Обсуждение: new maintainer for the ODBC driver?

> know about). Now that Byron is unable (I assume not unwilling) to
> contribute as he did, I think we need someone to take over
> responsibility or at least get heavily involved as soon as possible.
>
> I would volunteer myself, but my C is isn't really up to it and pgAdmin
> takes a fair bit of time... Anyone else?
> 
> Regards,
> Dave.
The ODBC driver still seems to be an important part of PostgreSQL. So I
really support your idea. Unfortunately I really have no time for this and
I'm rather new to the driver and Postgres internal communication. AFAIK
Thomas Lockhart offered his work in May but I don't think he had enough
time to check what happened on the list since then.

Regards, Zoltan

Whoever does do it, should have access to the Windows/Microsoft side of
things -- not just Visual C++/Studio so he/she can compile an executable
version of the driver,  but also the applications, such as Access 2000,
Borland <insert product here>, etc.    Since my old computer is gone, I
currently have none of the applications I need to do the job any longer. :(

It is very important that most of the applications people are using these
days with the driver be available to the person maintaining the driver so
that all the crazy odbc behaviors can be tested. I know I would rather not
have to go out and buy an application software package just to test it with
the odbc driver ( BTW,  I did this in the past with Borland c++ builder).

Also, it seems rather difficult to be maintaining a version that works well
with Unix and also works with Windows.

Any thoughts?

Byron

----- Original Message -----
From: Kovacs Zoltan Sandor <tip@pc10.radnoti-szeged.sulinet.hu>
To: <pgsql-interfaces@postgresql.org>
Cc: <dpage@vale-housing.co.uk>; <lockhart@alumni.caltech.edu>
Sent: Monday, October 23, 2000 7:16 AM
Subject: [INTERFACES] new maintainer for the ODBC driver?


> > know about). Now that Byron is unable (I assume not unwilling) to
> > contribute as he did, I think we need someone to take over
> > responsibility or at least get heavily involved as soon as possible.
> >
> > I would volunteer myself, but my C is isn't really up to it and pgAdmin
> > takes a fair bit of time... Anyone else?
> >
> > Regards,
> > Dave.
> The ODBC driver still seems to be an important part of PostgreSQL. So I
> really support your idea. Unfortunately I really have no time for this and
> I'm rather new to the driver and Postgres internal communication. AFAIK
> Thomas Lockhart offered his work in May but I don't think he had enough
> time to check what happened on the list since then.
>
> Regards, Zoltan
>

> The ODBC driver still seems to be an important part of PostgreSQL. So I
> really support your idea. Unfortunately I really have no time for this and
> I'm rather new to the driver and Postgres internal communication. AFAIK
> Thomas Lockhart offered his work in May but I don't think he had enough
> time to check what happened on the list since then.

:(

I did some work with ODBC to get it building for Unix boxes and to work
with Applixware. But afaik my environment does not allow me to build and
test other new features, so I can be of only limited help here.

That doesn't excuse my taking so long with Zoltan's fixes, but a new
ODBC maintainer is clearly needed. He/she would probably need some
familiarity with Windows tools such as Access to be most effective.

Zoltan, I've applied your patches to my tree, and will get them into the
development sources asap.
                    - Thomas

Perhaps the amount of work required to maintain the ODBC driver would be
lower if it was changed to sit on top of the libpq library. It is a large
thing to move it there, but once it's done it should be easier for it to
follow the upgrades without requiring large work.

Disclaimer: I haven't checked the code out really much, so I can't say if
this is true :-)

Other than that, perhaps some kind of "test suite" that should be "checked
off" by somebody before each relese - not necessarily checked by the person
maintaining it every time?

