post a syntactically correct ora_do command that calls a
stored procedure.
Thanks,
Royal Hansen
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12152-28@lion>;
Thu, 3 Aug 1995 07:05:25 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 03 Aug 95 02:24:31 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa20348;
3 Aug 95 3:24 +0100
Received: by gw.home.vix.com id AA26193; Wed, 2 Aug 95 09:48:51 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA26186; Wed, 2 Aug 95 09:48:47 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id RAA09877;
Wed, 2 Aug 1995 17:44:55 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id RAA27086; Wed, 2 Aug 1995 17:49:51 +0100
Message-Id: <[email protected]>
Subject: Re: DBD::Informix?
To: Chip Salzenberg <[email protected]>
Date: Wed, 2 Aug 1995 17:49:51 +0100 (BST)
Cc: [email protected]
In-Reply-To: <[email protected]> from "Chip Salzenberg" at Aug 2, 95 12:00:23 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 229

> Would anyone know if a DBD::Informix exists or is in progress?
> adTHANKvance

I'm in the process of doing one! Don't expect any great shakes for a while
yet, though.....

A.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23877-23@lion>;
Fri, 4 Aug 1995 07:08:43 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 04 Aug 95 01:23:25 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22318;
4 Aug 95 2:22 +0100
Received: by gw.home.vix.com id AA06323; Thu, 3 Aug 95 12:02:47 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA06319; Thu, 3 Aug 95 12:02:46 -0700
Received: from morpheus.cis.yale.edu (morpheus) by minerva.cis.yale.edu
with SMTP id AA10311 (5.67a/IDA-1.5 for <[email protected]>);
Thu, 3 Aug 1995 15:02:43 -0400
Received: by morpheus.cis.yale.edu id AA01244 (5.67a/IDA-1.5
for interest perl <[email protected]>);
Thu, 3 Aug 1995 15:02:40 -0400
Date: Thu, 3 Aug 1995 15:02:40 -0400 (EDT)
From: Royal Hansen <[email protected]>
X-Sender: rivory@morpheus
To: interest perl <[email protected]>
Subject: $lda pass
In-Reply-To: <Pine.SOL.3.91.950802123905.2383A-100000@morpheus>
Message-Id: <Pine.SOL.3.91.950803144646.398A-100000@morpheus>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

I make an initial ora_logon call after receiving the user/password data
from an html post submission. I then attempt to continue to pass the
$lda of that logon through a hidden variable from one html screen to a
corresponding perl script until the user finishes a session. The literal
string value ie. "DBI::db=HASH(....)" of the $lda comes through but when
I try to open a $csr with it, I am told that it can't call method prepare
in an empty package and it prints the $lda string. How can I pass this
$lda so that I don't have to logon for every action made by the user?
Thanks,
Royal Hansen
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <26885-3@lion>;
Fri, 4 Aug 1995 10:34:52 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 04 Aug 95 07:03:20 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22586;
4 Aug 95 8:02 +0100
Received: by gw.home.vix.com id AA07694; Thu, 3 Aug 95 20:05:53 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA07689; Thu, 3 Aug 95 20:05:51 -0700
Received: by metronet.com id AA18183 (5.67a/IDA1.5hp
for [email protected]); Thu, 3 Aug 1995 22:04:37 -0500
From: Tye McQueen <[email protected]>
Message-Id: <[email protected]>
Subject: Re: $lda pass
To: Royal Hansen <[email protected]>
Date: Thu, 3 Aug 1995 22:04:33 -0500 (CDT)
Cc: [email protected]
In-Reply-To: <Pine.SOL.3.91.950803144646.398A-100000@morpheus> from "Royal Hansen" at Aug 3, 95 03:02:40 pm
Content-Transfer-Encoding: 8bit
X-Mailer: ELM [version 2.4 PL24]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1732

Excerpts from the mail message of Royal Hansen:
) I make an initial ora_logon call after receiving the user/password data
) from an html post submission. I then attempt to continue to pass the
) $lda of that logon through a hidden variable from one html screen to a
) corresponding perl script until the user finishes a session. The literal
) string value ie. "DBI::db=HASH(....)" of the $lda comes through but when
) I try to open a $csr with it, I am told that it can't call method prepare
) in an empty package and it prints the $lda string. How can I pass this
) $lda so that I don't have to logon for every action made by the user?

A ora_login is specific to _one_ process (and _one_ invocation
of perl as well) so the only way to do this is to start a new
process that logs in to Oracle and hangs around, waiting for
requests to be sent to it by other perl scripts that get called
from the HTTP daemon.

This can be a global daemon that just runs forever and processes
the requests from all sessions. Or you can create a new daemon
for each session and have it exit when the session ends (I don't
know how to ensure that so I'd also/just have it exit if it
doesn't receive any requests after a certain length of time).
Or you can have a global daemon that times out and exits after
a certain length of time with no requests and that each session
starts a new only if there isn't one already running.

You may have to read up on interprocess communication (IPC) methods
(pipes, sockets, message queues, shared memory, etc.) so the scripts
and the daemon can communicate.

--
Tye McQueen [email protected] || [email protected]
Nothing is obvious unless you are overlooking something
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12177-13@lion>;
Mon, 7 Aug 1995 00:06:01 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 04 Aug 95 21:12:57 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa09181;
4 Aug 95 22:12 +0100
Received: by gw.home.vix.com id AA06972; Fri, 4 Aug 95 04:03:06 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA06968; Fri, 4 Aug 95 04:03:05 -0700
Received: from tilde.csc.ti.com ([128.247.160.56]) by gate.ti.com (8.6.12/)
with ESMTP id GAA25302 for <[email protected]>;
Fri, 4 Aug 1995 06:03:04 -0500
Received: from galaxy.mktg.tide.ti.com (galaxy.mktg.tide.ti.com [134.183.37.60])
by tilde.csc.ti.com (8.6.12/8.6.12) with SMTP id GAA12567
for <[email protected]>; Fri, 4 Aug 1995 06:02:33 -0500
Received: from darkstar.mktg.tide.ti.com.
by galaxy.mktg.tide.ti.com (4.1/SMI-4.1) id AA19927;
Fri, 4 Aug 95 12:59:56 +0200
From: Alexander Strauss <[email protected]>
Message-Id: <[email protected]>
Subject: perl-sql interface
To: [email protected]
Date: Fri, 4 Aug 1995 12:59:32 +0200 (MET DST)
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 387

--- hi there ---

My name is Alex and I've the following problem:

I'm writing at a perl script to generate multible system files
for UNIX administration.
The hostdata I want to request from the microsoft sql server,
which supports ANSI SQL.
Which implementations of a perl / sql interface are there:
is there a dynamic module for perl 5?

Which possibilities we have ???

Thanks Alex

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12010-17@lion>;
Mon, 7 Aug 1995 00:08:14 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 04 Aug 95 23:40:35 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa13874;
5 Aug 95 0:39 +0100
Received: by gw.home.vix.com id AA05528; Fri, 4 Aug 95 12:15:01 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA05518; Fri, 4 Aug 95 12:14:56 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa02223;
4 Aug 95 20:02 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa04565;
4 Aug 95 19:59 +0100
Received: from ig.co.uk by lion id <02304-0@lion>;
Fri, 4 Aug 1995 20:00:11 +0100
To: [email protected]
Subject: Short term plans - internal DBI/DBD interface changes
Date: Fri, 4 Aug 1995 20:00:11 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

This is just a quick note to let you know that I plan to implement some
internal changes soon. If you're not developing a DBD driver module you
can ignore this message.

I have never really felt comfortable with the way the DBI attaches data
to the handles. It's big, ugly and slow.

(a) RV the handle ref the application holds
|
v
HV the outer hash of the tie
|
| Perl 'magic' link to inner hash
v
(b) HV the inner hash of the tie
|
| DBI 'magic' link to handle data
v
RV ref to array (this is redundant and should have gone ages ago)
|
v
AV array holding handle data
|
~~~ array elements holding assorted data including the...
* ->
* -> pointer to 'dbi common handle data'
(c) * -> ---. pointer to 'implementors private data'
* -> |
... |
~~~ |
|
(d) { <-----'
....
}

The idea was that since (almost) everything was a perl datatype
everything would automatically get tidied up when the applications
reference was destroyed. The price is that it costs a lot to get from
point (b) to point (d). A knock-on effect was that I was having to
pass both (b) and (d) to functions within a DBD.

The new scheme, from point (b), will look more like this:

(b) HV the inner hash of the tie
|
| DBI 'magic' link to blessed SV holding data
v
(d) {
....
}

I think you'll agree that that's a little better :-)

Two key points make this practical:

1. The 'implementors private data' and the 'dbi common handle data'
both live at the same address! The trick is for the DBD to declare
its private data structure like this:

struct foo {
struct dbih_com_st com;
...
...
}

Since I know that my magic will be the first magic on the inner hash HV
I can optimise the (b) to (d) mapping into (simplifying greatly):

SvPVX( hv -> xmg_magic -> mg_obj )

2. The SV holding the structure is blessed into a 'parallel' class
specifically for the management of this memory. Generally the
class will only contain a DESTROY method. The DBI will provide
a pointer in dbistate to a function which knows how to clear
the contents of the 'dbi common handle data' structure.
A default DESTROY method will simply call that function.

The existing use of macros should greatly reduce the pain of any required code
changes.

In addition to the above I plan to migrate some of the implementors private data
and row/field buffer management developed for DBD::Oracle into the DBI. This
will enable the DBI to provide support for bindparam() and bindcol() methods
thus reducing the per-driver code required.

I hope to get into this next week but no promises!

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12177-35@lion>;
Mon, 7 Aug 1995 00:08:29 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 05 Aug 95 01:14:30 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17830;
5 Aug 95 2:13 +0100
Received: by gw.home.vix.com id AA06294; Fri, 4 Aug 95 12:23:50 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA06290; Fri, 4 Aug 95 12:23:49 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa02464;
4 Aug 95 20:05 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa04936;
4 Aug 95 20:02 +0100
Received: from ig.co.uk by lion id <02809-0@lion>;
Fri, 4 Aug 1995 20:05:20 +0100
To: [email protected]
Subject: ... also
Date: Fri, 4 Aug 1995 20:05:20 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

I also plan to put some effort into cleaning things up to make it
easier to implement drivers.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25385-6@lion>;
Mon, 7 Aug 1995 19:55:09 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 07 Aug 95 16:49:21 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa21665;
7 Aug 95 17:49 +0100
Received: by gw.home.vix.com id AA28591; Mon, 7 Aug 95 05:03:33 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA28587; Mon, 7 Aug 95 05:03:31 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa05395;
7 Aug 95 12:47 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa29664;
7 Aug 95 12:44 +0100
Received: from ig.co.uk by lion id <21401-0@lion>;
Mon, 7 Aug 1995 12:47:24 +0100
To: [email protected]
Subject: Re: Quick synopsis please.
Cc: [email protected]
Date: Mon, 7 Aug 1995 12:47:24 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]

In relation to:

|> > Tim,
|> > Do you have time to give a quick synopsis of what all the non-oracle
|> > variables you defined in dbdimp.h are used for? I finally have some
|> > time to work on DB2Perl and want to make sure that despite all the
|> > differences my version returns the same information your version
|> > returns.

> Basically, the two fields in question are dbh_generation and flags.
>
Ah, okay. Basically dbh_generation is (was) part of the work required to
deal with a disconnect/destroy of a dbh while prepared statements still
existed (such as the random destruction order when perl exists).

The fix/workaround was to never free the imp_dbh struct just
reuse it after incrementing dbh_generation. The sth code would check
the dbh_generation to see it was the same before using it.

All this code _will_ change in the next release (see my recent big
message). On the up side the new code will be much simpler!

Tim.

p.s. Oh yes, flags. That is/was used to check in DESTROY to see if a
finish/disconnect was required etc. Again, this will change.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <28733-31@lion>;
Tue, 8 Aug 1995 07:05:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 07 Aug 95 23:09:13 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa08495;
8 Aug 95 0:09 +0100
Received: by gw.home.vix.com id AA19603; Mon, 7 Aug 95 10:58:45 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA19599; Mon, 7 Aug 95 10:58:44 -0700
Received: from morpheus.cis.yale.edu (morpheus) by minerva.cis.yale.edu
with SMTP id AA29620 (5.67a/IDA-1.5 for <[email protected]>);
Mon, 7 Aug 1995 13:58:33 -0400
Received: by morpheus.cis.yale.edu id AA16846 (5.67a/IDA-1.5);
Mon, 7 Aug 1995 13:58:33 -0400
Date: Mon, 7 Aug 1995 13:58:32 -0400 (EDT)
From: Royal Hansen <[email protected]>
X-Sender: rivory@morpheus
To: Tim Bunce <[email protected]>
Cc: [email protected]
Subject: Oracle long
In-Reply-To: <[email protected]>
Message-Id: <Pine.SOL.3.91.950807135422.16451A-100000@morpheus>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

I am having problems reading in item of type long from an Oracle table to
a perl variable - $whatever -. Can DBD oraperl handle a long? If so, how?
Royal Hansen
[email protected]
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <29307-16@lion>;
Tue, 8 Aug 1995 07:07:02 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 08 Aug 95 01:23:01 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17219;
8 Aug 95 2:22 +0100
Received: by gw.home.vix.com id AA25623; Mon, 7 Aug 95 12:38:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA25619; Mon, 7 Aug 95 12:38:39 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id ab12005;
7 Aug 95 20:11 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aj19471;
7 Aug 95 20:08 +0100
Received: from ig.co.uk by lion id <26209-0@lion>;
Mon, 7 Aug 1995 20:00:59 +0100
To: [email protected]
Subject: Re: Oracle long
Cc: [email protected]
Date: Mon, 7 Aug 1995 20:00:59 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Royal Hansen <[email protected]>
>
> I am having problems reading in item of type long from an Oracle table to
> a perl variable - $whatever -. Can DBD oraperl handle a long? If so, how?
> Royal Hansen
> [email protected]
>
Try
$sth->fetchrow;
...
$sth->readblob($field, $offset, $length [, \$dest])

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <00365-0@lion>;
Tue, 8 Aug 1995 07:19:24 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 08 Aug 95 06:15:06 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17438;
8 Aug 95 7:14 +0100
Received: by gw.home.vix.com id AA12959; Mon, 7 Aug 95 17:01:50 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA12955; Mon, 7 Aug 95 17:01:48 -0700
Received: from hometown.com by metronet.com with SMTP
id AA11590 (5.67a/IDA1.5hp for <[email protected]>);
Mon, 7 Aug 1995 19:00:29 -0500
Message-Id: <[email protected]>
X-Sender: [email protected]
X-Mailer: Windows Eudora Version 2.1
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Mon, 07 Aug 1995 18:59:04 -0500
To: [email protected]
From: Steve Livingston <[email protected]>
Subject: DBD::Oracle make fails
Cc: [email protected]

Oh, my...DBD::Oracle make fails:

- Oracle 7.1.6
- Sun 4.1.3
- Perl 5.001l

What to do?

Cheers,
Steve

% make realclean
rm -rf *~ t/*~ *.o *.a mon.out core so_locations Oracle.bs Oracle.bso
Oracle.exp Oracle.c ./blib
mv Makefile Makefile.old 2>/dev/null
rm -rf ./blib/auto/DBD/Oracle ./blib/auto/DBD/Oracle
rm -f ./blib/auto/DBD/Oracle/Oracle.so ./blib/auto/DBD/Oracle/Oracle.bs
rm -f ./blib/auto/DBD/Oracle/Oracle.a ./blib/Oraperl.pm ./blib/DBD/Oracle.pm
rm -rf Makefile Makefile.old
rm -f ./perlmain.o ./perlmain.c Makefile extralibs.ld

elcaro!steve> (csh) /private1/usr/local/bin/ext/Oracle
% perl Makefile.PL

Configuring DBD::Oracle version 0.20...
Remember to actually read the README file!

Using Oracle in /private1/o7
Oracle sysliblist: -lm
Using /private1/o7/proc/lib/proc.mk version 1.17
System: perl5.001 sunos elcaro 4.1.3 1 sun4c sun4- dl_dlopen.xs
Compiler: gcc -O
Oracle proc.mk would have used these values but we override them:
CC: /usr/5bin/cc
CFLAGS: -I. -O
LDFLAGS: -Bdynamic -L$(LIBHOME)

Checking if your kit is complete...
Looks good
Writing Makefile for DBD::Oracle

elcaro!steve> (csh) /private1/usr/local/bin/ext/Oracle
% make
mkdir ./blib
mkdir ./blib/DBD
mkdir ./blib/auto
mkdir ./blib/auto/DBD
mkdir ./blib/auto/DBD/Oracle
/usr/local/bin/perl -I/usr/local/lib/perl5/sun4- -I/usr/local/lib/perl5
/usr/local/lib/perl5/ExtUtils/xsubpp -typemap
/usr/local/lib/perl5/ExtUtils/typemap Oracle.xs >Oracle.tc && mv Oracle.tc
Oracle.c
gcc -c -O -fpic -I/usr/local/lib/perl5/sun4-/CORE -Wall -pedantic
-Wno-comment -Wpointer-arith -Wcast-align -Wconversion -Wtraditional
-Wpointer-arith -Wbad-function-cast -Wcast-qual -I/private1/o7/rdbms/demo
-I/usr/local/lib/perl5/sun4-/DBI Oracle.c
gcc -c -O -fpic -I/usr/local/lib/perl5/sun4-/CORE -Wall -pedantic
-Wno-comment -Wpointer-arith -Wcast-align -Wconversion -Wtraditional
-Wpointer-arith -Wbad-function-cast -Wcast-qual -I/private1/o7/rdbms/demo
-I/usr/local/lib/perl5/sun4-/DBI dbdimp.c
dbdimp.c: In function `do_error':
dbdimp.c:50: warning: passing arg 2 of `oerhms' with different width due to
prototype
dbdimp.c:65: warning: implicit declaration of function `fprintf'
dbdimp.c: In function `fbh_dump':
dbdimp.c:76: warning: function `fprintf' was previously declared within a block
dbdimp.c: In function `alloc_imp_dbh':
dbdimp.c:102: warning: unused variable `sv'
dbdimp.c: In function `dump_error_status':
dbdimp.c:131: warning: function `fprintf' was previously declared within a block
dbdimp.c: In function `dbd_dr_login':
dbdimp.c:155: warning: function `fprintf' was previously declared within a block
dbdimp.c:146: warning: unused variable `items'
dbdimp.c: In function `dbd_db_commit':
dbdimp.c:173: warning: unused variable `items'
dbdimp.c: In function `dbd_db_rollback':
dbdimp.c:188: warning: unused variable `items'
dbdimp.c: In function `dbd_db_disconnect':
dbdimp.c:203: warning: unused variable `items'
dbdimp.c: In function `dbd_db_prepare':
dbdimp.c:256: warning: unused variable `items'
dbdimp.c: In function `dbd_preparse':
dbdimp.c:374: warning: function `fprintf' was previously declared within a block
Running Mkbootstrap for DBD::Oracle ()
chmod 644 Oracle.bs
LD_RUN_PATH="" ld -o blib/auto/DBD/Oracle/Oracle.so Oracle.o dbdimp.o
-L/private1/o7/lib -locic /private1/o7/lib/osntab.o -lsqlnet -lora
-lsqlnet -lora -lnlsrtl -lcv6 -lcore -lnlsrtl -lcv6 -lcore `cat
/private1/o7/rdbms/lib/sysliblist` -L/usr/5lib -lc -lm
ld: Undefined symbol
_environ
_dlopen
_dlclose
_dlsym
*** Error code 2
make: Fatal error: Command failed for target `blib/auto/DBD/Oracle/Oracle.so'

------------------

% perl -e 'use Config; print Config::myconfig()'
Summary of my perl5 (patchlevel 1) configuration:
Platform:
osname=, osver=1, archname=sun4-
uname='sunos elcaro 4.1.3 1 sun4c '
hint=recommended
Compiler:
cc='gcc', optimize='-O'
cppflags=''
ccflags =''
ldflags =''
stdchar='unsigned char', d_stdstdio=define, usevfork=false
voidflags=15, castflags=0, d_casti32=define, d_castneg=define
intsize=4, alignbytes=8, usemymalloc=y, randbits=31
Libraries:
so=so
libpth=/lib /usr/lib /usr/ucblib /usr/local/lib
libs=-lnsl -ldbm -ldl -lm -lc -lposix
libc=/usr/lib/libc.so.1.8.1
Dynamic Linking:
dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=
cccdlflags='-fpic', ccdlflags=' ', lddlflags=' '

---------------------

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <02123-2@lion>;
Tue, 8 Aug 1995 13:34:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 08 Aug 95 06:33:23 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa21330;
8 Aug 95 7:33 +0100
Received: by gw.home.vix.com id AA20011; Mon, 7 Aug 95 19:19:17 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA20006; Mon, 7 Aug 95 19:19:09 -0700
Received: from jazz.iinet.com.au ([email protected] [203.0.178.3])
by classic.iinet.com.au (8.6.12/8.6.9) with ESMTP id KAA10898
for <[email protected]>; Tue, 8 Aug 1995 10:19:01 +0800
From: Adrian Smith <[email protected]>
Received: (fon@localhost) by jazz.iinet.com.au (8.6.10/8.6.9) id KAA17136
for [email protected]; Tue, 8 Aug 1995 10:18:58 +0800
Message-Id: <[email protected]>
Subject: PERLDB for WWW question!
To: [email protected]
Date: Tue, 8 Aug 1995 10:18:56 +0800 (WST)
X-Mailer: ELM [version 2.4 PL23]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 621

>
> Hi, can anyone help me with this:
> > > Two, I would like a couple of recommendations for a
> > > database to use on Linux with perl. It will serve as
> > > the back-end to a dynamic web-page generator. I would
> > > expect only a few thousand fairly low-impact hits a day.
> > > I'm looking at mSql .. anything else? Freeware? Or should
> > > I be using something like Oracle (on Linux?).
> >
> I would also _really_ like to see an example (or working code)
> of using perl to query a database and generate html.
>
> Ideally, the whole WWW Form <-> perldb <-> sql server.
>
>
> Adrian
> [email protected]
>
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <08186-9@lion>;
Wed, 9 Aug 1995 03:55:02 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 02:03:07 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa27672;
9 Aug 95 3:02 +0100
Received: by gw.home.vix.com id AA27848; Tue, 8 Aug 95 13:23:12 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA27843; Tue, 8 Aug 95 13:23:11 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa08660;
8 Aug 95 20:31 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac00557;
8 Aug 95 20:28 +0100
Received: from ig.co.uk by lion id <04971-0@lion>;
Tue, 8 Aug 1995 18:24:08 +0100
To: [email protected]
Subject: Re: strdup on NeXTStep..?
Cc: [email protected]
Date: Tue, 8 Aug 1995 18:24:08 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim, can you think of where strdup might be causing autoloading problems
> in perl5/DBI/DBD::mSQL?
>
> strdup doesn't exist in DBD::mSQL.....
>
> The person's trying to run it on NeXTStep.....
>
Umm, wild guess... some linkers only show the first undefined symbol.
strdup might just be the first of many.

Beyond that I've no particular ideas.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <08186-10@lion>;
Wed, 9 Aug 1995 03:55:08 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 02:03:27 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa27781;
9 Aug 95 3:03 +0100
Received: by gw.home.vix.com id AA28665; Tue, 8 Aug 95 13:38:45 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA28660; Tue, 8 Aug 95 13:38:41 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa08702;
8 Aug 95 20:31 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ai00557;
8 Aug 95 20:29 +0100
Received: from ig.co.uk by lion id <05590-0@lion>;
Tue, 8 Aug 1995 20:29:00 +0100
To: [email protected], [email protected]
Subject: Re: perl-sql interface
Date: Tue, 8 Aug 1995 20:29:00 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Alexander Strauss <[email protected]>
>
> --- hi there ---
>
> My name is Alex and I've the following problem:
>
> I'm writing at a perl script to generate multible system files
> for UNIX administration.
> The hostdata I want to request from the microsoft sql server,
> which supports ANSI SQL.
> Which implementations of a perl / sql interface are there:
> is there a dynamic module for perl 5?
>
> Which possibilities we have ???
>
None specifically for microsoft sql server.
Perhaps the Sybperl module could be made to work.
I don't know how hard that would be.

> Thanks Alex
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <08525-0@lion>;
Wed, 9 Aug 1995 04:16:08 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 03:03:36 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa13317;
9 Aug 95 4:03 +0100
Received: by gw.home.vix.com id AA03800; Tue, 8 Aug 95 14:51:02 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA03792; Tue, 8 Aug 95 14:50:55 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa08670;
8 Aug 95 20:31 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id af00557;
8 Aug 95 20:28 +0100
Received: from ig.co.uk by lion id <05535-0@lion>;
Tue, 8 Aug 1995 20:22:42 +0100
To: [email protected], [email protected]
Mmdf-Warning: Unable to confirm address in preceding line at
disperse.demon.co.uk
Subject: Re: DBD::Oracle make fails
Date: Tue, 8 Aug 1995 20:22:42 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Steve Livingston <[email protected]>
>
> Oh, my...DBD::Oracle make fails:
>
> - Oracle 7.1.6
> - Sun 4.1.3
> - Perl 5.001l

> Summary of my perl5 (patchlevel 1) configuration:
> Platform:
> uname='sunos elcaro 4.1.3 1 sun4c '
> Libraries:
> libpth=/lib /usr/lib /usr/ucblib /usr/local/lib
> libs=-lnsl -ldbm -ldl -lm -lc -lposix
> libc=/usr/lib/libc.so.1.8.1
> Dynamic Linking:
> dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=
> cccdlflags='-fpic', ccdlflags=' ', lddlflags=' '

> LD_RUN_PATH="" ld -o blib/auto/DBD/Oracle/Oracle.so Oracle.o dbdimp.o
> -L/private1/o7/lib -locic /private1/o7/lib/osntab.o -lsqlnet -lora
> -lsqlnet -lora -lnlsrtl -lcv6 -lcore -lnlsrtl -lcv6 -lcore `cat
> /private1/o7/rdbms/lib/sysliblist` -L/usr/5lib -lc -lm
> ld: Undefined symbol
> _environ
> _dlopen

Any idea why the link line end with '-L/usr/5lib -lc -lm' while your perl
config doesn't mention it?

> Cheers,
> Steve
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09211-3@lion>;
Wed, 9 Aug 1995 07:01:22 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 04:30:39 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa04894;
9 Aug 95 5:30 +0100
Received: by gw.home.vix.com id AA14285; Tue, 8 Aug 95 17:38:37 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA14281; Tue, 8 Aug 95 17:38:35 -0700
Received: from billberry (billberry.ai.mit.edu)
by life.ai.mit.edu (4.1/AI-4.10) for [email protected]
id AA04782; Tue, 8 Aug 95 20:38:32 EDT
From: Sundar Narasimhan <[email protected]>
Received: by billberry (4.1/AI-4.10) id AA11857; Tue, 8 Aug 95 20:38:29 EDT
Date: Tue, 8 Aug 95 20:38:29 EDT
Message-Id: <9508090038.AA11857@billberry>
To: [email protected]
Subject: question re. perl-4.036, oraperl and Solaris 2.4

Hi, For various reasons, I am in need of a perl-4.036 and oraperl that
work on Solaris 2.4. Does any one know of such beasts?

Thanks.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <07849-39@lion>;
Wed, 9 Aug 1995 03:53:46 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 08 Aug 95 23:54:33 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26583;
9 Aug 95 0:54 +0100
Received: by gw.home.vix.com id AA07595; Tue, 8 Aug 95 00:39:43 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA07589; Tue, 8 Aug 95 00:39:40 -0700
Received: from localhost by uranus.dtv.dk; (5.65/1.1.8.2/07Jul95-1033AM)
id AA18759; Tue, 8 Aug 1995 09:43:30 +0200
Message-Id: <[email protected]>
To: Adrian Smith <[email protected]>
Cc: [email protected]
Subject: Re: PERLDB for WWW question!
Date: Tue, 08 Aug 95 09:43:30 +0200
From: Bo Frese Rasmussen <[email protected]>
X-Mts: smtp

> Hi, can anyone help me with this:
> > > Two, I would like a couple of recommendations for a
> > > database to use on Linux with perl. It will serve as
> > > the back-end to a dynamic web-page generator. I would
> > > expect only a few thousand fairly low-impact hits a day.
> > > I'm looking at mSql .. anything else? Freeware? Or should
> > > I be using something like Oracle (on Linux?).
> >
> I would also _really_ like to see an example (or working code)
> of using perl to query a database and generate html.
>
> Ideally, the whole WWW Form <-> perldb <-> sql server.
>
>
> Adrian
> [email protected]

Try to have a look at WDB.

http://arch-http.hq.eso.org/bfrasmus/wdb/

BTW: It supports mSQL, Sybase, Informix, Ingres95 and Oracle. At the
moment it doesn't use the DBD interfaces, but it should be rather trivial to
convert it to support that too.

\\\\\// Cheers
| | Bo
(.) (.)
==========oOO==(_)==OOo=====================================================
Bo Frese Rasmussen | E-Mail: [email protected]
DANMARKS TEKNISKE VIDENCENTER, Udviklingslab. | Phone : +45 45 25 73 03
Postbox 777, DK-2800 Lyngby | Fax : +45 45 88 89 84
___ | http://www.dtv.dk/~bfr/
============( )-___=======================================================
\ ( ( )
\_) ) /
(_/
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <12573-5@lion>;
Wed, 9 Aug 1995 12:33:18 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 11:18:57 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa01662;
9 Aug 95 12:18 +0100
Received: by gw.home.vix.com id AA22792; Tue, 8 Aug 95 20:11:07 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA22788; Tue, 8 Aug 95 20:11:05 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id UAA28363; Tue, 8 Aug 1995 20:11:27 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma028339; Tue Aug 8 20:11:15 1995
Received: by bonnie (5.x/SMI-SVR4) id AA01005; Tue, 8 Aug 1995 20:11:11 -0700
Date: Tue, 8 Aug 1995 20:11:11 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508090311.AA01005@bonnie>
To: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
X-Sun-Charset: US-ASCII

>From: [email protected] (Sundar Narasimhan)
>Date: Tue, 8 Aug 95 20:38:29 EDT
>To: [email protected]
>Subject: question re. perl-4.036, oraperl and Solaris 2.4
>
>Hi, For various reasons, I am in need of a perl-4.036 and oraperl that
>work on Solaris 2.4. Does any one know of such beasts?
>
>Thanks.
>

We have solaris 2.4 and are running perl-4.036 and oraperl-v2.4. Aside
from occasional segfaults, it works great. Did you want binaries, or just the
configs in order to get the thing to build?

Note that if you want to use syslog with Solaris, you will have to hack
up perl a bit. (Standard perl uses BSD sockets for syslog. However, Solaris,
being SYSV, uses streams.) If you don't care about syslog, then disregard
this comment.

Cheers,
Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <15834-0@lion>;
Wed, 9 Aug 1995 16:26:27 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 15:22:24 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa05487;
9 Aug 95 16:22 +0100
Received: by gw.home.vix.com id AA03634; Tue, 8 Aug 95 23:48:05 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA03624; Tue, 8 Aug 95 23:47:57 -0700
Received: by anna.mind.de (940816.SGI.8.6.9/940406.SGI) id IAA10796;
Wed, 9 Aug 1995 08:46:55 +0200
Date: Wed, 9 Aug 1995 08:46:55 +0200
Message-Id: <[email protected]>
From: Andreas Koenig <[email protected]>
To: [email protected]
Cc: [email protected], [email protected]
In-Reply-To: <[email protected]> (message from Tim Bunce on Tue, 8 Aug 1995 18:24:08 +0100)
Subject: Re: strdup on NeXTStep..?
Reply-To: [email protected]

>From: Tim Bunce <[email protected]>
>
>> From: [email protected]
>>
>> Tim, can you think of where strdup might be causing autoloading problems
>> in perl5/DBI/DBD::mSQL?
>>
>> strdup doesn't exist in DBD::mSQL.....
>>
>> The person's trying to run it on NeXTStep.....
>>

Alligator, CC such questions to me, please. You get the answer much
quicker.

strdup is missing in the C libraries of the NeXT and on Ultrix. That
is why Bambi wrote some utility functions into a small library
libmisc.a.

On these systems You have to add -lmisc to the ld line (by adding
-lmisc to the LIBS argument in your Makefile.PL). Unfortunately Bambi
never cared to install libmisc.a, so people have to copy it by
hand. So it has a fixed place in my README file. (In version 1.00 I
erroneous deleted that sentence)

andreas

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18120-20@lion>;
Wed, 9 Aug 1995 19:27:28 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 18:04:37 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa18973;
9 Aug 95 19:04 +0100
Received: by gw.home.vix.com id AA19637; Wed, 9 Aug 95 03:25:40 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA19624; Wed, 9 Aug 95 03:25:32 -0700
Received: from sunnyboy.informatik.tu-chemnitz.de by obelix.hrz.tu-chemnitz.de
with Local SMTP (PP) id <[email protected]>;
Wed, 9 Aug 1995 12:25:14 +0200
Received: from jonathan.csn.tu-chemnitz.de
by sunnyboy.informatik.tu-chemnitz.de (4.1/SMI-4.1) id AA09124;
Wed, 9 Aug 95 12:25:11 +0200
Message-Id: <[email protected]>
Received: by jonathan.csn.tu-chemnitz.de (NX5.67e/NX3.0X) id AA00554;
Wed, 9 Aug 95 12:23:57 GMT
Content-Type: text/plain
Mime-Version: 1.0 (NeXT Mail 3.3 v118.2)
Original-Received: by NeXT.Mailer (1.118.2)
Pp-Warning: Illegal Received field on preceding line
From: Karsten Heinze <[email protected]>
Date: Wed, 9 Aug 95 12:23:56 GMT
To: [email protected]
Subject: msql connect problems
Reply-To: [email protected]

Hi,

I have problems to connect to msql database via DBI-0.58 and
DBD-mSQL-0_59pl1.
In dbispec.v05 I read:
$dbh = &db'connect([$database [, $username [, $password
[, $module [, %attr]]]]]);

but
$drh->connect( 'test' ); or
$drh->connect( 'test' , 'khe' );
in test2.pl from the DBD-mSQL-0_59pl1 package didn't work:

khe@jonathan:/tmp/O_DBD-mSQL-0_59pl1-NS/DBD-mSQL-0_59pl1-NS/DBD/mSQL
> perl test2.pl
Bootstrapping mSQL-0.58 ** Disney release
(c)1995 Alligator Descartes <[email protected]>
(c)1994-1995 Portions by Tim Bunce <[email protected]>

DBI dispatch debug level set to 2
** Connecting to first instance....
>> connect DISPATCH (DBI::dr=HASH(0xf69a4) @3 g0 a10bc5c r1)
-> connect for DBD::mSQL::dr (DBI::dr=HASH(0xf69a4)~0x11e8cc
'test' 'khe')
rc: 0 what: login failed
<- connect= undef

Where is my error ?

Thanks,
Karsten

---
*****************************************************************
Karsten Heinze
Tel.: +49/172/3763092
Fax: +49/37296/15056
e-mail: [email protected] (NeXT-Mail/MIME)
*****************************************************************
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <19456-33@lion>;
Wed, 9 Aug 1995 22:43:08 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 20:46:03 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa03160;
9 Aug 95 21:45 +0100
Received: by gw.home.vix.com id AA01639; Wed, 9 Aug 95 06:57:47 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA01635; Wed, 9 Aug 95 06:57:46 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id IAA106401;
Wed, 9 Aug 1995 08:57:40 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA18178;
Wed, 9 Aug 1995 08:57:30 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: AvFILL - where is it?
Date: Wed, 09 Aug 95 08:57:29 -0500

Tim,

In dbdimp.c - alloc_imp_dbh(), you call AvFILL(imp_dbh_cache_av).
What does it do? I can't find it documented anywhere....wonder if
the pods on my machine are out of date.

Thanks.
Mike
[email protected]
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20961-4@lion>;
Thu, 10 Aug 1995 02:21:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 22:38:43 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa21176;
9 Aug 95 23:38 +0100
Received: by gw.home.vix.com id AA02099; Wed, 9 Aug 95 07:00:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA02095; Wed, 9 Aug 95 07:00:43 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id JAA100528;
Wed, 9 Aug 1995 09:00:41 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA18978;
Wed, 9 Aug 1995 09:00:40 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: AvFILL - where is it?
Date: Wed, 09 Aug 95 09:00:40 -0500

Tim,

In dbdimp.c - alloc_imp_dbh(), you call AvFILL(imp_dbh_cache_av).
What does it do? I can't find it documented anywhere....wonder if
the pods on my machine are out of date.

Thanks.
Mike
[email protected]

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20961-6@lion>;
Thu, 10 Aug 1995 02:21:19 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 09 Aug 95 22:42:27 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa23061;
9 Aug 95 23:42 +0100
Received: by gw.home.vix.com id AA06025; Wed, 9 Aug 95 08:06:35 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA06021; Wed, 9 Aug 95 08:06:34 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id KAA84403;
Wed, 9 Aug 1995 10:06:33 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA02558;
Wed, 9 Aug 1995 10:06:31 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: AvFILL - where is it?
Date: Wed, 09 Aug 95 10:06:31 -0500

Tim,

In dbdimp.c - alloc_imp_dbh(), you call AvFILL(imp_dbh_cache_av).
What does it do? I can't find it documented anywhere....wonder if
the pods on my machine are out of date.

Thanks.
Mike
[email protected]

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20961-30@lion>;
Thu, 10 Aug 1995 02:23:01 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 10 Aug 95 00:02:46 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa23158;
10 Aug 95 1:02 +0100
Received: by gw.home.vix.com id AA15582; Wed, 9 Aug 95 10:23:14 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA15576; Wed, 9 Aug 95 10:23:10 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa02227;
9 Aug 95 17:17 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa10983;
9 Aug 95 17:14 +0100
Received: from ig.co.uk by lion id <16907-0@lion>;
Wed, 9 Aug 1995 17:17:21 +0100
To: [email protected]
Subject: Re: AvFILL - where is it?
Cc: [email protected]
Date: Wed, 9 Aug 1995 17:17:21 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim,
>
> In dbdimp.c - alloc_imp_dbh(), you call AvFILL(imp_dbh_cache_av).
> What does it do? I can't find it documented anywhere....wonder if
> the pods on my machine are out of date.
>
I assume you're refering to AvFILL().

It's exactly the same as $#ary in perl.

> Thanks.
> Mike
> [email protected]
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20961-34@lion>;
Thu, 10 Aug 1995 02:23:21 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 10 Aug 95 00:27:26 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00854;
10 Aug 95 1:27 +0100
Received: by gw.home.vix.com id AA19793; Wed, 9 Aug 95 11:24:09 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA19789; Wed, 9 Aug 95 11:24:08 -0700
Received: from hpcc123.corp.hp.com by hp.com
with ESMTP (1.37.109.16/15.5+ECS 3.3) id AA213542646;
Wed, 9 Aug 1995 11:24:07 -0700
Received: by hpcc123.corp.hp.com (1.37.109.16/15.5+ECS 3.3) id AA061712646;
Wed, 9 Aug 1995 11:24:06 -0700
From: Jeff Okamoto <[email protected]>
Message-Id: <[email protected]>
Subject: Re: AvFILL - where is it?
To: [email protected]
Date: Wed, 9 Aug 95 11:24:06 PDT
Cc: [email protected], [email protected]
In-Reply-To: <[email protected]>; from "[email protected]" at Aug 09, 95 8:57 am
Mailer: Elm [revision: 70.85]

> In dbdimp.c - alloc_imp_dbh(), you call AvFILL(imp_dbh_cache_av).
> What does it do? I can't find it documented anywhere....wonder if
> the pods on my machine are out of date.

AvFILL is not described in any pod -- it's a macro in av.h that returns
the value of the highest index in the array minus the base value of the
array (thus, it's $#array - $[).