//Magnus

> -----Original Message-----
> From: Byron Nikolaidis [mailto:byron.nikolaidis@home.com]
> Sent: den 23 oktober 2000 17:17
> To: Kovacs Zoltan Sandor; pgsql-interfaces@postgreSQL.org
> Cc: dpage@vale-housing.co.uk; lockhart@alumni.caltech.edu
> Subject: Re: [INTERFACES] new maintainer for the ODBC driver?
> 
> 
> 
> Whoever does do it, should have access to the 
> Windows/Microsoft side of
> things -- not just Visual C++/Studio so he/she can compile an 
> executable
> version of the driver,  but also the applications, such as 
> Access 2000,
> Borland <insert product here>, etc.    Since my old computer 
> is gone, I
> currently have none of the applications I need to do the job 
> any longer. :(
> 
> It is very important that most of the applications people are 
> using these
> days with the driver be available to the person maintaining 
> the driver so
> that all the crazy odbc behaviors can be tested. I know I 
> would rather not
> have to go out and buy an application software package just 
> to test it with
> the odbc driver ( BTW,  I did this in the past with Borland 
> c++ builder).
> 
> Also, it seems rather difficult to be maintaining a version 
> that works well
> with Unix and also works with Windows.
> 
> Any thoughts?
> 
> Byron
> 
> ----- Original Message -----
> From: Kovacs Zoltan Sandor <tip@pc10.radnoti-szeged.sulinet.hu>
> To: <pgsql-interfaces@postgresql.org>
> Cc: <dpage@vale-housing.co.uk>; <lockhart@alumni.caltech.edu>
> Sent: Monday, October 23, 2000 7:16 AM
> Subject: [INTERFACES] new maintainer for the ODBC driver?
> 
> 
> > > know about). Now that Byron is unable (I assume not unwilling) to
> > > contribute as he did, I think we need someone to take over
> > > responsibility or at least get heavily involved as soon 
> as possible.
> > >
> > > I would volunteer myself, but my C is isn't really up to 
> it and pgAdmin
> > > takes a fair bit of time... Anyone else?
> > >
> > > Regards,
> > > Dave.
> > The ODBC driver still seems to be an important part of 
> PostgreSQL. So I
> > really support your idea. Unfortunately I really have no 
> time for this and
> > I'm rather new to the driver and Postgres internal 
> communication. AFAIK
> > Thomas Lockhart offered his work in May but I don't think 
> he had enough
> > time to check what happened on the list since then.
> >
> > Regards, Zoltan
> >
> 

Where is, if there is, a list of things that need to be fixed in the ODBC
driver?  As far as I'm concerned (with Access) it works reasonably well.  
There are only a couple of things I can think of off hand, like RI
reporting and the nested function bug,'hi' & ' ' & 'there' != textcat( fn( 'hi', ' ') 'there')
which are ones I've come across myself (which is probably why I do
remember them).

Also, has anyone had any success building the ODBC driver under windows
with free tools?  Hmm, wait a second, come to think of it, I do have
something that might work, Borland C++ 4.0 (for Dos, Windows, and Windows
NT).  It was given to me recently, but I never bothered to install it as
I've never done any C outside of Linux.  Is this good enough to compile
the ODBC driver?  I'd like to at least be able to try out some patches and
provide feedback to whoever does work on the driver.

-Cedar

Be wary... I'm not sure hat I'm getting into... but I want to help...

I would volunteer for the ODBC help, but I am not knowledgeable with C (I
can manipulate data and print, but that is about it. :P).  I can code a
decent amount in VB though, so if people help me on what to test on the ODBC
driver, I can run up a type of "test app" where it tries to connect via ODBC
to a postgres database and run a series of queries and see what happens.

I assume that might help, because then it can report back what type of
queries fail and such.  Also, if the ODBC runs in VB, odds are it should
work in the vast majority of other Windows apps.

Only requirement would be patience with me. :)

Adam Lang
Systems Engineer
Rutgers Casualty Insurance Company
----- Original Message -----
From: "Magnus Hagander" <mha@sollentuna.net>
To: "'Byron Nikolaidis'" <byron.nikolaidis@home.com>; "Kovacs Zoltan Sandor"
<tip@pc10.radnoti-szeged.sulinet.hu>; <pgsql-interfaces@postgresql.org>
Cc: <dpage@vale-housing.co.uk>; <lockhart@alumni.caltech.edu>
Sent: Monday, October 23, 2000 1:12 PM
Subject: RE: [INTERFACES] new maintainer for the ODBC driver?


> Perhaps the amount of work required to maintain the ODBC driver would be
> lower if it was changed to sit on top of the libpq library. It is a large
> thing to move it there, but once it's done it should be easier for it to
> follow the upgrades without requiring large work.
>
> Disclaimer: I haven't checked the code out really much, so I can't say if
> this is true :-)
>
> Other than that, perhaps some kind of "test suite" that should be "checked
> off" by somebody before each relese - not necessarily checked by the
person
> maintaining it every time?
>
> //Magnus
>
> > -----Original Message-----
> > From: Byron Nikolaidis [mailto:byron.nikolaidis@home.com]
> > Sent: den 23 oktober 2000 17:17
> > To: Kovacs Zoltan Sandor; pgsql-interfaces@postgreSQL.org
> > Cc: dpage@vale-housing.co.uk; lockhart@alumni.caltech.edu
> > Subject: Re: [INTERFACES] new maintainer for the ODBC driver?
> >
> >
> >
> > Whoever does do it, should have access to the
> > Windows/Microsoft side of
> > things -- not just Visual C++/Studio so he/she can compile an
> > executable
> > version of the driver,  but also the applications, such as
> > Access 2000,
> > Borland <insert product here>, etc.    Since my old computer
> > is gone, I
> > currently have none of the applications I need to do the job
> > any longer. :(
> >
> > It is very important that most of the applications people are
> > using these
> > days with the driver be available to the person maintaining
> > the driver so
> > that all the crazy odbc behaviors can be tested. I know I
> > would rather not
> > have to go out and buy an application software package just
> > to test it with
> > the odbc driver ( BTW,  I did this in the past with Borland
> > c++ builder).
> >
> > Also, it seems rather difficult to be maintaining a version
> > that works well
> > with Unix and also works with Windows.
> >
> > Any thoughts?
> >
> > Byron
> >
> > ----- Original Message -----
> > From: Kovacs Zoltan Sandor <tip@pc10.radnoti-szeged.sulinet.hu>
> > To: <pgsql-interfaces@postgresql.org>
> > Cc: <dpage@vale-housing.co.uk>; <lockhart@alumni.caltech.edu>
> > Sent: Monday, October 23, 2000 7:16 AM
> > Subject: [INTERFACES] new maintainer for the ODBC driver?
> >
> >
> > > > know about). Now that Byron is unable (I assume not unwilling) to
> > > > contribute as he did, I think we need someone to take over
> > > > responsibility or at least get heavily involved as soon
> > as possible.
> > > >
> > > > I would volunteer myself, but my C is isn't really up to
> > it and pgAdmin
> > > > takes a fair bit of time... Anyone else?
> > > >
> > > > Regards,
> > > > Dave.
> > > The ODBC driver still seems to be an important part of
> > PostgreSQL. So I
> > > really support your idea. Unfortunately I really have no
> > time for this and
> > > I'm rather new to the driver and Postgres internal
> > communication. AFAIK
> > > Thomas Lockhart offered his work in May but I don't think
> > he had enough
> > > time to check what happened on the list since then.
> > >
> > > Regards, Zoltan
> > >
> >

> The ODBC driver still seems to be an important part of PostgreSQL. So I
> really support your idea. Unfortunately I really have no time for this and
> I'm rather new to the driver and Postgres internal communication. AFAIK
> Thomas Lockhart offered his work in May but I don't think he had enough
> time to check what happened on the list since then.

Hi Zoltan. I *finally* committed your patches to the main tree a few
minutes ago. Sorry for the huge delay. afaik you are the defacto
maintainer of ODBC :)
                      - Thomas

On Wed, 25 Oct 2000, Thomas Lockhart wrote:

> Hi Zoltan. I *finally* committed your patches to the main tree a few
> minutes ago. Sorry for the huge delay. afaik you are the defacto
> maintainer of ODBC :)

Huh... What to say now... :-?

Well, Byron's comment about THE maintainer (he/she should have a large
amount of M$ application) points out that we could only continue using
ODBC in future versions of PostgreSQL if all of us test the driver
periodically with our installed Windows applications. I.e., Cedar might
test the Access side, Dave might test the pgadmin side (I don't remember
if it is written in VB or VC), Thomas might test the Applixware side and
we test it with our BCB/ODBCExpress application as well. As Magnus and
Adam wrote, the best thing would be to write something like Postgres'
regression tests but I think it would be very much time---the applications
already working with earlier versions of the driver might be used instead.
I think that among the above things pgadmin is the most important one
because---checking the mailing lists---it is the most widely used Windows
application for Postgres. But I'm not sure.

There must be (at least---and in my opinion at most) one man ("the
patcher") who is responsible for patching the source with the new codes.
As Byron wrote, each change should be tested by at least 4 people. So I
don't think it would be useful to change the CVS immediately, only after
every tester reported that the new version works well.

Respecting to Cedar's opinion about the free C compilers, I know that the
Borland C++ 5.5 command-line compiler is free, but I never used it. Other
possibilites are DJGPP which perhaps could create DLLs (but I also never
used it for making any code for Windows). CygWin can be also a solution.
This week I would like to try all these (th/f)ree compilers with the ODBC
source.