Jeff
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23030-12@lion>;
Thu, 10 Aug 1995 07:02:19 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 10 Aug 95 03:15:16 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa09294;
10 Aug 95 4:15 +0100
Received: by gw.home.vix.com id AA22025; Wed, 9 Aug 95 11:58:29 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA22021; Wed, 9 Aug 95 11:58:28 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id LAA17044; Wed, 9 Aug 1995 11:59:06 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma017025; Wed Aug 9 11:58:49 1995
Received: by bonnie (5.x/SMI-SVR4) id AA08350; Wed, 9 Aug 1995 11:58:48 -0700
Date: Wed, 9 Aug 1995 11:58:48 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508091858.AA08350@bonnie>
To: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
X-Sun-Charset: US-ASCII

>One Wed, Aug 9, Ron Hall wrote:
>
>On Tue, 8 Aug 1995, Kevin Kunzelman wrote:
>
>>
>> We have solaris 2.4 and are running perl-4.036 and oraperl-v2.4. Aside
>> from occasional segfaults, it works great. Did you want binaries, or just the
>> configs in order to get the thing to build?
>
>
> This works with remote databases OK?
> How do you specify them?

With oraperl, remote databases is not a problem. When logging in to a DB,
you can specify a TWO_TASK (an ORACLE-ism). For example a TWO_TASK of
"t:my_host:my_db_name" will connect to a particular DB on a specified host.

>
> I'll take a copy of the configs if you don't mind.
> Do you use syslog? If so so you have the hacks? If so,
> I would appreciate them also.

We use syslog quite a bit, in fact.

>
> Thanks in Advance.
>

Since several people currently and in the past have expressed interest in
getting perl/oraperl to work on Solaris, perhaps I should prepare an area
somewhere where people can FTP both the Solaris 2.4 configs for
perl-4.036/oraperl-2.4 and for the Solaris 2.4 syslog perl usubs.

Unfortunately, my company does not have an anonymous FTP site. However, I do
have accounts on external machines that might do. Maybe in a day or so I can
make these available (I'll want to clean them up a bit first).

Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23030-19@lion>;
Thu, 10 Aug 1995 07:03:50 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 10 Aug 95 04:26:17 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26030;
10 Aug 95 5:26 +0100
Received: by gw.home.vix.com id AA09004; Wed, 9 Aug 95 16:05:35 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA09000; Wed, 9 Aug 95 16:05:30 -0700
X400-Received: by mta bnr.ca in /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed;
Wed, 9 Aug 1995 19:02:09 -0400
X400-Received: by /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed;
Wed, 9 Aug 1995 15:58:19 -0400
X400-Received: by /PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/; Relayed;
Wed, 9 Aug 1995 15:57:00 -0400
Date: Wed, 9 Aug 1995 14:57:00 -0500
X400-Originator: /dd.id=0245837/g=robert/i=rk/s=moore/@bnr.ca
X400-Mts-Identifier: [/PRMD=BNR/ADMD=TELECOM.CANADA/C=CA/;bcars735.b.595:09.07.95.19.58.19]
X400-Content-Type: P2-1984 (2)
Content-Identifier: Problems buil...
From: "robert (r.k.) moore" <[email protected]>
Sender: "robert (r.k.) moore" <[email protected]>
Message-Id: <"10617 Wed Aug 9 15:58:32 1995"@bnr.ca>
To: [email protected]
Cc: [email protected]
Subject: Problems building DBD::mSQL

I am having problems building DBD::mSQL and was wondering
if anyonemight be willing to help me out. I'm trying to
make it on a Sun 4.1.3 and have Perl 5 (with patch level m).

Here is the log for make:

cc -c -O -pic -I/usr/local/lib/perl5/sun4-sunos/CORE -I/usr/local/Minerva/include -I/usr/local/lib/perl5/sun4-sunos/DBI dbdimp.c
"./dbdimp.h", line 84: syntax error at or near variable name "signed"
"./dbdimp.h", line 94: syntax error at or near variable name "signed"
"./dbdimp.h", line 118: syntax error at or near variable name "signed"
"./dbdimp.h", line 125: zero sized structure
"dbdimp.c", line 42: syntax error at or near variable name "signed"
"dbdimp.c", line 55: cbuf undefined
"dbdimp.c", line 55: nullok undefined
"dbdimp.c", line 57: dbtype undefined
"dbdimp.c", line 57: dbsize undefined
"dbdimp.c", line 57: dsize undefined
"dbdimp.c", line 57: prec undefined
"dbdimp.c", line 57: scale undefined
"dbdimp.c", line 59: ftype undefined
"dbdimp.c", line 59: indp undefined
"dbdimp.c", line 59: bufl undefined
"dbdimp.c", line 59: rlen undefined
"dbdimp.c", line 59: rcode undefined
"dbdimp.c", line 327: undefined structure or union
"dbdimp.c", line 327: undefined structure or union
"dbdimp.c", line 336: undefined structure or union
"dbdimp.c", line 337: ftype undefined
"dbdimp.c", line 381: sv undefined
"dbdimp.c", line 382: rv undefined
"dbdimp.c", line 383: undefined structure or union
"dbdimp.c", line 419: sv undefined
"dbdimp.c", line 420: indp undefined
"dbdimp.c", line 421: ftype undefined
"dbdimp.c", line 444: signed undefined
"dbdimp.c", line 444: syntax error at or near type word "char"
"dbdimp.c", line 461: t_cbufl undefined
"dbdimp.c", line 470: field_info_loop undefined
"dbdimp.c", line 470: fatal error: too many errors
*** Error code 1
make: Fatal error: Command failed for target `dbdimp.o'

and also my perl config info:

Summary of my perl5 (patchlevel 1) configuration:
Platform:
osname=sunos, osver=4.1.3_u1, archname=sun4-sunos
uname='sunos crchs32 4.1.3_u1 1 sun4m '
hint=recommended
Compiler:
cc='cc', optimize='-O'
cppflags=''
ccflags =''
ldflags =''
stdchar='unsigned char', d_stdstdio=define, usevfork=true
voidflags=15, castflags=0, d_casti32=define, d_castneg=define
intsize=4, alignbytes=8, usemymalloc=y, randbits=31
Libraries:
so=so
libpth=/lib /usr/lib /usr/ucblib /usr/local/lib
libs=-lnsl -ldbm -ldl -lm -lc -lposix
libc=/usr/lib/libc.so.1.9.1
Dynamic Linking:
dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=
cccdlflags='-pic', ccdlflags=' ', lddlflags='-assert nodefinitions'

Is there something I'm overlooking? I made DBI and it tested out okay.
I'd appreciate any help anyonemight be willing to proffer.

Thanks!

Regards,

Rob
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Robert K. Moore
Information Resource Center
Bell Northern Research
Richardson, Texas
[email protected]
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05151-6@lion>;
Fri, 11 Aug 1995 07:01:24 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 00:20:56 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa15940;
11 Aug 95 1:20 +0100
Received: by gw.home.vix.com id AA11423; Thu, 10 Aug 95 08:57:58 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA11419; Thu, 10 Aug 95 08:57:57 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id IAA05207; Thu, 10 Aug 1995 08:58:25 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma005205; Thu Aug 10 08:58:25 1995
Received: by bonnie (5.x/SMI-SVR4) id AA05152; Thu, 10 Aug 1995 08:58:20 -0700
Date: Thu, 10 Aug 1995 08:58:20 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508101558.AA05152@bonnie>
To: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
X-Sun-Charset: US-ASCII

>On Thu, Aug 10 1995 07:28 PDT 1995 Tungning Cherng wrote
>
>
>Would you send the pointer to me when you are done? I am waiting for
>and urging to have the correct configs. I compiled perl-4.036 with
>Solaris 2.3 and I got
> ./perl installperl
> ld.so.1: ./perl: fatal: libucb.so.1: can't open file: errno=2
> *** Signal 9
> make: Fatal error: Command failed for target `install'
>
>when I did "make install".
>
>I know I must have some parameter wrong in the config file ..

I have had this problem as well. There is a libucb.so.1 file in
/usr/ucblib. What you may try doing to fix this is adding a symbolic
link:
/usr/lib/libucb.so.1 -> /usr/ucblib/libucb.so.1

Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05151-21@lion>;
Fri, 11 Aug 1995 07:03:04 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 01:44:27 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07564;
11 Aug 95 2:44 +0100
Received: by gw.home.vix.com id AA14596; Thu, 10 Aug 95 09:45:56 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA14592; Thu, 10 Aug 95 09:45:54 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa21692;
10 Aug 95 13:37 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa14877;
10 Aug 95 13:34 +0100
Received: from ig.co.uk by lion id <26918-0@lion>;
Thu, 10 Aug 1995 13:37:17 +0100
To: [email protected]
Subject: Re: DBD::Oracle make fails
Cc: [email protected]
Date: Thu, 10 Aug 1995 13:37:17 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Steve Livingston <[email protected]>
>
[ CC'd to perldb-interest regarding problems with:

- Oracle 7.1.6
- Sun 4.1.3
- Perl 5.001l
]

> At 03:15 AM 8/10/95 +0100, you wrote:
> >
> >> From: Steve Livingston <[email protected]>
> >>
> >> - My "make" problem went away when I commented out this line from Makefile:
> >>
> >> #CLIBS= $(OTHERLIBS) -L/usr/5lib -lc
> >>
> >Umm, looking at your proc.mk file it looks like someone has added that line
> >by hand.
>
> Nope...it's right off the Oracle 7.1.6 CD-ROM.
> I installed the O7, and didn't touch the proc.mk
>
Damn.

Looking at the proc.mk file that line is obviously a total hack.
Oracle *really* bugs me sometimes.

I'll have to add some explicit workaround for this. Sigh.

> Cheers,
> Steve
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05151-30@lion>;
Fri, 11 Aug 1995 07:04:05 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 02:56:42 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22503;
11 Aug 95 3:56 +0100
Received: by gw.home.vix.com id AA24308; Thu, 10 Aug 95 12:16:22 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA24302; Thu, 10 Aug 95 12:16:21 -0700
Received: from uuman.near.net by poblano.bbnplanet.com id aa20121;
10 Aug 95 15:14 EDT
To: Kevin Kunzelman <[email protected]>
Cc: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
In-Reply-To: <9508101558.AA05152@bonnie>
Date: Thu, 10 Aug 1995 15:12:44 -0400
From: Tungning Cherng <[email protected]>
Message-Id: <[email protected]>

First, thanks for the help.
Now I got:
make: Warning: Both `makefile' and `Makefile' exist
/perl installperl
cp perl /usr/planet/local/bin/perl4.036
cp taintperl /usr/planet/local/bin/tperl4.036
chmod 755 /usr/planet/local/bin/tperl4.036
ln /usr/planet/local/bin/perl4.036 /usr/planet/local/bin/perl
ln /usr/planet/local/bin/tperl4.036 /usr/planet/local/bin/taintperl
cp x2p/a2p /usr/planet/local/bin/a2p
chmod 755 /usr/planet/local/bin/a2p
cp cppstdin /usr/planet/local/bin
chmod 755 /usr/planet/local/bin/cppstdin
cp h2ph /usr/planet/local/bin
chmod 755 /usr/planet/local/bin/h2ph
cp c2ph /usr/planet/local/bin
chmod 755 /usr/planet/local/bin/c2ph
cp pstruct /usr/planet/local/bin
chmod 755 /usr/planet/local/bin/pstruct
cp x2p/s2p /usr/planet/local/bin
chmod 755 /usr/planet/local/bin/s2p
cp x2p/find2perl /usr/planet/local/bin
chmod 755 /usr/planet/local/bin/find2perl
Installing /usr/planet/local/man/manl/perl.l
Installing /usr/planet/local/man/manl/h2ph.l
Installing /usr/planet/local/man/manl/a2p.l
Installing /usr/planet/local/man/manl/s2p.l
*** Signal 11 - core dumped
make: Fatal error: Command failed for target `install'

Thanks,
Tungning

On Thu, 10 Aug 1995 08:58:20 -0700 Kevin Kunzelman wrote:

>>On Thu, Aug 10 1995 07:28 PDT 1995 Tungning Cherng wrote
>>
>>
>>Would you send the pointer to me when you are done? I am waiting for
>>and urging to have the correct configs. I compiled perl-4.036 with
>>Solaris 2.3 and I got
>> ./perl installperl
>> ld.so.1: ./perl: fatal: libucb.so.1: can't open file: errno=2
>> *** Signal 9
>> make: Fatal error: Command failed for target `install'
>>
>>when I did "make install".
>>
>>I know I must have some parameter wrong in the config file ..
>
>I have had this problem as well. There is a libucb.so.1 file in
>/usr/ucblib. What you may try doing to fix this is adding a symbolic
>link:
> /usr/lib/libucb.so.1 -> /usr/ucblib/libucb.so.1
>
>Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05736-1@lion>;
Fri, 11 Aug 1995 07:04:24 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 05:13:57 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa19676;
11 Aug 95 6:13 +0100
Received: by gw.home.vix.com id AA17325; Thu, 10 Aug 95 18:24:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA17321; Thu, 10 Aug 95 18:24:43 -0700
Received: from billberry (billberry.ai.mit.edu)
by life.ai.mit.edu (4.1/AI-4.10) for [email protected]
id AA20053; Thu, 10 Aug 95 21:24:38 EDT
From: Sundar Narasimhan <[email protected]>
Received: by billberry (4.1/AI-4.10) id AA12439; Thu, 10 Aug 95 21:24:36 EDT
Date: Thu, 10 Aug 95 21:24:36 EDT
Message-Id: <9508110124.AA12439@billberry>
To: [email protected]
Cc: [email protected], [email protected]
In-Reply-To: Kevin Kunzelman's message of Thu, 10 Aug 1995 08:58:20 -0700 <9508101558.AA05152@bonnie>
Subject: question re. perl-4.036, oraperl and Solaris 2.4

X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Date: Thu, 10 Aug 1995 08:58:20 -0700
From: [email protected] (Kevin Kunzelman)
X-Sun-Charset: US-ASCII

>On Thu, Aug 10 1995 07:28 PDT 1995 Tungning Cherng wrote
>
>
>Would you send the pointer to me when you are done? I am waiting for
>and urging to have the correct configs. I compiled perl-4.036 with
>Solaris 2.3 and I got
> ./perl installperl
> ld.so.1: ./perl: fatal: libucb.so.1: can't open file: errno=2
> *** Signal 9
> make: Fatal error: Command failed for target `install'
>
>when I did "make install".
>
>I know I must have some parameter wrong in the config file ..

I have had this problem as well. There is a libucb.so.1 file in
/usr/ucblib. What you may try doing to fix this is adding a symbolic
link:
/usr/lib/libucb.so.1 -> /usr/ucblib/libucb.so.1

Kevin

Please don't do this. When Configure asks you what extra flags to
specify to ld add "-L/usr/ucblib", otherwise you risk annoying
other users on your system who may be quite surprised because of the
link.

When I compiled perl and oraperl on Solaris 2.3 and 2.4 - I found that
with the Lucid compiler I had to manually edit config.sh to make sure
that volatile was not defined. Specifying -DDEBUGGING was also helpful.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <07499-1@lion>;
Fri, 11 Aug 1995 11:34:28 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 09:31:13 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07688;
11 Aug 95 10:31 +0100
Received: by gw.home.vix.com id AA25941; Thu, 10 Aug 95 21:03:42 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA25937; Thu, 10 Aug 95 21:03:41 -0700
Received: from hometown.com by metronet.com with SMTP
id AA03436 (5.67a/IDA1.5hp for <[email protected]>);
Thu, 10 Aug 1995 23:02:21 -0500
Message-Id: <[email protected]>
X-Sender: [email protected]
X-Mailer: Windows Eudora Version 2.1
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Thu, 10 Aug 1995 23:00:33 -0500
To: [email protected]
From: Steve Livingston <[email protected]>
Subject: Usage documentation for DBD::Oracle?

- Is there any documentation on the usage of DBD::Oracle?
- Other than test.pl?

- Particularly the oraperl emulation functions
- Does oraperl have documentation somewhere?

Cheers,
Steve

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <07499-2@lion>;
Fri, 11 Aug 1995 11:34:35 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 09:31:55 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07950;
11 Aug 95 10:31 +0100
Received: by gw.home.vix.com id AA27072; Thu, 10 Aug 95 21:26:06 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA27068; Thu, 10 Aug 95 21:26:05 -0700
Received: from gonzo.tamu.edu (pckizer@localhost [127.0.0.1])
by gonzo.tamu.edu (8.6.12/8.6.12) with ESMTP id XAA18765;
Thu, 10 Aug 1995 23:25:45 -0500
From: Philip Kizer <[email protected]>
To: Kevin Kunzelman <[email protected]>
Cc: [email protected], [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Reply-To: [email protected]
X-Office: 326E Teague
X-Office-Phone: 409.862.4120
X-Organization: TAMU CIS Operating Systems Group, Unix
X-Pgp-Key: Get key 0x11989061 by sending message with subject: "get pgp key"
X-Url: http://gonzo.tamu.edu/
X-Face: _B<mzj1U/]zP}h)]5#BSta@y22b`"<mJtYKCPN{4BF^@S<[rOZL*pyXkh3jeqjeJHJrb%ld
f'dU)hBy9
In-Reply-To: Your message of "Thu, 10 Aug 1995 08:58:20 PDT." <9508101558.AA05152@bonnie>
Mime-Version: 1.0
Content-Type: multipart/mixed; boundary="----- =_aaaaaaaaaa0"
Content-Id: <[email protected]>
Date: Thu, 10 Aug 1995 23:25:44 -0500
Message-Id: <[email protected]>
Sender: [email protected]

------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <[email protected]>

Continuing inclusion of perldb-interest because or the oraperl link, since
it is still the current "production" version...

[email protected] (Kevin Kunzelman) wrote:
>>On Thu, Aug 10 1995 07:28 PDT 1995 Tungning Cherng wrote
>> ./perl installperl
>> ld.so.1: ./perl: fatal: libucb.so.1: can't open file: errno=2
>> *** Signal 9

Aack...this means perl was compiled wrong...here's the directions I always
forward to people trying to build perl4.036...you need to remove or put at
the end all possible references to /usr/ucb* from your PATH and definitely
_remove_ all reference to /usr/ucb* from LD_LIBRARY_PATH. LD_RUN_PATH is
OK to have set, but don't include /usr/ucb* in there either!

I really only give these out as general guidelines...definitely follow your
own local preferences, but I prefer to build with the SparcCompiler, use
-fast for optimizations, and stick to the standard /usr/local structure...:

------- =_aaaaaaaaaa0
MIME-Version: 1.0
Content-Type: message/rfc822

From: [email protected] (Michael D'Errico)
Newsgroups: comp.lang.perl,comp.unix.solaris
Subject: Instructions for compiling Perl (Re: Solaris 2.2 and Perl)
Date: 30 Jun 1993 23:12:06 -0700
Organization: Software Now, Santa Barbara, CA
Message-ID: <[email protected]>

[email protected] (Pete Hartman) writes:

>I have read a few different recommendations about how to build perl
>under Sun's Solaris 2.X operating systems.

>Has *anyone* actually compiled a working version of perl4.036 on a Solaris
>machine *without* using the UCB "compatability mode" stuff? If so, what
>was your secret?

I have built Perl on a Sparc Classic running Solaris 2.1 (after several
tries and reading a bunch of posts :). It needs none of the 'ucb' stuff
and I even figured out how to make signals reliable under Solaris. The
complete instructions are given below (I used gcc 2.3.3).

Michael D'Errico
[email protected]

=======================================================================
[ftp.uu.net:/systems/gnu/perl-4.036.tar.gz]

Configure (only add things in []'s if you have GNU gdbm)

{Return} (cursor position)
{Return}
{Return}
{Return} (defaults for Solaris_2_1 is wrong!)
/opt/perl/man/manl
{Return} (/usr/include)
{Return} ('none' for memory models)
gcc
n
-O2
-Dvolatile=__volatile__ [-DHAS_GDBM]
{Return} ('none' for ld flags)
-lsocket -lnsl [-lgdbm] -lm
{Return} (for 8-byte boundaries)
/opt/perl/bin
{Return}
{Return} (4321)
{Return} (libc.so)
y (setuid/setgid emulation, if you want)
y (Must use Perl's malloc routines!)
{Return} (gid_t)
{Return} (int)
{Return} (4)
/opt/perl/lib
{Return}
{Return} (15)
/opt/perl/scripts
{Return}
{Return} (uid_t)
{Return} (7)
{Return} (yacc)

sh (for shell escape from Configure)
vi config.sh
remove '-D__GNUC__' from cppflags

make depend (Configure will do this if you tell it to)

[for correct implementation of reliable signals]
vi doarg.c eval.c stab.c util.c
add the following line under '#include <signal.h>'
#define signal(x,y) sigset((x),(y))

make
make install \__ kind of a catch-22 -- you need to install
make test / to pass all tests, but it will complain
that you haven't tested it yet.

NOTE: if you are using GNU gdbm, the op/dbm test will fail.
---- To fix the test, change <Op.dbmx.*> to 'Op.dbmx' and
remove the '.pag' extensions from the filenames in
t/op/dbm.t and change the last line of the test to
"unlink 'Op.dbmx';". All of the tests will now pass.

------- =_aaaaaaaaaa0
Content-Type: text/plain; charset="us-ascii"
Content-ID: <[email protected]>

In my opinion, if LD_LIBRARY_PATH is set at all for typical internet
servers, the binaries requiring said LD_..._PATH have been compiled
incorrectly. I have yet to find any public program that took more than
minimal effort to make portable. Though I do understand (albeit with
grumblings) that it is required for programs like wordperfect where they
cannot be sure where the customer has installed their preferred X and Motif
libraries.

Good luck,
philip

____________________________________________________________ Philip Kizer ___
Texas A&M CIS Operating Systems Group, Unix ( 409.862.4120 ) [email protected]

"Relying on the government to protect your privacy is like asking a peeping
tom to install your window blinds." -John Perry Barlow, EFF co-founder

------- =_aaaaaaaaaa0--
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <07499-6@lion>;
Fri, 11 Aug 1995 11:34:56 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 10:06:13 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa19831;
11 Aug 95 11:06 +0100
Received: by gw.home.vix.com id AA26939; Thu, 10 Aug 95 21:20:59 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA26935; Thu, 10 Aug 95 21:20:58 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id VAA20450; Thu, 10 Aug 1995 21:21:35 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma020433; Thu Aug 10 21:21:33 1995
Received: by bonnie (5.x/SMI-SVR4) id AA12123; Thu, 10 Aug 1995 21:21:30 -0700
Date: Thu, 10 Aug 1995 21:21:30 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508110421.AA12123@bonnie>
To: [email protected], [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
X-Sun-Charset: US-ASCII

>From [email protected] Thu Aug 10 18:25 PDT 1995
>To: [email protected]
>
> X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
> Date: Thu, 10 Aug 1995 08:58:20 -0700
> From: [email protected] (Kevin Kunzelman)
> X-Sun-Charset: US-ASCII
>
> >On Thu, Aug 10 1995 07:28 PDT 1995 Tungning Cherng wrote
> >
> >
> >Would you send the pointer to me when you are done? I am waiting for
> >and urging to have the correct configs. I compiled perl-4.036 with
> >Solaris 2.3 and I got
> > ./perl installperl
> > ld.so.1: ./perl: fatal: libucb.so.1: can't open file: errno=2
> > *** Signal 9
> > make: Fatal error: Command failed for target `install'
> >
> >when I did "make install".
> >
> >I know I must have some parameter wrong in the config file ..
>
> I have had this problem as well. There is a libucb.so.1 file in
> /usr/ucblib. What you may try doing to fix this is adding a symbolic
> link:
> /usr/lib/libucb.so.1 -> /usr/ucblib/libucb.so.1
>
> Kevin
>
>Please don't do this. When Configure asks you what extra flags to
>specify to ld add "-L/usr/ucblib", otherwise you risk annoying
>other users on your system who may be quite surprised because of the
>link.
>
>When I compiled perl and oraperl on Solaris 2.3 and 2.4 - I found that
>with the Lucid compiler I had to manually edit config.sh to make sure
>that volatile was not defined. Specifying -DDEBUGGING was also helpful.

This problem happens not with a static library, but a dynamic one
(/usr/ucblib/libucb.so.1). As far as I know, ldflags can affect static
linking, but I don't believe that they can affect dynamic linking (at least
from what I can discern from the ld man page). Therefore, you would have to
statically link libucb (which isn't such a big deal because it's only like 96K).

It seems like a good suggestion to hack your config.sh file so that

ldflags='-L/usr/ucblib/'

I'll probably recompile our stuff here to do that. Then I won't have to futz
with dynamic loading of libucb.

Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-19@lion>;
Mon, 14 Aug 1995 07:02:47 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 20:15:49 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa27471;
11 Aug 95 21:15 +0100
Received: by gw.home.vix.com id AA26402; Fri, 11 Aug 95 05:08:38 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA26397; Fri, 11 Aug 95 05:08:35 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id NAA06469;
Fri, 11 Aug 1995 13:05:20 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id NAA23225; Fri, 11 Aug 1995 13:10:33 +0100
Message-Id: <[email protected]>
Subject: Re: Problems building DBD::mSQL (fwd)
To: [email protected]
Date: Fri, 11 Aug 1995 13:10:33 +0100 (BST)
Cc: [email protected]
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 1223

> > Thanks for your response. I tried removing the comment
> > and then making it and got:
> >
> > LD_RUN_PATH="/usr/local/Minerva/lib" ld -o blib/auto/DBD/mSQL/mSQL.so -assert nodefinitions dbdimp.o mSQL.o -L/usr/local/Minerva/lib-L/usr/local/Minerva/lib -R/usr/local/Minerva/lib -L/usr/local/Minerva/lib -lmsql -lm
> > ld: -lmsql: No such file or directory
> > *** Error code 4
>
> > Any further suggestions? Seems like I'm having a bad day compiling or
> > something. I'd appreciate any pointers you might have. Thanks again
> > for the help!
> >
> Since I'm not the author and I don't use mSQL I can't help further.
> I've CC'd this to perldb-interest.

Hi. Sorry about the delay on replies. I've been out of circulation for a
day or two on this.

There's a couple of Makefile.PL bugs I've patched up which may get rid of
a couple of glitchy things.

ftp://ftp.mcqueen.com/pub/languages/perl5/DBPerl/

Similarly, can anyone that's downloaded the module please email on what
their experience with it is? I'm having trouble working out whether
or not it works on anything, some things, or nothing except my machine!

Thanks!

cc:d to perldb-interest.

> Tim.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-25@lion>;
Mon, 14 Aug 1995 07:03:40 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 20:41:11 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07382;
11 Aug 95 21:41 +0100
Received: by gw.home.vix.com id AA10170; Fri, 11 Aug 95 01:13:36 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA10166; Fri, 11 Aug 95 01:13:34 -0700
Received: from hometown.com by metronet.com with SMTP
id AA23137 (5.67a/IDA1.5hp for <[email protected]>);
Fri, 11 Aug 1995 03:12:15 -0500
Message-Id: <[email protected]>
X-Sender: [email protected]
X-Mailer: Windows Eudora Version 2.1
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Fri, 11 Aug 1995 03:10:26 -0500
To: [email protected]
From: Steve Livingston <[email protected]>
Subject: Re: Usage documentation for DBD::Oracle?

- I have downloaded oraperl, just to get the man pages

- Thanks for your many contributions

Cheers,
Steve

At 09:42 AM 8/11/95 +0200, you wrote:
>> - Particularly the oraperl emulation functions
>> - Does oraperl have documentation somewhere?
>
>Would you believe, "man oraperl"? Twelve pages on my system.
>
> Kevin
>
>

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-34@lion>;
Mon, 14 Aug 1995 07:04:47 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 22:22:53 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa15329;
11 Aug 95 23:22 +0100
Received: by gw.home.vix.com id AA23980; Fri, 11 Aug 95 04:11:14 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA23976; Fri, 11 Aug 95 04:11:12 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa20904;
11 Aug 95 11:33 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac17547;
11 Aug 95 11:31 +0100
Received: from ig.co.uk by lion id <06944-0@lion>;
Fri, 11 Aug 1995 10:38:35 +0100
To: [email protected]
Subject: Re: Problems building DBD::mSQL
Cc: [email protected]
Date: Fri, 11 Aug 1995 10:38:35 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: "robert (r.k.) moore" <[email protected]>
>
> Tim,
>
> Thanks for your response. I tried removing the comment
> and then making it and got:
>
> LD_RUN_PATH="/usr/local/Minerva/lib" ld -o blib/auto/DBD/mSQL/mSQL.so -assert nodefinitions dbdimp.o mSQL.o -L/usr/local/Minerva/lib-L/usr/local/Minerva/lib -R/usr/local/Minerva/lib -L/usr/local/Minerva/lib -lmsql -lm
> ld: -lmsql: No such file or directory
> *** Error code 4

> Any further suggestions? Seems like I'm having a bad day compiling or
> something. I'd appreciate any pointers you might have. Thanks again
> for the help!
>
Since I'm not the author and I don't use mSQL I can't help further.
I've CC'd this to perldb-interest.

> Rob

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-38@lion>;
Mon, 14 Aug 1995 07:05:13 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 11 Aug 95 23:59:29 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa20005;
12 Aug 95 0:59 +0100
Received: by gw.home.vix.com id AA02358; Fri, 11 Aug 95 06:35:50 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA02354; Fri, 11 Aug 95 06:35:47 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa20891;
11 Aug 95 11:34 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa17547;
11 Aug 95 11:31 +0100
Received: from ig.co.uk by lion id <06910-0@lion>;
Fri, 11 Aug 1995 10:34:44 +0100
To: [email protected], [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected], [email protected]
Date: Fri, 11 Aug 1995 10:34:44 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Sundar Narasimhan <[email protected]>
>
> From: [email protected] (Kevin Kunzelman)
>
> >On Thu, Aug 10 1995 07:28 PDT 1995 Tungning Cherng wrote
> >
> >Would you send the pointer to me when you are done? I am waiting for
> >and urging to have the correct configs. I compiled perl-4.036 with
> >Solaris 2.3 and I got
> > ./perl installperl
> > ld.so.1: ./perl: fatal: libucb.so.1: can't open file: errno=2
> > *** Signal 9
> > make: Fatal error: Command failed for target `install'
> >
> >when I did "make install".
> >
> >I know I must have some parameter wrong in the config file ..
>
> I have had this problem as well. There is a libucb.so.1 file in
> /usr/ucblib. What you may try doing to fix this is adding a symbolic
> link:
> /usr/lib/libucb.so.1 -> /usr/ucblib/libucb.so.1
>
> Kevin
>
> Please don't do this. When Configure asks you what extra flags to
> specify to ld add "-L/usr/ucblib", otherwise you risk annoying
> other users on your system who may be quite surprised because of the
> link.
>
Please don't even do this. To build a reliable perl on Solaris 2 you
must *not* use *any* udb libraries (they are busted in various ways).

For Perl 4 the FAQ says:

1.17) How do I get Perl to compile on Solaris?

The following directions are for perl, version 4. Perl, version 5,
should compile more easily. If not, send mail to The Perl Porters
Mailing List ([email protected])

John Lees* reports:

I have built perl on Solaris 2.1, 2.2 beta, and 2.2 FCS. Take
/usr/ucb out of your path and do not use any BSD/UCB libraries.
Only -lsocket, -lnsl, and -lm are needed. You can use the hint for
Solaris 2.0, but the one for 2.1 is wrong. Do not use vfork. Do not
use -I/usr/ucbinclude. The result works fine for me, but of couse
does not support a couple of BSDism's.

Casper H.S. Dik* reports

You must remove all the references to /usr/ucblib AND
/usr/ucbinclude. And ignore the Solaris_2.1 hints. They are wrong.
The undefining of vfork() probably has to do with the confusion it
gives to the compilers. If you use cc, you mustn't compile
util.c/tutil.c with -O. I only used the following libs: -lsocket
-lnsl -lm (there is a problem with -lmalloc)

Michael D'Errico* reports:

If you are using Solaris 2.x, the signal handling is broken. If
you set up a signal handler such as 'ripper' it will be forgotten
after the first time the signal is caught. To fix this, you need
to recompile Perl. Just add '#define signal(x,y) sigset((x),(y))'
after the '#include <signal.h>' directive in each file that it
occurs, then make it again.

For Perl 5 there is a similar set of instructions but I don't have a copy.
Ask [email protected] if you're still stuck.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-39@lion>;
Mon, 14 Aug 1995 07:05:21 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 00:06:58 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22527;
12 Aug 95 1:06 +0100
Received: by gw.home.vix.com id AA02374; Fri, 11 Aug 95 06:36:08 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA02369; Fri, 11 Aug 95 06:36:03 -0700
Received: by cpuv1.net.kiae.su id AA00122 (5.65.kiae-2 );
Fri, 11 Aug 1995 17:21:57 +0400
Message-Id: <[email protected]>
Subject: Re: Problems building DBD::mSQL (fwd)
To: [email protected]
Date: Fri, 11 Aug 95 17:21:57 GMT+3:00
Cc: [email protected], [email protected]
In-Reply-To: <[email protected]>; from "[email protected]" at Aug 11, 95 1:10 pm
From: "Andrew N. Chernoff" <[email protected]>
Reply-To: "Andrew N. Chernoff" <[email protected]>
X-Class: Fast
X-Priority: fast
Precedence: Special-Delivery
X-Mailer: ELM [version 2.3 PL11]

hi,
help me!
how can I unsubscribe from this list?

I've try send:

unsubscribe

to [email protected]

but I'ven't result :(

With respect, Andrew N. Chernoff
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-43@lion>;
Mon, 14 Aug 1995 07:05:51 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 00:25:30 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa28431;
12 Aug 95 1:25 +0100
Received: by gw.home.vix.com id AA02338; Fri, 11 Aug 95 06:34:10 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA02334; Fri, 11 Aug 95 06:34:08 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa20925;
11 Aug 95 11:34 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ae17547;
11 Aug 95 11:31 +0100
Received: from ig.co.uk by lion id <07286-0@lion>;
Fri, 11 Aug 1995 11:15:22 +0100
To: [email protected]
Subject: Re: Oracle long
Cc: [email protected]
Date: Fri, 11 Aug 1995 11:15:22 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Jonathan Olson <[email protected]>
>
> I still cannot say that I am expert in the matters of Oraperl, but I have
> learned some things. I would like you to grant some insights on how to
> tackle the problem we have encountered.
>
> We are selecting something from the Oracle database that looks like this :
>
> $select="select varchar2_datatype, long_datatype from table";
> $cursor=&ora_open($lda, $select) || die .....;
> @fetch = &ora_fetch($cursor);
>
> Now, as far as I can tell, it is being fetched (but truncated as seen by
> the /logs/error file) :
>
> ERROR EVENT: DBI::st=HASH(0x1401c0a48)->DBI::default_handler:
> 1406 ORA-01406: fetched column value was truncated (DBD: ofetch
> rcode)
>
Set $ora_trunc = 1 before the ora_open() to silence that error.

> Furthermore, when you do :
>
> while (@fetch = &ora_fetch($cursor)) { do stuff }
>
> it does strange things when it tumbles onto a variable of the long datatype.
>
> Now you mentioned using the $sth->fetchrow ... $sth->readblob(...) approach.
> Could you elaborate on exactly how to implement it in this scenario?
>
I posted posted this to comp.lang.perl.misc recently:

---

Newsgroups: comp.lang.perl.misc
Subject: Re: Perl and Oracle long
Message-ID: <[email protected]>
References: <[email protected]>
Date: Tue, 8 Aug 1995 19:59:14 GMT

In article <[email protected]>,
Royal Hansen <[email protected]> wrote:
> Can perl handle a string like the Oracle type long - which handles up to
> two gigabytes of text data. If it can, how do you do it? If it can't,
> is there a way around it?

Basically perl is limited only by the malloc it uses and the memory you
have available.

It's generally a good idea to work in chunks and stream the data through
whatever processes you require as you read it.

Here's a totally untested piece of code off the top of my head for
DBD::Oracle. This uses the new and untested readblob method.

my $long_field = 1; # which field in the query is of type long.
my $length = 32*1024;

while(@row = $sth->fetch) {
my $offset = 0;
do {
my $chunk = $sth->readblob($long_field, $offset, $length);
last unless length $chunk;
$offset += $length;
print OUTPUT $chunk;
} while 1;
}

Note that readblob is experimental. Also note the comment in the source:
/* Sadly, even though retl is a ub4, oracle will cap the */
/* value of retl at 65535 even if more was returned! */
/* This is according to the OCI manual for Oracle 7.0. */
/* Once again Oracle causes us grief. How can we tell what */
/* length to assign to destsv? We do have a compromise: if */
/* retl is exactly 65535 we assume that all data was read. */
Sigh.

---

> Also, do you have any other insights as to how DBD handles Oracle longs?
>
Take a look in the code.

> As always, thanks.
>
> Jonathan.
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-44@lion>;
Mon, 14 Aug 1995 07:06:00 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 00:33:25 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00619;
12 Aug 95 1:33 +0100
Received: by gw.home.vix.com id AA25882; Fri, 11 Aug 95 04:56:56 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA25876; Fri, 11 Aug 95 04:56:55 -0700
Received: from billberry (billberry.ai.mit.edu)
by life.ai.mit.edu (4.1/AI-4.10) for [email protected]
id AA08468; Fri, 11 Aug 95 07:56:17 EDT
From: Sundar Narasimhan <[email protected]>
Received: by billberry (4.1/AI-4.10) id AA12528; Fri, 11 Aug 95 07:56:16 EDT
Date: Fri, 11 Aug 95 07:56:16 EDT
Message-Id: <9508111156.AA12528@billberry>
To: [email protected]
Cc: [email protected], [email protected]
In-Reply-To: Kevin Kunzelman's message of Thu, 10 Aug 1995 21:21:30 -0700 <9508110421.AA12123@bonnie>
Subject: question re. perl-4.036, oraperl and Solaris 2.4

>From [email protected] Thu Aug 10 18:25 PDT 1995
>To: [email protected]
>Please don't do this. When Configure asks you what extra flags to
>specify to ld add "-L/usr/ucblib", otherwise you risk annoying
>other users on your system who may be quite surprised because of the
>link.
>
>When I compiled perl and oraperl on Solaris 2.3 and 2.4 - I found that
>with the Lucid compiler I had to manually edit config.sh to make sure
>that volatile was not defined. Specifying -DDEBUGGING was also helpful.

This problem happens not with a static library, but a dynamic one
(/usr/ucblib/libucb.so.1). As far as I know, ldflags can affect static
linking, but I don't believe that they can affect dynamic linking (at least
from what I can discern from the ld man page). Therefore, you would have to
statically link libucb (which isn't such a big deal because it's only like 96K).

It seems like a good suggestion to hack your config.sh file so that

ldflags='-L/usr/ucblib/'

I'll probably recompile our stuff here to do that. Then I won't have to futz
with dynamic loading of libucb.
Dynamic loading of libraries is affected by your environment variable
LD_LIBRARY_PATH. I am not on my Solaris machine right now, but if I
remember right, the only library needed out of ucblib is -ldbm. If you
add -lucb to your linkstream (and yes, Configure does discover this),
then be prepared to deal with a whole host of other problems.

Another command that you may want to know about is "ldd".
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <19515-13@lion>;
Mon, 14 Aug 1995 07:06:18 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 00:53:34 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06218;
12 Aug 95 1:53 +0100
Received: by gw.home.vix.com id AA19774; Fri, 11 Aug 95 11:13:47 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA19765; Fri, 11 Aug 95 11:13:43 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id ac19710;
11 Aug 95 17:11 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ae12234;
11 Aug 95 17:08 +0100
Received: from ig.co.uk by lion id <10049-0@lion>;
Fri, 11 Aug 1995 16:29:03 +0100
To: [email protected], [email protected]
Subject: Re: Usage documentation for DBD::Oracle?
Date: Fri, 11 Aug 1995 16:29:03 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Steve Livingston <[email protected]>
>
> - Is there any documentation on the usage of DBD::Oracle?
> - Other than test.pl?
>
Only the oraperl documentation which is not supplied with DBD::Oracle.

> - Particularly the oraperl emulation functions
> - Does oraperl have documentation somewhere?
>
ftp.demon.co.uk:/pub/perl/db/perl4/oraperl/

I'll ship the oraperl manual with DBD::Oracle in future.