I know that the current source can be compiled with VC and BCB 4/5 as
well, the compiling process is almost automatic, but these compilers are
not free. In a need I can compile the DLL with our registered BCB 4.

Finally, a question to Byron: what are your future plans with the driver?
Are you going to have some time for checking the new parts of the code or
answerring questions about the working mechanism of the old parts?

IMHO, the ODBC driver at the current state is a great one, so it's worth
to continue the work on it. Have you got any statistics about the number
of people using the driver at the moment? This year most of the mails
on the INTERFACES list are about the JDBC driver. So, sometimes I don't
think that the ODBC driver has any future. But now, maybe... :-)

Regards, Zoltan

I have a "learning" edition of VC++ 4.  Never really used it.... but I could
try to compile if necessary. (Never to late to learn, right?)

As for driver support, I think the reason there is not much static on the
ODBC Windows end is that most coders that write for Windows (that use VB,
etc.) are not using Linux Application Servers for backends.  Sort of a
different mentality than Java developers.  Java typically has the mindset of
"any platform" and VB, VC, Access, tends to have the mindset of "Microsoft
Platform".

For Postgres to make inroads on Windows based apps, an ODBC driver has to be
supported... or even better, an OLE DB driver (for ADO connections).

Adam Lang
Systems Engineer
Rutgers Casualty Insurance Company
----- Original Message -----
From: "kovacsz" <zoli@pc10.radnoti-szeged.sulinet.hu>
To: "Thomas Lockhart" <lockhart@alumni.caltech.edu>
Cc: <pgsql-interfaces@postgresql.org>; <dpage@vale-housing.co.uk>;
<mha@sollentuna.net>; <cedarc@visionforisrael.com>;
<aalang@rutgersinsurance.com>; <byron.nikolaidis@home.com>
Sent: Thursday, October 26, 2000 11:00 AM
Subject: Re: new maintainer for the ODBC driver?


>
> On Wed, 25 Oct 2000, Thomas Lockhart wrote:
>
> > Hi Zoltan. I *finally* committed your patches to the main tree a few
> > minutes ago. Sorry for the huge delay. afaik you are the defacto
> > maintainer of ODBC :)
>
> Huh... What to say now... :-?
>
> Well, Byron's comment about THE maintainer (he/she should have a large
> amount of M$ application) points out that we could only continue using
> ODBC in future versions of PostgreSQL if all of us test the driver
> periodically with our installed Windows applications. I.e., Cedar might
> test the Access side, Dave might test the pgadmin side (I don't remember
> if it is written in VB or VC), Thomas might test the Applixware side and
> we test it with our BCB/ODBCExpress application as well. As Magnus and
> Adam wrote, the best thing would be to write something like Postgres'
> regression tests but I think it would be very much time---the applications
> already working with earlier versions of the driver might be used instead.
> I think that among the above things pgadmin is the most important one
> because---checking the mailing lists---it is the most widely used Windows
> application for Postgres. But I'm not sure.
>
> There must be (at least---and in my opinion at most) one man ("the
> patcher") who is responsible for patching the source with the new codes.
> As Byron wrote, each change should be tested by at least 4 people. So I
> don't think it would be useful to change the CVS immediately, only after
> every tester reported that the new version works well.
>
> Respecting to Cedar's opinion about the free C compilers, I know that the
> Borland C++ 5.5 command-line compiler is free, but I never used it. Other
> possibilites are DJGPP which perhaps could create DLLs (but I also never
> used it for making any code for Windows). CygWin can be also a solution.
> This week I would like to try all these (th/f)ree compilers with the ODBC
> source.
>
> I know that the current source can be compiled with VC and BCB 4/5 as
> well, the compiling process is almost automatic, but these compilers are
> not free. In a need I can compile the DLL with our registered BCB 4.
>
> Finally, a question to Byron: what are your future plans with the driver?
> Are you going to have some time for checking the new parts of the code or
> answerring questions about the working mechanism of the old parts?
>
> IMHO, the ODBC driver at the current state is a great one, so it's worth
> to continue the work on it. Have you got any statistics about the number
> of people using the driver at the moment? This year most of the mails
> on the INTERFACES list are about the JDBC driver. So, sometimes I don't
> think that the ODBC driver has any future. But now, maybe... :-)
>
> Regards, Zoltan
>

-- 
Peter Mount
Enterprise Support Officer, Maidstone Borough Council
Email: petermount@maidstone.gov.uk
WWW: http://www.maidstone.gov.uk
All views expressed within this email are not the views of Maidstone Borough
Council


-----Original Message-----
From: kovacsz [mailto:zoli@pc10.radnoti-szeged.sulinet.hu]
Sent: Thursday, October 26, 2000 4:01 PM
To: Thomas Lockhart
Cc: pgsql-interfaces@postgresql.org; dpage@vale-housing.co.uk;
mha@sollentuna.net; cedarc@visionforisrael.com;
aalang@rutgersinsurance.com; byron.nikolaidis@home.com
Subject: [INTERFACES] Re: new maintainer for the ODBC driver?