> Cheers,
> Steve
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-47@lion>;
Mon, 14 Aug 1995 07:06:26 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 01:30:56 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa13971;
12 Aug 95 2:30 +0100
Received: by gw.home.vix.com id AA25756; Fri, 11 Aug 95 12:33:34 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA25752; Fri, 11 Aug 95 12:33:29 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa26102;
11 Aug 95 18:51 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa28550;
11 Aug 95 18:49 +0100
Received: from ig.co.uk by lion id <11969-0@lion>;
Fri, 11 Aug 1995 18:50:45 +0100
To: [email protected]
Subject: ... also
Date: Fri, 11 Aug 1995 18:50:45 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

also, please tell me if you're doing a DBD and what stage your at
even if you think I already know!

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <19515-18@lion>;
Mon, 14 Aug 1995 07:07:01 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 03:29:55 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa04266;
12 Aug 95 4:29 +0100
Received: by gw.home.vix.com id AA10517; Fri, 11 Aug 95 16:15:53 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA10512; Fri, 11 Aug 95 16:15:49 -0700
Received: by disperse.demon.co.uk id aa12926; 11 Aug 95 23:32 +0100
Received: from post.demon.co.uk by disperse.demon.co.uk id aa25011;
11 Aug 95 18:32 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa25521;
11 Aug 95 18:29 +0100
Received: from ig.co.uk by lion id <11719-0@lion>;
Fri, 11 Aug 1995 18:31:55 +0100
To: [email protected]
Subject: New release next week
Date: Fri, 11 Aug 1995 18:31:55 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

Just a quick note to follow up on my last big message.
I've found some time today and I've implemented the initial changes.

After a spot of polishing early next week I should be able to make
a release. That release will be for DBD developers only.

I will expect them to catch-up with the DBI changes quickly since I
plan to make further DBI interface changes and want to release them in
bite-size chunks.

Can everyone actually working on a DBD module please tell me so
I can track what's happening out there and keep you in sync.

Thanks.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <19515-22@lion>;
Mon, 14 Aug 1995 07:07:25 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 06:44:57 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa02164;
12 Aug 95 7:44 +0100
Received: by gw.home.vix.com id AA14021; Fri, 11 Aug 95 17:12:22 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA14017; Fri, 11 Aug 95 17:12:21 -0700
Received: from localhost (mellon@localhost) by toccata.fugue.com (8.6.9/8.6.11)
with SMTP id RAA12151; Fri, 11 Aug 1995 17:08:18 -0700
Message-Id: <[email protected]>
To: "Andrew N. Chernoff" <[email protected]>
Cc: [email protected], [email protected], [email protected]
Subject: Re: Problems building DBD::mSQL (fwd)
In-Reply-To: Your message of "Fri, 11 Aug 1995 17:21:57 GMT." <[email protected]>
Date: Fri, 11 Aug 1995 17:08:16 -0700
From: Ted Lemon <[email protected]>

You're already unsubscribed from the list. I've received both of your
unsubscribe requests, and I removed you from the list upon receiving
the first one. The reason you're still getting mail is that your
mailer is hard to reach, so some messages were queued up for you
before I processed your unsubscribe request, and delivered to you
afterwards.

For future reference to other readers on this list, to unsubscribe,
send mail to [email protected]. Please have a little
patience - we do not run listserv or majordomo, so your request must
be processed by hand.

_MelloN_
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18984-64@lion>;
Mon, 14 Aug 1995 07:08:44 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 12 Aug 95 19:46:51 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa10532;
12 Aug 95 20:46 +0100
Received: by gw.home.vix.com id AA08008; Sat, 12 Aug 95 08:50:04 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA07997; Sat, 12 Aug 95 08:50:02 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id QAA01234;
Sat, 12 Aug 1995 16:46:38 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id QAA09787; Sat, 12 Aug 1995 16:51:53 +0100
Message-Id: <[email protected]>
Subject: Re: ... also
To: Tim Bunce <[email protected]>
Date: Sat, 12 Aug 1995 16:51:53 +0100 (BST)
Cc: [email protected]
In-Reply-To: <[email protected]> from "Tim Bunce" at Aug 11, 95 06:50:45 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 276

> also, please tell me if you're doing a DBD and what stage your at
> even if you think I already know!

DBD::mSQL -> pre-Alpha released to pl2. Waiting on DBI release....

DBD::Informix -> Getting the code tohether

> Tim.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20478-7@lion>;
Mon, 14 Aug 1995 07:10:43 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 14 Aug 95 03:28:20 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa05793;
14 Aug 95 4:28 +0100
Received: by gw.home.vix.com id AA08025; Sun, 13 Aug 95 10:40:42 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA08021; Sun, 13 Aug 95 10:40:41 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id SAA02767;
Sun, 13 Aug 1995 18:37:14 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id SAA14156; Sun, 13 Aug 1995 18:42:32 +0100
Message-Id: <[email protected]>
Subject: DBD::mSQL-0.59pl2 out.
To: Andreas Koenig <[email protected]>,
Tim Bunce <[email protected]>, [email protected]
Date: Sun, 13 Aug 1995 18:42:32 +0100 (BST)
Cc: [email protected]
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 440

Hopefully this may fix some more bugs here and there.

There's been not much activity on this since I've been off with a throat
infection for the last week, but hopefully I'll get back into it later this
week.

Again, if you could let me know what happens on comilation, I'd appreciate
it. There's been very little feedback, all-in-all.

ftp://ftp.mcqueen.com/pub/languages/perl5/DBPerl

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <19515-43@lion>;
Mon, 14 Aug 1995 07:11:20 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 14 Aug 95 06:00:45 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa27910;
14 Aug 95 7:00 +0100
Received: by gw.home.vix.com id AA21959; Sun, 13 Aug 95 20:51:18 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA21954; Sun, 13 Aug 95 20:51:16 -0700
Received: from ms.mimos.my by mimos.my with SMTP id AA13493 (5.67a/IDA-1.5
for <[email protected]>); Mon, 14 Aug 1995 11:51:12 +0800
Received: by ms.mimos.my (5.64/7.0) id AA19349; Mon, 14 Aug 95 11:51:12 +0800
Date: Mon, 14 Aug 1995 11:51:11 +0800
From: Rosehaslina Razali <[email protected]>
To: [email protected]
Subject: I'm just started....
Message-Id: <[email protected]>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

I'm just started to use DBperl to interface with Oracle.
The problem is I don't know anything about DBperl and where to
get the sources for it. I have read the DBperl API specification
but I can't get any informations about DBperl except the available
functions, variablesand how to use it in the program.
Can you provide or tell me which site I can go to get
the basic information about DBperl and also where can I get the
sources for it. I went through the demon.co.uk site but can't
understand the README and the files available and the connection
to DBperl.
Please help me as soon as possible since I need to use
it as soon as possible.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23425-36@lion>;
Mon, 14 Aug 1995 14:31:13 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 14 Aug 95 10:23:44 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa15964;
14 Aug 95 11:23 +0100
Received: by gw.home.vix.com id AA01146; Sun, 13 Aug 95 22:39:09 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA01141; Sun, 13 Aug 95 22:39:05 -0700
Received: by anna.mind.de (940816.SGI.8.6.9/940406.SGI) id HAA00970;
Mon, 14 Aug 1995 07:36:49 +0200
Date: Mon, 14 Aug 1995 07:36:49 +0200
Message-Id: <[email protected]>
From: Andreas Koenig <[email protected]>
To: [email protected]
Cc: [email protected]
In-Reply-To: <[email protected]> ([email protected])
Subject: Re: DBD::mSQL-0.59pl2 out.
Reply-To: [email protected]

>From: [email protected]
>
>Again, if you could let me know what happens on comilation, I'd appreciate
>it. There's been very little feedback, all-in-all.

There seem to be many busy people out there :)

The mileage seems to be improved. Here's what I get from make:

cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP22-irix/CORE -I/usr/local/Minerva/include -I/usr/local/lib/perl5/IP22-irix/DBI dbdimp.c
/usr/local/bin/perl -I/usr/local/lib/perl5/IP22-irix -I/usr/local/lib/perl5 /usr/local/lib/perl5/ExtUtils/xsubpp -typemap /usr/local/lib/perl5/ExtUtils/typemap mSQL.xs >mSQL.tc && mv mSQL.tc mSQL.c
cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP22-irix/CORE -I/usr/local/Minerva/include -I/usr/local/lib/perl5/IP22-irix/DBI mSQL.c
cfe: Error: mSQL.c, line 475: Incompatible pointer type assignment (3.3.16.1, 3.5.4.1(21))
sv = sv_2mortal( newSVpv( fbh->cbuf, fbh->rlen ) );
-------------------------------------^
cfe: Error: mSQL.c, line 475: Type char * of this argument is incompatible with type unsigned char * of function prototype description (3.3.16.1)
sv = sv_2mortal( newSVpv( fbh->cbuf, fbh->rlen ) );
-------------------------------------^
cfe: Error: mSQL.c, line 694: Incompatible pointer type assignment (3.3.16.1, 3.5.4.1(21))
av_store(av, i, newSVpv(imp_sth->fbh[i].cbuf,0));
-------------------------------------------^
cfe: Error: mSQL.c, line 694: Type char * of this argument is incompatible with type unsigned char * of function prototype description (3.3.16.1)
av_store(av, i, newSVpv(imp_sth->fbh[i].cbuf,0));
-------------------------------------------^
make: *** [mSQL.o] Error 1

Hope that helps,
andreas

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23425-40@lion>;
Mon, 14 Aug 1995 14:31:31 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 14 Aug 95 11:06:38 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa01598;
14 Aug 95 12:06 +0100
Received: by gw.home.vix.com id AA05528; Sun, 13 Aug 95 23:47:39 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA05524; Sun, 13 Aug 95 23:47:37 -0700
Received: (from kstock@localhost) by nova.auspex.fr (8.6.9/8.6.9) id IAA00631;
Robinson France, Mon, 14 Aug 1995 08:48:54 +0200
Date: Mon, 14 Aug 1995 08:48:54 +0200
From: Kevin Stock <[email protected]>
Message-Id: <[email protected]>
To: [email protected], [email protected]
Subject: Re: I'm just started....

Hello Rosehaslina,

> I'm just started to use DBperl to interface with Oracle.
> The problem is I don't know anything about DBperl and where to
> get the sources for it. I have read the DBperl API specification
> but I can't get any informations about DBperl except the available
> functions, variablesand how to use it in the program.
> Can you provide or tell me which site I can go to get
> the basic information about DBperl and also where can I get the
> sources for it. I went through the demon.co.uk site but can't
> understand the README and the files available and the connection
> to DBperl.
> Please help me as soon as possible since I need to use
> it as soon as possible.

The programming interface of DBperl for Oracle (now called DBI::Oracle)
is the same as that of Oraperl. If you download Oraperl (in the Perl4
area of ftp.demon.co.uk) its manual page may help.

Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <24157-3@lion>;
Mon, 14 Aug 1995 14:32:08 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 14 Aug 95 13:07:31 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa18234;
14 Aug 95 14:07 +0100
Received: by gw.home.vix.com id AA12943; Mon, 14 Aug 95 01:59:32 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA12939; Mon, 14 Aug 95 01:59:30 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id JAA03671;
Mon, 14 Aug 1995 09:56:07 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id KAA05083; Mon, 14 Aug 1995 10:01:27 +0100
Message-Id: <[email protected]>
Subject: Re: DBD::mSQL-0.59pl2 out.
To: [email protected]
Date: Mon, 14 Aug 1995 10:01:26 +0100 (BST)
Cc: [email protected]
In-Reply-To: <[email protected]> from "Andreas Koenig" at Aug 14, 95 07:36:49 am
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 2152

> >Again, if you could let me know what happens on comilation, I'd appreciate
> >it. There's been very little feedback, all-in-all.
>
> There seem to be many busy people out there :)

Tell me about it......

8-)

> The mileage seems to be improved. Here's what I get from make:
>
> cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP22-irix/CORE -I/usr/local/Minerva/include -I/usr/local/lib/perl5/IP22-irix/DBI dbdimp.c
> /usr/local/bin/perl -I/usr/local/lib/perl5/IP22-irix -I/usr/local/lib/perl5 /usr/local/lib/perl5/ExtUtils/xsubpp -typemap /usr/local/lib/perl5/ExtUtils/typemap mSQL.xs >mSQL.tc && mv mSQL.tc mSQL.c
> cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP22-irix/CORE -I/usr/local/Minerva/include -I/usr/local/lib/perl5/IP22-irix/DBI mSQL.c
> cfe: Error: mSQL.c, line 475: Incompatible pointer type assignment (3.3.16.1, 3.5.4.1(21))
> sv = sv_2mortal( newSVpv( fbh->cbuf, fbh->rlen ) );
> -------------------------------------^
> cfe: Error: mSQL.c, line 475: Type char * of this argument is incompatible with type unsigned char * of function prototype description (3.3.16.1)
> sv = sv_2mortal( newSVpv( fbh->cbuf, fbh->rlen ) );
> -------------------------------------^
> cfe: Error: mSQL.c, line 694: Incompatible pointer type assignment (3.3.16.1, 3.5.4.1(21))
> av_store(av, i, newSVpv(imp_sth->fbh[i].cbuf,0));
> -------------------------------------------^
> cfe: Error: mSQL.c, line 694: Type char * of this argument is incompatible with type unsigned char * of function prototype description (3.3.16.1)
> av_store(av, i, newSVpv(imp_sth->fbh[i].cbuf,0));
> -------------------------------------------^
> make: *** [mSQL.o] Error 1
>
> Hope that helps,

Well, considering I'm not ecstatic about it failing *again*, yeah, that's
a lot better. It's the same problem as before, just not cleaned up in the
XS file.

Isn't the SGI awful picky about signedness?

Thanks Andreas, most useful. I value your input greatly.

> andreas

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04757-26@lion>;
Tue, 15 Aug 1995 04:15:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 00:05:04 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06207;
15 Aug 95 1:04 +0100
Received: by gw.home.vix.com id AA26332; Mon, 14 Aug 95 04:52:45 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA26328; Mon, 14 Aug 95 04:52:43 -0700
Message-Id: <[email protected]>
Received: from mip.lasc.lockheed.com by netserve.lasc.lockheed.com
with SMTP (1.37.109.4/16.2) id AA24775; Mon, 14 Aug 95 07:52:37 -0400
Received: by mip.lasc.lockheed.com (1.38.193.4/16.2) id AA10621;
Mon, 14 Aug 1995 07:53:13 -0400
Date: Mon, 14 Aug 1995 07:53:13 -0400
From: Terry Greenlaw <[email protected]>
To: [email protected]
Subject: Re: Oracle long

Tim mentioned:

> Note that readblob is experimental. Also note the comment in the source:
> /* Sadly, even though retl is a ub4, oracle will cap the */
> /* value of retl at 65535 even if more was returned! */
> /* This is according to the OCI manual for Oracle 7.0. */
> /* Once again Oracle causes us grief. How can we tell what */
> /* length to assign to destsv? We do have a compromise: if */
> /* retl is exactly 65535 we assume that all data was read. */
> Sigh.

> Tim.

The 64K max data transmission frame is due to Oracle's two task architecture.
Basically, everything acts like it is going over SQL*Net. This greatly
simplifies their client-server model, but to maintain portability, they have
to restict their frame to a 2^16 offset (for 16-bit OS's).
They rely on the underlying network protocol to split the packets into
deliverable chunks and reassemble them on the other side. OCI should have
functions to split the blobs into bite-sized chunks and reassemble them
automatically, as well.

The Oracle7 OCI manual says "If more than 65535 bytes were requested and
returned...". Is it possible to reliably get > 64K chunks byte chunks in
the first place? I haven't had much luck with single blocks that size. Of
course, it was a lot harder in (early??) V6 when the return code was SIGNED!
(That may have been a platform-specific bug. My memory ain't what is used
to be)

A possible workaround would be to model the retrieval like long
assignments. If the chunk is > 64K, set retl to -1, and set the first
4 bytes to the actual length. Of course, just fixing retl to actually get
the full ub4 it is declared to be would be better ;-}

tog
---+++---
Received: from punt2.demon.co.uk by lion with SMTP (PP) id <04756-20@lion>;
Tue, 15 Aug 1995 04:15:06 +0100
Received: from punt2.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 02:52:39 GMT
Received: from gw.home.vix.com by punt2.demon.co.uk id aa06297;
15 Aug 95 3:50 +0100
Received: by gw.home.vix.com id AA20825; Mon, 14 Aug 95 10:38:16 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA20820; Mon, 14 Aug 95 10:38:08 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa28261;
14 Aug 95 14:28 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac07241;
14 Aug 95 14:25 +0100
Received: from ig.co.uk by lion id <22392-0@lion>;
Mon, 14 Aug 1995 11:38:17 +0100
To: [email protected], [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected], [email protected]
Content-Type: X-sun-attachment
Date: Mon, 14 Aug 1995 11:38:17 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

----------
X-Sun-Data-Type: text
X-Sun-Data-Description: text
X-Sun-Data-Name: text
X-Sun-Content-Lines: 39

> From: Sundar Narasimhan <[email protected]>
>
> >From [email protected] Thu Aug 10 18:25 PDT 1995
> >To: [email protected]
> >Please don't do this. When Configure asks you what extra flags to
> >specify to ld add "-L/usr/ucblib", otherwise you risk annoying
> >other users on your system who may be quite surprised because of the
> >link.
> >
> >When I compiled perl and oraperl on Solaris 2.3 and 2.4 - I found that
> >with the Lucid compiler I had to manually edit config.sh to make sure
> >that volatile was not defined. Specifying -DDEBUGGING was also helpful.
>
> This problem happens not with a static library, but a dynamic one
> (/usr/ucblib/libucb.so.1). As far as I know, ldflags can affect static
> linking, but I don't believe that they can affect dynamic linking (at least
> from what I can discern from the ld man page). Therefore, you would have to
> statically link libucb (which isn't such a big deal because it's only like 96K).
>
> It seems like a good suggestion to hack your config.sh file so that
>
> ldflags='-L/usr/ucblib/'
>
> I'll probably recompile our stuff here to do that. Then I won't have to futz
> with dynamic loading of libucb.
> Dynamic loading of libraries is affected by your environment variable
> LD_LIBRARY_PATH. I am not on my Solaris machine right now, but if I
> remember right, the only library needed out of ucblib is -ldbm. If you
> add -lucb to your linkstream (and yes, Configure does discover this),
> then be prepared to deal with a whole host of other problems.
>
> Another command that you may want to know about is "ldd".
>
I have attached the semi-official 'Generic notes about building Perl5 on
Solaris'.

Tim.

----------
X-Sun-Data-Type: sun-deskset-message
X-Sun-Data-Name: sun-deskset-message
X-Sun-Encoding-Info: uuencode
X-Sun-Content-Lines: 140

begin 600 sun-deskset-message
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2;W5T(&]F('1H92!B;W@N"@H*

end
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05191-0@lion>;
Tue, 15 Aug 1995 04:15:19 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 00:05:05 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06252;
15 Aug 95 1:05 +0100
Received: by gw.home.vix.com id AA08648; Mon, 14 Aug 95 08:04:51 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA08644; Mon, 14 Aug 95 08:04:50 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa01996;
14 Aug 95 15:20 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa17999;
14 Aug 95 15:17 +0100
Received: from ig.co.uk by lion id <25144-0@lion>;
Mon, 14 Aug 1995 15:20:25 +0100
To: [email protected]
Subject: Re: New DBI version's availability?
Cc: [email protected]
Date: Mon, 14 Aug 1995 15:20:25 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim, have you a date that the new DBI will be released?
>
Tuesday looks possible.

> Could you email me a copy since I cannot get access to ftp.demon.co.uk at
> the moment?
>
Okay.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04757-31@lion>;
Tue, 15 Aug 1995 04:15:38 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 00:44:52 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa19593;
15 Aug 95 1:44 +0100
Received: by gw.home.vix.com id AA03228; Mon, 14 Aug 95 06:50:11 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA03195; Mon, 14 Aug 95 06:50:08 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa28303;
14 Aug 95 14:28 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id af07241;
14 Aug 95 14:26 +0100
Received: from ig.co.uk by lion id <22561-0@lion>;
Mon, 14 Aug 1995 12:23:06 +0100
To: [email protected], [email protected]
Subject: Re: I'm just started....
Date: Mon, 14 Aug 1995 12:23:06 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Rosehaslina Razali <[email protected]>
>
> I'm just started to use DBperl to interface with Oracle.
> The problem is I don't know anything about DBperl and where to
> get the sources for it. I have read the DBperl API specification
> but I can't get any informations about DBperl except the available
> functions, variablesand how to use it in the program.
> Can you provide or tell me which site I can go to get
> the basic information about DBperl and also where can I get the
> sources for it. I went through the demon.co.uk site but can't
> understand the README and the files available and the connection
> to DBperl.
> Please help me as soon as possible since I need to use
> it as soon as possible.
>
Sigh.

Sorry to be blunt but 'DBperl' (or 'the DBI and DBD::Oracle modules'
to be exact) are NOT ready for general use. If you can't build and
work with incomplete, undocumented and unstable code you should not
be trying to use it.

I would love to give a helping hand to everyone but I simply don't
have the time. I must concentrate any spare time I have on actually
progressing the development.

Perhaps others on this mailing list will contribute some help to
get you started.

(In the long term, if the DBI/DBD modules are to succeed others must
contribute to its development, documentation and support. That'll
be easier once we're over the ciurrent round of API changes.)

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05191-5@lion>;
Tue, 15 Aug 1995 04:15:43 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 01:57:58 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07275;
15 Aug 95 2:57 +0100
Received: by gw.home.vix.com id AA01680; Mon, 14 Aug 95 06:20:07 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA01660; Mon, 14 Aug 95 06:20:05 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id IAA139817
for <[email protected]>; Mon, 14 Aug 1995 08:20:05 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA18510;
Mon, 14 Aug 1995 08:20:03 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Subject: help with parameter binding in Oracle
Date: Mon, 14 Aug 95 08:20:03 -0500

I need a quick synopsis of how parameter binding works in OCI. I
have been sitting here for the past hour trying to understand what
Tim's code for binding parameters is doing and it just doesn't make
sense.

That of course, leaves three possible reasons for my confusion:
1) its Monday, 2) I got up too early - 4 am or 3) I don't know
enough about how parameter binding in Oracle works. I'm hoping
it is reason 3 :-)

Thanks.

Mike Moran
[email protected]
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04757-33@lion>;
Tue, 15 Aug 1995 04:15:51 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 01:58:49 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07491;
15 Aug 95 2:58 +0100
Received: by gw.home.vix.com id AA20522; Mon, 14 Aug 95 10:33:01 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA20513; Mon, 14 Aug 95 10:32:58 -0700
Received: (thorn@localhost) by amaretto.cc.mcgill.ca (8.6.10/8.6.5) id NAA25609;
Mon, 14 Aug 1995 13:32:40 -0400
Date: Mon, 14 Aug 1995 13:32:40 -0400 (EDT)
From: Ron Hall <[email protected]>
To: Kevin Kunzelman <[email protected]>
Cc: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4 (long)
In-Reply-To: <9508091858.AA08350@bonnie>
Message-Id: <[email protected]>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

On Wed, 9 Aug 1995, Kevin Kunzelman wrote:

> Since several people currently and in the past have expressed interest in
> getting perl/oraperl to work on Solaris, perhaps I should prepare an area
> somewhere where people can FTP both the Solaris 2.4 configs for
> perl-4.036/oraperl-2.4 and for the Solaris 2.4 syslog perl usubs.
>
Has this happened yet, i.e. upload to ftp.demon.co.uk?
Sorry for my impatience :)

On a separate and related note I`m trying to compile oraperl
with the following bits of software

Solaris 2.4
perl-4.036 (Many thanks to Micheal D'errico for his
Timely posting, it really works and I wasn't
even close :) )
ora-perl2.4
Oracle 7.1.6

Included below are the log of the make and the makefile, separated
by a line of =s

Ron

==============================================================================
I get the following when compiling:
cc -o oraperl /opt/src/perl-4.036/uperl.o oracle.o orafns.o getcursor.o colons.o debug.o strtoul.o usersub.o \
-lm -L/opt/oracle/lib -locic /opt/oracle/lib/osntab.o -lsqlnet -lora -lcv6 -lcore -lsqlnet `. /opt/src/perl-4.036/config.sh; echo $libs`
Undefined first referenced
symbol in file
limuppx /opt/oracle/lib/libcore.a(ldtsnf.o)
lmsrcin /opt/oracle/lib/libsqlnet.a(nlem.o)
liiccp /opt/oracle/lib/libcore.a(ldtdts.o)
lihmupr /opt/oracle/lib/libora.a(kzsrcr.o)
lihcrenv /opt/oracle/lib/libora.a(upilog.o)
lispln /opt/oracle/lib/libcore.a(ldtdts.o)
lisply /opt/oracle/lib/libcore.a(ldtdts.o)
liisch /opt/oracle/lib/libsqlnet.a(naum.o)
ligcnv /opt/oracle/lib/libora.a(kzsrcr.o)
limdigx /opt/oracle/lib/libora.a(kpuscn.o)
lmsrcbn /opt/oracle/lib/libsqlnet.a(nlem.o)
lmsrgbf /opt/oracle/lib/libsqlnet.a(nlem.o)
liicml /opt/oracle/lib/libsqlnet.a(nau.o)
limnceq /opt/oracle/lib/libora.a(kpuscn.o)
liicop /opt/oracle/lib/libcore.a(ldtdts.o)
lihcsn /opt/oracle/lib/libora.a(upiprv.o)
liicpc /opt/oracle/lib/libcore.a(ldtdts.o)
ligstm /opt/oracle/lib/libora.a(ttcx2y.o)
liicps /opt/oracle/lib/libcore.a(ldtdts.o)
liicpu /opt/oracle/lib/libcore.a(ldtdts.o)
ligratio /opt/oracle/lib/libora.a(ttcdty.o)
liipad /opt/oracle/lib/libcore.a(ldtdts.o)
limr2w /opt/oracle/lib/libora.a(kpg.o)
lii42b /opt/oracle/lib/libcore.a(ldtdts.o)
linbyte /opt/oracle/lib/libcore.a(ldtdts.o)
liwalpx /opt/oracle/lib/libora.a(upilog.o)
lmsrip /opt/oracle/lib/libsqlnet.a(nlem.o)
lirnorm /opt/oracle/lib/libcore.a(lmfe.o)
limc2wx /opt/oracle/lib/libora.a(kpg.o)
limlowx /opt/oracle/lib/libcore.a(ldtsnf.o)
liwspax /opt/oracle/lib/libora.a(kpuscn.o)
liwctex /opt/oracle/lib/libora.a(kpuscn.o)
licbas /opt/oracle/lib/libora.a(upilog.o)
lihmnorm /opt/oracle/lib/libcore.a(lmfe.o)
limcpl /opt/oracle/lib/libcore.a(ldtmfn.o)
ligpth /opt/oracle/lib/libora.a(ttcpro.o)
lincml /opt/oracle/lib/libsqlnet.a(naum.o)
limopen /opt/oracle/lib/libsqlnet.a(nau.o)
lisplth /opt/oracle/lib/libcore.a(ldtdts.o)
lihninfo /opt/oracle/lib/libcore.a(ldtdts.o)
liwdigx /opt/oracle/lib/libora.a(kpuscn.o)
lihci2h /opt/oracle/lib/libora.a(kzsrcr.o)
ligncc /opt/oracle/lib/libora.a(upiprv.o)
lmsrtrm /opt/oracle/lib/libsqlnet.a(nlem.o)
lihasc /opt/oracle/lib/libcore.a(ldtsnf.o)
lihlinfo /opt/oracle/lib/libora.a(upilog.o)
lircop /opt/oracle/lib/libora.a(kpg.o)
litnum /opt/oracle/lib/libcore.a(ldtdts.o)
lilinfo /opt/oracle/lib/libcore.a(ldtdts.o)
lihcurrlangid /opt/oracle/lib/libsqlnet.a(nau.o)
liwc2mx /opt/oracle/lib/libora.a(kpuscn.o)
lircpu /opt/oracle/lib/libora.a(kzsrcr.o)
limalpx /opt/oracle/lib/libora.a(kpuscn.o)
limfwdx /opt/oracle/lib/libora.a(kpg.o)
limalnx /opt/oracle/lib/libcore.a(ldtsnf.o)
limpuc /opt/oracle/lib/libora.a(kpg.o)
ligrls /opt/oracle/lib/libora.a(ttcpro.o)
limspax /opt/oracle/lib/libora.a(kpuscn.o)
limctex /opt/oracle/lib/libcore.a(ldtsnf.o)
lintrn /opt/oracle/lib/libora.a(ttcacs.o)
liwlowx /opt/oracle/lib/libora.a(kpuscn.o)
limcpen /opt/oracle/lib/libsqlnet.a(nau.o)
limcpcx /opt/oracle/lib/libora.a(kpg.o)
liwc2ux /opt/oracle/lib/libora.a(kpuscn.o)
liiskip /opt/oracle/lib/libora.a(kpuscn.o)
liclas /opt/oracle/lib/libora.a(upilog.o)
limcpnx /opt/oracle/lib/libcore.a(ldtsnf.o)
lihlmod /opt/oracle/lib/libora.a(upilog.o)
ld: fatal: Symbol referencing errors. No output written to oraperl
*** Error code 1
make: Fatal error: Command failed for target `oraperl'

==============================================================================
The makefile is the following, don't know that it's right, but there
appears to be a great decoupling of libraries in this version of
Oracle....

# Makefile for Oraperl and Coraperl

# Change these to your ORACLE installation directory and Perl source directory
#
ORACLE_HOME = /opt/oracle
SRC = /opt/src/perl-4.036

# Oracle Definitions, copied from $(ORACLE_HOME)/c/demo/proc.mk
# ALL_ORA_LIBS is the only entry that the Makefile actually uses;
# change it to whatever you need to link Pro*C programs
#
#OTHERLIBS = `cat $(ORACLE_HOME)/rdbms/lib/sysliblist`
#CLIBS = $(OTHERLIBS)
#OCILIB = $(ORACLE_HOME)/lib/libocic.a
#NETLIBS = $(ORACLE_HOME)/lib/osntab.o $(ORACLE_HOME)/lib/ociold.o\
# $(ORACLE_HOME)/lib/libsqlnet.a \
# $(ORACLE_HOME)/lib/libsql.a \
# $(ORACLE_HOME)/lib/libnetv2.a\
# $(ORACLE_HOME)/lib/libnetwork.a
#ORALIBS = $(ORACLE_HOME)/lib/libora.a\
# /opt/oracle/lib/libc3v6.a\
# /opt/oracle/lib/libcore.a\
# /opt/oracle/lib/libcore3.a\
# /opt/oracle/lib/libcv6.a\
# /opt/oracle/lib/libknlopt.a\
# /opt/oracle/lib/libnlsrtl.a\
# /opt/oracle/lib/libnlsrtl3.a\
# /opt/oracle/lib/libnsg.a\
# /opt/oracle/lib/libnttli.a\
# /opt/oracle/lib/libpcc.a\
# /opt/oracle/lib/libpcore.a\
# /opt/oracle/lib/libpls.a\
# /opt/oracle/lib/libproc.a\
# /opt/oracle/lib/libpsd.a\
# /opt/oracle/lib/libsql.a\
# /opt/oracle/lib/libsqlnet.a\
# /opt/oracle/lib/libstublm.a\
#
#ALL_ORA_LIBS = $(CLIBS) $(OCILIB) $(NETLIBS) $(ORALIBS)
OCILIB = -locic
NETLIBS = $(ORACLE_HOME)/lib/osntab.o -lsqlnet
ORALIBS = -lora -lcv6 -lcore -lsqlnet
ALL_ORA_LIBS = -L$(ORACLE_HOME)/lib $(CLIBS) $(OCILIB) $(NETLIBS) $(ORALIBS)

# Perl Definitions, taken from $SRC/usub/Makefile
# Don't include the curses libraries here - they go in CURSELIB
#
GLOBINCS =
LOCINCS =
LIBS = `. $(SRC)/config.sh; echo $$libs`

# Oraperl Definitions

# Set DEBUG to -DDEBUGGING, -DPERL_DEBUGGING or leave blank (see orafns.h)
# If it is not blank, uncomment the definition of DBUG_O
#
DEBUG =
#DBUG_O = dbug/dbug.o

# Curses libraries, only required if you want to build Coraperl
# You may also need -ltermlib or -ltermcap
#
CURSELIB = -lcurses

# Row cache size for SELECT statements.
# If you want to change the default, uncomment this and set the value you want
#
#CACHE = -DCACHE_SIZE=5

# Bind variable padding.
# If you want binding an empty string to provoke an error, uncomment this line.
#
#BIND = -DNO_BIND_PADDING

# Some system-specific things
#
# If your system library does not include strtoul, uncomment the next line
STRTOUL = strtoul.o
#
# If your malloc() returns anything other than a char *, set the appropriate
# type here (don't include the *)
# MALLOC_PTR_TYPE=void
#
# If you are using Perl v3 instead of v4, uncomment the next line
# STR_2MORTAL = -Dstr_2mortal=str_2static

# Database, username and password to use for testing
#
TESTDATA = t scott tiger

#
# Leave these blank lines so that patches to what is below
# won't be upset by your changes to the setups above.
#

# From here on, you shouldn't need to change anything. If you do, let me know.

SRCS = oracle.mus orafns.c getcursor.c colons.c usersub.c \
debug.c strtoul.c
OBJS = oracle.o orafns.o getcursor.o colons.o debug.o $(STRTOUL)
OOBJS = $(OBJS) usersub.o $(DBUG_O)
COBJS = $(OBJS) cusersub.o $(DBUG_O)
HDRS = patchlevel.h orafns.h
DEFS = $(STRTOL) $(PUTENV) $(STR_2MORTAL) $(DEBUG) $(CACHE) $(BIND)

CFLAGS = -Idbug -I$(SRC) $(GLOBINCS) $(LOCINCS) $(DEFS)

oraperl: $(SRC)/uperl.o $(OOBJS)
$(CC) -o oraperl $(SRC)/uperl.o $(OOBJS) \
-lm $(ALL_ORA_LIBS) $(LIBS)

coraperl: $(SRC)/uperl.o $(COBJS) $(SRC)/usub/curses.o
$(CC) -o coraperl $(SRC)/uperl.o $(COBJS) $(SRC)/usub/curses.o \
-lm $(ALL_ORA_LIBS) $(LIBS) $(CURSELIB)

all: oraperl coraperl

test: oraperl
@oraperl -e '&ora_version'
@(cd testdir ; \
rm -f My-Results ; \
echo "Testing oraperl, please wait ..." ; \
for i in *.pl ; do ../oraperl $$i $(TESTDATA) ; done > My-Results ; \
if cmp -s Standard-Results My-Results ; \
then echo "Test successful" ; \
else echo "Test failed - compare My-Results with Standard-Results" ;\
fi; echo)

# We use oraperl although perl would suffice, because we know where it is!

install: oraperl install.pl
@./oraperl ./install.pl ${SRC}

cusersub.c: usersub.c
@rm -f cusersub.c
ln usersub.c cusersub.c

cusersub.o: cusersub.c
$(CC) -c $(CFLAGS) -DCURSES cusersub.c

oracle.c: $(SRC)/usub/mus oracle.mus
$(SRC)/usub/mus oracle.mus >oracle.c

$(OOBJS) $(COBJS): $(HDRS)

dbug/dbug.o:
(cd dbug ; $(MAKE) dbug.o)
@echo " (back to main directory)"

clean:
(cd dbug ; $(MAKE) clean)
@echo " (back to main directory)"
rm -f nohup.out *.o oracle.c cusersub.c
rm -f testdir/My-Results listing tags core

realclean clobber: clean
(cd dbug ; $(MAKE) clobber)
(cd doc ; $(MAKE) clobber)
@echo " (back to main directory)"
rm -f oraperl coraperl

listing:
pr -fn Makefile $(HDRS) $(SRCS) >listing

docs:
(cd doc ; $(MAKE) docs)
@echo " (back to main directory)"

shar: clean
shar -n oraperl-v2 -a -s [email protected] -F -o :Part -l 64 \

Readme [C-Q]* Row* [S-z]*

=============================================================================

Again Thanks in Advance and sorry for the long message.

Ron Hall | DISCLAIMER: I said it. I must've meant it. There is
Analyst | nobody else to blame.
<[email protected]> | Clown:n: A man who acts too natural.
+1 514 398 3718 | "Whenever doubt can be called into question, do so."

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05191-8@lion>;
Tue, 15 Aug 1995 04:16:07 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 02:19:37 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa12894;
15 Aug 95 3:19 +0100
Received: by gw.home.vix.com id AA10140; Mon, 14 Aug 95 08:27:22 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA10135; Mon, 14 Aug 95 08:27:20 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa28244;
14 Aug 95 14:28 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab07241;
14 Aug 95 14:25 +0100
Received: from ig.co.uk by lion id <22382-0@lion>;
Mon, 14 Aug 1995 11:35:34 +0100
To: [email protected]
Subject: Re: Problems building DBD::mSQL (fwd)
Cc: [email protected]
Date: Mon, 14 Aug 1995 11:35:34 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: "Andrew N. Chernoff" <[email protected]>
>
> hi,
> help me!
> how can I unsubscribe from this list?
>
> I've try send:
>
> unsubscribe
>
> to [email protected]
>
> but I'ven't result :(
>
> With respect, Andrew N. Chernoff
>
It may take a few days. Tad Lemon, the list admin, may be away.

I've Bcc'd this to him.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05191-9@lion>;
Tue, 15 Aug 1995 04:16:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 02:50:04 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa19768;
15 Aug 95 3:50 +0100
Received: by gw.home.vix.com id AA05705; Mon, 14 Aug 95 07:28:17 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA05675; Mon, 14 Aug 95 07:28:13 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa28309;
14 Aug 95 14:29 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ag07241;
14 Aug 95 14:26 +0100
Received: from ig.co.uk by lion id <22583-0@lion>;
Mon, 14 Aug 1995 12:26:45 +0100
To: [email protected]
Subject: Re: User group for DBD::Oracle help?
Cc: [email protected]
Date: Mon, 14 Aug 1995 12:26:45 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Steve Livingston <[email protected]>
>
> - What's the best usenet group for posting DBD::Oracle stuff?
>
None. I specifically request that all discussions of the *APLHA*
software take place in the perldb-interest mailing list.

It's to immature for wide scale release.

> - I presume user-help is not appropriate on the developers mailing list.
>
> - Stuff like:
> - When I "use strict":
> - I get complaints about $ora_errno, $ora_errstr, ...
> - What's the fully-qualified name prefix for these variables?
>
It is very much appropriate. I hope that other people will contribute
help without relying on me all the time.

> Cheers,
> Steve
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05974-0@lion>;
Tue, 15 Aug 1995 04:51:12 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 03:41:31 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa29283;
15 Aug 95 4:41 +0100
Received: by gw.home.vix.com id AA03043; Mon, 14 Aug 95 06:47:25 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA03039; Mon, 14 Aug 95 06:47:23 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa28273;
14 Aug 95 14:28 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ad07241;
14 Aug 95 14:25 +0100
Received: from ig.co.uk by lion id <22463-0@lion>;
Mon, 14 Aug 1995 11:59:48 +0100
To: [email protected], [email protected]
Subject: Re: Problems building DBD::mSQL (fwd)
Cc: [email protected], [email protected], [email protected]
Date: Mon, 14 Aug 1995 11:59:48 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Ted Lemon <[email protected]>
>
>
> You're already unsubscribed from the list. I've received both of your
> unsubscribe requests, and I removed you from the list upon receiving
> the first one. The reason you're still getting mail is that your
> mailer is hard to reach, so some messages were queued up for you
> before I processed your unsubscribe request, and delivered to you
> afterwards.
>
> For future reference to other readers on this list, to unsubscribe,
> send mail to [email protected]. Please have a little
> patience - we do not run listserv or majordomo, so your request must
> be processed by hand.
>
> _MelloN_
>
Thanks Ted (the unsung hero of the perldb-interest list).

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06687-0@lion>;
Tue, 15 Aug 1995 06:35:14 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 04:00:16 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa03684;
15 Aug 95 5:00 +0100
Received: by gw.home.vix.com id AA06904; Mon, 14 Aug 95 14:24:53 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA06900; Mon, 14 Aug 95 14:24:53 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id OAA31302; Mon, 14 Aug 1995 14:25:30 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma031287; Mon Aug 14 14:25:13 1995
Received: by bonnie (5.x/SMI-SVR4) id AA04503; Mon, 14 Aug 1995 14:25:11 -0700
Date: Mon, 14 Aug 1995 14:25:11 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508142125.AA04503@bonnie>
To: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
X-Sun-Charset: US-ASCII

On Mon, 14 Aug 1995 13:32:40 (EDT), Ron Hall <[email protected]> wrote:

> I get the following when compiling:
>cc -o oraperl /opt/src/perl-4.036/uperl.o oracle.o orafns.o getcursor.o colons.o debug.o strtoul.o usersub.o \
> -lm -L/opt/oracle/lib -locic /opt/oracle/lib/osntab.o -lsqlnet -lora -lcv6 -lcore -lsqlnet `. /opt/src/perl-4.036/config.sh; echo $libs`
>Undefined first referenced
> symbol in file
>limuppx /opt/oracle/lib/libcore.a(ldtsnf.o)
>lmsrcin /opt/oracle/lib/libsqlnet.a(nlem.o)

etc., etc., etc.

>lihlmod /opt/oracle/lib/libora.a(upilog.o)
>ld: fatal: Symbol referencing errors. No output written to oraperl
>*** Error code 1
>make: Fatal error: Command failed for target `oraperl'
>

This is clearly a library problem. Try using this in your Makefile:

OTHERLIBS = `cat $(ORACLE_HOME)/rdbms/lib/sysliblist`
CLIBS = $(OTHERLIBS)
OCILIB = -locic
NETLIBS = $(ORACLE_HOME)/lib/osntab.o -lsqlnet
ORALIBS = -lora -lcore -lnlsrtl -lcore -lnetv2 -lnetwork -lcv6 \
$(ORACLE_HOME)/lib/__fstd.o

ALL_ORA_LIBS = -L$(ORACLE_HOME)/lib $(CLIBS) $(OCILIB) $(NETLIBS) $(ORALIBS)

You may also need to do this:

LIBS = -L/usr/ucblib `. $(SRC)/config.sh; echo $$libs`
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06687-2@lion>;
Tue, 15 Aug 1995 06:35:21 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 04:24:19 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa08851;
15 Aug 95 5:24 +0100
Received: by gw.home.vix.com id AA23441; Mon, 14 Aug 95 11:13:45 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA23437; Mon, 14 Aug 95 11:13:44 -0700
Received: by northshore.ecosoft.com id AA05641 (5.67a/IDA-1.5
for [email protected]); Mon, 14 Aug 1995 14:13:39 -0400
Date: Mon, 14 Aug 1995 14:13:39 -0400 (EDT)
From: "a j j. brown" <[email protected]>
X-Sender: ajbrown@northshore
To: [email protected]
Message-Id: <Pine.SUN.3.91.950814141255.5366A-100000@northshore>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

Please unsubscribe me for now.
Thanks,
Aj Brown
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06687-4@lion>;
Tue, 15 Aug 1995 06:35:28 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 04:53:34 GMT
Received: from frigate.doc.ic.ac.uk by punt.demon.co.uk id aa13650;
15 Aug 95 5:53 +0100
Received: from gw.home.vix.com by frigate.doc.ic.ac.uk with SMTP (PP)
id <[email protected]>; Tue, 15 Aug 1995 05:51:48 +0100
Received: by gw.home.vix.com id AA26489; Mon, 14 Aug 95 11:55:53 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA26485; Mon, 14 Aug 95 11:55:53 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id LAA21798; Mon, 14 Aug 1995 11:56:25 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma021754; Mon Aug 14 11:56:04 1995
Received: by bonnie (5.x/SMI-SVR4) id AA02559; Mon, 14 Aug 1995 11:55:57 -0700
Date: Mon, 14 Aug 1995 11:55:57 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508141855.AA02559@bonnie>
To: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
X-Sun-Charset: US-ASCII

>From [email protected] Mon Aug 14 10:33 PDT 1995
On Mon, 14 Aug 1995 13:32:40 (EDT), Ron Hall <[email protected]> wrote:

>> Since several people currently and in the past have expressed interest in
>> getting perl/oraperl to work on Solaris, perhaps I should prepare an area
>> somewhere where people can FTP both the Solaris 2.4 configs for
>> perl-4.036/oraperl-2.4 and for the Solaris 2.4 syslog perl usubs.
>>
> Has this happened yet, i.e. upload to ftp.demon.co.uk?
> Sorry for my impatience :)

This will happen today, probably. Over the weekend, I went through my
perl-4.036, oraperl-v.24, and perl-syslog for Solaris stuff and made something
that was presentable.

For those that are interested, I fixed the -lucb stuff (in other words, got
rid of it) without screwing up the DBM stuff. IF you are interested in my
environment, I have Solaris 2.4, Oracle 7.0.16, and am using gcc-2.6.3.

As for the oraperl-v2.4 stuff, I wrote a little How-To on building it for
Solaris 2.4. There's also an appendix on how to add syslog user subs.

I just want to do some final checking on some stuff, then I will upload
the stuff to /[email protected].

I will send a message to this list when I do so.

Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06687-6@lion>;
Tue, 15 Aug 1995 06:35:33 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 04:54:54 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa13879;
15 Aug 95 5:54 +0100
Received: by gw.home.vix.com id AA09999; Mon, 14 Aug 95 08:24:39 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: by gw.home.vix.com id AA09994; Mon, 14 Aug 95 08:24:36 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id QAA06352
for <[email protected]>; Mon, 14 Aug 1995 16:21:18 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id QAA18515 for [email protected];
Mon, 14 Aug 1995 16:26:38 +0100
Message-Id: <[email protected]>
Subject: Help Available if necessary
To: [email protected]
Date: Mon, 14 Aug 1995 16:26:37 +0100 (BST)
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 459

As a result of seeing a lot of requests being posted to perldb-interest
concerning help needed on DBDI/DBD related issues, I'm happy to help
people with problems on it ( for the moment! )

If you email your problems to:

[email protected]

I'll see what I can do. Please remember, I'm very busy, so don't email me
stupid questions if the answers can be found in the README's or documentation.

Thanks.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <07723-4@lion>;
Tue, 15 Aug 1995 09:38:41 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 07:38:12 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00752;
15 Aug 95 8:38 +0100
Received: by gw.home.vix.com id AA18371; Mon, 14 Aug 95 17:29:39 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA18366; Mon, 14 Aug 95 17:29:38 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id RAA13864 for <[email protected]>;
Mon, 14 Aug 1995 17:30:15 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma013858; Mon Aug 14 17:30:10 1995
Received: by bonnie (5.x/SMI-SVR4) id AA06859; Mon, 14 Aug 1995 17:30:09 -0700
Date: Mon, 14 Aug 1995 17:30:09 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508150030.AA06859@bonnie>
To: [email protected]
Subject: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris 2.4
X-Sun-Charset: US-ASCII

Instructions are currently in:
ftp://ftp.demon.co.uk/incoming/README.solaris2

And possibly may be moved to /pub/perl and/or /pub/perl/db/perl4/oraperl

Hope they are helpful,
Kevin

P.S. -- (These instructions do not use the UCB compatibility libraries in
any way.)
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <07723-6@lion>;
Tue, 15 Aug 1995 09:38:49 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 08:04:54 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11183;
15 Aug 95 9:04 +0100
Received: by gw.home.vix.com id AA24844; Mon, 14 Aug 95 19:19:50 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA24839; Mon, 14 Aug 95 19:19:49 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id TAA22208; Mon, 14 Aug 1995 19:20:26 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma022174; Mon Aug 14 19:19:57 1995
Received: by bonnie (5.x/SMI-SVR4) id AA08456; Mon, 14 Aug 1995 19:19:56 -0700
Date: Mon, 14 Aug 1995 19:19:56 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508150219.AA08456@bonnie>
To: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
X-Sun-Charset: US-ASCII

On Mon, 14 Aug 1995 14:25:11, Kevin Kunzelman wrote:
>
>You may also need to do this:
>
>LIBS = -L/usr/ucblib `. $(SRC)/config.sh; echo $$libs`
>

Actually, *don't* do this...
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <08376-10@lion>;
Tue, 15 Aug 1995 10:57:54 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 09:54:19 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22635;
15 Aug 95 10:54 +0100
Received: by gw.home.vix.com id AA20146; Mon, 14 Aug 95 17:50:33 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA20143; Mon, 14 Aug 95 17:50:33 -0700
Received: by gw.home.vix.com id AA18366; Mon, 14 Aug 95 17:29:38 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11)
id RAA13864 for <[email protected]>;
Mon, 14 Aug 1995 17:30:15 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma013858; Mon Aug 14 17:30:10 1995
Received: by bonnie (5.x/SMI-SVR4) id AA06859; Mon, 14 Aug 1995 17:30:09 -0700
Date: Mon, 14 Aug 1995 17:30:09 -0700
From: Kevin Kunzelman <[email protected]>
Message-Id: <9508150030.AA06859@bonnie>
To: [email protected]
Subject: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris 2.4
X-Sun-Charset: US-ASCII

Instructions are currently in:
ftp://ftp.demon.co.uk/incoming/README.solaris2

And possibly may be moved to /pub/perl and/or /pub/perl/db/perl4/oraperl

Hope they are helpful,
Kevin

P.S. -- (These instructions do not use the UCB compatibility libraries in
any way.)
---+++---
Received: from punt2.demon.co.uk by lion with SMTP (PP) id <17268-3@lion>;
Tue, 15 Aug 1995 17:15:39 +0100
Received: from punt2.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 14:56:00 GMT
Received: from gw.home.vix.com by punt2.demon.co.uk id aa11291;
15 Aug 95 15:54 +0100
Received: by gw.home.vix.com id AA17133; Tue, 15 Aug 95 01:59:59 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA17128; Tue, 15 Aug 95 01:59:55 -0700
Received: from ipgate.col.sw-ley.de (root@localhost)
by edina.xenologics.com (8.6.8.1/8.6.6) with UUCP id LAA21847;
Tue, 15 Aug 1995 11:01:09 +0200
Received: from HPLEY1.col.sw-ley.de by ipgate.col.sw-ley.de
with smtp (Smail3.1.28.1 #20) id m0si3Iz-0003hEC;
Mon, 14 Aug 95 17:28 GMT
Received: by HPLEY1.col.sw-ley.de (1.38.193.5/LEYCOL-V1.0) id AA29868;
Mon, 14 Aug 1995 18:14:46 +0200
Message-Id: <9508141614.AA29868@HPLEY1>
Date: Mon, 14 Aug 1995 18:14:46 +0200
From: [email protected]
To: [email protected]
Cc: [email protected]
Subject: problems with ora_do

Hello,

I've some problems with the usage of oraperl. I just want to insert a
new row in a table. What I tried was

...
&ora_do($lda, "insert into test values (1, 2)");

However, this doesn't work because the line
my $ret = $csr->execute;
is commented out in Oraperl.pm. A comment above it says that their is
a hard problem with this line. So, what other operation do I have
to use for my row insertion?

so long

Norbert

Norbert Kiesel Software-Ley GmbH
Lindenstr. 15 Phone: +49 2238 96600 Some say: "Microsoft is the answer."
D-50259 Pulheim Fax: +49 2238 50842 This is wrong. Microsoft is the
Germany Email: [email protected] question--and the answer is "No!".

---+++---
Received: from punt2.demon.co.uk by lion with SMTP (PP) id <17268-6@lion>;
Tue, 15 Aug 1995 17:15:54 +0100
Received: from punt2.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 15:02:19 GMT
Received: from gw.home.vix.com by punt2.demon.co.uk id aa14317;
15 Aug 95 16:00 +0100
Received: by gw.home.vix.com id AA13853; Tue, 15 Aug 95 00:46:32 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA13846; Tue, 15 Aug 95 00:46:27 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa12117;
15 Aug 95 4:51 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa00416;
15 Aug 95 4:48 +0100
Received: from ig.co.uk by lion id <05882-0@lion>;
Tue, 15 Aug 1995 04:28:45 +0100
To: [email protected]
Subject: Re: Problems building DBD::mSQL (fwd)
Date: Tue, 15 Aug 1995 04:28:45 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Tim Bunce <[email protected]>
>
> > [...]
>
> It may take a few days. Tad Lemon, the list admin, may be away.
>
Ooops, make that 'Ted Lemon'.
Sorry Ted.

> Tim.
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <19307-6@lion>;
Tue, 15 Aug 1995 19:32:27 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 18:17:22 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26179;
15 Aug 95 19:17 +0100
Received: by gw.home.vix.com id AA14297; Tue, 15 Aug 95 00:53:39 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA14283; Tue, 15 Aug 95 00:53:23 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa12123;
15 Aug 95 4:51 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab00416;
15 Aug 95 4:48 +0100
Received: from ig.co.uk by lion id <05944-0@lion>;
Tue, 15 Aug 1995 04:49:46 +0100
To: [email protected], [email protected]
Subject: Re: help with parameter binding in Oracle
Date: Tue, 15 Aug 1995 04:49:46 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> I need a quick synopsis of how parameter binding works in OCI. I
> have been sitting here for the past hour trying to understand what
> Tim's code for binding parameters is doing and it just doesn't make
> sense.
>
> That of course, leaves three possible reasons for my confusion:
> 1) its Monday, 2) I got up too early - 4 am or 3) I don't know
> enough about how parameter binding in Oracle works. I'm hoping
> it is reason 3 :-)
>
4) Oracle makes a meal of it.
5) It was 4am when I wrote it.
6) It's 4.30am now.

:-)

Here's a quick overview

prepare call preparse

preparse scans the sql statement looking for bind parameters.
Any it finds of the form ':1' or '?' are rewritten as ':p1' etc.
(This enables a newer form of OCI function to be used later.)

It creates a hash (imp_sth->bind_names) which maps bind parameters
(also called 'place holders') to C structures (phs_tpl).
Each struct holds attributes for the corresponding placeholder,
in the form that OCI want's, them plus a ref to a scalar to
store the bound value.

prepare then calls OCI oparse to process the sql statement.

prepare returns.

when execute is called it calls dbd_bind_ph for each bind var.
dbd_bind_ph sets up the phs_tpl C struct and calls OCI obndrv().

Having an OCI manual is vital for a detailed understanding since many
of the very terse structure field names come directly from the manual.

> Thanks.
>
I hope that helps.

> Mike Moran
> [email protected]
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20613-25@lion>;
Tue, 15 Aug 1995 21:24:35 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 20:20:44 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa21474;
15 Aug 95 21:20 +0100
Received: by gw.home.vix.com id AA04245; Tue, 15 Aug 95 05:53:16 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA04240; Tue, 15 Aug 95 05:53:14 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa05510;
15 Aug 95 12:47 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac22490;
15 Aug 95 12:45 +0100
Received: from ig.co.uk by lion id <06837-0@lion>;
Tue, 15 Aug 1995 06:47:00 +0100
To: [email protected], [email protected]
Subject: Re: Oracle long
Date: Tue, 15 Aug 1995 06:47:00 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Terry Greenlaw <[email protected]>
>
> Tim mentioned:
>
> > Note that readblob is experimental. Also note the comment in the source:
> > /* Sadly, even though retl is a ub4, oracle will cap the */
> > /* value of retl at 65535 even if more was returned! */
> > /* This is according to the OCI manual for Oracle 7.0. */
> > /* Once again Oracle causes us grief. How can we tell what */
> > /* length to assign to destsv? We do have a compromise: if */
> > /* retl is exactly 65535 we assume that all data was read. */
> > Sigh.
>
> The 64K max data transmission frame is due to Oracle's two task architecture.
> Basically, everything acts like it is going over SQL*Net. This greatly
> simplifies their client-server model, but to maintain portability, they have
> to restict their frame to a 2^16 offset (for 16-bit OS's).
> They rely on the underlying network protocol to split the packets into
> deliverable chunks and reassemble them on the other side. OCI should have
> functions to split the blobs into bite-sized chunks and reassemble them
> automatically, as well.
>
> The Oracle7 OCI manual says "If more than 65535 bytes were requested and
> returned...". Is it possible to reliably get > 64K chunks byte chunks in
> the first place?

That was my interpretation of that sentance.

> I haven't had much luck with single blocks that size. Of
> course, it was a lot harder in (early??) V6 when the return code was SIGNED!
> (That may have been a platform-specific bug. My memory ain't what is used
> to be)
>
:-)

> A possible workaround would be to model the retrieval like long
> assignments. If the chunk is > 64K, set retl to -1, and set the first
> 4 bytes to the actual length. Of course, just fixing retl to actually get
> the full ub4 it is declared to be would be better ;-}
>
I'm strongly tempted to just mandate that 64K is the portable max that
the DBI spec will define.

> tog
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <21219-0@lion>;
Tue, 15 Aug 1995 22:17:12 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 20:34:58 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa27825;
15 Aug 95 21:34 +0100
Received: by gw.home.vix.com id AA28126; Tue, 15 Aug 95 04:00:33 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA28121; Tue, 15 Aug 95 04:00:29 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id LAA11233;
Tue, 15 Aug 1995 11:57:06 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id MAA21464; Tue, 15 Aug 1995 12:02:28 +0100
Message-Id: <[email protected]>
Subject: DBI-0.a1
To: [email protected]
Date: Tue, 15 Aug 1995 12:02:27 +0100 (BST)
Cc: [email protected]
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 380

Tim, it seems to compile OK on Solaris 2.3/gcc2.5.6.

Plenty of warnings though that I can forward on if necessary.

Also, on running 'perl Makefile.PL' it appears 'lib/DBD/NullP.pm' is missing
from my kit

DBD::mSQL still runs under with no mods, although, I'll have a closer nose to
see what needs done later today.

Regards.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <21219-9@lion>;
Tue, 15 Aug 1995 22:17:40 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 20:53:47 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06403;
15 Aug 95 21:53 +0100
Received: by gw.home.vix.com id AA03286; Tue, 15 Aug 95 05:39:12 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA03281; Tue, 15 Aug 95 05:39:10 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa05501;
15 Aug 95 12:47 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa22490;
15 Aug 95 12:44 +0100
Received: from ig.co.uk by lion id <06808-0@lion>;
Tue, 15 Aug 1995 06:42:18 +0100
To: [email protected]
Subject: New DBD::Oracle release (for current DBD developers only)
Date: Tue, 15 Aug 1995 06:42:18 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

I've uploaded an alpha/experimental version of DBD::Oracle
to go along with DBI-0.59.

Once again, please ignore if you're not developing a driver right now.

It's in the usual place but called DBD-Oracle-0.a1.tar.Z

From the Changes file:

Changes in DBD::Oracle 0.21, 15 Aug 1995:

NOTE: THIS IS AN UNSTABLE RELEASE!
It requires the closely related DBI 0.59 release.
See Changes file in that release.
Major reworking of internal data management!

Only execute and fetchrow have yet to move into dbdimp.c.
These were kept back since I plan other associated changes.

Known problems:
Warning mode is on by default so certain oraperl coding
styles cause warnings such as:
Statement handle ... destroyed without finish() at ...
disconnect(...) invalidates 1 associated cursor(s) at ...
In future oraperl mode handles will have those warnings disabled.
...

Please note that I plan to make more changes this week (buffer
management etc) and that I'll have little time to hold anyone's hand in
changing their existing DBD code. You'll need to keep up or loose track.
I did warn you about the perils of starting DBD development early! :-)

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <21219-12@lion>;
Tue, 15 Aug 1995 22:17:54 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 21:02:27 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa10528;
15 Aug 95 22:02 +0100
Received: by gw.home.vix.com id AA16686; Tue, 15 Aug 95 01:48:58 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA16678; Tue, 15 Aug 95 01:48:55 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa11040;
15 Aug 95 4:18 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa27635;
15 Aug 95 4:15 +0100
Received: from ig.co.uk by lion id <05804-0@lion>;
Tue, 15 Aug 1995 04:18:08 +0100
To: [email protected]
Subject: New release of DBI module for DBD developers only
Date: Tue, 15 Aug 1995 04:18:08 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

This release of the DBI contains a heavily modified DBD interface.
This is only of interest to driver developers. If you are not
developing a driver (and I recommend you *don't* right now!)
you can ignore this.

The file is in the usual place but is called DBI-0.a1.tar.Z
to prevent the casual reader grabbing a copy.

From the Changes file:

DBI 0.59 15 Aug 1995

NOTE: THIS IS AN UNSTABLE RELEASE!

Major reworking of internal data management!
Performance improvements and memory leaks fixed.
Added a new NullP (empty) driver and a -m flag
to test.pl to help check for memory leaks.
Major changes to DBI/DBD interface include:

1. All imp_xxh_t structs must now have a dbih_ccc_t struct
as the first element in the structure.

2. The previous initialisation scheme of

# create and init private data structure
$imp_data = _private_func(...);
# create handle and reference private data structure
$xxh = DBD::_new_xxh({...}, $imp_data);

is now

# publish size of overall structure required
$imp_data_size = ...; # sizeof(imp_xxh_t)
# create 'blank' handle of required size
$xxh = DBD::_new_xxh({...});
# init private data into space in handle
_private_func($xxh, ...);

_private_func moves 'down' to the next level of handle class.
E.g. a previous
$imp_data = ...::dr::_login(...)
becomes
...::db::_login($dbh,...)

3. The local variable initialisation scheme of

foo(sth)
{
D_dbihcom(sth);
D_imp_sth;

is now

foo(sth)
{
D_imp_sth(sth);
also: D_imp_dbh_from_sth;
and: D_imp_drh_from_dbh;
if needed.

Study DBD::Oracle version 0.21 for more details.
(Comparing parts of v0.21 with v0.20 may be useful.)

Known Problems:

1. A Perl5.001m binary built with -DDEBUGGING seems to
cause (@row = $sth->fetchrow) to be called in a scalar
context. Works fine using a non -DDEBUGGING perl.
This might be a local problem. Please report to findings.

The Oracle module may follow soon (It's done, just needs a
change log written, that might take some time :-).

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-3@lion>;
Wed, 16 Aug 1995 07:01:48 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 22:49:23 GMT
Received: from [192.5.5.1] by punt.demon.co.uk id aa27716; 15 Aug 95 23:49 +0100
Received: by gw.home.vix.com id AA12318; Tue, 15 Aug 95 08:12:49 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA12313; Tue, 15 Aug 95 08:12:41 -0700
Received: from muskie.uhc.com by gar.uhc.com (AIX 3.2/UCB 5.64/4.03) id AB51860;
Tue, 15 Aug 1995 10:12:33 -0500
Received: by muskie.uhc.com (AIX 3.2/UCB 5.64/4.03) id AA16639;
Tue, 15 Aug 1995 10:11:25 -0500
From: Mack Amin <[email protected]>
Message-Id: <[email protected]>
Subject: inquiry
To: [email protected]
Date: Tue, 15 Aug 1995 10:11:25 -0500 (CDT)
Cc: Mack Amin <[email protected]>
In-Reply-To: <9508150030.AA06859@bonnie> from "Kevin Kunzelman" at Aug 14, 95 05:30:09 pm
X-Mailer: ELM [version 2.4 PL23]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 48

How can I get out of this list? Thanks!

-Mack

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-5@lion>;
Wed, 16 Aug 1995 07:01:57 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 22:53:12 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa29141;
15 Aug 95 23:53 +0100
Received: by gw.home.vix.com id AA01903; Tue, 15 Aug 95 05:23:45 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA01894; Tue, 15 Aug 95 05:23:35 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa05507;
15 Aug 95 12:47 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab22490;
15 Aug 95 12:45 +0100
Received: from ig.co.uk by lion id <06825-0@lion>;
Tue, 15 Aug 1995 06:44:26 +0100
To: [email protected], [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
Date: Tue, 15 Aug 1995 06:44:26 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Kevin Kunzelman <[email protected]>
>
> As for the oraperl-v2.4 stuff, I wrote a little How-To on building it for
> Solaris 2.4. There's also an appendix on how to add syslog user subs.
>
Great.

> I just want to do some final checking on some stuff, then I will upload
> the stuff to /[email protected].
>
Please email to me directly (or just post it here) and I'll upload it.
I manage the ftp.demon.co.uk:/pub/perl/db tree remotely.

Thanks.

> Kevin
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-14@lion>;
Wed, 16 Aug 1995 07:02:30 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 23:17:53 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa09527;
16 Aug 95 0:17 +0100
Received: by gw.home.vix.com id AA15741; Tue, 15 Aug 95 09:10:01 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA15726; Tue, 15 Aug 95 09:09:46 -0700
Received: from bssmlsa.nerc-bas.ac.uk.nerc-bas.ac.uk
by bsfiles.nerc-bas.ac.uk (8.6.10/NERC-1.2(Solaris 2.x) id RAA20308;
Tue, 15 Aug 1995 17:09:19 +0100
Date: Tue, 15 Aug 1995 17:09:18 +0100
Message-Id: <[email protected]>
From: Andrew G Wood <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: DBI-0.a1/DBD-Oracle-0.a1
Reply-To: [email protected]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII

Hello again,

Thought that as DBI-0.a1/DBD-Oracle-0.a1 were on ftp.demon.co.uk I would
give them a try.

-----------------------------------------------------------------------------

** Running Irix 5.3 on an Indigo with perl5 ......

% perl -e 'use Config; print Config::myconfig()'
Summary of my perl5 (patchlevel 1) configuration:
Platform:
osname=irix, osver=5, archname=IP20-irix
uname='irix bssmlsa 5.3 11091811 ip20 mips '
hint=recommended
Compiler:
cc='cc', optimize='-O'
cppflags='-D_POSIX_SOURCE -ansiposix -D_BSD_TYPES'
ccflags ='-D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000'
ldflags =''
stdchar='unsigned char', d_stdstdio=define, usevfork=false
voidflags=15, castflags=1, d_casti32=define, d_castneg=
intsize=4, alignbytes=8, usemymalloc=y, randbits=15
Libraries:
so=so
libpth=/usr/lib /usr/local/lib /lib
libs=-lgdbm -lsun -lm -lc -lcrypt -lbsd -lPW
libc=/usr/lib/libc.so
Dynamic Linking:
dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=
cccdlflags=' ', ccdlflags=' ', lddlflags='-shared'

** Then ......

% perl Makefile.PL

Build, TEST and INSTALL Perl 5 first. Use version 5.001m or later.

DO NOT build the DBI under the Perl source tree.

Use 'make test' to execute self tests.

MakeMaker (v4.16)
Checking if your kit is complete...
Warning: the following files are missing in your kit:
lib/DBD/Null.pm
Please inform the author.

NAME => 'DBI'
VERSION => '0.59'
Using PERL=/nerc/bin/perl
Writing Makefile for DBI

** The file that is in lib/DBD is NullP.pm
**
** .... and finally

% make
mkdir ./blib
mkdir ./blib/auto
mkdir ./blib/auto/DBI
/nerc/bin/perl -I/usr/local/lib/perl5/IP20-irix -I/usr/local/lib/perl5 /usr/local/lib/perl5/ExtUtils/xsubpp -typemap /usr/local/lib/perl5/ExtUtils/typemap DBI.xs >DBI.tc && mv DBI.tc DBI.c
cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP20-irix/CORE DBI.c
cfe: Error: DBI.c, line 789: Incompatible pointer type assignment (3.3.16.1, 3.5.4.1(21))
AV *ima_av = ((XRV*) (attribs)->sv_any )->xrv_rv ;
----^
cfe: Error: DBI.c, line 789: Type SV * of rhs of assignment expression is incompatible with type AV * of lhs (3.3.16.1)
AV *ima_av = ((XRV*) (attribs)->sv_any )->xrv_rv ;
----^
*** Error code 1 (bu21)

-----------------------------------------------------------------------------

** Then for good measure I had a go at the new Oracle DBD (DBD-Oracle-0.a1)
** ..... without installing the above DBI of course!

% perl Makefile.PL

Configuring DBD::Oracle version 0.21...
Remember to actually read the README file!

Using Oracle in /nerc/packages/oracle/product/7.1.4
Oracle sysliblist: -lc -lsocket -lnsl -lm
Using /nerc/packages/oracle/product/7.1.4/proc/lib/proc.mk version 1.17
System: perl5.001 irix bssmlsa 5.3 11091811 ip20 mips IP20-irix dl_dlopen.xs
Compiler: cc -O -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000
Oracle proc.mk would have used these values but we override them:
CC: cc
CFLAGS: -I. -O `if [ -d /usr/lib/cmplrs ] ; then echo "-systype svr4 -KPIC" ; fi `
LDFLAGS: -L$(LIBHOME) `if [ -x /usr/necbin/uversion ] ; then echo "-Zcommon_pic -znodefs" ; elif [ -d /usr/lib/cmplrs ] ; then echo "-systype svr4" ; fi `

Checking if your kit is complete...
Looks good
Writing Makefile for DBD::Oracle

** .... and finally, finally

% make
mkdir ./blib
mkdir ./blib/DBD
mkdir ./blib/auto
mkdir ./blib/auto/DBD
mkdir ./blib/auto/DBD/Oracle
/nerc/bin/perl -I/usr/local/lib/perl5/IP20-irix -I/usr/local/lib/perl5 /usr/local/lib/perl5/ExtUtils/xsubpp -typemap /usr/local/lib/perl5/ExtUtils/typemap Oracle.xs >Oracle.tc && mv Oracle.tc Oracle.c
cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP20-irix/CORE -I/nerc/packages/oracle/product/7.1.4/rdbms/demo -I/usr/local/lib/perl5/IP20-irix/DBI Oracle.c
cfe: Error: ./dbdimp.h, line 21: Syntax Error
dbih_drc_t dbihcom;
----^
cfe: Error: ./dbdimp.h, line 21: Empty declaration specifiers (3.5)
dbih_drc_t dbihcom;
---------------^
cfe: Warning 625: ./dbdimp.h, line 22: Empty declaration (3.5)
};
-^
cfe: Error: ./dbdimp.h, line 26: Syntax Error
dbih_dbc_t dbihcom;
----^
cfe: Error: ./dbdimp.h, line 26: Empty declaration specifiers (3.5)
dbih_dbc_t dbihcom;
---------------^
cfe: Error: ./dbdimp.h, line 30: Syntax Error
};
^
cfe: Warning 625: ./dbdimp.h, line 30: Empty declaration (3.5)
};
-^
cfe: Error: ./dbdimp.h, line 35: Syntax Error
dbih_stc_t dbihcom;
----^
cfe: Error: ./dbdimp.h, line 35: Empty declaration specifiers (3.5)
dbih_stc_t dbihcom;
---------------^
cfe: Error: ./dbdimp.h, line 51: Syntax Error
};
^
cfe: Warning 625: ./dbdimp.h, line 51: Empty declaration (3.5)
};
-^
*** Error code 1 (bu21)

-----------------------------------------------------------------------------

I guess this may be a problem with the "very picky" C complier which comes
with Irix ..... yeh I probably should use gcc, but what the hell, a poor
workman blames his tools ;-)

Andy.
*
Dr Andy Wood **
Database Administrator ***
***
****
[email protected] ****
[email protected] ****
[email protected] ****
****
[email protected] ****
...*****.____
Post: .::______.:~'
****
British Antarctic Survey ****
High Cross ****
Madingley Road ****
Cambridge, CB3 0ET, UK ***
**
+44 (0) 1223 61188 **
*

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-19@lion>;
Wed, 16 Aug 1995 07:03:10 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 15 Aug 95 23:50:02 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa21890;
16 Aug 95 0:49 +0100
Received: by gw.home.vix.com id AA03781; Tue, 15 Aug 95 05:46:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA03775; Tue, 15 Aug 95 05:46:42 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id HAA60927;
Tue, 15 Aug 1995 07:46:37 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA06744;
Tue, 15 Aug 1995 07:46:34 -0500
From: [email protected]
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected]
Subject: Re: help with parameter binding in Oracle
In-Reply-To: Your message of Tue, 15 Aug 95 04:49:46 +0100. <[email protected]>
Date: Tue, 15 Aug 95 07:46:33 -0500

Tim Bunce <[email protected]> writes:
|> Here's a quick overview
|>
|> prepare call preparse
|>
|> preparse scans the sql statement looking for bind parameters.
|> Any it finds of the form ':1' or '?' are rewritten as ':p1' etc.
|> (This enables a newer form of OCI function to be used later.)

Well, I can atleast skip doing the translation of place holders. DB2 only
supports '?'.

|> It creates a hash (imp_sth->bind_names) which maps bind parameters
|> (also called 'place holders') to C structures (phs_tpl).
|> Each struct holds attributes for the corresponding placeholder,
|> in the form that OCI want's, them plus a ref to a scalar to
|> store the bound value.

Next question. Are you assuming that all incoming parameters are
character strings? ( phs->ftype = 1 ) Does this mean that OCI will
translate the character string to the appriate sql data type?

|> prepare then calls OCI oparse to process the sql statement.
|>
|> prepare returns.
|>
|> when execute is called it calls dbd_bind_ph for each bind var.
|> dbd_bind_ph sets up the phs_tpl C struct and calls OCI obndrv().
|>
|> Having an OCI manual is vital for a detailed understanding since many
|> of the very terse structure field names come directly from the manual.

Unfortunately, the only OCI manual I have access to if a version 6
manual....not much help.

|> > Thanks.
|> >
|> I hope that helps.

Well, it atleast points out all the major problems I have to deal with
in trying to use DB2 :-)

Alternate subject, making any progress on the rewrite?

Thanks.
Mike
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-25@lion>;
Wed, 16 Aug 1995 07:03:43 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 00:42:53 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa09060;
16 Aug 95 1:42 +0100
Received: by gw.home.vix.com id AA14641; Tue, 15 Aug 95 15:14:04 -0700
Date: Tue, 15 Aug 95 15:14:04 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From [email protected] Tue Aug 15 15:14:03 1995
Received: by gw.home.vix.com id AA14634; Tue, 15 Aug 95 15:14:03 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11) id PAA30615; Tue, 15 Aug 1995 15:14:38 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma030597; Tue Aug 15 15:14:32 1995
Received: by bonnie (5.x/SMI-SVR4)
id AA20466; Tue, 15 Aug 1995 15:14:29 -0700
Date: Tue, 15 Aug 1995 15:14:29 -0700
Message-Id: <9508152214.AA20466@bonnie>
To: [email protected]
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Cc: [email protected]
From: [email protected]
X-Sun-Charset: US-ASCII

On Thu, 10 Aug 1995 18:39:23, "Sigurdur I. Bjornsson" <[email protected]> wrote:
>
>> We have solaris 2.4 and are running perl-4.036 and oraperl-v2.4. Aside
>> from occasional segfaults, it works great. Did you want binaries, or just the
>> configs in order to get the thing to build?
>
>I would like to get the binaries if is not to much trouble.

Actually, on second thought it probably is not such a great idea to give away
the binaries. This is because there are configurations in the build that
may not work for your system depending on where you have installed perl
and ORACLE and what version of ORACLE you have. The ORACLE libraries that
I have may not work for you. Also, it is questionable to redistribute
proprietary copyrighted software (i.e., the ORACLE libraries).

As an alternative, I would suggest getting
/pub/perl/perl-4.036.tar.gz
/pub/perl/db/perl4/oraperl/oraperl-v2.4.tar.gz
/incoming/README.solaris2

from ftp.demon.co.uk and building it yourself.

Cheers,
Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-27@lion>;
Wed, 16 Aug 1995 07:03:53 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 00:49:32 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa10843;
16 Aug 95 1:49 +0100
Received: by gw.home.vix.com id AA24204; Tue, 15 Aug 95 11:03:57 -0700
Date: Tue, 15 Aug 95 11:03:57 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From [email protected] Tue Aug 15 11:03:56 1995
Received: by gw.home.vix.com id AA24201; Tue, 15 Aug 95 11:03:56 -0700
Date: Tue, 15 Aug 95 11:03:56 -0700
From: [email protected]
Message-Id: <[email protected]>
Received: from gw.home.vix.com (gw.home.vix.com [192.5.5.1]) by toccata.fugue.c\

om (8.6.9/8.6.11) with ESMTP id KAA22394 for <[email protected]>; Tue, 15 Aug 19\
95 10:55:50 -0700
Received: by gw.home.vix.com id AA23582; Tue, 15 Aug 95 10:57:43 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: from mailgw.intellimatch.com by lucy.intellimatch.com (5.x/SMI-SVR4)
id AA06379; Tue, 15 Aug 1995 10:53:05 -0700
Received: by mailgw.intellimatch.com with Microsoft Mail
id <[email protected]>; Tue, 15 Aug 95 10:44:39 PDT
From: Igor Khurgin <[email protected]>
To: perldb-interest-request <[email protected]>
Subject: RE: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris
2.4
Date: Tue, 15 Aug 95 10:47:00 PDT
Message-Id: <[email protected]>
Return-Receipt-To: <[email protected]>
Encoding: 25 TEXT
X-Mailer: Microsoft Mail V3.0

I was not able to find anything in the /incoming directory. Would you please
let me no when are you going to make uploads.

Thank you very much

Igor Khurgin
[email protected]
----------
From: perldb-interest-request
To: perldb-interest
Subject: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris
2.4
Date: Monday, August 14, 1995 5:30PM

Instructions are currently in:
ftp://ftp.demon.co.uk/incoming/README.solaris2

And possibly may be moved to /pub/perl and/or /pub/perl/db/perl4/oraperl

Hope they are helpful,
Kevin

P.S. -- (These instructions do not use the UCB compatibility libraries in
any way.)
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-29@lion>;
Wed, 16 Aug 1995 07:04:02 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 00:52:17 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11685;
16 Aug 95 1:52 +0100
Received: by gw.home.vix.com id AA05620; Tue, 15 Aug 95 13:12:48 -0700
Date: Tue, 15 Aug 95 13:12:48 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From [email protected] Tue Aug 15 13:12:48 1995
Received: by gw.home.vix.com id AA05614; Tue, 15 Aug 95 13:12:48 -0700
Received: (from mix@localhost) by digirati.digirati.com (8.6.10/8.6.9) id NAA12815 for [email protected]; Tue, 15 Aug 1995 13:11:44 -0700
Date: Tue, 15 Aug 1995 13:11:44 -0700
From: Michael Shatsman <[email protected]>
Message-Id: <[email protected]>
To: [email protected]
Content-Length: 24

subscribe [email protected]
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-31@lion>;
Wed, 16 Aug 1995 07:04:13 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 00:55:06 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa12342;
16 Aug 95 1:54 +0100
Received: by gw.home.vix.com id AA24470; Tue, 15 Aug 95 11:08:56 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA24183; Tue, 15 Aug 95 11:03:42 -0700
Date: Tue, 15 Aug 95 11:03:42 -0700
From: [email protected]
Message-Id: <[email protected]>
Original-Received: from gw.home.vix.com
(gw.home.vix.com [192.5.5.1]) by toccata.fugue.c\
PP-warning: Illegal Received field on preceding line

om (8.6.9/8.6.11) with ESMTP id KAA22394 for <[email protected]>; Tue, 15 Aug 19\
95 10:55:50 -0700
Received: by gw.home.vix.com id AA23582; Tue, 15 Aug 95 10:57:43 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: from mailgw.intellimatch.com by lucy.intellimatch.com (5.x/SMI-SVR4)
id AA06379; Tue, 15 Aug 1995 10:53:05 -0700
Received: by mailgw.intellimatch.com with Microsoft Mail
id <[email protected]>; Tue, 15 Aug 95 10:44:39 PDT
From: Igor Khurgin <[email protected]>
To: perldb-interest-request <[email protected]>
Subject: RE: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris
2.4
Date: Tue, 15 Aug 95 10:47:00 PDT
Message-Id: <[email protected]>
Return-Receipt-To: <[email protected]>
Encoding: 25 TEXT
X-Mailer: Microsoft Mail V3.0

I was not able to find anything in the /incoming directory. Would you please
let me no when are you going to make uploads.