[snip]

IMHO, the ODBC driver at the current state is a great one, so it's worth
to continue the work on it. Have you got any statistics about the number
of people using the driver at the moment? This year most of the mails
on the INTERFACES list are about the JDBC driver. So, sometimes I don't
think that the ODBC driver has any future. But now, maybe... :-)

PM: I had noticed that the ODBC driver had gone quiet lately, but then even
JDBC has it's mad moments, then quiet spells. Both JDBC & ODBC are similar
api's, so there can be some overlap between them. Some time ago, I did
mention about copying the JDBC's crypt implementation over to ODBC (strip
some Java specific stuff out and you'll have pure C) to enable ODBC to use
crypt().

In article <Pine.LNX.4.10.10010261603500.3001-100000@tir>, kovacsz
<zoli@pc10.radnoti-szeged.sulinet.hu> writes

>There must be (at least---and in my opinion at most) one man ("the
>patcher") who is responsible for patching the source with the new codes.
>As Byron wrote, each change should be tested by at least 4 people. So I
>don't think it would be useful to change the CVS immediately, only after
>every tester reported that the new version works well.
I thought the whole point of a CVS tree was that it was current i.e. not
guaranteed to be bug free. The sources can be marked with release tags
when they are considered stable. This would mean that you can have a
virtually unlimited number of people testing the code.
>
>Respecting to Cedar's opinion about the free C compilers, I know that the
>Borland C++ 5.5 command-line compiler is free, but I never used it. Other
>possibilites are DJGPP which perhaps could create DLLs (but I also never
>used it for making any code for Windows). CygWin can be also a solution.
>This week I would like to try all these (th/f)ree compilers with the ODBC
>source.
There is another compiler called 'mingw32', which is a native GCC
distribution for Windows - I don't know if it is properly maintained.
The other option would be to use a cross compiler.

>IMHO, the ODBC driver at the current state is a great one, so it's worth
>to continue the work on it. Have you got any statistics about the number
>of people using the driver at the moment? 
Me me me! I use it for a hospital database system. And it works very
well!
-- 
Jamie Walker              "While there are no known bugs in it, it might 
jamie@sagaxis.co.uk        destroy your filesystems, eat your data and
http://www.sagaxis.co.uk/  start World War III. You have been warned."



________________________________________________________                          1stUp.com - Free the Web  Get your
freeInternet access at http://www.1stUp.com
 

Oh geez! I'm almost forgot, I do have a copy of Turbo C++ 4.5 (I think).
Wow... That C class I took really must have fallen into the back of my mind.
Again,  I have no clue if they are appropriate for compiling the ODBC
driver.

Adam Lang
Systems Engineer
Rutgers Casualty Insurance Company
----- Original Message -----
From: "Jamie Walker" <jamie@sagaxis.co.uk>
To: <pgsql-interfaces@postgresql.org>
Sent: Thursday, October 26, 2000 8:26 PM
Subject: Re: [INTERFACES] Re: new maintainer for the ODBC driver?


> In article <Pine.LNX.4.10.10010261603500.3001-100000@tir>, kovacsz
> <zoli@pc10.radnoti-szeged.sulinet.hu> writes
>
> >There must be (at least---and in my opinion at most) one man ("the
> >patcher") who is responsible for patching the source with the new codes.
> >As Byron wrote, each change should be tested by at least 4 people. So I
> >don't think it would be useful to change the CVS immediately, only after
> >every tester reported that the new version works well.
> I thought the whole point of a CVS tree was that it was current i.e. not
> guaranteed to be bug free. The sources can be marked with release tags
> when they are considered stable. This would mean that you can have a
> virtually unlimited number of people testing the code.
> >
> >Respecting to Cedar's opinion about the free C compilers, I know that the
> >Borland C++ 5.5 command-line compiler is free, but I never used it. Other
> >possibilites are DJGPP which perhaps could create DLLs (but I also never
> >used it for making any code for Windows). CygWin can be also a solution.
> >This week I would like to try all these (th/f)ree compilers with the ODBC
> >source.
> There is another compiler called 'mingw32', which is a native GCC
> distribution for Windows - I don't know if it is properly maintained.
> The other option would be to use a cross compiler.
>
> >IMHO, the ODBC driver at the current state is a great one, so it's worth
> >to continue the work on it. Have you got any statistics about the number
> >of people using the driver at the moment?
> Me me me! I use it for a hospital database system. And it works very
> well!
> --
> Jamie Walker              "While there are no known bugs in it, it might
> jamie@sagaxis.co.uk        destroy your filesystems, eat your data and
> http://www.sagaxis.co.uk/  start World War III. You have been warned."
>
>
>
> ________________________________________________________
>                            1stUp.com - Free the Web
>    Get your free Internet access at http://www.1stUp.com

> Well, Byron's comment about THE maintainer (he/she should have a large
> amount of M$ application) points out that we could only continue using
> ODBC in future versions of PostgreSQL if all of us test the driver
> periodically with our installed Windows applications. I.e., Cedar might
> test the Access side, Dave might test the pgadmin side (I don't remember
> if it is written in VB or VC), Thomas might test the Applixware side and
> we test it with our BCB/ODBCExpress application as well. As Magnus and
> Adam wrote, the best thing would be to write something like Postgres'
> regression tests but I think it would be very much time---the applications
> already working with earlier versions of the driver might be used instead.
> I think that among the above things pgadmin is the most important one
> because---checking the mailing lists---it is the most widely used Windows
> application for Postgres. But I'm not sure.
> There must be (at least---and in my opinion at most) one man ("the
> patcher") who is responsible for patching the source with the new codes.
> As Byron wrote, each change should be tested by at least 4 people. So I
> don't think it would be useful to change the CVS immediately, only after
> every tester reported that the new version works well.

I'll disagree, at least a little. The ODBC driver benefited tremendously
from Byron's dedicated work. However, since he held the source code
fairly closely, it was difficult for others to step in and do
significant work.

Another example applies to you, Zoltan. Let's consider me the maintainer
of ODBC for this argument. I sat on your patches for *months* before
applying them. That is both bad and unnecessary. imho there is no reason
why the ODBC driver can not be developed similarly to the rest of the
PostgreSQL code tree, with people contributing patches, other people
testing them, the tree varying from "slightly broken" to "rock solid",
etc etc.

The important thing is that people are interested, and that contributors
take care in what they are doing. If someone has a patch that needs
testing, post it and ask for help. If it is a "no brainer", or if you
have already been able to test at least a little, then ask for it to be
applied (and if someone has the time and interest to stay on top of it
then they will likely have committer's privilege to the tree ;)

> IMHO, the ODBC driver at the current state is a great one, so it's worth
> to continue the work on it. Have you got any statistics about the number
> of people using the driver at the moment? This year most of the mails
> on the INTERFACES list are about the JDBC driver. So, sometimes I don't
> think that the ODBC driver has any future. But now, maybe... :-)

Ah, I think it is more used, and more valuable, than we realize. When
things "just work" you don't hear much about it :)

The several people who are in this thread would be a great combination
to support the ODBC driver. Byron does not have a machine to do
development, and I'm sure he would welcome the sight of his ODBC code
thriving in the future.
                           - Thomas

kovacsz wrote:
> 
> On Wed, 25 Oct 2000, Thomas Lockhart wrote:
> 
> > Hi Zoltan. I *finally* committed your patches to the main tree a few
> > minutes ago. Sorry for the huge delay. 

Sorry, I don't have the originial mails to this thread (I've only been
subscribed to Interfaces since Thursday). I am currently trying to get
ODBC to work with 7.1. Does this (the fact that patches to the ODBC
driver had been waiting in the wings) mean that the ODBC driver in the
snapshot from last week (19th) - this is the version I am currently
using, which has been working great (including TOAST!) so far - won't
work?

Regards,
Frank

I having been lurking on this thread for a while and would like to make a 
suggestion.  Have you guys thought about using SourceForge to help maintain 
the
driver?  They have resources for compiling, testing, etc...  As well you might
be able to get more people involved by posting in the Help Wanted section.

Just a thought.