Thank you very much

Igor Khurgin
[email protected]
----------
From: perldb-interest-request
To: perldb-interest
Subject: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris
2.4
Date: Monday, August 14, 1995 5:30PM

Instructions are currently in:
ftp://ftp.demon.co.uk/incoming/README.solaris2

And possibly may be moved to /pub/perl and/or /pub/perl/db/perl4/oraperl

Hope they are helpful,
Kevin

P.S. -- (These instructions do not use the UCB compatibility libraries in
any way.)
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-37@lion>;
Wed, 16 Aug 1995 07:04:47 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 01:14:42 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17611;
16 Aug 95 2:14 +0100
Received: by gw.home.vix.com id AA04267; Tue, 15 Aug 95 12:51:52 -0700
Date: Tue, 15 Aug 95 12:51:52 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From @post.demon.co.uk,@ignite.demon.co.uk:[email protected] Tue Aug 15 12:51:47 1995
Received: by gw.home.vix.com id AA04256; Tue, 15 Aug 95 12:51:47 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa29889;
15 Aug 95 19:32 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa23364;
15 Aug 95 19:29 +0100
Received: from ig.co.uk by lion id <19285-0@lion>;
Tue, 15 Aug 1995 19:31:48 +0100
To: [email protected]
Subject: Re: DBI-0.a1
Cc: [email protected]
Date: Tue, 15 Aug 1995 19:31:48 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim, it seems to compile OK on Solaris 2.3/gcc2.5.6.
>
> Plenty of warnings though that I can forward on if necessary.
>
Yes please (I know of some but it's always handy to see them
from other compilers).

> Also, on running 'perl Makefile.PL' it appears 'lib/DBD/NullP.pm' is missing
> from my kit
>
Okay.

> DBD::mSQL still runs under with no mods, although, I'll have a closer nose to
> see what needs done later today.
>
Er, I don't think so.

If you installed the a1 DBI then a precompiled DBD would complain about
a version mismatch and trying to compile the DBD again would complain
about all sorts of things.

DBD code changes are required!

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-39@lion>;
Wed, 16 Aug 1995 07:04:57 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 01:33:06 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22042;
16 Aug 95 2:33 +0100
Received: by gw.home.vix.com id AA08421; Tue, 15 Aug 95 13:59:01 -0700
Date: Tue, 15 Aug 95 13:59:01 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From @post.demon.co.uk,@ignite.demon.co.uk:[email protected] Tue Aug 15 13:58:58 1995
Received: by gw.home.vix.com id AA08412; Tue, 15 Aug 95 13:58:58 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa24380;
15 Aug 95 18:06 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa02785;
15 Aug 95 18:03 +0100
Received: from ig.co.uk by lion id <18372-0@lion>;
Tue, 15 Aug 1995 17:33:22 +0100
To: [email protected]
Subject: Re: help with parameter binding in Oracle
Cc: [email protected]
Date: Tue, 15 Aug 1995 17:33:22 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim Bunce <[email protected]> writes:
>
> |> It creates a hash (imp_sth->bind_names) which maps bind parameters
> |> (also called 'place holders') to C structures (phs_tpl).
> |> Each struct holds attributes for the corresponding placeholder,
> |> in the form that OCI want's, them plus a ref to a scalar to
> |> store the bound value.
>
> Next question. Are you assuming that all incoming parameters are
> character strings? ( phs->ftype = 1 ) Does this mean that OCI will
> translate the character string to the appriate sql data type?
>
Currently yes and yes. I plan to leave numbers as numbers (IVs and NVs)
in a later version. The same goes for fetched data.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-41@lion>;
Wed, 16 Aug 1995 07:05:08 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 01:35:06 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22544;
16 Aug 95 2:35 +0100
Received: by gw.home.vix.com id AA24186; Tue, 15 Aug 95 11:03:43 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA24183; Tue, 15 Aug 95 11:03:42 -0700
Date: Tue, 15 Aug 95 11:03:42 -0700
From: [email protected]
Message-Id: <[email protected]>
Original-Received: from gw.home.vix.com
(gw.home.vix.com [192.5.5.1]) by toccata.fugue.c\
PP-warning: Illegal Received field on preceding line

om (8.6.9/8.6.11) with ESMTP id KAA22394 for <[email protected]>; Tue, 15 Aug 19\
95 10:55:50 -0700
Received: by gw.home.vix.com id AA23582; Tue, 15 Aug 95 10:57:43 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Received: from mailgw.intellimatch.com by lucy.intellimatch.com (5.x/SMI-SVR4)
id AA06379; Tue, 15 Aug 1995 10:53:05 -0700
Received: by mailgw.intellimatch.com with Microsoft Mail
id <[email protected]>; Tue, 15 Aug 95 10:44:39 PDT
From: Igor Khurgin <[email protected]>
To: perldb-interest-request <[email protected]>
Subject: RE: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris
2.4
Date: Tue, 15 Aug 95 10:47:00 PDT
Message-Id: <[email protected]>
Return-Receipt-To: <[email protected]>
Encoding: 25 TEXT
X-Mailer: Microsoft Mail V3.0

I was not able to find anything in the /incoming directory. Would you please
let me no when are you going to make uploads.

Thank you very much

Igor Khurgin
[email protected]
----------
From: perldb-interest-request
To: perldb-interest
Subject: Instructions for building perl-4.036 and oraperl-v2.4 for Solaris
2.4
Date: Monday, August 14, 1995 5:30PM

Instructions are currently in:
ftp://ftp.demon.co.uk/incoming/README.solaris2

And possibly may be moved to /pub/perl and/or /pub/perl/db/perl4/oraperl

Hope they are helpful,
Kevin

P.S. -- (These instructions do not use the UCB compatibility libraries in
any way.)
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-43@lion>;
Wed, 16 Aug 1995 07:05:20 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 02:16:51 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa02287;
16 Aug 95 3:16 +0100
Received: by gw.home.vix.com id AA13017; Tue, 15 Aug 95 14:54:58 -0700
Date: Tue, 15 Aug 95 14:54:58 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From [email protected] Tue Aug 15 14:54:57 1995
Received: by gw.home.vix.com id AA13012; Tue, 15 Aug 95 14:54:57 -0700
Received: (from tnprox@localhost) by marion.match.com (8.6.11/8.6.11) id OAA28842; Tue, 15 Aug 1995 14:55:28 -0700
Received: from unknown(204.177.149.200) by marion.match.com via smap (V1.3)
id sma028817; Tue Aug 15 14:55:12 1995
Received: by bonnie (5.x/SMI-SVR4)
id AA20311; Tue, 15 Aug 1995 14:55:05 -0700
Date: Tue, 15 Aug 1995 14:55:05 -0700
Message-Id: <9508152155.AA20311@bonnie>
To: [email protected]
Subject: Re: Installing Oraperl on Solaris 2.4
Cc: [email protected]
From: [email protected]
X-Sun-Charset: US-ASCII

On Tue, 15 Aug 95 14:16:00, Mike Emeigh <[email protected]> wrote:

>Kevin:
>
>I read your posts to perldb-interest indicating that you've developed a
>"how-to" on installing Oraperl on Solaris 2.4. I have been trying without
>much success to do so. Unfortunately, because my DNS does not support
>reverse DNS lookup, I'm blocked from accessing ftp.demon.co.uk. Could you
>please E-mail me a copy of the things you posted there, or indicate where
>else I might me able to retrtieve them via FTP? Thank you for your help.

OK, here it is. I promised these instructions for building perl-4.036
and oraperl-v2.4 a couple of days ago to some people on
perldb-interest. Also included are instructions on how to add syslog
support to oraperl for Solaris 2.4.

Let me know if there are any problems,

Kevin Kunzelman
([email protected])

-- cut here (solaris_perl.howto) --

------------------------------------------
How To Build perl-4.036 Under Solaris 2.4:
------------------------------------------

I spent last night (10-AUG-1995) and this morning going over this stuff
again. We've been using perl/oraperl for over a year now with very
occasional glitches. In entering a discussion on the perldb-interest
mailing list starting a few days ago ([email protected]), I
realized that our perl setup was not quite clean. I was committed to
cleaning the thing up.

With some hints from people involved in the discussion on
perldb-interest, and after running about 30 or so compilations of
perl-4.036 (I really wanted to get it right!), here's what I learned:

------------------------------------------------------
The Configure Utility and Some Notes About /usr/ucblib
------------------------------------------------------

If you use the Configure utility, it will not get you exactly what you
need.

The Configure utility wants to have you use the UCB (in?)compatibility
library libucb (located in /usr/ucblib). You do *not* want to use
this library because it will cause you some problems (as some people
in the discussion pointed out). In fact, it was the use of this
library that was causing some glitches that many months ago I got
around by hacking the perl source (not the best way to do it, to say
the least). A way to find out if you have these glitches is to test
the rand() function in perl:

print rand(500), "\n";

If this returns you a number that is way out of range like
16876.32987987 (a problem I had to hack around previously). Then you
know that you are using libucb and you will want to recompile perl
without it.

---------------
From perl1.faq:
---------------

>1.12) How do I get Perl to compile on Solaris?
>
> John Lees <[email protected]> reports:
>
> I have built perl on Solaris 2.1, 2.2 beta, and 2.2 FCS. Take /usr/ucb
> out of your path and do not use any BSD/UCB libraries. Only -lsocket,
> -lnsl, and -lm are needed. You can use the hint for Solaris 2.0, but
> the one for 2.1 is wrong.

The Solaris 2.1 hint certainly is wrong, but the Solaris 2.0 hint is
minimal. I would actually suggest trying to use the Solaris 2.4 hint
that I created (see below).

> Do not use vfork. Do not use
> -I/usr/ucbinclude. The result works fine for me, but of couse does
> not support a couple of BSDism's.
>
> Casper H.S. Dik <[email protected]> reports
>
> You must remove all the references to /usr/ucblib AND
> /usr/ucbinclude. And ignore the Solaris_2.1 hints. They are wrong.
> The undefining of vfork() probably ahs to do with the confusion it
> gives to the compilers.

In my hints file for Solaris 2.4, I opted to not use vfork (I undef'ed
it in the hint). I was able to compile things fine either way, but I
figured that I would not use it since some people had problems with it
(presumably when using a compiler other than gcc).

> If you use cc, you mustn't compile
> util.c/tutil.c with -O. I only used the following libs: -lsocket
> -lnsl -lm (there is a problem with -lmalloc)

I had no problem of this sort, but I use gcc.

>
> Michael D'Errico <[email protected]> reports:
>
> If you are using Solaris 2.x, the signal handling is broken. If you set
> up a signal handler such as 'ripper' it will be forgotten after the first
> time the signal is caught. To fix this, you need to recompile Perl. Just
> add '#define signal(x,y) sigset((x),(y))' after the '#include <signal.h>'
> directive in each file that it occurs, then make it again.

Hmm. I haven't noticed a problem with signal handling.

------------
Installation
------------

1) Get the perl-4.036.tar.gz distribution from somewhere (e.g.,
prep.ai.mit.edu), gunzip it, and untar it.

2) Take the next few lines and put them in
$PERL_BUILD_DIR/hints/solaris_2_4.sh :

d_odbm='undef' # The t/op/dbm test will fail otherwise
i_fcntl='define' # This is also necessary for t/op/dbm to work
d_vfork='undef' # I don't know if this is necessary,
# but others have mentioned it...
nativegcc='undef' # Using system libraries not compiled with GCC
libs='-lsocket -lnsl -lm' # Do *not* include -lucb, -ldbm, or -lmalloc

3) Run the Configure utility. Except in a few cases, you pretty much
want to accept whatever the Configure script suggests to you:

a) When it prompts you which of the "handy" defaults you want
to use, type "solaris_2_4".

b) You will have to make your own decisions as to where you
want to put the manual pages (the default is /usr/man/manl,
but I put mine in /usr/local/man/man1).

c) You will have to make your own decisions about which C
compiler to use (we use gcc), and which additional cc flags to use.
We ended up using:

-fpcc-struct-return -traditional -Dvolatile=__volatile__
-I/usr/local/lib/gcc-lib/sparc-sun-solaris2.3/2.6.3/include

d) When it asks which additional libraries you need to compile
perl, make sure that you do *not* include -lucb, -ldbm, or
-lmalloc. You probably want to enter:

-lsocket -lnsl -lm

e) Make your own decisions about where you want the
executables to reside (the default, /usr/local/bin, worked
fine for me).

4) Run "make depend"

5) Run "make"

6) Run "make test" to make sure everything works right.

7) Run "make install" as root. You're done!

Notes: After completing the perl installation, I made symbolic links
from /usr/bin/perl to /usr/local/bin/perl as well.
Incidentally, in Solaris 2.4, /bin is linked to /usr/bin, so
this means that you can access perl by /bin/perl,
/usr/bin/perl, or /usr/local/bin/perl. Depending on your
needs, you may or may not want to make these links).

If you find any problems or have corrections to this document, please
send them to my email address below. Thanks.

Kevin Kunzelman
([email protected])

-- cut here (solaris_oraperl.howto) --

--------------------------------------------
How To Build oraperl-v2.4 Under Solaris 2.4:
--------------------------------------------

To build oraperl all you should have to do is edit the Makefile
appropriately. I wish I could give a lot of easy diffs for everyone
to use, but most likely the changes you need to make will depend on
how you installed ORACLE, what version of ORACLE you have (we have
7.0.16), and where you built perl.

-------------
Instructions:
-------------

1) Before building oraperl, you will need to have built perl-4.036
(see "How To Build perl-4.036 Under Solaris 2.4"). This is
because oraperl is a C add-on which uses the perl 4 standard
user subs interface and ORACLE OCI (Oracle Call Interface).
You will also need to make sure that all appropriate ORACLE
libraries are available (see list).

2) Make appropriate changes to your Makefile:

a) If you plan to use gcc, make sure you put this near the top
of the Makefile:

CC = gcc

b) Set ORACLE_HOME to whatever your ORACLE_HOME environment
variable is. For example,

ORACLE_HOME = /opt/oracle

c) Set SRC to wherever you built perl-4.036. For example,

SRC = /usr/local/pub/perl-4.036

d) Comment out DEBUG and DBUG_O (unless you really want them).

#DEBUG = -DPERL_DEBUGGING
#DBUG_O = dbug/dbug.o

e) Make sure all of your Oracle library defines are set up
correctly. I had to change some of them and add a few. This
might be because I'm using Oracle version 7, not version 6. I
changed

OCILIB = $(ORACLE_HOME)/rdbms/lib/libocic.a
NETLIBS = $(ORACLE_HOME)/rdbms/lib/osntab.o \
$(ORACLE_HOME)/rdbms/lib/libsqlnet.a
ORALIBS = $(ORACLE_HOME)/rdbms/lib/libora.a

ALL_ORA_LIBS = $(CLIBS) $(OCILIB) $(NETLIBS) $(ORALIBS)

to

OCILIB = -locic
NETLIBS = $(ORACLE_HOME)/lib/osntab.o -lsqlnet
ORALIBS = -lora -lcore -lnlsrtl -lcore \
-lnetv2 -lnetwork -lcv6 \
$(ORACLE_HOME)/lib/__fstd.o

ALL_ORA_LIBS = -L$(ORACLE_HOME)/lib \
$(CLIBS) $(OCILIB) $(NETLIBS) $(ORALIBS)

In the above change for ALL_ORA_LIBS, you might notice that I
listed -lcore twice. This is intentional -- I couldn't get
oraperl to compile otherwise.

3) If you don't like compiler warnings, then you may optionally edit a
couple of the source files (oracle.mus and orafns.c cause
warnings, but probably won't hurt anything if used in a
build). Here are the diffs:

% diff oracle.mus.old oracle.mus
23c23
< static enum uservars {
---
> enum uservars {
33c33
< static enum usersubs {
---
> enum usersubs {
214c214
< csrp = (struct cursor *) strtoul(csr, (char *) NULL, 0);
---
> csrp = (struct cursor *) strtoul(csr, (char **) NULL, 0);

% diff orafns.c.old orafns.c
13a14
> #include <string.h>

4) Run "make".

5) Run "make install". You're done!

-------------------------------------------
Additional Instructions for Solaris syslog:
-------------------------------------------

If you are interested in adding syslog support to oraperl, then you
might want to add a few additional user subs. Normally, perl (and
thus oraperl) supports syslog. However, the syslog library for perl
(written in Perl) uses BSD sockets (udp) and Solaris syslog normally
uses SysV streams (ugh!). The syslogd man page says,

syslogd reads from the STREAMS log driver, /dev/log, and
from any transport provider specified in /etc/netconfig,
/etc/net/transport/hosts, and /etc/net/transport/services.

However, even after some fiddling with transport providers, I could
not get syslogd to accept messages via udp. If some wizard knows how
to set this up, let me know!

About a year ago it was fairly critical that we got syslog stuff to
work with perl, so we found a workaround with user subs which is
described as follows:

1) Perform steps 1-3 in the above instructions if you have not already
done so.

2) Add a new file to your oraperl build directory called syslog.mus
(syslog.mus is attached to the end of this message).

3) Modify usersub.c to add initialization for the syslog user subs:
Here's the diff:

% diff usersub.c.old usersub.c
29a30
> init_syslog();

4) Modify the Makefile so that we build the new user subs into oraperl.
(Make sure we build syslog.mus->syslog.c->syslog.o. Add syslog.o to
the build for oraperl, and rm syslog.c when we do a "make clean".)
Here are my diffs:

% diff Makefile.old Makefile
78c88
< OBJS = oracle.o orafns.o getcursor.o colons.o debug.o $(STRTOUL)
---
> OBJS = syslog.o oracle.o orafns.o getcursor.o colons.o debug.o $(STRTOUL)
121a132,134
> syslog.c: $(SRC)/usub/mus syslog.mus
> $(SRC)/usub/mus syslog.mus >syslog.c
>
131c144
< rm -f nohup.out *.o oracle.c cusersub.c
---
> rm -f nohup.out *.o oracle.c cusersub.c syslog.c

5) Run "make".

6) Run "make install". You're done!

---------------------------
About the syslog User Subs:
---------------------------

Four user subs are added to support syslog: openlog, syslog, closelog,
and setlogmask. These work just like the ones in the syslog (3)
manpage.

In order to get all of the appropriate logging options, facilities, and
levels (such as LOG_PID, LOG_USER, LOG_CRIT, etc.) for the syslog routines, you
will want to put a "require 'syslog.ph';" somewhere near the top of
your oraperl program.

For example,

require 'syslog.ph';

&openlog ("My ID", &LOG_PID, &LOG_USER);
&syslog (&LOG_CRIT, "Critical Condition: OH NO!");
&closelog ();

NOTE: These user subs are not compatible with the perl subs in
'syslog.pl'. They behave much more like the C versions. With some
hacking, they could be made compatible, but I never got around to doing this.

If you find any problems or have corrections to this document, please
send them to my email address below. Thanks.

Kevin Kunzelman
([email protected])

-- cut here (syslog.mus) --

/* syslog.mus
*
* User subroutine interface to SYSLOG functions
* Kevin Kunzelman ([email protected]) and Scott Fraize ([email protected])
* syslog.mus is in the public domain.
* Last Modified, 12-AUG-1995
*/

#include <ctype.h>

#include "EXTERN.h"
#include "perl.h"

#include <syslog.h>

#include "patchlevel.h"

enum usersubs {
US_openlog,
US_syslog,
US_closelog,
US_setlogmask,
};

static int usersub();

int
init_syslog()
{
char *filename = "syslog.c";

make_usub("openlog", US_openlog, usersub, filename);
make_usub("syslog", US_syslog, usersub, filename);
make_usub("closelog", US_closelog, usersub, filename);
make_usub("setlogmask", US_setlogmask, usersub, filename);
}

static int
usersub(ix, sp, items)
int ix;
register int sp;
register int items;
{
STR **st = stack->ary_array + sp;
register int i;
register char *tmps;
register STR *Str; /* used in str_get and str_gnum macros */

switch (ix)
{

/*
NOTE: It is *critical* that the following fields are
separated by tabs (not just whitespace). Otherwise,
perl's mus (make user subs) program will not parse them
correctly.
*/

CASE void openlog
I char * ident
I int logopt
I int facility
END

CASE void syslog
I int priority
I char * message
END

CASE void closelog
END

CASE int setlogmask
I int maskpri
END

default:
fatal("Unimplemented user-defined subroutine");
break;
}

return sp;
}

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-46@lion>;
Wed, 16 Aug 1995 07:06:02 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 03:06:58 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11701;
16 Aug 95 4:06 +0100
Received: by gw.home.vix.com id AA05541; Tue, 15 Aug 95 13:11:15 -0700
Date: Tue, 15 Aug 95 13:11:15 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From [email protected] Tue Aug 15 13:11:14 1995
Received: by gw.home.vix.com id AA05535; Tue, 15 Aug 95 13:11:14 -0700
Received: (from mix@localhost) by digirati.digirati.com (8.6.10/8.6.9) id NAA12795 for [email protected]; Tue, 15 Aug 1995 13:10:11 -0700
Date: Tue, 15 Aug 1995 13:10:11 -0700
From: Michael Shatsman <[email protected]>
Message-Id: <[email protected]>
To: [email protected]
Subject: Can't install mSQL DBD under Solaris2.4
Content-Length: 396

While trying to recompile Perl with DBD driver for mSQL,
I've received next error message:

Undefined first referenced
symbol in file
ocof ./blib/auto/DBD/mSQL/mSQL.a(mSQL.o)
ocon ./blib/auto/DBD/mSQL/mSQL.a(mSQL.o)
ld: fatal: Symbol referencing errors. No output written to perl
make: *** [perl] Error 1

Thanks,

Michael
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-48@lion>;
Wed, 16 Aug 1995 07:06:13 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 03:31:04 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa16445;
16 Aug 95 4:30 +0100
Received: by gw.home.vix.com id AA00256; Tue, 15 Aug 95 12:13:16 -0700
Date: Tue, 15 Aug 95 12:13:16 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From @post.demon.co.uk,@ignite.demon.co.uk:[email protected] Tue Aug 15 12:13:02 1995
Received: by gw.home.vix.com id AA00232; Tue, 15 Aug 95 12:13:02 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id ab27728;
15 Aug 95 18:55 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa15178;
15 Aug 95 18:53 +0100
Received: from ig.co.uk by lion id <18987-0@lion>;
Tue, 15 Aug 1995 18:55:27 +0100
To: [email protected], [email protected]
Subject: Re: DBI-0.a1/DBD-Oracle-0.a1
Date: Tue, 15 Aug 1995 18:55:27 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Andrew G Wood <[email protected]>
>
> Hello again,
>
> Thought that as DBI-0.a1/DBD-Oracle-0.a1 were on ftp.demon.co.uk I would
> give them a try.
>
> Warning: the following files are missing in your kit:
> lib/DBD/Null.pm

Thanks.

> cfe: Error: DBI.c, line 789: Incompatible pointer type assignment (3.3.16.1, 3.5.4.1(21))
> AV *ima_av = ((XRV*) (attribs)->sv_any )->xrv_rv ;
> ----^
> cfe: Error: DBI.c, line 789: Type SV * of rhs of assignment expression is incompatible with type AV * of lhs (3.3.16.1)
> AV *ima_av = ((XRV*) (attribs)->sv_any )->xrv_rv ;
> ----^
> *** Error code 1 (bu21)
>
Thanks.

> ** Then for good measure I had a go at the new Oracle DBD (DBD-Oracle-0.a1)
> ** ..... without installing the above DBI of course!
>
> cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP20-irix/CORE -I/nerc/packages/oracle/product/7.1.4/rdbms/demo -I/usr/local/lib/perl5/IP20-irix/DBI Oracle.c
> cfe: Error: ./dbdimp.h, line 21: Syntax Error
> dbih_drc_t dbihcom;
> ----^
> cfe: Error: ./dbdimp.h, line 21: Empty declaration specifiers (3.5)

You need to build the a1 DBD against the a1 DBI.

> I guess this may be a problem with the "very picky" C complier which comes
> with Irix ..... yeh I probably should use gcc, but what the hell, a poor
> workman blames his tools ;-)
>
:-)

> Andy.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-52@lion>;
Wed, 16 Aug 1995 07:07:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 03:59:45 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa21639;
16 Aug 95 4:59 +0100
Received: by gw.home.vix.com id AA09667; Tue, 15 Aug 95 14:11:42 -0700
Date: Tue, 15 Aug 95 14:11:42 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From [email protected] Tue Aug 15 14:11:40 1995
Received: by gw.home.vix.com id AA09661; Tue, 15 Aug 95 14:11:40 -0700
Received: (thorn@localhost) by amaretto.cc.mcgill.ca (8.6.10/8.6.5) id RAA29405; Tue, 15 Aug 1995 17:11:31 -0400
Date: Tue, 15 Aug 1995 17:11:30 -0400 (EDT)
From: Ron Hall <[email protected]>
To: [email protected]
Subject: Hopelessly lost
Message-Id: <[email protected]>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

Oraperl compiled (finally)
Now here is the goal

machine 132.206.35.11
DB Name MAIN
user zot
pass zotme

What is the ora_logon call that gets me connected

ora_login('t:132.206.35.11:MAIN', "zot", "zotme")

or what?

Solaris 2.4
Oraperl 2.4
Oracle 7.1.6

All help appreciated....

Ron Hall | DISCLAIMER: I said it. I must've meant it. There is
Analyst | nobody else to blame.
<[email protected]> | Clown:n: A man who acts too natural.
+1 514 398 3718 | "Whenever doubt can be called into question, do so."

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <23025-58@lion>;
Wed, 16 Aug 1995 07:07:46 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 05:34:36 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa10447;
16 Aug 95 6:34 +0100
Received: by gw.home.vix.com id AA20538; Tue, 15 Aug 95 16:36:24 -0700
Date: Tue, 15 Aug 95 16:36:24 -0700
From: [email protected]
Message-Id: <[email protected]>
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk

From @post.demon.co.uk,@ignite.demon.co.uk:[email protected] Tue Aug 15 16:36:22 1995
Received: by gw.home.vix.com id AA20532; Tue, 15 Aug 95 16:36:22 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa05637;
15 Aug 95 21:26 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa16929;
15 Aug 95 21:24 +0100
Received: from ig.co.uk by lion id <21027-0@lion>;
Tue, 15 Aug 1995 21:26:45 +0100
To: [email protected], [email protected]
Subject: $dbh->do($statement) anyone? (was: problems with ora_do)
Date: Tue, 15 Aug 1995 21:26:45 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: nk@hpley1
>
> Hello,
>
> I've some problems with the usage of oraperl. I just want to insert a
> new row in a table. What I tried was
>
> ....
> &ora_do($lda, "insert into test values (1, 2)");
>
> However, this doesn't work because the line
> my $ret = $csr->execute;
> is commented out in Oraperl.pm. A comment above it says that their is
> a hard problem with this line.

I've pretty much decided that the DBI needs an 'execute immediate'
method (like ora_do) and that an emulation via prepare+execute
is not sufficient because of anomalies on some vendors API's
(who shall remain nameless :-).

How about

$dbh->do($statement)

as an interface?

> So, what other operation do I have to use for my row insertion?
>
If you don't do any Data Definition statements (create table etc)
then you could just uncomment that line.

A better alternative is to use prepare+execute directly (write
yourself a utility function just like ora_do.

> so long
>
> Norbert
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <24566-5@lion>;
Wed, 16 Aug 1995 08:30:30 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 07:15:21 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa08094;
16 Aug 95 8:15 +0100
Received: by gw.home.vix.com id AA06273; Tue, 15 Aug 95 19:51:27 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA06265; Tue, 15 Aug 95 19:51:24 -0700
Received: by presence.COM (4.1/SMI-4.1+TAS1.54) id AA14356;
Tue, 15 Aug 95 19:52:51 PDT
From: Damian Martinez <[email protected]>
Message-Id: <[email protected]>
Subject: Mailing List Problems?
To: [email protected], [email protected]
Date: Tue, 15 Aug 1995 19:52:50 -0700 (PDT)
X-Mailer: ELM [version 2.4 PL24]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 242

Is anyone else seeing wackiness with the messages from this
list? It appears that they are being forwarded from the
perldb-interest-request alias instead of coming from the
person who actually mailed it.

Maybe it's just me...
??

Damian
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <01525-3@lion>;
Wed, 16 Aug 1995 19:23:01 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 16:58:49 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa27466;
16 Aug 95 17:58 +0100
Received: by gw.home.vix.com id AA16392; Tue, 15 Aug 95 23:01:24 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA16387; Tue, 15 Aug 95 23:01:23 -0700
Received: from localhost (mellon@localhost) by toccata.fugue.com (8.6.9/8.6.11)
with SMTP id WAA23466; Tue, 15 Aug 1995 22:59:20 -0700
Message-Id: <[email protected]>
To: Damian Martinez <[email protected]>
Cc: [email protected]
Subject: Re: Mailing List Problems?
In-Reply-To: Your message of "Tue, 15 Aug 1995 19:52:50 PDT." <[email protected]>
Date: Tue, 15 Aug 1995 22:59:19 -0700
From: Ted Lemon <[email protected]>

> Is anyone else seeing wackiness with the messages from this
> list? It appears that they are being forwarded from the
> perldb-interest-request alias instead of coming from the
> person who actually mailed it.

It's very complicated and unpleasant. The good news is that the
problem is fixed. I've been trying to reduce the number of spurious
``your message hasn't been delivered for 30 seconds - we will continue
to try to deliver it for 110 days'' bounces that I get by adding a
``Precedence: junk'' line to each outgoing message.

I actually had this working just fine, as follows:

perldb-interest: "|(/bin/echo 'Precedence: junk'; /usr/bin/sed -e '1d') |/usr/sbin/sendmail -fperldb-interest-request -oi -oem perldb-interest-real"

The above complicated hair puts in the Precedence: junk line, strips
off the leading unix-style From line, sets the envelope sender to
perldb-interest-request so that I get the bounces instead of all of
you, and sends it all out through the perldb-interest-real mailing
list.

Unfortunately, I saw a message go by shortly after I made the above
change which suggested to me that sendmail was not putting in the
unix-style From line, so I took out the sed command and replaced it
with cat. This produced the Apparently-To: headers - sendmail sees
the unix-style From line as a bad header, and treats it and everything
past it (including the To: line) as part of the message body.

Anyway, I've put it back, and it should be fine now. I do apologize
for the gnarly-looking messages that you've seen, though.

_MelloN_
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04034-12@lion>;
Thu, 17 Aug 1995 07:02:09 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 20:19:40 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa02447;
16 Aug 95 21:19 +0100
Received: by gw.home.vix.com id AA04246; Wed, 16 Aug 95 03:09:26 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA04241; Wed, 16 Aug 95 03:09:21 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa19731;
16 Aug 95 10:30 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac10677;
16 Aug 95 10:28 +0100
Received: from ig.co.uk by lion id <25417-0@lion>;
Wed, 16 Aug 1995 10:29:38 +0100
To: [email protected], [email protected], [email protected]
Subject: Re: Mailing List Problems?
Cc: [email protected]
Date: Wed, 16 Aug 1995 10:29:38 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Damian Martinez <[email protected]>
>
> Is anyone else seeing wackiness with the messages from this
> list? It appears that they are being forwarded from the
> perldb-interest-request alias instead of coming from the
> person who actually mailed it.
>
> Maybe it's just me...
> ??
>
I'm also seeing this on recent messages (except for your message!)

> Damian
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04034-32@lion>;
Thu, 17 Aug 1995 07:04:44 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 21:37:37 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa08493;
16 Aug 95 22:37 +0100
Received: by gw.home.vix.com id AA19789; Wed, 16 Aug 95 07:41:18 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA19784; Wed, 16 Aug 95 07:41:16 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id JAA25621
for <[email protected]>; Wed, 16 Aug 1995 09:41:15 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA19412;
Wed, 16 Aug 1995 09:41:13 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Subject: question about prepare
Date: Wed, 16 Aug 95 09:41:12 -0500

The description for prepare is

$sth = $dbh->prepare($statement[,%args])

while the current implementation seems to be

$sth = $dbh->prepare($dbh,$statement)

or am I misinterpreting things?

Now to the real question. In DB2, inorder to bind values to the
prepared statement I need to know the expected type of the parameter.
This is fairly simple if the parameter is going to be used in conjuction
with a result column and basically impossible if used in any other
context. So are we going to attempt to support the original definition
or are we moving towards the current definition?

Mike

------------------------------------------------------------------------------
Michael H. Moran | Standard Disclaimer: The content of
[email protected] | this posting is independent of
Commercial Performance | official IBM position.
IBM Corporation, Austin, Texas |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04034-45@lion>;
Thu, 17 Aug 1995 07:06:06 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 22:29:23 GMT
Received: from frigate.doc.ic.ac.uk by punt.demon.co.uk id aa01332;
16 Aug 95 23:29 +0100
Received: from gw.home.vix.com by frigate.doc.ic.ac.uk with SMTP (PP)
id <[email protected]>; Wed, 16 Aug 1995 23:29:08 +0100
Received: by gw.home.vix.com id AA12535; Wed, 16 Aug 95 05:39:03 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA12521; Wed, 16 Aug 95 05:38:57 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa19792;
16 Aug 95 10:31 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa10838;
16 Aug 95 10:28 +0100
Received: from ig.co.uk by lion id <25553-0@lion>;
Wed, 16 Aug 1995 10:31:13 +0100
To: [email protected]
Mmdf-Warning: Unable to confirm address in preceding line at
disperse.demon.co.uk
Subject: Re: question re. perl-4.036, oraperl and Solaris 2.4
Date: Wed, 16 Aug 1995 10:31:13 +0100
From: Tim Bunce <[email protected]>
Sender: [email protected]
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> As an alternative, I would suggest getting
> /pub/perl/perl-4.036.tar.gz
> /pub/perl/db/perl4/oraperl/oraperl-v2.4.tar.gz
> /incoming/README.solaris2
>
> from ftp.demon.co.uk and building it yourself.
>
README.solaris2 is now installed at /pub/perl/db/perl4/oraperl/README.solaris2

> Cheers,
> Kevin
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04034-51@lion>;
Thu, 17 Aug 1995 07:06:40 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 23:01:55 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa14494;
17 Aug 95 0:01 +0100
Received: by gw.home.vix.com id AA21143; Wed, 16 Aug 95 08:07:57 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Received: by gw.home.vix.com id AA21138; Wed, 16 Aug 95 08:07:55 -0700
Received: from biad38.uthscsa.edu.ric.uthscsa.edu (biad38.uthscsa.edu)
by uthscsa.edu (PMDF V4.3-13 #8477)
id <[email protected]>;
Wed, 16 Aug 1995 10:12:32 -0500 (CDT)
Received: by biad38.uthscsa.edu.ric.uthscsa.edu (4.1/SMI-4.1) id AA00917;
Wed, 16 Aug 95 10:06:50 CDT
Date: Wed, 16 Aug 1995 10:06:49 -0500 (CDT)
From: Thomas Ganslandt <[email protected]>
Subject: Debug output in Oracle DBD
To: [email protected]
Cc: [email protected]
Message-Id: <[email protected]>
X-Envelope-To: [email protected]
Mime-Version: 1.0
X-Mailer: ELM [version 2.4 PL22]
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 768

I am using the Oracle DBD 1.17 (with oraperl emulation) to access an oracle
database through the world-wide-web. Sometimes, but not always (it seems
to depend on how long the script takes to run), I get a debug message
into my output:

Bad free() ignored at [path of script] line 221.

The given line contains the ora_logoff statement.

Modifying Oraperl.pm as John Carlson ([email protected]) suggested on this
list (setting $debug = 0) did not resolve the problem.
Am I doing something wrong or is there a way to shut the errormessage
off or at least direct it not to stdout but to stderr, so it doesn't
get displayed in the output?

Btw. I am running perl5.000m on SunOS 5.3 with Oracle 7.0.16.

Thanks,

Thomas ([email protected])
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04034-57@lion>;
Thu, 17 Aug 1995 07:07:08 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 16 Aug 95 23:58:58 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06878;
17 Aug 95 0:57 +0100
Received: by gw.home.vix.com id AA10629; Wed, 16 Aug 95 12:50:11 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA10580; Wed, 16 Aug 95 12:50:06 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa26495;
16 Aug 95 19:44 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab07475;
16 Aug 95 19:42 +0100
Received: from ig.co.uk by lion id <02214-0@lion>;
Wed, 16 Aug 1995 19:44:38 +0100
To: [email protected]
Subject: Help requested for ora_do behaviour testing
Date: Wed, 16 Aug 1995 19:44:38 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> I've pretty much decided that the DBI needs an 'execute immediate'
> method (like ora_do) and that an emulation via prepare+execute
> is not sufficient because of anomalies on some vendors API's
> (who shall remain nameless :-).
>
> How about
>
> $dbh->do($statement)
>
> as an interface?
>
I've added a 'do' method now but I've left the default implementation as:

sub do {
my($dbh, $statement) = @_;
my $sth = $dbh->prepare($statement) or return undef;
$sth->execute or return undef;
my $rows = $sth->rows;
($rows == 0) ? "OK" : $rows;
}

I seem to recall that many months ago someone sent me a note saying that

&ora_do($lda, "insert into foo values 1");

was inserting the data twice. That's why I commented out the execute
in ora_do. Now I'm not so sure about this (I never was really but never
had the time to look into it).

oraperl-2.4 ora_do seems to do: csr=ora_open; ora_close(csr); and ora_open
always does an execute OCI call (unless bind vars are present).

So, the question is, why doesn't the ora_do(...insert...) example above
execute twice in oraperl-2.4? Or, perhaps it doesn't actually do it
in either and the original message was mistaken.

If anyone has some spare time I'd appreciate it if you could uncomment
the execute line from ora_do in Oraperl.pm are try out things like
insert and create table statements (where double execution is easy
to spot).

It doesn't matter which version of DBD:Oracle you use.

Thanks in advance.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05243-1@lion>;
Thu, 17 Aug 1995 07:07:17 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 00:06:47 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa09947;
17 Aug 95 1:06 +0100
Received: by gw.home.vix.com id AA10885; Wed, 16 Aug 95 12:51:04 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA10875; Wed, 16 Aug 95 12:50:58 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa17639;
16 Aug 95 17:29 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa06347;
16 Aug 95 17:26 +0100
Received: from ig.co.uk by lion id <00980-0@lion>;
Wed, 16 Aug 1995 17:28:35 +0100
To: [email protected]
Subject: Status report
Cc: [email protected], [email protected]
Date: Wed, 16 Aug 1995 17:28:35 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

Just to let you know that I've had rather a good day.

DBI and DBD::Oracle have progressed well. DBD::Oracle is now working
better than the last major release and the architecture is developing well.
I might even encourage some DBD::Oracle vs oraperl benchmarks :-)

I hope to make another (much more complete and stable) release tomorrow.

*Please* feedback any comments you have on the last (a1) release ASAP.

Thanks.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04034-64@lion>;
Thu, 17 Aug 1995 07:08:12 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 02:36:28 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17549;
17 Aug 95 3:36 +0100
Received: by gw.home.vix.com id AA04139; Wed, 16 Aug 95 11:13:54 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA04133; Wed, 16 Aug 95 11:13:54 -0700
Message-Id: <[email protected]>
To: [email protected]
Subject: Testing...
Date: Wed, 16 Aug 1995 11:13:54 -0700
From: Ted Lemon <[email protected]>

At the request of a couple of members, I've made the incredibly stupid choice
of mucking further with the mailing list redistribution process. This is a
test message to see if my foolishness had its expected bad consqeuences.

_MelloN_
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05243-11@lion>;
Thu, 17 Aug 1995 07:08:38 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 03:23:19 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26428;
17 Aug 95 4:23 +0100
Received: by gw.home.vix.com id AA14188; Wed, 16 Aug 95 13:41:22 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA14183; Wed, 16 Aug 95 13:41:21 -0700
Received: from localhost (mellon@localhost) by toccata.fugue.com (8.6.9/8.6.11)
with SMTP id NAA25762; Wed, 16 Aug 1995 13:38:38 -0700
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected], [email protected]
Subject: Re: Mailing List Problems?
In-Reply-To: Your message of "Wed, 16 Aug 1995 19:33:32 BST." <[email protected]>
Date: Wed, 16 Aug 1995 13:38:37 -0700
From: Ted Lemon <[email protected]>

> Is setting up majordomo a practical option for you?

I really don't want to use majordomo - I'm not comfortable with the
security risk, although it is definitely handy. I'd be more likely to
set up a web page, actually, although I don't want to disenfranchise
those poor benighted souls who can't run any kind of http client...