On Fri, 27 Oct 2000 15:07:49 +0000, Thomas Lockhart 
<lockhart@alumni.caltech.edu> wrote:

> > Well, Byron's comment about THE maintainer (he/she should have a large
> > amount of M$ application) points out that we could only continue using
> > ODBC in future versions of PostgreSQL if all of us test the driver
> > periodically with our installed Windows applications. I.e., Cedar might
> > test the Access side, Dave might test the pgadmin side (I don't remember
> > if it is written in VB or VC), Thomas might test the Applixware side and
> > we test it with our BCB/ODBCExpress application as well. As Magnus and
> > Adam wrote, the best thing would be to write something like Postgres'
> > regression tests but I think it would be very much time---the applications
> > already working with earlier versions of the driver might be used instead.
> > I think that among the above things pgadmin is the most important one
> > because---checking the mailing lists---it is the most widely used Windows
> > application for Postgres. But I'm not sure.
> > There must be (at least---and in my opinion at most) one man ("the
> > patcher") who is responsible for patching the source with the new codes.
> > As Byron wrote, each change should be tested by at least 4 people. So I
> > don't think it would be useful to change the CVS immediately, only after
> > every tester reported that the new version works well.
> 
> I'll disagree, at least a little. The ODBC driver benefited tremendously
> from Byron's dedicated work. However, since he held the source code
> fairly closely, it was difficult for others to step in and do
> significant work.
> 
> Another example applies to you, Zoltan. Let's consider me the maintainer
> of ODBC for this argument. I sat on your patches for *months* before
> applying them. That is both bad and unnecessary. imho there is no reason
> why the ODBC driver can not be developed similarly to the rest of the
> PostgreSQL code tree, with people contributing patches, other people
> testing them, the tree varying from "slightly broken" to "rock solid",
> etc etc.
> 
> The important thing is that people are interested, and that contributors
> take care in what they are doing. If someone has a patch that needs
> testing, post it and ask for help. If it is a "no brainer", or if you
> have already been able to test at least a little, then ask for it to be
> applied (and if someone has the time and interest to stay on top of it
> then they will likely have committer's privilege to the tree ;)
> 
> > IMHO, the ODBC driver at the current state is a great one, so it's worth
> > to continue the work on it. Have you got any statistics about the number
> > of people using the driver at the moment? This year most of the mails
> > on the INTERFACES list are about the JDBC driver. So, sometimes I don't
> > think that the ODBC driver has any future. But now, maybe... :-)
> 
> Ah, I think it is more used, and more valuable, than we realize. When
> things "just work" you don't hear much about it :)
> 
> The several people who are in this thread would be a great combination
> to support the ODBC driver. Byron does not have a machine to do
> development, and I'm sure he would welcome the sight of his ODBC code
> thriving in the future.
> 
>                             - Thomas
> 

> > > Hi Zoltan. I *finally* committed your patches to the main tree a few
> > > minutes ago. Sorry for the huge delay.
> Sorry, I don't have the originial mails to this thread (I've only been
> subscribed to Interfaces since Thursday). I am currently trying to get
> ODBC to work with 7.1. Does this (the fact that patches to the ODBC
> driver had been waiting in the wings) mean that the ODBC driver in the
> snapshot from last week (19th) - this is the version I am currently
> using, which has been working great (including TOAST!) so far - won't
> work?

??

If the driver is working for you now, it will continue to work. Zoltan
fixed a couple of features related to error reporting and to referential
integrity. If you need those then the patches might help.

If the driver is not working for you now, be sure you rtfm. Setting it
up for the first time is usually not trivial, though it is if a miracle
happens and you do every step in exactly the right way (even the steps
which are not well described, hence the miracle ;)
                    - Thomas

> I having been lurking on this thread for a while and would like to make a
> suggestion.  Have you guys thought about using SourceForge to help maintain
> the driver?  They have resources for compiling, testing, etc...  As well you
> might be able to get more people involved by posting in the Help Wanted section.