_MelloN_
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05243-12@lion>;
Thu, 17 Aug 1995 07:08:45 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 03:25:00 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26763;
17 Aug 95 4:24 +0100
Received: by gw.home.vix.com id AA13725; Wed, 16 Aug 95 13:37:06 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA13716; Wed, 16 Aug 95 13:37:03 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id PAA44852
for <[email protected]>; Wed, 16 Aug 1995 15:36:51 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA25022;
Wed, 16 Aug 1995 14:46:07 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: Getting the DBD loaded...problem
Date: Wed, 16 Aug 95 14:46:07 -0500

Ok, I finally got enough code rewritten to make it worth trying to
test DB2perl. The DBI (DBI-0.58) test shows success, but when I
attempt to run the test program for DB2 I get the following error from perl:

Can't call method "import" in empty package "DB2" at test2.pl line 4.

I understand what the error message is trying to tell me, but I haven't
got a clue as to why or how to fix it.

Thanks.
Mike

------------------------------------------------------------------------------
Michael H. Moran | Standard Disclaimer: The content of
[email protected] | this posting is independent of
Commercial Performance | official IBM position.
IBM Corporation, Austin, Texas |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <04034-74@lion>;
Thu, 17 Aug 1995 07:09:23 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 04:46:49 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11522;
17 Aug 95 5:46 +0100
Received: by gw.home.vix.com id AA12332; Wed, 16 Aug 95 13:13:24 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA12326; Wed, 16 Aug 95 13:13:20 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa25785;
16 Aug 95 19:34 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac05049;
16 Aug 95 19:31 +0100
Received: from ig.co.uk by lion id <02073-0@lion>;
Wed, 16 Aug 1995 19:33:32 +0100
To: [email protected], [email protected]
Subject: Re: Mailing List Problems?
Cc: [email protected]
Date: Wed, 16 Aug 1995 19:33:32 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Ted Lemon <[email protected]>
>
> > Is anyone else seeing wackiness with the messages from this
> > list? It appears that they are being forwarded from the
> > perldb-interest-request alias instead of coming from the
> > person who actually mailed it.
>
> It's very complicated and unpleasant. The good news is that the
> problem is fixed. I've been trying to reduce the number of spurious
> ``your message hasn't been delivered for 30 seconds - we will continue
> to try to deliver it for 110 days'' bounces that I get by adding a
> ``Precedence: junk'' line to each outgoing message.
>
Thanks Ted.

Is setting up majordomo a practical option for you?

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06621-9@lion>;
Thu, 17 Aug 1995 10:23:51 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 08:08:08 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07706;
17 Aug 95 9:08 +0100
Received: by gw.home.vix.com id AA08398; Wed, 16 Aug 95 19:47:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA08393; Wed, 16 Aug 95 19:47:42 -0700
Received: by condor.messaging.cs.mci.com; id AA12586;
Wed, 16 Aug 1995 20:45:08 -0600
Date: Wed, 16 Aug 1995 20:45:08 -0600
Message-Id: <[email protected]>
X-Sender: [email protected]
X-Mailer: Windows Eudora Version 2.1.1
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
To: Ted Lemon <[email protected]>, Tim Bunce <[email protected]>
From: Patrick Crawford <[email protected]>
Subject: Re: Mailing List Problems?
Cc: [email protected], [email protected]

>
>I really don't want to use majordomo - I'm not comfortable with the
>security risk, although it is definitely handy. I'd be more likely to
>set up a web page, actually, although I don't want to disenfranchise
>those poor benighted souls who can't run any kind of http client...
>

How would you categorize the security risk of majordomo? Of all the software
out there to perform this function it is, IMHO, the one most clearly
designed with security in mind. I have installed and re-installed it several
times through three revisions. If anything, it is too paranoid about security.

Pat

--
Patrick Crawford, Senior Engineer MCI Information Systems Corporation
Phone: +1 719-535-4255 Messaging Development 1819/117
Fax: +1 719-535-4678 2424 Garden of the Gods Road
[email protected] Colorado Springs, CO 80919

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <08908-1@lion>;
Thu, 17 Aug 1995 13:46:31 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 12:24:55 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17427;
17 Aug 95 13:24 +0100
Received: by gw.home.vix.com id AA16857; Wed, 16 Aug 95 22:23:14 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA16851; Wed, 16 Aug 95 22:23:11 -0700
Received: by anna.mind.de (940816.SGI.8.6.9/940406.SGI) id HAA26709;
Thu, 17 Aug 1995 07:21:54 +0200
Date: Thu, 17 Aug 1995 07:21:54 +0200
Message-Id: <[email protected]>
From: Andreas Koenig <[email protected]>
To: [email protected]
Cc: [email protected], [email protected]
In-Reply-To: <[email protected]> ([email protected])
Subject: Re: Getting the DBD loaded...problem
Reply-To: [email protected]

>Can't call method "import" in empty package "DB2" at test2.pl line 4.

Your DB2.pm file seems to need the

package DB2;

line.

andreas
---+++---
Received: from punt2.demon.co.uk by lion with SMTP (PP) id <14887-0@lion>;
Thu, 17 Aug 1995 22:39:18 +0100
Received: from punt2.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 21:14:27 GMT
Received: from gw.home.vix.com by punt2.demon.co.uk id aa13502;
17 Aug 95 22:13 +0100
Received: by gw.home.vix.com id AA18923; Thu, 17 Aug 95 06:57:45 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA18918; Thu, 17 Aug 95 06:57:37 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id OAA04269;
Thu, 17 Aug 1995 14:54:14 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id OAA24769; Thu, 17 Aug 1995 14:59:40 +0100
Message-Id: <[email protected]>
Subject: Re: Getting the DBD loaded...problem
To: [email protected]
Date: Thu, 17 Aug 1995 14:59:39 +0100 (BST)
Cc: [email protected]
In-Reply-To: <[email protected]> from "[email protected]" at Aug 17, 95 07:24:05 am
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 1399

> Just discovered something rather interesting. I had been assuming that
> Tim's Oracle.pm or Oraperl.pm was doing a "use Oracle;" type call
> to load that particular interface using some form of automagic. So
> I went and looked ... wrong.
>
> Neither module every does a "use Oracle;". Rather, Tim does a
> $drh->install_driver('Oracle'), which is I change Oracle to DB2 works
> on my system as does use DBD::DB2.
>
> So that brings up the next question, is there any difference between
> the two methods? Does the DBI need former as opposed to the latter?

Not knowing much about perl's loading stuff, but having worked on DBI
and some DBDs a bit, I can say this ( which may be wrong! ):

The main DBI handler is loaded with 'use DBI'. The DBDs are loaded by doing
the install_driver function because I may want to, eg, load an Oracle and
mSQL driver, then take data from one and pass it straight into the other.

If I did a 'use Oracle' if wouldn't specifically get a handle ( $drh ) to
use simultaneously with another $drh, eg:

use DBI;
$oradrh = DBI->install_driver('Oracle');
$msqldrh = DBI->install_driver('mSQL');

$oradbh = connect.....;
$msqldbh = connect....;

and so on.

No? This wau I can slurp data into a perl 'structure' ( of some nature ),
then stuff it into whatever driver I want to. No intermediate dumping.

> Mike

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <14886-6@lion>;
Thu, 17 Aug 1995 22:40:15 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 18:57:21 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07647;
17 Aug 95 19:57 +0100
Received: by gw.home.vix.com id AA13326; Thu, 17 Aug 95 05:25:48 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA13321; Thu, 17 Aug 95 05:25:46 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id HAA176881
for <[email protected]>; Thu, 17 Aug 1995 07:25:45 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA24308;
Thu, 17 Aug 1995 07:25:46 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Subject: Is the list having problems? Software change?
Date: Thu, 17 Aug 95 07:25:45 -0500

The last three items I have sent to the list where not echoed
back to me and only one has been answered.

Mike

------------------------------------------------------------------------------
Michael H. Moran | Standard Disclaimer: The content of
[email protected] | this posting is independent of
Commercial Performance | official IBM position.
IBM Corporation, Austin, Texas |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <14886-10@lion>;
Thu, 17 Aug 1995 22:40:59 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 19:13:53 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa15188;
17 Aug 95 20:13 +0100
Received: by gw.home.vix.com id AA16717; Thu, 17 Aug 95 06:19:02 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA16712; Thu, 17 Aug 95 06:19:00 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa07121;
17 Aug 95 13:19 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ad25177;
17 Aug 95 13:16 +0100
Received: from ig.co.uk by lion id <07486-0@lion>;
Thu, 17 Aug 1995 11:54:35 +0100
To: [email protected], [email protected]
Subject: Re: Debug output in Oracle DBD
Cc: [email protected]
Date: Thu, 17 Aug 1995 11:54:35 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Thomas Ganslandt <[email protected]>
>
> I am using the Oracle DBD 1.17 (with oraperl emulation) to access an oracle
> database through the world-wide-web. Sometimes, but not always (it seems
> to depend on how long the script takes to run), I get a debug message
> into my output:
>
> Bad free() ignored at [path of script] line 221.
>
> The given line contains the ora_logoff statement.
>
> Modifying Oraperl.pm as John Carlson ([email protected]) suggested on this
> list (setting $debug = 0) did not resolve the problem.
> Am I doing something wrong

No. It's an internal problem with DBI, DBD::Oracle or Oracle itself.

> or is there a way to shut the errormessage off

No. Not without relinking perl with a different malloc.

> or at least direct it not to stdout but to stderr, so it doesn't
> get displayed in the output?
>
It does get directed to stderr already (see line 280 in malloc.c of
Perl5.001m). Perhaps your stderr has been redirected to stdout.

It would be *very* helpful if you could try to send me a small
self-contained reproducable test case.

> Thanks,
>
> Thomas ([email protected])
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <14886-12@lion>;
Thu, 17 Aug 1995 22:41:21 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 19:40:31 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa26547;
17 Aug 95 20:40 +0100
Received: by gw.home.vix.com id AA13310; Thu, 17 Aug 95 05:24:57 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA13305; Thu, 17 Aug 95 05:24:56 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id HAA60927;
Thu, 17 Aug 1995 07:24:13 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA24280;
Thu, 17 Aug 1995 07:24:06 -0500
From: [email protected]
Message-Id: <[email protected]>
To: [email protected]
Cc: [email protected], [email protected], [email protected]
Subject: Re: Getting the DBD loaded...problem
In-Reply-To: Your message of Thu, 17 Aug 95 07:21:54 +0200. <[email protected]>
Date: Thu, 17 Aug 95 07:24:05 -0500

Andreas Koenig <[email protected]> writes:
|> >Can't call method "import" in empty package "DB2" at test2.pl line 4.
|>
|> Your DB2.pm file seems to need the
|>
|> package DB2;
|>
|> line.
|>
|> andreas

Just discovered something rather interesting. I had been assuming that
Tim's Oracle.pm or Oraperl.pm was doing a "use Oracle;" type call
to load that particular interface using some form of automagic. So
I went and looked ... wrong.

Neither module every does a "use Oracle;". Rather, Tim does a
$drh->install_driver('Oracle'), which is I change Oracle to DB2 works
on my system as does use DBD::DB2.

So that brings up the next question, is there any difference between
the two methods? Does the DBI need former as opposed to the latter?

Thanks
Mike
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <14886-13@lion>;
Thu, 17 Aug 1995 22:41:30 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 20:08:10 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa08896;
17 Aug 95 21:07 +0100
Received: by gw.home.vix.com id AA10762; Thu, 17 Aug 95 04:32:13 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA10757; Thu, 17 Aug 95 04:32:11 -0700
Message-Id: <[email protected]>
Received: from mip.lasc.lockheed.com by netserve.lasc.lockheed.com
with SMTP (1.37.109.4/16.2) id AA04656; Thu, 17 Aug 95 07:32:03 -0400
Received: by mip.lasc.lockheed.com (1.38.193.4/16.2) id AA15019;
Thu, 17 Aug 1995 07:32:19 -0400
Date: Thu, 17 Aug 1995 07:32:19 -0400
From: Terry Greenlaw <[email protected]>
To: [email protected]
Subject: Re: Help requested for ora_do behaviour testing

> From [email protected] Wed Aug 16 22:34 EDT 1995
> Received: from marble.litc.lockheed.com by mip.lasc.lockheed.com with SMTP
> (1.38.193.4/16.2) id AA09400; Wed, 16 Aug 1995 22:34:08 -0400
> Return-Path: <[email protected]>
> Received: from gw.home.vix.com (gw.home.vix.com [192.5.5.1]) by marble.litc.lockheed.com (8.6.12/8.6.9) with ESMTP id TAA16689 for <[email protected]>; Wed, 16 Aug 1995 19:37:23 -0700
> Received: by gw.home.vix.com id AA10629; Wed, 16 Aug 95 12:50:11 -0700
> X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
> Precedence: junk
> Sender: [email protected]
> Received: by gw.home.vix.com id AA10580; Wed, 16 Aug 95 12:50:06 -0700
> Received: from post.demon.co.uk by disperse.demon.co.uk id aa26495;
> 16 Aug 95 19:44 +0100
> Received: from ignite.demon.co.uk by post.demon.co.uk id ab07475;
> 16 Aug 95 19:42 +0100
> Received: from ig.co.uk by lion id <02214-0@lion>;
> Wed, 16 Aug 1995 19:44:38 +0100
> To: [email protected]
> Subject: Help requested for ora_do behaviour testing
> Date: Wed, 16 Aug 1995 19:44:38 +0100
> From: Tim Bunce <[email protected]>
> Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
> Message-Id: <[email protected]>
> Status: R

> I seem to recall that many months ago someone sent me a note saying that

> &ora_do($lda, "insert into foo values 1");

> was inserting the data twice. That's why I commented out the execute
> in ora_do. Now I'm not so sure about this (I never was really but never
> had the time to look into it).

> oraperl-2.4 ora_do seems to do: csr=ora_open; ora_close(csr); and ora_open
> always does an execute OCI call (unless bind vars are present).

> So, the question is, why doesn't the ora_do(...insert...) example above
> execute twice in oraperl-2.4? Or, perhaps it doesn't actually do it
> in either and the original message was mistaken.

> If anyone has some spare time I'd appreciate it if you could uncomment
> the execute line from ora_do in Oraperl.pm are try out things like
> insert and create table statements (where double execution is easy
> to spot).

> It doesn't matter which version of DBD:Oracle you use.

> Thanks in advance.

> Tim.

This was related to the delayed parse issue. I was definitely getting
double inserts before commenting out the execute in ora_do. We had
discussed the switching of the defflg var in oparse to non-zero to
delay the DDL until oexec. Seems like years ago ;-} Has Oracle cleaned
the bugs out of it yet?

Does oraperl-2.4 check the SQL statement to see if it is DDL? A
func = CDA.ft"; call right after oparse plus a compare of func
against the values in ocidem FT_* would be a good giveaway.

Terry Greenlaw (on-site @ Lockheed) Encompass Technologies
[email protected] [email protected]

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <14886-14@lion>;
Thu, 17 Aug 1995 22:41:42 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 20:36:01 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa21158;
17 Aug 95 21:35 +0100
Received: by gw.home.vix.com id AA20836; Thu, 17 Aug 95 07:28:38 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA20831; Thu, 17 Aug 95 07:28:33 -0700
Received: from bssmlsa.nerc-bas.ac.uk.nerc-bas.ac.uk
by bsfiles.nerc-bas.ac.uk (8.6.10/NERC-1.2(Solaris 2.x) id PAA20547;
Thu, 17 Aug 1995 15:28:29 +0100
Date: Thu, 17 Aug 1995 15:28:28 +0100
Message-Id: <[email protected]>
From: Andrew G Wood <[email protected]>
To: [email protected]
In-Reply-To: <[email protected]> ([email protected])
Subject: Re: Is the list having problems? Software change?
Reply-To: [email protected]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII

>>>>> "mhm" == mhm <[email protected]> writes:

> The last three items I have sent to the list where not echoed
> back to me and only one has been answered.

Me too .... only I just sent out one item to the list ;-)

Andy.
*
Dr Andy Wood **
Database Administrator ***
***
****
[email protected] ****
[email protected] ****
[email protected] ****
****
[email protected] ****
...*****.____
Post: .::______.:~'
****
British Antarctic Survey ****
High Cross ****
Madingley Road ****
Cambridge, CB3 0ET, UK ***
**
+44 (0) 1223 61188 **
*

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <15704-2@lion>;
Thu, 17 Aug 1995 23:03:10 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 21:53:44 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa24952;
17 Aug 95 22:51 +0100
Received: by gw.home.vix.com id AA18856; Thu, 17 Aug 95 06:52:08 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA18851; Thu, 17 Aug 95 06:52:05 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id IAA64605;
Thu, 17 Aug 1995 08:51:29 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA21214;
Thu, 17 Aug 1995 08:49:18 -0500
From: [email protected]
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected]
Subject: Re: question about prepare
In-Reply-To: Your message of Thu, 17 Aug 95 11:49:23 +0100. <[email protected]>
Date: Thu, 17 Aug 95 08:49:18 -0500

Tim Bunce <[email protected]> writes:
|>
|> > Now to the real question. In DB2, inorder to bind values to the
|> > prepared statement I need to know the expected type of the parameter.
|> > This is fairly simple if the parameter is going to be used in
|> > conjuction with a result column and basically impossible if used
|> > in any other
|> > context. So are we going to attempt to support the original
|> > definition or are we moving towards the current definition?
|> >
|> Most databases provide a way to describe the bind parameters in addition
|> to, and separate from, a way to describe output parameters.
|>
|> Are you saying that DB2 does not provide such a mechanism? That would
|> seem very strange.

No. Lets see if I can explain it.

int var1, var3;
char *var2;
struct type var4;

create table tab (
a decimal(3,2);
b smallint;
c timestamp;
d blob;
}

statement = "select a, b, c, d from tab where (a = :1 and b= :2) or
(c = :3 and d = :4);"

for each variable to bound prior to execution the function
requires that I specify the type of the incoming variable (ie, var[1,2,3,4]
and the type of its intended use.

The call looks like this:

SQLBindParameter(hstmt,ipar,fParmType,fCType,fSqlType,cbParamDef,ibScale,
rgbValue,cbValueMax,pcbValue);

hstmt - DB2 statement handle
ipar - integer parameter number
fParmType - Input, Output or Input-Output
fCType - C datatype of the paramter
fSqlType - sql data type of the parameter
cbParamDef - precision of the parameter marker's SqlType
ibScale - scale of parameter type (only for Decimal, Numeric, and
time stamp)
rgbValue - point to the actual variable use for substitution
cbValueMax - max lenght of rgbValue buffer if char or binary type
pcbValue - pointer to variable containing length of string info

The incoming variable type is easy because DB2 provides a default type
which forces translation if possible to the intended use type. My problem
is that since, I never get to see the actual table definition I don't
now what types of the columns are. It gets more complex when you start
trying to deal with SQL level functions, in which I have no way of
knowing what type of parameter is exepected short of writing and embedding
an SQL parser. :-(

The post execution binding is quite trivial compared to parameter binding.

Mike
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <15704-4@lion>;
Thu, 17 Aug 1995 23:03:16 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 22:02:44 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00335;
17 Aug 95 23:02 +0100
Received: by gw.home.vix.com id AA16635; Thu, 17 Aug 95 06:16:22 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA16630; Thu, 17 Aug 95 06:16:13 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id ab07149;
17 Aug 95 13:20 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab25295;
17 Aug 95 13:17 +0100
Received: from ig.co.uk by lion id <07667-0@lion>;
Thu, 17 Aug 1995 12:14:49 +0100
To: [email protected]
Subject: Re: Getting the DBD loaded...problem
Cc: [email protected]
Date: Thu, 17 Aug 1995 12:14:49 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Ok, I finally got enough code rewritten to make it worth trying to
> test DB2perl. The DBI (DBI-0.58) test shows success, but when I
> attempt to run the test program for DB2 I get the following error from perl:
>
> Can't call method "import" in empty package "DB2" at test2.pl line 4.
>
> I understand what the error message is trying to tell me, but I haven't
> got a clue as to why or how to fix it.
>
Er, why do you have a package called 'DB2' ?

If you're still stuck send me a copy of everything.

> Thanks.
> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18112-0@lion>;
Fri, 18 Aug 1995 07:00:56 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 22:35:56 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa15025;
17 Aug 95 23:35 +0100
Received: by gw.home.vix.com id AA27730; Thu, 17 Aug 95 01:51:55 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA27725; Thu, 17 Aug 95 01:51:51 -0700
Received: from bssmlsa.nerc-bas.ac.uk.nerc-bas.ac.uk
by bsfiles.nerc-bas.ac.uk (8.6.10/NERC-1.2(Solaris 2.x) id JAA18055;
Thu, 17 Aug 1995 09:51:38 +0100
Date: Thu, 17 Aug 1995 09:51:37 +0100
Message-Id: <[email protected]>
From: Andrew G Wood <[email protected]>
To: [email protected]
Cc: [email protected]
In-Reply-To: <[email protected]> (message from Tim Bunce on Wed, 16 Aug 1995 19:44:38 +0100)
Subject: Re: Help requested for ora_do behaviour testing
Reply-To: [email protected]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII

>>>>> "Tim" == Tim Bunce <[email protected]> writes:

> So, the question is, why doesn't the ora_do(...insert...)
> example above execute twice in oraperl-2.4? Or, perhaps it
> doesn't actually do it in either and the original message was
> mistaken.

> If anyone has some spare time I'd appreciate it if you could
> uncomment the execute line from ora_do in Oraperl.pm are try out
> things like insert and create table statements (where double
> execution is easy to spot).

> It doesn't matter which version of DBD:Oracle you use.

> Thanks in advance.

OK so here goes .... uncommented the line in Oraperl.pm .....

---------------------------------------------------------------------

# Oracle OCI will automatically execute DDL statements in prepare()!
# We must be carefull not to execute them again! This needs careful
# examination and thought.
# **** the following line was uncommented as a test
my $ret = $csr->execute;

---------------------------------------------------------------------

created the following test to add into Tim's test script ......

---------------------------------------------------------------------
...

&test2();

print "\nTest complete.\n";

exit 0;

sub test2 {
local($lda) = &ora_login($dbname, $dbuser, '')
|| die "ora_login: $ora_errno: $ora_errstr\n";

&ora_do($lda, "create table oraperl_test
( test1 char(1) not null,
test2 number not null)")
|| die "ora_do: $ora_errno: $ora_errstr\n";

&ora_do($lda, "insert into oraperl_test values
( 'A', 1 )")
|| die "ora_do: $ora_errno: $ora_errstr\n";

&ora_do($lda, "insert into oraperl_test values
( 'B', 2 )")
|| die "ora_do: $ora_errno: $ora_errstr\n";

local($csr) = &ora_open($lda,
"select test1, test2 from oraperl_test")
|| die "ora_open: $ora_errno: $ora_errstr\n";

while(@fields = &ora_fetch($csr)){
print "Fetched: "; print "@fields\n";
}

&ora_do($lda, "drop table oraperl_test")
|| die "ora_do: $ora_errno: $ora_errstr\n";

print "ora_logoff...\n";

&ora_logoff($lda) || warn "ora_logoff($lda): $ora_errno: $ora_errstr\n";

}

---------------------------------------------------------------------

The resuts of this as follows .....

% ora_test.pl

Oraperl test application $Revision: 1.14 $
DBD::Oracle driver installed as DBI::dr=HASH(0x100b73e0)
DBI::dr=HASH(0x100cada0) debug level set to 1

Oraperl Emulation Interface version 1.17
Oracle Driver 1.17
DBI Switch by Tim Bunce, version 1.39

Fetched: A 1
Fetched: B 2
ora_logoff...

Test complete.

---------------------------------------------------------------------

To me that seems to indicate that the extra execute line is not causing
any problems ..... but don't quote me on that :-)

Andy.
*
Dr Andy Wood **
Database Administrator ***
***
****
[email protected] ****
[email protected] ****
[email protected] ****
****
[email protected] ****
...*****.____
Post: .::______.:~'
****
British Antarctic Survey ****
High Cross ****
Madingley Road ****
Cambridge, CB3 0ET, UK ***
**
+44 (0) 1223 61188 **
*

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18112-7@lion>;
Fri, 18 Aug 1995 07:01:22 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 23:49:20 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa13720;
18 Aug 95 0:49 +0100
Received: by gw.home.vix.com id AA02850; Thu, 17 Aug 95 10:26:18 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA02837; Thu, 17 Aug 95 10:25:54 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa24703;
17 Aug 95 17:15 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa10033;
17 Aug 95 17:12 +0100
Received: from ig.co.uk by lion id <11026-0@lion>;
Thu, 17 Aug 1995 17:15:19 +0100
To: [email protected]
Subject: Re: hints file for Oracle
Cc: [email protected]
Date: Thu, 17 Aug 1995 17:15:19 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Alan Burlison <[email protected]>
>
> Tim,
>
> The following hints file is required for DBD::Oracle on svr4, you might
> like to add it to the next release :-)
>
> File: Oracle/hints/svr4.pl
>
> # Some SVR4 systems may need to link against -lc to pick up things like
> $att{LIBS} = [ '-lsocket -lnsl -lm -ldl -lc' ];
>
> This works fine with DBI v0.58 and DBD::Oracle v0.20.

Okay. Thanks.

> The only problem
> I found was that the file 'test.pl' in Oracle failed - I think due to
> the user/password being incorrect.
>
I ought to find a way to prompt for these at Makefile.PL time.

> A quick question - is it possible to install an event handler *before*
> opening a database - the default one upchucks stuff if the initial
> database login fails, but the $dbh->handler() method requires a
> database handle - so you can't install a handler until after the
> database has been opened. Also, the params list for the handler in
> dbispec.v06 takes 4 params, wheras the example in dbi.pm only takes 3 -
> I assume the version in dbi.pm is correct?
>
No, the spec is right. It's one of those things that's not been done yet.
That third arg is where you'd specify the handler.

> Alan Burlison [email protected]
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18112-9@lion>;
Fri, 18 Aug 1995 07:01:28 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 23:52:50 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa14858;
18 Aug 95 0:52 +0100
Received: by gw.home.vix.com id AA16026; Thu, 17 Aug 95 06:01:02 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA16014; Thu, 17 Aug 95 06:00:56 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa07117;
17 Aug 95 13:19 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac25177;
17 Aug 95 13:16 +0100
Received: from ig.co.uk by lion id <07457-0@lion>;
Thu, 17 Aug 1995 11:49:23 +0100
To: [email protected], [email protected]
Subject: Re: question about prepare
Date: Thu, 17 Aug 1995 11:49:23 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> The description for prepare is
>
> $sth = $dbh->prepare($statement[,%args])
>
> while the current implementation seems to be
>
> $sth = $dbh->prepare($dbh,$statement)
>
> or am I misinterpreting things?
>
The current is actually

$sth = $dbh->prepare($statement)

I'm ignoring the optional args for now. This also applies to some other
methods such as connect.

> Now to the real question. In DB2, inorder to bind values to the
> prepared statement I need to know the expected type of the parameter.
> This is fairly simple if the parameter is going to be used in conjuction
> with a result column and basically impossible if used in any other
> context. So are we going to attempt to support the original definition
> or are we moving towards the current definition?
>
Most databases provide a way to describe the bind parameters in addition
to, and separate from, a way to describe output parameters.

Are you saying that DB2 does not provide such a mechanism? That would
seem very strange.

> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18112-11@lion>;
Fri, 18 Aug 1995 07:01:38 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Thu, 17 Aug 95 23:55:21 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa15733;
18 Aug 95 0:55 +0100
Received: by gw.home.vix.com id AA19365; Thu, 17 Aug 95 07:01:41 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA19358; Thu, 17 Aug 95 07:01:36 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa09369;
17 Aug 95 13:53 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa02485;
17 Aug 95 13:50 +0100
Received: from ig.co.uk by lion id <08984-0@lion>;
Thu, 17 Aug 1995 13:53:26 +0100
To: [email protected], [email protected]
Subject: Re: Getting the DBD loaded...problem
Cc: [email protected]
Date: Thu, 17 Aug 1995 13:53:26 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Andreas Koenig <[email protected]> writes:
> |> >Can't call method "import" in empty package "DB2" at test2.pl line 4.
> |>
> |> Your DB2.pm file seems to need the
> |>
> |> package DB2;
> |>
> |> line.
> |>
> |> andreas
>
> Just discovered something rather interesting. I had been assuming that
> Tim's Oracle.pm or Oraperl.pm was doing a "use Oracle;" type call
> to load that particular interface using some form of automagic. So
> I went and looked

Always a good idea!

> ... wrong.
>
See! :-)

Be very wary of assuming anything. Follow the Oracle.* files closely.
Especially the next one I release!

> Neither module every does a "use Oracle;". Rather, Tim does a
> $drh->install_driver('Oracle'), which is I change Oracle to DB2 works
> on my system as does use DBD::DB2.
>
> So that brings up the next question, is there any difference between
> the two methods? Does the DBI need former as opposed to the latter?
>
The application *never* does a 'use DBD::*' only a 'require DBI'.

The DBI install_driver *must* be used and it's best to call it via:

DBI->connect(..., 'DB2');

> Thanks
> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18112-16@lion>;
Fri, 18 Aug 1995 07:02:03 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 00:44:48 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00619;
18 Aug 95 1:44 +0100
Received: by gw.home.vix.com id AA02017; Thu, 17 Aug 95 10:11:32 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA02010; Thu, 17 Aug 95 10:11:29 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa24276;
17 Aug 95 17:11 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab09061;
17 Aug 95 17:08 +0100
Received: from ig.co.uk by lion id <10707-0@lion>;
Thu, 17 Aug 1995 16:51:18 +0100
To: [email protected]
Mmdf-Warning: Unable to confirm address in preceding line at
disperse.demon.co.uk
Subject: Re: question about prepare
Cc: [email protected]
Date: Thu, 17 Aug 1995 16:51:18 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim Bunce <[email protected]> writes:
> |>
> |> > Now to the real question. In DB2, inorder to bind values to the
> |> > prepared statement I need to know the expected type of the parameter.
> |> > This is fairly simple if the parameter is going to be used in
> |> > conjuction with a result column and basically impossible if used
> |> > in any other
> |> > context. So are we going to attempt to support the original
> |> > definition or are we moving towards the current definition?
> |> >
> |> Most databases provide a way to describe the bind parameters in addition
> |> to, and separate from, a way to describe output parameters.
> |>
> |> Are you saying that DB2 does not provide such a mechanism? That would
> |> seem very strange.
>
> No. Lets see if I can explain it.
>
> int var1, var3;
> char *var2;
> struct type var4;
>
> create table tab (
> a decimal(3,2);
> b smallint;
> c timestamp;
> d blob;
> }
>
> statement = "select a, b, c, d from tab where (a = :1 and b= :2) or
> (c = :3 and d = :4);"
>
> for each variable to bound prior to execution the function
> requires that I specify the type of the incoming variable (ie, var[1,2,3,4]
> and the type of its intended use.
>
> The call looks like this:
>
> SQLBindParameter(hstmt,ipar,fParmType,fCType,fSqlType,cbParamDef,ibScale,
> rgbValue,cbValueMax,pcbValue);
>
> hstmt - DB2 statement handle
> ipar - integer parameter number
> fParmType - Input, Output or Input-Output
> fCType - C datatype of the paramter
> fSqlType - sql data type of the parameter
> cbParamDef - precision of the parameter marker's SqlType
> ibScale - scale of parameter type (only for Decimal, Numeric, and
> time stamp)
> rgbValue - point to the actual variable use for substitution
> cbValueMax - max lenght of rgbValue buffer if char or binary type
> pcbValue - pointer to variable containing length of string info
>
> The incoming variable type is easy because DB2 provides a default type
> which forces translation if possible to the intended use type. My problem
> is that since, I never get to see the actual table definition I don't
> now what types of the columns are. It gets more complex when you start
> trying to deal with SQL level functions, in which I have no way of
> knowing what type of parameter is exepected short of writing and embedding
> an SQL parser. :-(
>
> The post execution binding is quite trivial compared to parameter binding.
>
Scour your manuals looking for something like a GetStmtAttr or
DescribeParam function.

Other people using the DB2 API must have had the same problem.

Ask IBM :-)

> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18112-22@lion>;
Fri, 18 Aug 1995 07:02:59 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 01:08:14 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06357;
18 Aug 95 2:08 +0100
Received: by gw.home.vix.com id AA09259; Thu, 17 Aug 95 11:49:41 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA09253; Thu, 17 Aug 95 11:49:40 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id NAA20363;
Thu, 17 Aug 1995 13:49:35 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA21076;
Thu, 17 Aug 1995 13:49:28 -0500
From: [email protected]
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected], [email protected]
Subject: Re: question about prepare
In-Reply-To: Your message of Thu, 17 Aug 95 16:51:18 +0100. <[email protected]>
Date: Thu, 17 Aug 95 13:49:28 -0500

Tim Bunce <[email protected]> writes:
|> Scour your manuals looking for something like a GetStmtAttr or
|> DescribeParam function.
|>
|> Other people using the DB2 API must have had the same problem.
|>
|> Ask IBM :-)
|>

Already did. The answer was "The function you are looking for
is SQLDescribeParam, which is not in plan to be supported".

Translation, the old f*rts in charge still have the mainframe
batch job is the only thing mentality.

Any other ideas?

Mike
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18418-2@lion>;
Fri, 18 Aug 1995 07:03:30 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 02:10:41 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa19364;
18 Aug 95 3:10 +0100
Received: by gw.home.vix.com id AA24862; Thu, 17 Aug 95 15:35:54 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA24856; Thu, 17 Aug 95 15:35:39 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa23069;
17 Aug 95 23:29 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa13501;
17 Aug 95 23:26 +0100
Received: from ig.co.uk by lion id <15625-0@lion>;
Thu, 17 Aug 1995 22:55:11 +0100
To: [email protected], [email protected]
Subject: Re: Help requested for ora_do behaviour testing
Date: Thu, 17 Aug 1995 22:55:11 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Terry Greenlaw <[email protected]>
>
> > I seem to recall that many months ago someone sent me a note saying that
>
> > &ora_do($lda, "insert into foo values 1");
>
> > was inserting the data twice. That's why I commented out the execute
> > in ora_do. Now I'm not so sure about this (I never was really but never
> > had the time to look into it).
>
> > oraperl-2.4 ora_do seems to do: csr=ora_open; ora_close(csr); and ora_open
> > always does an execute OCI call (unless bind vars are present).
>
> > So, the question is, why doesn't the ora_do(...insert...) example above
> > execute twice in oraperl-2.4? Or, perhaps it doesn't actually do it
> > in either and the original message was mistaken.
>
> > If anyone has some spare time I'd appreciate it if you could uncomment
> > the execute line from ora_do in Oraperl.pm are try out things like
> > insert and create table statements (where double execution is easy
> > to spot).
>
> This was related to the delayed parse issue.

We are presuming it is/was (I'm wearing my deerstalker hat now :-).

> I was definitely getting
> double inserts before commenting out the execute in ora_do.

And once commented out you only got one?

Could you experiment with the new release (0.22) and try changing
the values of oparse_defer in dbdimp.c (any anything else that comes
to mind).

> We had discussed the switching of the defflg var in oparse to non-zero to
> to delay the DDL until oexec.

> Seems like years ago ;-}

It was awhile back. I've just avoided the issue. But now people are
complaining that ora_do() does not work at all for them.

We need to get to the bottom of this soon.

> Has Oracle cleaned the bugs out of it yet?

What version of Oracle are you using?

> Does oraperl-2.4 check the SQL statement to see if it is DDL?
> A func = CDA.ft"; call right after oparse plus a compare of func
> against the values in ocidem FT_* would be a good giveaway.
>
Nope. It doesn't.

> Terry Greenlaw (on-site @ Lockheed) Encompass Technologies
> [email protected] [email protected]
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18418-4@lion>;
Fri, 18 Aug 1995 07:03:42 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 03:12:21 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00804;
18 Aug 95 4:11 +0100
Received: by gw.home.vix.com id AA28707; Thu, 17 Aug 95 16:37:17 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA28701; Thu, 17 Aug 95 16:37:14 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa20025;
17 Aug 95 22:39 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa06883;
17 Aug 95 22:37 +0100
Received: from ig.co.uk by lion id <14882-0@lion>;
Thu, 17 Aug 1995 22:39:06 +0100
To: [email protected]
Subject: THE BIG ONE! Announce DBI-0.61 and DBD-Oracle-0.22
Date: Thu, 17 Aug 1995 22:39:06 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

Now the fun really starts!

I hereby declare that anyone interested in writing a DBI database driver
can reliably base their work on this release of these two modules.

There will, of course, be changes in future releases but nothing as
dramatic as the last two!

Anyone who has started a driver is advised to copy the new DBD::Oracle
module files and merge their work in rather than trying to merge in
this release of DBD::Oracle into their code. Trust me on this!

For those too young to remember, the DBperl effort will have its
third birthday on September 29th. It would be kind'a nice to have
something to show for it after three years :-)

I hope to have an Ingres driver working soon.

Tim.

--- Change notes ---

Changes in DBI 0.60, 17 Aug 1995

This release has significant code changes but much less
dramatic than the previous release. The new implementors data
handling mechanism has matured significantly (don't be put off
by all the struct typedefs in DBIXS.h, there's just to make it
easier for drivers while keeping things type-safe).

The DBI now includes two new methods:

do $dbh->do($statement)

This method prepares, executes and finishes a statement. It is
designed to be used for executing one-off non-select statements
where there is no benefit in reusing a prepared statement handle.

fetch $array_ref = $sth->fetch;

This method is the new 'lowest-level' row fetching method. The
previous @row = $sth->fetchrow method now defaults to calling
the fetch method and expanding the returned array reference.

The DBI now provides fallback attribute FETCH and STORE functions
which drivers should call if they don't recognise an attribute.

THIS RELEASE IS A GOOD STARTING POINT FOR DRIVER DEVELOPERS!
Study DBIXS.h from the DBI and Oracle.xs etc from DBD::Oracle.
There will be further changes in the interface but nothing
as dramatic as these last two releases! (I hope :-)

Changes in DBD::Oracle 0.22, 17 Aug 1995:

Much more maturity, in line with the required DBI-0.60 release.
The .xs file is now very 'clean'. It's an excellent base for
developing other drivers. See notes in the DBI Changes file.

The oraperl &ora_do should be working now (along with $sth->do()).
It does not leak (as far as I can tell).

This release is stable enough for general use again (like 0.20).
Please test heavily.

End.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18112-30@lion>;
Fri, 18 Aug 1995 07:03:56 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 04:08:01 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa10314;
18 Aug 95 5:07 +0100
Received: by gw.home.vix.com id AA28210; Thu, 17 Aug 95 16:26:34 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA28205; Thu, 17 Aug 95 16:26:32 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa23072;
17 Aug 95 23:29 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab13501;
17 Aug 95 23:26 +0100
Received: from ig.co.uk by lion id <15639-0@lion>;
Thu, 17 Aug 1995 22:58:12 +0100
To: [email protected], [email protected]
Subject: Re: Getting the DBD loaded...problem
Cc: [email protected]
Date: Thu, 17 Aug 1995 22:58:12 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> The main DBI handler is loaded with 'use DBI'. The DBDs are loaded by doing
> the install_driver function because I may want to, eg, load an Oracle and
> mSQL driver, then take data from one and pass it straight into the other.
>
> If I did a 'use Oracle' if wouldn't specifically get a handle ( $drh ) to
> use simultaneously with another $drh, eg:
>
> use DBI;
> $oradrh = DBI->install_driver('Oracle');
> $msqldrh = DBI->install_driver('mSQL');
>
> $oradbh = connect.....;
> $msqldbh = connect....;
>
> and so on.
>
True, but you'd rarely want to fiddle with driver handles.
Best just to use connect and specify the driver there:

$oradbh = DBI->connect(..., 'Oracle');
$msqldbh = DBI->connect(..., 'mSQL');

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18418-7@lion>;
Fri, 18 Aug 1995 07:04:04 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 04:19:59 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa12420;
18 Aug 95 5:19 +0100
Received: by gw.home.vix.com id AA24874; Thu, 17 Aug 95 15:36:11 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA24869; Thu, 17 Aug 95 15:36:08 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa23074;
17 Aug 95 23:29 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac13501;
17 Aug 95 23:26 +0100
Received: from ig.co.uk by lion id <15651-0@lion>;
Thu, 17 Aug 1995 22:59:37 +0100
To: [email protected]
Subject: Re: hints file for Oracle
Cc: [email protected]
Date: Thu, 17 Aug 1995 22:59:37 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Alan Burlison <[email protected]>
>
> In-Reply-To: <[email protected]>
> Tim,
>
> > > A quick question - is it possible to install an event handler
> > > *before* opening a database - the default one upchucks stuff if the
> > > initial database login fails, but the $dbh->handler() method
> > > requires a database handle - so you can't install a handler until
> > > after the database has been opened. Also, the params list for the
> > > handler in dbispec.v06 takes 4 params, wheras the example in dbi.pm
> > > only takes 3 - I assume the version in dbi.pm is correct?
> > >
> > No, the spec is right. It's one of those things that's not been done
> > yet. That third arg is where you'd specify the handler.
>
> Does this mean I should forget about setting a handler for the present?

Yes, for a little while at least.

> I've tried calling $dbh->handler(), but I just get a run-time error on
> the method lookup as it does not seem to exist.
>
> Alan Burlison [email protected]
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18418-8@lion>;
Fri, 18 Aug 1995 07:04:10 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 04:29:06 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa14025;
18 Aug 95 5:28 +0100
Received: by gw.home.vix.com id AA00935; Thu, 17 Aug 95 17:10:30 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA00928; Thu, 17 Aug 95 17:10:28 -0700
Received: from behemoth.illustra.com (behemoth.illustra.com)
by angular.illustra.com (4.1/smail2.5/06-10-94/M) id AA07747;
Thu, 17 Aug 95 17:07:48 PDT
Received: from jones.illustra.com
by behemoth.illustra.com (4.1/smail2.5/06-10-94/S) id AA04793;
Thu, 17 Aug 95 17:10:43 PDT
Received: by jones.illustra.com (931110.SGI/930416.SGI)
for @behemoth.illustra.com:[email protected] id AA04826;
Thu, 17 Aug 95 17:09:22 -0700
From: Martin Siegenthaler <[email protected]>
Message-Id: <[email protected]>
Date: Thu, 17 Aug 1995 17:09:20 -0700
In-Reply-To: [email protected] "Re: question about prepare" (Aug 17, 1:49pm)
References: <[email protected]>
X-Mailer: Z-Mail (3.1.0 22feb94 MediaMail)
To: [email protected]
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0

unsubscribe

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <18418-9@lion>;
Fri, 18 Aug 1995 07:04:18 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 04:36:26 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa15171;
18 Aug 95 5:35 +0100
Received: by gw.home.vix.com id AA23093; Thu, 17 Aug 95 15:07:59 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA23087; Thu, 17 Aug 95 15:07:56 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id ab20227;
17 Aug 95 22:43 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa07351;
17 Aug 95 22:40 +0100
Received: from ig.co.uk by lion id <15543-0@lion>;
Thu, 17 Aug 1995 22:42:40 +0100
To: [email protected]
Subject: Re: question about prepare
Cc: [email protected]
Date: Thu, 17 Aug 1995 22:42:40 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim Bunce <[email protected]> writes:
> |> Scour your manuals looking for something like a GetStmtAttr or
> |> DescribeParam function.
> |>
> |> Other people using the DB2 API must have had the same problem.
> |>
> |> Ask IBM :-)
> |>
>
> Already did. The answer was "The function you are looking for
> is SQLDescribeParam, which is not in plan to be supported".
>
> Translation, the old f*rts in charge still have the mainframe
> batch job is the only thing mentality.
>
Sigh, so much for IBM's changed approach to things.

I trust you've pointed out that their API is therefore broken.

> Any other ideas?
>
Would the oraperl approach of always using strings work?
(eg let the database convert as required).

> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20184-0@lion>;
Fri, 18 Aug 1995 10:25:09 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 08:37:14 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa16344;
18 Aug 95 9:36 +0100
Received: by gw.home.vix.com id AA14271; Thu, 17 Aug 95 21:07:05 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA14264; Thu, 17 Aug 95 21:07:04 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id XAA173690;
Thu, 17 Aug 1995 23:07:03 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA26316;
Thu, 17 Aug 1995 23:07:01 -0500
From: [email protected]
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected]
Subject: Re: question about prepare
In-Reply-To: Your message of Thu, 17 Aug 95 22:42:40 +0100. <[email protected]>
Date: Thu, 17 Aug 95 23:07:00 -0500

Tim Bunce <[email protected]> writes:
|> > Already did. The answer was "The function you are looking for
|> > is SQLDescribeParam, which is not in plan to be supported".
|> >
|> > Translation, the old f*rts in charge still have the mainframe
|> > batch job is the only thing mentality.
|> >
|> Sigh, so much for IBM's changed approach to things.
|>
|> I trust you've pointed out that their API is therefore broken.

Actually, from what I can tell it is only broken as far as MicroSoft's
ODBC specification, but not broken as far as the ISO and ANSI SQL
standards are concerned. Now, if you could provide me with a reference
or references which indicates otherwise ... ;-)

Well, the good and the bad of it is that I'm providing real life
ammunition for the product architect to push the powers that be toward
doing the right thing.

|> > Any other ideas?
|> >
|> Would the oraperl approach of always using strings work?
|> (eg let the database convert as required).

It would work to a point. Unfortunately, there are a lot of type
conversions which are not allowed because none of the specs explicitly
say they are to be allowed. Not to mention SQL functions which get
eval'ed prior to making the trip to the engine, thus the type has to be
correct at the function call.

The only things that come to mind are:

a) Embedding a lexical scanner for SQL which knows about all the
functions. While this would pretty much solve all the problems the
performance penalties and maintainance problems make it unfeasable.

b) Doing everything in char and designing some method of making
special cases of everthing that makes it break. Yuck.

c) Mutating the parameters to the exec call. Instead of doing
execute($sth,@exec_values) where each item in the array is a single
value, allowing each item to be a list containg the value and the
type. The question then becomes is it really a list or is it a
value<symbol>type string.

Well, which of the above sounds less horrid?

Mike
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <28913-7@lion>;
Fri, 18 Aug 1995 19:51:21 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 18:32:50 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11787;
18 Aug 95 19:32 +0100
Received: by gw.home.vix.com id AA04136; Fri, 18 Aug 95 02:26:21 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA04131; Fri, 18 Aug 95 02:26:19 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa18496;
18 Aug 95 10:24 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa20272;
18 Aug 95 10:21 +0100
Received: from ig.co.uk by lion id <20123-0@lion>;
Fri, 18 Aug 1995 10:22:59 +0100
To: [email protected]
Subject: Re: question about prepare
Cc: [email protected]
Date: Fri, 18 Aug 1995 10:22:59 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim Bunce <[email protected]> writes:
> |> > Already did. The answer was "The function you are looking for
> |> > is SQLDescribeParam, which is not in plan to be supported".
> |> >
> |> > Translation, the old f*rts in charge still have the mainframe
> |> > batch job is the only thing mentality.
> |> >
> |> Sigh, so much for IBM's changed approach to things.
> |>
> |> I trust you've pointed out that their API is therefore broken.
>
> Actually, from what I can tell it is only broken as far as MicroSoft's
> ODBC specification, but not broken as far as the ISO and ANSI SQL
> standards are concerned. Now, if you could provide me with a reference
> or references which indicates otherwise ... ;-)
>
The GetStmtAttr() function is defined by the X/Open + SQL Access Group spec

Data Management: SQL Call Level Interface (CLI)

The version I have is a little dated. I'm awaiting a newly released one.

> Well, the good and the bad of it is that I'm providing real life
> ammunition for the product architect to push the powers that be toward
> doing the right thing.
>
> |> > Any other ideas?
> |> >
> |> Would the oraperl approach of always using strings work?
> |> (eg let the database convert as required).
>
> It would work to a point. Unfortunately, there are a lot of type
> conversions which are not allowed because none of the specs explicitly
> say they are to be allowed. Not to mention SQL functions which get
> eval'ed prior to making the trip to the engine, thus the type has to be
> correct at the function call.
>
> The only things that come to mind are:
>
> a) Embedding a lexical scanner for SQL which knows about all the
> functions. While this would pretty much solve all the problems the
> performance penalties and maintainance problems make it unfeasable.
>
Agreed.

> b) Doing everything in char and designing some method of making
> special cases of everthing that makes it break. Yuck.
>
Yuck but a worthwhile short term get-it-working measure.

> c) Mutating the parameters to the exec call. Instead of doing
> execute($sth,@exec_values) where each item in the array is a single
> value, allowing each item to be a list containg the value and the
> type. The question then becomes is it really a list or is it a
> value<symbol>type string.
>
Eeek. Don't mess with the DBI spec. The interface definition (such as
it is) is sacrosanct.

> Well, which of the above sounds less horrid?
>
The 'right' approach is probably to use the recently discussed
bind_param call. Something like:

$sth->bind_param($param_index_or_name, \$value [, \%attribs]);

where \%attribs would allow you to specify type. I'm not settled
on the interface spec yet but it should be something like this.

I'd suggest going with (b) in the short term till I/we agree a
spec and DBI support for some form of bind_param call.

> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <29565-3@lion>;
Fri, 18 Aug 1995 20:21:39 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 19:21:06 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa01649;
18 Aug 95 20:20 +0100
Received: by gw.home.vix.com id AA22278; Fri, 18 Aug 95 07:23:39 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA22273; Fri, 18 Aug 95 07:23:36 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id PAA08236;
Fri, 18 Aug 1995 15:20:03 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id PAA06956; Fri, 18 Aug 1995 15:25:32 +0100
Message-Id: <[email protected]>
Subject: Re: THE BIG ONE! Announce DBI-0.61 and DBD-Oracle-0.22
To: Tim Bunce <[email protected]>
Date: Fri, 18 Aug 1995 15:25:31 +0100 (BST)
Cc: [email protected]
In-Reply-To: <[email protected]> from "Tim Bunce" at Aug 18, 95 01:56:15 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 1666

> > > For those too young to remember, the DBperl effort will have its
> > > third birthday on September 29th. It would be kind'a nice to have
> > > something to show for it after three years :-)
> >
> > Hey, do you think we ought to arrange a party?........
> >
> :-)

Well, if anyone in Scotland is interested!....Drop me a line!

> > Ditto Informix. I found that moving all the vendor code into dbdimp.c
> > and leaving vendor.xs as a generic appeared to let preprocessors work fairly
> > happilly along with xsubpp too.......Just a thought.....
> >
> That reminds me...

Oh no..........*covers head*

> I forgot to say that it's very important *NOT* to change your copy of
> Oracle.xs in any way other than changing all instances of 'Oracle' into
> 'vendor' (as appropriate). That also applies to the prototypes of the
> dbd_xx_*() functions.
>
> If you (that is, any driver developer) feels a need to change Oracle.xs
> in any other way please talk to me first.
>
> This is all part of a long term strategy.
>
> In the short term I forsee a period of driver development followed by a
> period of consolidating the DBI spec around the work we have done and
> planning our way forward.
>
> The recent publication of the updated X/Open specs and Microsoft's
> statements that ODBC 3.0 will comply with it gives us a clear target.

Sounds good. Fortunatly, you seem to have done the moving of vendor code out
of the vendor.xs file before me which is what I'd been occassionally nagging
you about anyway after seeing how trivial adding new drivers can be.

Here's to a prosperous future on this..!

> Tim.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09987-6@lion>;
Mon, 21 Aug 1995 07:01:50 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 19:57:27 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa16076;
18 Aug 95 20:56 +0100
Received: by gw.home.vix.com id AA21806; Fri, 18 Aug 95 07:18:26 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA21795; Fri, 18 Aug 95 07:18:21 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa15176;
18 Aug 95 15:17 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa06800;
18 Aug 95 15:14 +0100
Received: from ig.co.uk by lion id <22429-0@lion>;
Fri, 18 Aug 1995 13:56:15 +0100
To: [email protected]
Subject: Re: THE BIG ONE! Announce DBI-0.61 and DBD-Oracle-0.22
Cc: [email protected]
Date: Fri, 18 Aug 1995 13:56:15 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> > Now the fun really starts!
> >
> > I hereby declare that anyone interested in writing a DBI database driver
> > can reliably base their work on this release of these two modules.
> >
> > There will, of course, be changes in future releases but nothing as
> > dramatic as the last two!
> >
> > Anyone who has started a driver is advised to copy the new DBD::Oracle
> > module files and merge their work in rather than trying to merge in
> > this release of DBD::Oracle into their code. Trust me on this!
>
> Right, I'm planning to update the DBD::mSQL driver onto the new DBI
> at the beginning of next week, hopefully......
>
Great.

> > For those too young to remember, the DBperl effort will have its
> > third birthday on September 29th. It would be kind'a nice to have
> > something to show for it after three years :-)
>
> Hey, do you think we ought to arrange a party?........
>
:-)

> > I hope to have an Ingres driver working soon.
>
> Ditto Informix. I found that moving all the vendor code into dbdimp.c
> and leaving vendor.xs as a generic appeared to let preprocessors work fairly
> happilly along with xsubpp too.......Just a thought.....
>
> 8-)
>
That reminds me...

I forgot to say that it's very important *NOT* to change your copy of
Oracle.xs in any way other than changing all instances of 'Oracle' into
'vendor' (as appropriate). That also applies to the prototypes of the
dbd_xx_*() functions.

If you (that is, any driver developer) feels a need to change Oracle.xs
in any other way please talk to me first.

This is all part of a long term strategy.

In the short term I forsee a period of driver development followed by a
period of consolidating the DBI spec around the work we have done and
planning our way forward.

The recent publication of the updated X/Open specs and Microsoft's
statements that ODBC 3.0 will comply with it gives us a clear target.

> Alligator Descartes |
> [email protected] |
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09987-10@lion>;
Mon, 21 Aug 1995 07:02:16 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 20:31:09 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa29519;
18 Aug 95 21:30 +0100
Received: by gw.home.vix.com id AA23838; Fri, 18 Aug 95 00:00:56 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA23833; Fri, 18 Aug 95 00:00:54 -0700
Received: (from kstock@localhost) by nova.auspex.fr (8.6.9/8.6.9) id JAA00609;
Robinson France, Fri, 18 Aug 1995 09:02:09 +0200
Date: Fri, 18 Aug 1995 09:02:09 +0200
From: Kevin Stock <[email protected]>
Message-Id: <[email protected]>
To: [email protected], [email protected]
Subject: Re: Help requested for ora_do behaviour testing

Hi,

This is to clarify exactly how Oraperl handles &ora_do().

> > I seem to recall that many months ago someone sent me a note saying that
>
> > &ora_do($lda, "insert into foo values 1");
>
> > was inserting the data twice. That's why I commented out the execute
> > in ora_do. Now I'm not so sure about this (I never was really but never
> > had the time to look into it).
>
> > oraperl-2.4 ora_do seems to do: csr=ora_open; ora_close(csr); and ora_open
> > always does an execute OCI call (unless bind vars are present).
>
> > So, the question is, why doesn't the ora_do(...insert...) example above
> > execute twice in oraperl-2.4? Or, perhaps it doesn't actually do it
> > in either and the original message was mistaken.
>
> > If anyone has some spare time I'd appreciate it if you could uncomment
> > the execute line from ora_do in Oraperl.pm are try out things like
> > insert and create table statements (where double execution is easy
> > to spot).
>
> > It doesn't matter which version of DBD:Oracle you use.
>
> > Thanks in advance.
>
> > Tim.
>
>
> This was related to the delayed parse issue. I was definitely getting
> double inserts before commenting out the execute in ora_do. We had
> discussed the switching of the defflg var in oparse to non-zero to
> delay the DDL until oexec. Seems like years ago ;-} Has Oracle cleaned
> the bugs out of it yet?
>
> Does oraperl-2.4 check the SQL statement to see if it is DDL? A
> func = CDA.ft"; call right after oparse plus a compare of func
> against the values in ocidem FT_* would be a good giveaway.

I deliberately kept Oraperl's "knowledge" of SQL to a minimum. The
only thing that it checks for is substitution variables, which it
requires to be in order so as to simplify binding to them later.

In Oraperl, the task of processing a statement is split up as follows:

&ora_open()
count substitution variables
open a cursor (OCI: oopen)
parse the statement (OCI: osql3)
if there are no substitution variables
execute the statement (OCI: oexec)
count the number of return values (OCI: odsc)
set up a data cache to receive returned values (OCI: odsc)

&ora_bind()
bind values to variables (OCI: obndrn)
execute the statement (OCI: oexec)

&ora_fetch()
fetch data returned (OCI: ofen)

&ora_close()
close the cursor (OCI: oclose)
discard the data cache

A DDL statement evidently contains no substitution variables, so the
statement will be directly executed by &ora_open().

As you said, &ora_do() is just a wrapper that is equivalent to

&ora_close(&ora_open(...));

with a little error checking on the way. In a sensible case, with no
substitution variables and no values to return (such as a DDL statement
or the example at the start of this message), the sequence of calls is
therefore:

count substitution variables (= 0)
open a cursor (OCI: oopen)
parse the statement (OCI: osql3)
execute the statement (OCI: oexec)
count the number of return values (= 0) (OCI: odsc)
close the cursor (OCI: oclose)

I don't see where in that sequence it's possible for a statement of any
kind to be executed twice.

Supposing someone is foolish enough to call &ora_d() with a statement
which isn't suitable, the result will be the following:

substitution variables the statement will not be executed
return values a data cache will be allocated and
then discarded

Hope that helps to clarify the issue.

Kevin
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09987-18@lion>;
Mon, 21 Aug 1995 07:03:49 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 21:15:45 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17054;
18 Aug 95 22:15 +0100
Received: by gw.home.vix.com id AA13719; Fri, 18 Aug 95 05:16:07 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA13714; Fri, 18 Aug 95 05:16:04 -0700
Received: from barbera.inferentia.it (barbera.inferentia.it [151.99.220.61])
by relay.smau.it (8.6.12/8.6.12) with SMTP id OAA20036
for <[email protected]>; Fri, 18 Aug 1995 14:14:54 +0200
Posted-Date: Fri, 18 Aug 1995 14:14:54 +0200
Received: by barbera.inferentia.it with Microsoft Mail
id <[email protected]>; Fri, 18 Aug 95 14:22:18 PDT
From: Shari <[email protected]>
To: "'[email protected]'" <[email protected]>
Subject: DBD::Oracle and DDL statements
Date: Fri, 18 Aug 95 14:24:00 PDT
Message-Id: <[email protected]>
Encoding: 40 TEXT
X-Mailer: Microsoft Mail V3.0

In the DBI-Oracle 0.22 package Todo file Tim Bounce writes:

>Careful! Since your message I have updated the ora_do comment to read:
>
> my $csr = $lda->prepare($stmt) or return undef;
>
> # Oracle OCI will automatically execute DDL statements in prepare()!
> # We must be carefull not to execute them again! This needs careful
> # examination and thought.
># my $ret = $csr->execute;
>
>I've not yet got to the bottom of this. I need a way to tell if oparse()
>has actually executed the statement. Do any OCI gurus out there know of
>a way? (Sometime I'll get around to asking Oracle.)

You may find the struct cda "ft" field useful.
It should countain the "Oracle function code" and it is valid (to the
very limited tests I have done):

(assuming non-deferred mode)
1) if the oparse() was a DML/session control/transaction control/system
control, it is valid if cda.rc == 0
2) if the oparse() was a DDL, it is valid even if the rc != 0.

So, if cda.rc == 0 after the oparse, you can inspect cda.ft and determine
the statement. Tabel 2-1 on page 2-6 of the OCI 7.1 Reference Manual
lists che SQL function codes. From a simple table you could detemine the
statement and, if defer was false, avoid the exec if it was DDL.
Of course, a more general solution would be deferred mode, but I
understand that would complicate matters a lot in the code (it is
feasible in an application, because an error will generally determine a
definitive action in whatever point of the program it happens).

I have made only simple tests on AIX 3.2.5 and DEC OSF/1 (Digital Unix
:-) 3.2 using Oracle 7.1.6.2.
If nobody else out there is able to test more thoroughly, I'll do it but
it will take some time.

Davide Migliavacca ([email protected])
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09987-20@lion>;
Mon, 21 Aug 1995 07:04:16 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 21:15:55 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa16896;
18 Aug 95 22:14 +0100
Received: by gw.home.vix.com id AA10869; Fri, 18 Aug 95 12:06:51 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA10862; Fri, 18 Aug 95 12:06:49 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id ab10998;
18 Aug 95 20:02 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa22526;
18 Aug 95 19:59 +0100
Received: from ig.co.uk by lion id <29408-0@lion>;
Fri, 18 Aug 1995 20:02:09 +0100
To: [email protected]
Subject: Minor release: DBD-Oracle-0.23
Cc: [email protected]
Date: Fri, 18 Aug 1995 20:02:09 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

A few loose ends:

Changes in DBD::Oracle 0.23, 18 Aug 1995:

Fixed Oraperl.pm debug which was left on by default by accident.
Added small patches from [email protected] for
longs (dbtype_is_long, dbd_describe and dbd_st_readblob).
Added svr4 hints from Alan Burlison <[email protected]>
Changed dbd_describe to return true for success (to be consistent).

Available from the usual place.

Davide, can you check that the way I've implemented the changes your
patches implied suits you.

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <10541-0@lion>;
Mon, 21 Aug 1995 07:05:12 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 21:30:36 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa22957;
18 Aug 95 22:30 +0100
Received: by gw.home.vix.com id AA07170; Fri, 18 Aug 95 11:07:38 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA07165; Fri, 18 Aug 95 11:07:33 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa05763;
18 Aug 95 18:55 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa12333;
18 Aug 95 18:53 +0100
Received: from ig.co.uk by lion id <28190-0@lion>;
Fri, 18 Aug 1995 18:48:42 +0100
To: [email protected]
Subject: Re: Help requested for ora_do behaviour testing
Cc: [email protected]
Date: Fri, 18 Aug 1995 18:48:42 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Andrew G Wood <[email protected]>
>
> [...]
>
> To me that seems to indicate that the extra execute line is not causing
> any problems ..... but don't quote me on that :-)
>
Ooops, just did!

:-)

Thanks.

> Andy.
> *
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <10541-1@lion>;
Mon, 21 Aug 1995 07:05:26 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 21:31:09 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa23187;
18 Aug 95 22:30 +0100
Received: by gw.home.vix.com id AA17330; Fri, 18 Aug 95 06:07:18 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA17325; Fri, 18 Aug 95 06:07:16 -0700
Received: from pobox.mot.com (pobox.mot.com [129.188.137.100])
by motgate.mot.com (8.6.11/8.6.10/MOT-3.7) with ESMTP id IAA13101
for <[email protected]>; Fri, 18 Aug 1995 08:07:13 -0500
Received: from comm.mot.com (il02dns1.comm.mot.com [145.1.3.2])
by pobox.mot.com (8.6.11/8.6.10/MOT-3.8) with SMTP id IAA08393
for <[email protected]>; Fri, 18 Aug 1995 08:07:12 -0500
Received: from tor.comm.mot.com by comm.mot.com (4.1/SMI-4.0) id AA11671;
Fri, 18 Aug 95 08:07:12 CDT
Received: from tdc12 by tor.comm.mot.com with SMTP (1.38.193.4/16.2) id AA09180;
Fri, 18 Aug 1995 09:02:21 -0400
Received-Date: Fri, 18 Aug 1995 09:02:21 -0400
Received: by tdc12 (4.1) id AA04257; Fri, 18 Aug 95 09:05:29 EDT
Date: Fri, 18 Aug 1995 09:05:28 -0400 (EDT)
From: "Rizwan Premji x.5616" <[email protected]>
Subject:
To: [email protected]
Message-Id: <Pine.3.89.9508180918.A3628-0100000@tdc12>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII

Need to unsubscribe for now,

Thanks!

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09987-30@lion>;
Mon, 21 Aug 1995 07:06:16 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 22:11:00 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa09125;
18 Aug 95 23:10 +0100
Received: by gw.home.vix.com id AA03373; Fri, 18 Aug 95 10:10:30 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA03368; Fri, 18 Aug 95 10:10:28 -0700
Received: from summer.illustra.com
by angular.illustra.com (4.1/smail2.5/06-10-94/M) id AA09855;
Fri, 18 Aug 95 10:07:51 PDT
Received: by summer.illustra.com (4.1/smail2.5/06-10-94/S) id AA26771;
Fri, 18 Aug 95 10:14:13 PDT
Date: Fri, 18 Aug 95 10:14:13 PDT
From: John Gaffney <[email protected]>
Message-Id: <[email protected]>
To: [email protected]
Subject: unsubscribe

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09987-31@lion>;
Mon, 21 Aug 1995 07:06:26 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 22:17:22 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11640;
18 Aug 95 23:17 +0100
Received: by gw.home.vix.com id AA07580; Fri, 18 Aug 95 11:11:35 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA07575; Fri, 18 Aug 95 11:11:25 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa05765;
18 Aug 95 18:55 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ab12333;
18 Aug 95 18:53 +0100
Received: from ig.co.uk by lion id <28218-0@lion>;
Fri, 18 Aug 1995 18:54:42 +0100
To: [email protected]
Subject: Re: hints file for Oracle
Cc: [email protected], [email protected]
Date: Fri, 18 Aug 1995 18:54:42 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Alan Burlison <[email protected]>
>
> Tim,
>
> The following hints file is required for DBD::Oracle on svr4, you might
> like to add it to the next release :-)
>
> File: Oracle/hints/svr4.pl
>
> # Some SVR4 systems may need to link against -lc to pick up things like
> $att{LIBS} = [ '-lsocket -lnsl -lm -ldl -lc' ];
>
Umm, 'some', 'may', 'things like'. Care to clarify?

Why _exactly_ is this needed, and why doesn't MakeMaker do this already?
(CC'd to the MakeMaker mailing list.)

Is anyone else using DBD::Oracle on an svr4 system (not solaris 2)?

> Alan Burlison [email protected]
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <10541-8@lion>;
Mon, 21 Aug 1995 07:06:49 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 22:49:51 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa24039;
18 Aug 95 23:49 +0100
Received: by gw.home.vix.com id AA12225; Fri, 18 Aug 95 12:26:18 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA12219; Fri, 18 Aug 95 12:26:15 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa13176;
18 Aug 95 20:25 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa26018;
18 Aug 95 20:23 +0100
Received: from ig.co.uk by lion id <29647-0@lion>;
Fri, 18 Aug 1995 20:25:33 +0100
To: [email protected]
Subject: Re: THE BIG ONE! Announce DBI-0.61 and DBD-Oracle-0.22
Cc: [email protected]
Date: Fri, 18 Aug 1995 20:25:33 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Sounds good. Fortunatly, you seem to have done the moving of vendor code out
> of the vendor.xs file before me which is what I'd been occassionally nagging
> you about anyway after seeing how trivial adding new drivers can be.
>
I'd been meaning to do it ever since I created dbdimp.c.

One slight fly in the ointment is execute() it still has the binding
loop code in *.xs. I'll fix that sometime soon. Probably as a higher
level loop than the one that's there now (which is rather oracle specific).

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <10541-10@lion>;
Mon, 21 Aug 1995 07:07:11 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 22:50:33 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa24100;
18 Aug 95 23:49 +0100
Received: by gw.home.vix.com id AA08426; Fri, 18 Aug 95 11:29:29 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA08421; Fri, 18 Aug 95 11:29:27 -0700
Received: from alma.mcqueen.com (alma.mcqueen.com [193.133.160.249])
by fruitbat.mcqueen.com (8.6.12/8.6.12) with ESMTP id TAA09812;
Fri, 18 Aug 1995 19:25:57 +0100
From: [email protected]
Received: (from descarte@localhost) by alma.mcqueen.com (8.6.12/8.6.12)
id TAA13721; Fri, 18 Aug 1995 19:31:26 +0100
Message-Id: <[email protected]>
Subject: Re: hints file for Oracle
To: Tim Bunce <[email protected]>
Date: Fri, 18 Aug 1995 19:31:25 +0100 (BST)
Cc: [email protected]
In-Reply-To: <[email protected]> from "Tim Bunce" at Aug 18, 95 06:54:42 pm
X-Mailer: ELM [version 2.4 PL24]
Content-Type: text
Content-Length: 644

> > File: Oracle/hints/svr4.pl
> > $att{LIBS} = [ '-lsocket -lnsl -lm -ldl -lc' ];
> >
> Umm, 'some', 'may', 'things like'. Care to clarify?
>
> Why _exactly_ is this needed, and why doesn't MakeMaker do this already?
> (CC'd to the MakeMaker mailing list.)
>
> Is anyone else using DBD::Oracle on an svr4 system (not solaris 2)?

If Dynix/ptx can ever be classed as SVR4, yes, I use it. I can also test
against HP-UX if necessary.

( Classing either Dynix or HP-UX as an OS may be taken with a big pinch of
salt! )

DYNIX worked fine with no twiddlings, if memory serves......

> Tim.

--
Alligator Descartes |
[email protected] |
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <09987-35@lion>;
Mon, 21 Aug 1995 07:07:42 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Fri, 18 Aug 95 23:41:38 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa14847;
19 Aug 95 0:41 +0100
Received: by gw.home.vix.com id AA12138; Fri, 18 Aug 95 12:24:02 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA12122; Fri, 18 Aug 95 12:23:57 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa12788;
18 Aug 95 20:21 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa25374;
18 Aug 95 20:18 +0100
Received: from ig.co.uk by lion id <29545-0@lion>;
Fri, 18 Aug 1995 20:20:39 +0100
To: [email protected]
Subject: Anyone fancy benchmarking DBD::Oracle against oraperl?
Date: Fri, 18 Aug 1995 20:20:39 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

I know DBD::Oracle will be slower right now, I'm just curious
about how much.

For a worthwhile comparison there are several permutations to test:

oraperl cache: cache=1 and 5
Oraperl mode: $Oraperl::safe=1 and 0
row count: 1 and 1000

The cache and the row count can be done at runtime. The $Oraperl::safe
is a load-time setting. Ideally I'd like to see a script that you run as:

orabench loop_count safe_mode

and does soemthing like:

eval '$Oraperl::safe=$safe; use Oraperl; 1' || die $@ if $] >= 5;
$lda = &ora_login(...);
# ... create table and load with 1000 records ...
foreach $rows (1, 1000) {
foreach $cache (1, 5) {
$start = time;
for($i=0; i < $loop_count; ++$i) {
&run_test($cache, $rows);
}
$duration = time - $start;
print "... $safe, $rows, $cache = $duration ...
}
}

where run_test does something like:

local($csr) = &ora_open("select int1,int2,float1,float2,str1,str2 ...
@row = &ora_fetch while $rows--;
&ora_close($csr);

Anyone fancy doing it?

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <10541-16@lion>;
Mon, 21 Aug 1995 07:08:00 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 19 Aug 95 01:07:42 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11988;
19 Aug 95 2:07 +0100
Received: by gw.home.vix.com id AA22108; Fri, 18 Aug 95 14:57:15 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA22103; Fri, 18 Aug 95 14:57:13 -0700
Message-Id: <[email protected]>
Received: by kato.den.csci.csc.com (1.37.109.15/16.2) id AA185723366;
Fri, 18 Aug 1995 16:02:46 -0600
From: Tomasz Wardega <[email protected]>
Subject:
To: [email protected]
Date: Fri, 18 Aug 95 16:02:45 MDT
Mailer: Elm [revision: 70.85.2.1]

unsubscribe
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <10541-18@lion>;
Mon, 21 Aug 1995 07:08:15 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 19 Aug 95 01:56:27 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa23573;
19 Aug 95 2:55 +0100
Received: by gw.home.vix.com id AA18914; Fri, 18 Aug 95 14:02:37 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA18908; Fri, 18 Aug 95 14:02:33 -0700
Received: by metronet.com id AA03795 (5.67a/IDA1.5hp
for [email protected]); Fri, 18 Aug 1995 16:01:40 -0500
From: Tye McQueen <[email protected]>
Message-Id: <[email protected]>
Subject: Re: [MM] Re: hints file for Oracle
To: Tim Bunce <[email protected]>
Date: Fri, 18 Aug 1995 16:01:39 -0500 (CDT)
Cc: [email protected], [email protected],
[email protected]
In-Reply-To: <[email protected]> from "Tim Bunce" at Aug 18, 95 06:54:42 pm
Content-Transfer-Encoding: 8bit
X-Mailer: ELM [version 2.4 PL24]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1870

Excerpts from the mail message of Tim Bunce:
) > From: Alan Burlison <[email protected]>
) >
) > The following hints file is required for DBD::Oracle on svr4, you might
) > like to add it to the next release :-)
) >
) > File: Oracle/hints/svr4.pl
) >
) > # Some SVR4 systems may need to link against -lc to pick up things like
) > $att{LIBS} = [ '-lsocket -lnsl -lm -ldl -lc' ];
)
) Umm, 'some', 'may', 'things like'. Care to clarify?
)
) Why _exactly_ is this needed, and why doesn't MakeMaker do this already?
) (CC'd to the MakeMaker mailing list.)
)
) Is anyone else using DBD::Oracle on an svr4 system (not solaris 2)?

That looks like something I wrote. I'll take credit and blame
for it at least for the sake of the next paragraph.

So far "some" is only whatever Unisys system Alan and one other
person have used. "may" is because, as far as I could tell from
my end, some of the dynamically loaded extensions worked okay
before this fix but one of them didn't. "thinks like" must be
because I couldn't remember which routine was not being found
and then forgot to finish my sentence. I think it was _ecvt().

The description is very vague because it doesn't make sense to
me why it is needed and I don't have access to a system to play
around with it if I really wanted to try to figure it out. But
it seems to fix the few problems it addresses and have not heard
of it hurting anything yet (and I've tested it on my machines).

I'm putting together a README.svr4 for Perl that will describe this
and many other things in case people are curious or run into a
problem and need to know why some of the strange things were done.
--
Tye McQueen [email protected] || [email protected]
Nothing is obvious unless you are overlooking something
http://www.metronet.com/~tye/ (scripts, links, nothing fancy)
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <10541-22@lion>;
Mon, 21 Aug 1995 07:09:02 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Sat, 19 Aug 95 03:05:37 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa07219;
19 Aug 95 4:05 +0100
Received: by gw.home.vix.com id AA18914; Fri, 18 Aug 95 14:02:37 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA18908; Fri, 18 Aug 95 14:02:33 -0700
Received: by metronet.com id AA03795 (5.67a/IDA1.5hp
for [email protected]); Fri, 18 Aug 1995 16:01:40 -0500
From: Tye McQueen <[email protected]>
Message-Id: <[email protected]>
Subject: Re: [MM] Re: hints file for Oracle
To: Tim Bunce <[email protected]>
Date: Fri, 18 Aug 1995 16:01:39 -0500 (CDT)
Cc: [email protected], [email protected],
[email protected]
In-Reply-To: <[email protected]> from "Tim Bunce" at Aug 18, 95 06:54:42 pm
Content-Transfer-Encoding: 8bit
X-Mailer: ELM [version 2.4 PL24]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 1870

Excerpts from the mail message of Tim Bunce:
) > From: Alan Burlison <[email protected]>
) >
) > The following hints file is required for DBD::Oracle on svr4, you might
) > like to add it to the next release :-)
) >
) > File: Oracle/hints/svr4.pl
) >
) > # Some SVR4 systems may need to link against -lc to pick up things like
) > $att{LIBS} = [ '-lsocket -lnsl -lm -ldl -lc' ];
)
) Umm, 'some', 'may', 'things like'. Care to clarify?
)
) Why _exactly_ is this needed, and why doesn't MakeMaker do this already?
) (CC'd to the MakeMaker mailing list.)
)
) Is anyone else using DBD::Oracle on an svr4 system (not solaris 2)?

That looks like something I wrote. I'll take credit and blame
for it at least for the sake of the next paragraph.

So far "some" is only whatever Unisys system Alan and one other
person have used. "may" is because, as far as I could tell from
my end, some of the dynamically loaded extensions worked okay
before this fix but one of them didn't. "thinks like" must be
because I couldn't remember which routine was not being found
and then forgot to finish my sentence. I think it was _ecvt().

The description is very vague because it doesn't make sense to
me why it is needed and I don't have access to a system to play
around with it if I really wanted to try to figure it out. But
it seems to fix the few problems it addresses and have not heard
of it hurting anything yet (and I've tested it on my machines).

I'm putting together a README.svr4 for Perl that will describe this
and many other things in case people are curious or run into a
problem and need to know why some of the strange things were done.
--
Tye McQueen [email protected] || [email protected]
Nothing is obvious unless you are overlooking something
http://www.metronet.com/~tye/ (scripts, links, nothing fancy)
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <20943-0@lion>;
Mon, 21 Aug 1995 20:59:48 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 19:52:41 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa01185;
21 Aug 95 20:52 +0100
Received: by gw.home.vix.com id AA23736; Mon, 21 Aug 95 05:34:35 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA23731; Mon, 21 Aug 95 05:34:33 -0700
Message-Id: <[email protected]>
Received: from mip.lasc.lockheed.com by netserve.lasc.lockheed.com
with SMTP (1.37.109.4/16.2) id AA15420; Mon, 21 Aug 95 08:34:24 -0400
Received: by mip.lasc.lockheed.com (1.38.193.4/16.2) id AA20456;
Mon, 21 Aug 1995 08:34:58 -0400
Date: Mon, 21 Aug 1995 08:34:58 -0400
From: Terry Greenlaw <[email protected]>
To: [email protected]
Subject: Re: Help requested for ora_do behaviour testing

> Hi,
> This is to clarify exactly how Oraperl handles &ora_do().
> In Oraperl, the task of processing a statement is split up as follows:

[stuff deleted]

> with a little error checking on the way. In a sensible case, with no
> substitution variables and no values to return (such as a DDL statement
> or the example at the start of this message), the sequence of calls is
> therefore:

> count substitution variables (= 0)
> open a cursor (OCI: oopen)
> parse the statement (OCI: osql3)
> execute the statement (OCI: oexec)
> count the number of return values (= 0) (OCI: odsc)
> close the cursor (OCI: oclose)

> I don't see where in that sequence it's possible for a statement of any
> kind to be executed twice.

> Hope that helps to clarify the issue.

> Kevin

The key difference is that you used the osql3 call, which is deprecated in
Oracle7 OCI. The replacement oparse supports a deferred execution mode which
can reduce traffic between the client and server by waiting until oexec to
actually do the execution. An oparse on a non-select statement in non-deferred
mode followed by an oexec was executing the command twice. That behavior does
not occur in the later versions of DBI/DBD, but I don't know why, yet. Tim did
mention that there were bugs with oparse in OCI.

Thanks for the input into the internals of Oraperl. Without it, I'd still be
writing text processing scripts in SQL*Plus and hating life in general.

tog

Terry Greenlaw (on-site @ Lockheed) Encompass Technologies
[email protected] [email protected]

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25844-3@lion>;
Tue, 22 Aug 1995 04:20:53 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 20:52:33 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa29495;
21 Aug 95 21:51 +0100
Received: by gw.home.vix.com id AA24793; Mon, 21 Aug 95 05:54:48 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA24788; Mon, 21 Aug 95 05:54:46 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail1.austin.ibm.com (8.6.11/8.6.11) with SMTP id HAA165543;
Mon, 21 Aug 1995 07:54:43 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA19900;
Mon, 21 Aug 1995 07:54:41 -0500
From: [email protected]
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected]
Subject: Re: question about prepare
In-Reply-To: Your message of Fri, 18 Aug 95 10:22:59 +0100. <[email protected]>
Date: Mon, 21 Aug 95 07:54:40 -0500

Tim Bunce <[email protected]> writes:
|> The 'right' approach is probably to use the recently discussed
|> bind_param call. Something like:
|>
|> $sth->bind_param($param_index_or_name, \$value [, \%attribs]);
|>
|> where \%attribs would allow you to specify type. I'm not settled
|> on the interface spec yet but it should be something like this.
|>

Where did this dicussion take place, as I didn't get through
perldb-interest.

Thanks.

Mike
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25844-5@lion>;
Tue, 22 Aug 1995 04:21:02 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 20:54:00 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00209;
21 Aug 95 21:53 +0100
Received: by gw.home.vix.com id AA21432; Mon, 21 Aug 95 05:05:11 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA21427; Mon, 21 Aug 95 05:05:08 -0700
Message-Id: <[email protected]>
Received: from mip.lasc.lockheed.com by netserve.lasc.lockheed.com
with SMTP (1.37.109.4/16.2) id AA15299; Mon, 21 Aug 95 08:04:59 -0400
Received: by mip.lasc.lockheed.com (1.38.193.4/16.2) id AA17886;
Mon, 21 Aug 1995 08:05:32 -0400
Date: Mon, 21 Aug 1995 08:05:32 -0400
From: Terry Greenlaw <[email protected]>
To: [email protected]
Subject: Re: Help requested for ora_do behaviour testing

[ Assorted chunks deleted for brevity ]
> > > If anyone has some spare time I'd appreciate it if you could uncomment
> > > the execute line from ora_do in Oraperl.pm are try out things like
> > > insert and create table statements (where double execution is easy
> > > to spot).

The execute line does need to be uncommented now to work. With DBD 14 (the
last version I checked), leaving it uncommented produced two inserts.
The DBD version we are using now is 19.

I just discovered that someone here replaced most of the Oracle 7.0.15
files we were using with a link to the 7.1.6 directory not too long ago!!
That pretty much rules out any assumptions here of dealing with a stable
environment... After I kill him, I'll look for changes in Oracle's stuff.

It may also be related to the changes in DBD. Now that bind variables are
allowed, the code would have to change in ora_do to only execute if no bind
variables are present. Could those changes have eliminated the problem?

> > I was definitely getting
> > double inserts before commenting out the execute in ora_do.

> And once commented out you only got one?
Yup.

> Could you experiment with the new release (0.22) and try changing
> the values of oparse_defer in dbdimp.c (any anything else that comes
> to mind).
>
> It was awhile back. I've just avoided the issue. But now people are
> complaining that ora_do() does not work at all for them.

> We need to get to the bottom of this soon.

> > Has Oracle cleaned the bugs out of it yet?

> What version of Oracle are you using?

What day of the week is it ;-}
We have 7.1.6 currently in production.

Terry Greenlaw (on-site @ Lockheed) Encompass Technologies
[email protected] [email protected]

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25844-10@lion>;
Tue, 22 Aug 1995 04:21:17 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 21:05:49 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa05846;
21 Aug 95 22:04 +0100
Received: by gw.home.vix.com id AA21154; Mon, 21 Aug 95 04:59:23 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA21149; Mon, 21 Aug 95 04:59:22 -0700
Received: from dax.austin.ibm.com (dax.austin.ibm.com [129.35.112.95])
by netmail.austin.ibm.com (8.6.11/8.6.11) with SMTP id GAA71674;
Mon, 21 Aug 1995 06:59:21 -0500
Received: by dax.austin.ibm.com (AIX 4.1/UCB 5.64/4.03-client-2.6)
for [email protected] at austin.ibm.com; id AA17674;
Mon, 21 Aug 1995 06:59:14 -0500
From: [email protected]
Message-Id: <[email protected]>
To: Tim Bunce <[email protected]>
Cc: [email protected]
Subject: Re: question about prepare
In-Reply-To: Your message of Fri, 18 Aug 95 10:22:59 +0100. <[email protected]>
Date: Mon, 21 Aug 95 06:59:13 -0500

Tim Bunce <[email protected]> writes:
|> > b) Doing everything in char and designing some method of making
|> > special cases of everthing that makes it break. Yuck.
|> >
|> Yuck but a worthwhile short term get-it-working measure.
|>
|> > c) Mutating the parameters to the exec call. Instead of doing
|> > execute($sth,@exec_values) where each item in the array is a single
|> > value, allowing each item to be a list containg the value and the
|> > type. The question then becomes is it really a list or is it a
|> > value<symbol>type string.
|> >
|> Eeek. Don't mess with the DBI spec. The interface definition (such as
|> it is) is sacrosanct.
|>
|> > Well, which of the above sounds less horrid?
|> >
|> The 'right' approach is probably to use the recently discussed
|> bind_param call. Something like:
|>
|> $sth->bind_param($param_index_or_name, \$value [, \%attribs]);
|>
|> where \%attribs would allow you to specify type. I'm not settled
|> on the interface spec yet but it should be something like this.
|>
|> I'd suggest going with (b) in the short term till I/we agree a
|> spec and DBI support for some form of bind_param call.

What if, we allow the elements of @exec_values to include hashes. Then
I could provide a user callable function to to create the hash for
the more exotic sql types. Something along the lines of what
Shigeya Suzuki <[email protected]> was talking about in his
"perl or sybperl problem??" message on perl5-porters to which you
respond, today.

If it is not a hash, then I can assume that the sql type is either
char (or string), integer, or double; otherwise, I assume it is a
more exotic sql type and process accordingly.

Thoughts, complaints, anything?

Mike

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25844-12@lion>;
Tue, 22 Aug 1995 04:21:24 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 21:27:58 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa16632;
21 Aug 95 22:27 +0100
Received: by gw.home.vix.com id AA03885; Mon, 21 Aug 95 08:09:41 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA03880; Mon, 21 Aug 95 08:09:38 -0700
Received: from barbera.inferentia.it (barbera.inferentia.it [151.99.220.61])
by relay.smau.it (8.6.12/8.6.12) with SMTP id RAA25343
for <[email protected]>; Mon, 21 Aug 1995 17:08:21 +0200
Posted-Date: Mon, 21 Aug 1995 17:08:21 +0200
Received: by barbera.inferentia.it with Microsoft Mail
id <[email protected]>; Mon, 21 Aug 95 17:15:46 PDT
From: Shari <[email protected]>
To: "'[email protected]'" <[email protected]>
Subject: DBD::Oracle How to handle RAW and LONG RAW parameters?
Date: Mon, 21 Aug 95 17:17:00 PDT
Message-Id: <[email protected]>
Encoding: 48 TEXT
X-Mailer: Microsoft Mail V3.0

I think I found a problem in the current DBD::Oracle implementation
(rumble).

Since DBD::Oracle 0.23 we are able to fetch LONG and LONG RAW columns
correctly (and even efficiently :-).
But how can one insert a LONG/LONG RAW column?
The current implementation does not allow a programmer to specify the
type of the bind parameter when passing the argument to do or execute
(unfortunately, since this comes from DBI, am I actually referring to
something which must be done at that level?)
In the current DBD implementation, this means using type 1 (VARCHAR2) or
8 (LONG) when binding a scalar basing on its length (dbd_bind_ph() in
dbdimp.c).
This is OK if the destination column is of a corresponding type, but due
to OCI limitations, you cannot bind a non-RAW LONG to a LONG RAW, so
BLOBs are out of reach for the DBI programmer.
Even if that would be possibile, there is a fundamental difference
between RAW and non-RAW, in that the second type(s) get "massaged" by
Oracle and may possibly undergo character translation, while RAW are left
ountouched (and of course, BLOBs must be left untouched in a typical
application).

CLI defines the BindParam function, which should allow for specification
of input datatype for parameters. Unfortunately, I have seen no trace of
this in DBI. The use of BLOB in databases should be a strong argument in
favor of a different implementation of execute and do, for instance,
which could accept a hash instead of arrays in order to specify, for each
parameter, the desidered binding type.
Still better would be a complete BindParam implementation, since it would
be possible to avoid copying parameter values when binding (which becomes
an issue if you have large BLOBs to insert).

If I am missing something, please tell me! I am currently developing
something which fetches BLOBs from a remote database and the inserts them
in a local database. While the first task, with DBI and the Oraperl
emulation, is solved with < 200 lines of perl, the second is (to my
limited skill) out of reach.

And, if there is a "simple" way to work around this problem (other than
my rewriting core parts of DBD, which I will not do in order not to miss
future releases), please suggest. As of now I am about to write an
external program to do this job on behalf of the Perl replicator, but of
course this is suboptimal (I really want to do everything in Perl).

(If you managed to read to this point: thank you!)

Davide Migliavacca
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25844-21@lion>;
Tue, 22 Aug 1995 04:21:57 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 23:03:47 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa00312;
22 Aug 95 0:03 +0100
Received: by gw.home.vix.com id AA23832; Mon, 21 Aug 95 05:39:11 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA23826; Mon, 21 Aug 95 05:39:09 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa19065;
21 Aug 95 13:38 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa04016;
21 Aug 95 13:34 +0100
Received: from ig.co.uk by lion id <15452-0@lion>;
Mon, 21 Aug 1995 13:36:16 +0100
To: [email protected]
Subject: Re: question about prepare
Cc: [email protected]
Date: Mon, 21 Aug 1995 13:36:16 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: [email protected]
>
> Tim Bunce <[email protected]> writes:
> |> > b) Doing everything in char and designing some method of making
> |> > special cases of everthing that makes it break. Yuck.
> |> >
> |> Yuck but a worthwhile short term get-it-working measure.
> |>
> |> > c) Mutating the parameters to the exec call. Instead of doing
> |> > execute($sth,@exec_values) where each item in the array is a single
> |> > value, allowing each item to be a list containg the value and the
> |> > type. The question then becomes is it really a list or is it a
> |> > value<symbol>type string.
> |> >
> |> Eeek. Don't mess with the DBI spec. The interface definition (such as
> |> it is) is sacrosanct.
> |>
> |> > Well, which of the above sounds less horrid?
> |> >
> |> The 'right' approach is probably to use the recently discussed
> |> bind_param call. Something like:
> |>
> |> $sth->bind_param($param_index_or_name, \$value [, \%attribs]);
> |>
> |> where \%attribs would allow you to specify type. I'm not settled
> |> on the interface spec yet but it should be something like this.
> |>
> |> I'd suggest going with (b) in the short term till I/we agree a
> |> spec and DBI support for some form of bind_param call.
>
> What if, we allow the elements of @exec_values to include hashes. Then
> I could provide a user callable function to to create the hash for
> the more exotic sql types. Something along the lines of what
> Shigeya Suzuki <[email protected]> was talking about in his
> "perl or sybperl problem??" message on perl5-porters to which you
> respond, today.
>
> If it is not a hash, then I can assume that the sql type is either
> char (or string), integer, or double; otherwise, I assume it is a
> more exotic sql type and process accordingly.
>
> Thoughts, complaints, anything?
>
My previous message basically said "don't mess with the arguments to
execute, use bind_param()" and "use chars for now till we agree the spec".

You'll need to tell my why that's not accaptable before I'll devote much
time to considering alternatives.

> Mike
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25844-23@lion>;
Tue, 22 Aug 1995 04:22:05 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 23:18:42 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06473;
22 Aug 95 0:18 +0100
Received: by gw.home.vix.com id AA23601; Mon, 21 Aug 95 13:03:06 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA23572; Mon, 21 Aug 95 13:02:42 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa25961;
21 Aug 95 20:59 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa18242;
21 Aug 95 20:57 +0100
Received: from ig.co.uk by lion id <20923-0@lion>;
Mon, 21 Aug 1995 20:59:22 +0100
To: [email protected]
Subject: Re: (Little) Bug in DBD::Oracle 0.23, Oracle.xs
Cc: [email protected]
Date: Mon, 21 Aug 1995 20:59:22 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Shari <[email protected]>
>
> There is a bug in _execute in file Oracle.xs (starting at line 179 in the
> 0.23 sources) which forces all parameters to being named :pn with n being
> its position,

It's not a bug.

> while dbd_preparse (dbdimp.c) accepts ?, :1 and :symbolic_name.
> Thus if parameter are named :p1, :p2, etc, or :1, :2, etc., the code
> works; in any other situation dbd_bind_ph fails not finding the
> corresponding ":pn" parameter.
> Unfortunately I am not sufficiently fluent in perl extensions to patch
> the code myself (it has taken a while to track the spot down...)
>
The SQL *STANDARD* only allows placeholders of the form '?'.
Oracle's named placeholders are very nice, but non-standard.

The execute method's optional bind values are designed for
a) the standard, and b) simplicity.

Optional bind values given to the execute function are assigned in
sequence to the 'corresponding' placeholders.

If you want/need to use named placeholders then you need to use
the (non existent) bind_param call and not give extra args to execute.

> P.S.
> I am very interested in the bind_param call since it's one of the
> features I terribly need (working most of my time with BLOBs, that is
> LONG RAW which are not usable with the current implementation of
> dbd_bind_ph).

In what way 'not usable'?

If I don't know I can't fix it. (Then again, if I do know I might not
fix it, or be able to fix it, anyway :-)

> I have seen it mentioned many times in the mailing list
> messages, but didn't find it in DBI nor in the specs. Is it a proposed
> enhancement?
>
Yes.

I'll look into it again. Nag me from time to time if it doesn't show up.

> Bye
> Davide Migliavacca
>
Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <25844-25@lion>;
Tue, 22 Aug 1995 04:22:13 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Mon, 21 Aug 95 23:46:38 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa17688;
22 Aug 95 0:46 +0100
Received: by gw.home.vix.com id AA13454; Mon, 21 Aug 95 10:36:08 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA13440; Mon, 21 Aug 95 10:36:06 -0700
Received: from leafusa.UUCP by ileaf.com (4.1/SMI-4.1) id AA15072;
Mon, 21 Aug 95 13:35:43 EDT
Received: from cartier.HQ.Ileaf.COM by HQ.Ileaf.COM (4.1/SMI-4.0-leafusa)
for [email protected] id AA13831; Mon, 21 Aug 95 13:33:50 EDT
Received: by cartier.HQ.Ileaf.COM (4.1/SMI-4.0-hq) for [email protected]
id AA21064; Mon, 21 Aug 95 13:33:35 EDT
Date: Mon, 21 Aug 95 13:33:35 EDT
From: Jason Cheung <[email protected]>
Message-Id: <[email protected]>
To: [email protected]
Subject: unsubscribe

unsubscribe
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <28369-2@lion>;
Tue, 22 Aug 1995 07:01:02 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 04:19:35 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa24778;
22 Aug 95 5:19 +0100
Received: by gw.home.vix.com id AA24893; Mon, 21 Aug 95 13:24:13 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA24888; Mon, 21 Aug 95 13:24:11 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa27332;
21 Aug 95 21:16 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa21254;
21 Aug 95 21:13 +0100
Received: from ig.co.uk by lion id <21230-0@lion>;
Mon, 21 Aug 1995 21:15:07 +0100
To: [email protected], [email protected]
Subject: Re: Help requested for ora_do behaviour testing
Date: Mon, 21 Aug 1995 21:15:07 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Terry Greenlaw <[email protected]>
>
> > Hi,
> > This is to clarify exactly how Oraperl handles &ora_do().
> > In Oraperl, the task of processing a statement is split up as follows:
>
> [stuff deleted]
>
> > with a little error checking on the way. In a sensible case, with no
> > substitution variables and no values to return (such as a DDL statement
> > or the example at the start of this message), the sequence of calls is
> > therefore:
>
> > count substitution variables (= 0)
> > open a cursor (OCI: oopen)
> > parse the statement (OCI: osql3)
> > execute the statement (OCI: oexec)
> > count the number of return values (= 0) (OCI: odsc)
> > close the cursor (OCI: oclose)
>
> > I don't see where in that sequence it's possible for a statement of any
> > kind to be executed twice.
>
> > Hope that helps to clarify the issue.
>
> > Kevin
>
> The key difference is that you used the osql3 call, which is deprecated in
> Oracle7 OCI. The replacement oparse supports a deferred execution mode which
> can reduce traffic between the client and server by waiting until oexec to
> actually do the execution.

I think this is a red herring (to use an English colloquialism).

> An oparse on a non-select statement in non-deferred
> mode followed by an oexec was executing the command twice.

Terry, so far you're the only person who has reported this to me.
Has *anyone* else seen this?

> That behavior does
> not occur in the later versions of DBI/DBD, but I don't know why, yet. Tim did
> mention that there were bugs with oparse in OCI.
>
Off hand I don't recall anything directly related to this.

DBD::Oracle does not now, and never has, deferred parsing (unless you
edit the code by hand. It'll be an option later).

The documentation for both osql3 and oparse states that "Data Definition
Language statements are executed when * is called. DML statements and
queries are not executed until you call the oexec function."

(It seems that a subsequent oexec is silently ignored.)

Terry, your original message referred to double inserts. Since inserts
are DML not DDL statements I think the problem must have been (past tense)
elsewhere.

It may well have been a bug elsewhere in the DBI or DBD::Oracle which
has since been fixed. If you can't reproduce it now let's just drop it.

> Thanks for the input into the internals of Oraperl. Without it, I'd still be
> writing text processing scripts in SQL*Plus and hating life in general.
>
Ditto.

> tog
>
> Terry Greenlaw (on-site @ Lockheed) Encompass Technologies
> [email protected] [email protected]
>
Tim.
---+++---
Received: from punt2.demon.co.uk by lion with SMTP (PP) id <01663-4@lion>;
Tue, 22 Aug 1995 17:12:04 +0100
Received: from punt2.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 14:03:09 GMT
Received: from gw.home.vix.com by punt2.demon.co.uk id aa20504;
22 Aug 95 15:02 +0100
Received: by gw.home.vix.com id AA24658; Mon, 21 Aug 95 21:35:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA24653; Mon, 21 Aug 95 21:35:42 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa06026;
22 Aug 95 5:05 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id ac13569;
22 Aug 95 5:03 +0100
Received: from ig.co.uk by lion id <27281-0@lion>;
Tue, 22 Aug 1995 05:04:36 +0100
To: [email protected], [email protected]
Subject: Re: DBD::Oracle How to handle RAW and LONG RAW parameters?
Date: Tue, 22 Aug 1995 05:04:36 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

> From: Shari <[email protected]>
>
> I think I found a problem in the current DBD::Oracle implementation

> The current implementation does not allow a programmer to specify the
> type of the bind parameter [...]

True. Just not yet implemented.

> CLI defines the BindParam function, which should allow for specification
> of input datatype for parameters. Unfortunately, I have seen no trace of
> this in DBI.

That's what the 'WARNING: THIS IS ALPHA SOFTWARE' means.

> Still better would be a complete BindParam implementation, since it would
> be possible to avoid copying parameter values when binding (which becomes
> an issue if you have large BLOBs to insert).
>
A bind_param is on it's way.

> If I am missing something, please tell me!

Only a little patience and understanding :-)

> And, if there is a "simple" way to work around this problem (other than
> my rewriting core parts of DBD, which I will not do in order not to miss
> future releases), please suggest.

I'm happy for people to work on the DBI and DBD modules if they follow
a few very simple rules:

1) Tell me what you want to do and why *before* you do it.
2) Listen to my suggestions (ignore at your peril :-).
3) Send me a patch *quickly* based on the last release.

The DBD/DBD modules progress in fits and starts. I may not work on them
for days (or months :-) but when I do things can move very fast. A patch
that's useful at 9am much be much less useful to me by 9pm.

> As of now I am about to write an
> external program to do this job on behalf of the Perl replicator, but of
> course this is suboptimal (I really want to do everything in Perl).
>
As of now (4:50am) I have a working bind_param. I also have a
working bind_col and bind_columns. A release should follow shortly.
(along with cries of anguish from driver developers :-)

> Davide Migliavacca
>
Tim.

p.s. Davide, my last message to you (CC'd to the list) bounced:

From: Mail Delivery Subsystem <[email protected]>
Subject: Returned mail: Local configuration error

----- The following addresses had delivery problems -----
<[email protected]> (unrecoverable error)

----- Transcript of session follows -----
.. while talking to relay.inferentia.it.:
>>> MAIL From:<@post.demon.co.uk,@ignite.demon.co.uk:[email protected]> SIZE=1752
<<< 452 <@post.demon.co.uk,@ignite.demon.co.uk:[email protected]>... Insufficient disk space; try again later
553 interbusiness.it config error: mail loops back to myself
554 <[email protected]>... Local configuration error
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <01665-18@lion>;
Tue, 22 Aug 1995 17:14:58 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 11:38:28 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa16000;
22 Aug 95 12:38 +0100
Received: by gw.home.vix.com id AA01559; Mon, 21 Aug 95 23:06:25 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA01554; Mon, 21 Aug 95 23:06:17 -0700
Received: from post.demon.co.uk by disperse.demon.co.uk id aa13805;
22 Aug 95 7:04 +0100
Received: from ignite.demon.co.uk by post.demon.co.uk id aa20207;
22 Aug 95 7:02 +0100
Received: from ig.co.uk by lion id <28645-0@lion>;
Tue, 22 Aug 1995 07:04:12 +0100
To: [email protected]
Subject: Announce: DBI-0.61 and DBD-Oracle-0.24
Date: Tue, 22 Aug 1995 07:04:12 +0100
From: Tim Bunce <[email protected]>
Organisation: Paul Ingram Group, Software Systems, +44 1 483 424424
Message-Id: <[email protected]>

Remember that old adage about 'the only time you'll find a programer
at work an 7 in the morning is when ...'.

Well, I took pity on the pleas for some form on bind_param method so
here's a new release with some whiz-bang features. Available from the
usual spot.

Current DBI Implementation Notes:

$ary_ref = $sth->fetch; currently always returns the same array ref.
This may be a feature that says, then again it may not.

Some confusion exists in the indexing of columns and parameters.
Perl starts at zero, SQL standards start at 1. I've probably been
inconsistent in my usage. We need to resolve this one way or other.

Changes in DBI 0.61, 22 Aug 1995

Added $sth->bind_col($column, \$var [, \%attribs ]);

This method enables perl variables to be directly and automatically
updated when a row is fetched. It requires no driver support
(if the driver has been written to use DBIS->get_fbav).
Currently \%attribs is unused.

Added $sth->bind_columns(\$var [, \$var , ...]);

This method is a short-cut for bind_col which binds all the
columns of a query in one go (with no attributes). It also
requires no driver support.

Added $sth->bind_param($parameter, $var [, \%attribs ]);

This method enables attributes to be specified when values are
bound to placeholders. It also enables binding to occur away
from the execute method to improve execute efficiency.
The DBI does not provide a default implementation of this.
See the DBD::Oracle module for a detailed example.

The DBI now provides default implementations of both fetch and
fetchrow. Each is written in terms of the other. A driver is
expected to implement at least one of them.

More macro and assorted structure changes in DBDXS.h. Sorry!
The old dbihcom definitions have gone. All fields have macros.
The imp_xxh_t type is now used within the DBI as well as drivers.
Drivers must set DBIc_NUM_FIELDS(imp_sth) and DBIc_NUM_PARAMS(imp_sth).

test.pl includes a trivial test of bind_param and bind_columns.

Changes in DBD::Oracle 0.24, 22 Aug 1995:

User visible changes:

$sth->bind_param method implemented (needs testing :-)
and new DBI $sth->bind_col* methods work for DBD::Oracle.

Changes to be propogated to other drivers:

- Slight change to STORE functions to allow them to return status
if called as $h->STORE(...).

- Revised signature of dbd_bind_ph function and added a bind_param
method for it. Prototype moved from dbdimp.h to Oracle.h

- Revised execute method to use modified dbd_bind_ph function and
remove remaining oracle specifics.

- imp_xxh_t structures first element now called 'com' not 'dbihcom'
and new DBI macros used to access fields.

Test and enjoy :-)

Tim.
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05893-15@lion>;
Wed, 23 Aug 1995 07:02:41 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 18:26:27 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa11340;
22 Aug 95 19:26 +0100
Received: by gw.home.vix.com id AA09510; Tue, 22 Aug 95 01:43:17 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA09504; Tue, 22 Aug 95 01:43:11 -0700
Received: from bssmlsa.nerc-bas.ac.uk.nerc-bas.ac.uk
by bsfiles.nerc-bas.ac.uk (8.6.10/NERC-1.2(Solaris 2.x) id JAA19100;
Tue, 22 Aug 1995 09:43:02 +0100
Date: Tue, 22 Aug 1995 09:42:59 +0100
Message-Id: <[email protected]>
From: Andrew G Wood <[email protected]>
To: [email protected]
Cc: [email protected]
Subject: DBI-0.61
Reply-To: [email protected]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII

Well you have not heard from me for a few days ... my compiler has been
off on holiday. But as there seemed to be a few new versions flying
around he could not wait to get back into it .....

So the obligatory ....

% perl -e 'use Config; print Config::myconfig()'
Summary of my perl5 (patchlevel 1) configuration:
Platform:
osname=irix, osver=5, archname=IP20-irix
uname='irix bssmlsa 5.3 11091811 ip20 mips '
hint=recommended
Compiler:
cc='cc', optimize='-O'
cppflags='-D_POSIX_SOURCE -ansiposix -D_BSD_TYPES'
ccflags ='-D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000'
ldflags =''
stdchar='unsigned char', d_stdstdio=define, usevfork=false
voidflags=15, castflags=1, d_casti32=define, d_castneg=
intsize=4, alignbytes=8, usemymalloc=y, randbits=15
Libraries:
so=so
libpth=/usr/lib /usr/local/lib /lib
libs=-lgdbm -lsun -lm -lc -lcrypt -lbsd -lPW
libc=/usr/lib/libc.so
Dynamic Linking:
dlsrc=dl_dlopen.xs, dlext=so, d_dlsymun=
cccdlflags=' ', ccdlflags=' ', lddlflags='-shared'

Then the usual make problems ....

% perl Makefile.PL

Build, TEST and INSTALL Perl 5 first. Use version 5.001m or later.

DO NOT build the DBI under the Perl source tree.

Use 'make test' to execute self tests.

MakeMaker (v4.16)
Checking if your kit is complete...
Looks good
NAME => 'DBI'
VERSION => '0.61'
Using PERL=/nerc/bin/perl
Writing Makefile for DBI

% make
mkdir ./blib
mkdir ./blib/auto
mkdir ./blib/auto/DBI
/nerc/bin/perl -I/usr/local/lib/perl5/IP20-irix -I/usr/local/lib/perl5 /usr/local/lib/perl5/ExtUtils/xsubpp -typemap /usr/local/lib/perl5/ExtUtils/typemap DBI.xs >DBI.tc && mv DBI.tc DBI.c
cc -c -D_POSIX_SOURCE -ansiposix -D_BSD_TYPES -Olimit 3000 -O -I/usr/local/lib/perl5/IP20-irix/CORE DBI.c
cfe: Error: DBI.c, line 1411: Incompatible pointer type assignment (3.3.16.1, 3.5.4.1(21))
AV *av = (SV*)((XRV*) (retsv)->sv_any )->xrv_rv ;
----^
cfe: Error: DBI.c, line 1411: Type SV * of rhs of assignment expression is incompatible with type AV * of lhs (3.3.16.1)
AV *av = (SV*)((XRV*) (retsv)->sv_any )->xrv_rv ;
----^
*** Error code 1 (bu21)

C'est la vie!

Andy.
*
Dr Andy Wood **
Database Administrator ***
***
****
[email protected] ****
[email protected] ****
[email protected] ****
****
[email protected] ****
...*****.____
Post: .::______.:~'
****
British Antarctic Survey ****
High Cross ****
Madingley Road ****
Cambridge, CB3 0ET, UK ***
**
+44 (0) 1223 61188 **
*

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06633-5@lion>;
Wed, 23 Aug 1995 07:05:04 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 21:11:10 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa29826;
22 Aug 95 22:10 +0100
Received: by gw.home.vix.com id AA14347; Tue, 22 Aug 95 02:14:35 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA14342; Tue, 22 Aug 95 02:14:32 -0700
Received: from barbera.inferentia.it (barbera.inferentia.it [151.99.220.61])
by relay.smau.it (8.6.12/8.6.12) with SMTP id LAA24810;
Tue, 22 Aug 1995 11:13:11 +0200
Posted-Date: Tue, 22 Aug 1995 11:13:11 +0200
Received: by barbera.inferentia.it with Microsoft Mail
id <[email protected]>; Tue, 22 Aug 95 11:20:43 PDT
From: Shari <[email protected]>
To: 'PerlDB-interest mailing list' <[email protected]>
Cc: 'Tim Bunce' <[email protected]>
Subject: RE: (Little) Bug in DBD::Oracle 0.23, Oracle.xs
Date: Tue, 22 Aug 95 11:23:00 PDT
Message-Id: <[email protected]>
Encoding: 80 TEXT
X-Mailer: Microsoft Mail V3.0

From Tim Bunce:

>> From: Shari <[email protected]>
>>
>> There is a bug in _execute in file Oracle.xs (starting at line 179 in
the
>> 0.23 sources) which forces all parameters to being named :pn with n
being
>> its position,

> It's not a bug.

Sorry, I must disagree. It IS a bug in that ? placeholders are not
handled correctly.
In fact, they are transformed in :n, while _execute looks for :pn and
doesn't find them.
Or at least, this happens to me:
$cursor = ora_open($lda, "select * from table where key=?");
ora_bind($cursor, '756');
results in:
dbd_bind_ph placeholder ':p1' unknown at /home/shari/test.pl line 21.

dbd_bind_ph transforms ? in :n. Either the bug is there, or in execute,
but they sure don't seem to understand each other.

> The SQL *STANDARD* only allows placeholders of the form '?'.
> Oracle's named placeholders are very nice, but non-standard.

I totally agree, but see above.

>> P.S.
>> I am very interested in the bind_param call since it's one of the
>> features I terribly need (working most of my time with BLOBs, that is

>> LONG RAW which are not usable with the current implementation of
>> dbd_bind_ph).

> In what way 'not usable'?
In that the automatic type binding in dbd_bind_ph (see my other
message to the mailing list) uses only type 1 and 8. To use LONG RAW you
need type 24, which is not easily deductible from a variable. RAW and
CHAR get a very diffrerent treatment in Oracle. so you will not be able
to bind a LONG variable to a LONG RAW column. So, I need a way to tell
DBI/DBD to bind as a LONG RAW (or a VARBINARY or whatever the name would
be). I used to work on SQL Server some time ago and if I'm not wrong, it
too required different types when using a BLOB (non-interpreted byte)
column as opposed to a very large (but still of string type) column.
The ridiculous thing is that according to the OCI manual, you SHOULD be
able to bind a LONG to a LONG RAW, the obndrv goes fine, but in execution
Oracle screams at you.

> If I don't know I can't fix it. (Then again, if I do know I might not
> fix it, or be able to fix it, anyway :-)
Sorry, I broke the problem in two, but my messages to the mailing list
come back after a very long time (dunno if it happens only to me or to
others also).

>> I have seen it mentioned many times in the mailing list
>> messages, but didn't find it in DBI nor in the specs. Is it a proposed

>> enhancement?
>>
>Yes.

> I'll look into it again. Nag me from time to time if it doesn't show
up.

I suppose I'll have to wait for a DBI 0.7 specs announcement?

> Bye
> Davide Migliavacca
>
Tim.

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06633-8@lion>;
Wed, 23 Aug 1995 07:05:31 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 21:20:02 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa04318;
22 Aug 95 22:19 +0100
Received: by gw.home.vix.com id AA04453; Tue, 22 Aug 95 07:53:30 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA04448; Tue, 22 Aug 95 07:53:26 -0700
Received: (from uucp@localhost) by kirin.clinicom.com (8.6.9/8.6.6) id LAA01395
for <[email protected]>; Tue, 22 Aug 1995 11:08:38 -0600
Received: from clinicom.clinicom.com(10.0.0.60) by kirin.clinicom.com
via smap (V1.3) id sma001393; Tue Aug 22 11:08:34 1995
Received: (from spc@localhost) by clinicom.com (8.6.9/8.6.9) id IAA219316
for [email protected]; Tue, 22 Aug 1995 08:54:21 -0600
From: Stephen Carley <[email protected]>
Message-Id: <[email protected]>
Subject: unsubscribe
To: [email protected]
Date: Tue, 22 Aug 1995 08:54:21 -0600 (MDT)
X-Mailer: ELM [version 2.4 PL24 PGP3 *ALPHA*]
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 12

unsubscribe
---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05893-44@lion>;
Wed, 23 Aug 1995 07:06:35 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 23:17:50 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa28924;
23 Aug 95 0:17 +0100
Received: by gw.home.vix.com id AA04991; Tue, 22 Aug 95 08:01:40 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA04985; Tue, 22 Aug 95 08:01:38 -0700
Received: from barbera.inferentia.it (barbera.inferentia.it [151.99.220.61])
by relay.smau.it (8.6.12/8.6.12) with SMTP id RAA15762;
Tue, 22 Aug 1995 17:00:27 +0200
Posted-Date: Tue, 22 Aug 1995 17:00:27 +0200
Received: by barbera.inferentia.it with Microsoft Mail
id <[email protected]>; Tue, 22 Aug 95 17:07:59 PDT
From: Shari <[email protected]>
To: 'Tim Bunce' <[email protected]>
Cc: 'PerlDB-interest mailing list' <[email protected]>
Subject: RE: DBD::Oracle How to handle RAW and LONG RAW parameters?
Date: Tue, 22 Aug 95 17:07:00 PDT
Message-Id: <[email protected]>
Encoding: 37 TEXT
X-Mailer: Microsoft Mail V3.0

> That's what the 'WARNING: THIS IS ALPHA SOFTWARE' means.

You know, when alpha software is this quality, I sometimes forget the
block-lettered warning :-)

>> If I am missing something, please tell me!

> Only a little patience and understanding :-)

Ehr... well, I suppose I'll go and fetch them in the Perl Archives :-)

> And, if there is a "simple" way to work around this problem (other than

> my rewriting core parts of DBD, which I will not do in order not to
miss
> future releases), please suggest.

> I'm happy for people to work on the DBI and DBD modules if they follow
> a few very simple rules:

> 1) Tell me what you want to do and why *before* you do it.
> 2) Listen to my suggestions (ignore at your peril :-).
> 3) Send me a patch *quickly* based on the last release.

OK. Message follows.

> Tim.

> p.s. Davide, my last message to you (CC'd to the list) bounced:

Yes, I was too absorbed in DBD to notice half my filesystems were full
:-)

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <06633-18@lion>;
Wed, 23 Aug 1995 07:06:45 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Tue, 22 Aug 95 23:37:42 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa06675;
23 Aug 95 0:37 +0100
Received: by gw.home.vix.com id AA06937; Tue, 22 Aug 95 08:36:04 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA06930; Tue, 22 Aug 95 08:36:01 -0700
Received: from barbera.inferentia.it (barbera.inferentia.it [151.99.220.61])
by relay.smau.it (8.6.12/8.6.12) with SMTP id RAA25267;
Tue, 22 Aug 1995 17:34:53 +0200
Posted-Date: Tue, 22 Aug 1995 17:34:53 +0200
Received: by barbera.inferentia.it with Microsoft Mail
id <[email protected]>; Tue, 22 Aug 95 17:42:25 PDT
From: Shari <[email protected]>
To: 'Tim Bunce' <[email protected]>
Cc: 'PerlDB-interest mailing list' <[email protected]>
Subject: Proposed "extension": LONG RAW parameters in DBD::Oracle
Date: Tue, 22 Aug 95 17:42:00 PDT
Message-Id: <[email protected]>
Encoding: 37 TEXT
X-Mailer: Microsoft Mail V3.0

What I want to do:
Insert a very big BLOB in a LONG RAW column in an Oracle table
How:
implementing support for the LONG VARRAW datatype in
bind_param/dbd_bind_ph
Why:
Because I like to be able to do it :-)

Apparently, the most general way of doing it is by using a LONG VARRAW,
which is a buffer with 4 bytes (sb4) prepended to specify the length of
data. In fact, using the length parameter of the obdrn calls one may be
limited to 64k (according to OCI docs).
To use a LONG VARRAW, in the obdrn call, type is 95 (BTW, in
~oracle/rdbms/demo/ocidfn.h there are symbolic #defines SQLT_xxx for
Oracle external datatypes) and the length parameter should be specified
as -1.

Using bind_param this should become a trivial :-) modification to
dbd_bind_ph in dbdimp.c.

There is one additional problem (which of course may be only my scarce
knowledge of Perl).
Using a LONG VARx in C/C++ generally means using a large buffer which is
then accessed (when filling data in) starting at offset sizeof(sb4) so
that no big block moves are necessary aftewards.
Is it possible to do something like this in Perl?
If yes, enlighten me (please). If not, a programmer will generally do the
big block copy leading to inefficiency and memory waste (we're talking of
-maybe- megabytes).

Any suggestion greatly appreciated.

Davide Migliavacca
([email protected])

---+++---
Received: from punt.demon.co.uk by lion with SMTP (PP) id <05893-52@lion>;
Wed, 23 Aug 1995 07:07:59 +0100
Received: from punt.demon.co.uk via puntmail for [email protected];
Wed, 23 Aug 95 00:53:48 GMT
Received: from gw.home.vix.com by punt.demon.co.uk id aa02628;
23 Aug 95 1:52 +0100
Received: by gw.home.vix.com id AA29911; Tue, 22 Aug 95 13:50:44 -0700
X-Btw: vix.com is also gw.home.vix.com and vixie.sf.ca.us
Precedence: junk
Sender: [email protected]
Received: by gw.home.vix.com id AA29905; Tue, 22 Aug 95 13:50:42 -0700
Received: from centhub ([151.116.23.137])
by uswat.advtech.uswest.com (8.6.12/8.6.12) with SMTP id OAA21355;
Tue, 22 Aug 1995 14:50:39 -0600
Original-Received: by centhub.mnet.uswest.com (M-Net
Hub.950111)
PP-warning: Illegal Received field on preceding line
Received: from coefmd3 by coefmd3.uswest.com. (5.0/SMI-SVR4) id AA25843;
Tue, 22 Aug 1995 14:16:57 +0700
Message-Id: <[email protected].>
X-Mailer: exmh version 1.6 4/21/95
To: [email protected]
Cc: [email protected]
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Date: Tue, 22 Aug 1995 14:16:55 -0600
From: John Liptak <[email protected]>
Content-Length: 0

I've never gotten into the perl MakeMaker stuff, what's the problem here?

I am trying to skip from 0.58, I've been busy with other things :-(

$ perl Makefile.PL

Build, TEST and INSTALL Perl 5 first. Use version 5.001m or later.

DO NOT build the DBI under the Perl source tree.

Use 'make test' to execute self tests.

MakeMaker
NAME => 'DBI'
VERSION => '0.61'
Using header files found in /usr/local/lib/perl5/sun4-solaris/CORE
Looking for perl 5 by these names:
miniperl perl,
in these dirs:
/usr/bin /usr/openwin/bin /home/jliptak/bin /opt/IXImd12s/bin /usr/dt/bin
/opt/internet/bin /usr/lib/nis /usr/ucb /usr/sbin /usr/local/CenterLine/bin
/usr/local/CenterLine/sparc-solaris2/bin /usr/local/bin /opt/SUNWspro/bin
/usr/ccs/bin /usr/local/bin
Executing /home/jliptak/bin/perl
Writing Makefile for DBI
$ make
mkdir /home/jliptak/DBI/blib
mkdir /home/jliptak/DBI/blib/auto
mkdir /home/jliptak/DBI/blib/auto/DBI
/home/jliptak/bin/perl /usr/local/lib/perl5/ExtUtils/xsubpp -typemap
/usr/local/lib/perl5/ExtUtils/typemap DBI.xs >DBI.tc && mv DBI.tc DBI.c
Error: no tab in DBD::_::st::fetchrow argument declaration ' SV *sth'
Error: no tab in DBD::_::st::fetch argument declaration ' SV *sth'
Error: no tab in DBD::_::common::STORE argument declaration ' SV * h'
'SV * ' not in typemap at /usr/local/lib/perl5/ExtUtils/xsubpp line
527, <F> line 1198.
Error: no tab in DBD::_::common::STORE argument declaration ' SV *
keysv'
Error: no tab in DBD::_::common::STORE argument declaration ' SV *
valuesv'
clcc -c -O -PIC -I/usr/local/lib/perl5/sun4-solaris/CORE DBI.c
"./DBIXS.h", line 35: undefined control
*** Error code 2
make: Fatal error: Command failed for target `DBI.o'

John
--
John Liptak - MTS U S WEST Communications
WK (303) 624-0140 931 14TH Street 9TH FL.
PG (303) 820-9284 Denver, Colorado 80202
[email protected] [email protected]

---+++---