That is a possibility. We have been a "PostgreSQL community" for several
years, and imho it would be nice to support the core and related
products through that community. If we fail to do that over time, then
perhaps SourceForge would be a help. Hmm, maybe if they see the light
and run the One True Database... ;)
                        - Thomas

> -----Original Message-----
> From: kovacsz [mailto:zoli@pc10.radnoti-szeged.sulinet.hu]
> Sent: 26 October 2000 15:01
> To: Thomas Lockhart
> Cc: pgsql-interfaces@postgresql.org; dpage@vale-housing.co.uk;
> mha@sollentuna.net; cedarc@visionforisrael.com;
> aalang@rutgersinsurance.com; byron.nikolaidis@home.com
> Subject: Re: new maintainer for the ODBC driver?
> 
> 
> 
> On Wed, 25 Oct 2000, Thomas Lockhart wrote:
> 
> > Hi Zoltan. I *finally* committed your patches to the main tree a few
> > minutes ago. Sorry for the huge delay. afaik you are the defacto
> > maintainer of ODBC :)
> 
> Huh... What to say now... :-?
> 
> Well, Byron's comment about THE maintainer (he/she should have a large
> amount of M$ application) points out that we could only continue using
> ODBC in future versions of PostgreSQL if all of us test the driver
> periodically with our installed Windows applications. I.e., 
> Cedar might
> test the Access side, Dave might test the pgadmin side (I 
> don't remember
> if it is written in VB or VC), Thomas might test the 
> Applixware side and
> we test it with our BCB/ODBCExpress application as well. As Magnus and
> Adam wrote, the best thing would be to write something like Postgres'
> regression tests but I think it would be very much time---the 
> applications
> already working with earlier versions of the driver might be 
> used instead.
> I think that among the above things pgadmin is the most important one
> because---checking the mailing lists---it is the most widely 
> used Windows
> application for Postgres. But I'm not sure.

(pgAdmin is written in VB) I certainly can help with testing. I could
obviously test with pgAdmin, but can also test other functions that pgAdmin
doesn't use - knocking up test apps in VB is generally very quick.

> IMHO, the ODBC driver at the current state is a great one, so 
> it's worth
> to continue the work on it. 

I agree. If anyone has ever messed around getting Merant/Data
Direct/Informix's drivers working with Informix they would agree whole
heartedly as well!! (in my opinion of course!)

That said, there are still things missing (don't ask me what, I haven't
looked into it yet). The biggest giveaway is when using ADO in VB - the
message 'Driver does not support requested properties' is not uncommon. This
does tend to be when messing around with the nut and bolts though, not in
general day to day use (fetching data, executing queries etc.). The
OpenSchema method of the ADODB.Connection object really shows this
problem....

> Have you got any statistics about 
> the number
> of people using the driver at the moment? This year most of the mails
> on the INTERFACES list are about the JDBC driver. So, 
> sometimes I don't
> think that the ODBC driver has any future. But now, maybe... :-)

I suspect it is more likely the case that the driver meets *most* people's
requirements in it's current form and is very stable. In the last 2 weeks,
pgAdmin v7.0.2 has been downloaded about 400 times from my site alone - if
you assume that only developers/admins use pgAdmin and that even in our 160
staff company that accounts for 2 of our 50 or so PostgreSQL/ODBC users,
then you are looking at huge numbers world wide... ;-)

Regards,
Dave.

> ODBC to work with 7.1. Does this (the fact that patches to the ODBC
> driver had been waiting in the wings) mean that the ODBC driver in the
> snapshot from last week (19th) - this is the version I am currently
> using, which has been working great (including TOAST!) so far - won't
> work?
I'm almost sure that the current WinODBC driver (v6.5) will work with the
new PG (7.1) but I didn't try it. The meaning of "we currently haven't got
a maintainer" is that if some development is required on the connection
part of the whole thing for 7.1, it is almost surely not written yet. In
other words, no new features are included in the driver since 7.0 beta
releases. AFAIK the TOAST implementation didn't change anything in the
connection part, there are only changes in backend internals.

So don't panic about the version number of the driver which is still 6.5.

Regards, Zoltan

So, so far we have a few people with access to a compiler (including myself
- VC++ 6 Enterprise) but not necessarily the required C skills. 

Is there any one who knows the driver and can maintain it but doesn't have a
compiler? If so perhaps between us (the readers/contributers to this list)
we can get something sorted out.

On a slightly different note, what about the InstallShield code for the
setup program - do we have access to that (Byron?), it works so well it
would be a shame to have to re-write it.

Any ideas? :-)

Regards,
Dave.

Yes, the installshield code should be in the same directory as the driver
code in the cvs repository.   It has an odd extension I think.

Byron

----- Original Message -----
From: Dave Page <dpage@vale-housing.co.uk>
To: 'kovacsz' <zoli@pc10.radnoti-szeged.sulinet.hu>; Thomas Lockhart
<lockhart@alumni.caltech.edu>
Cc: <pgsql-interfaces@postgresql.org>; <byron.nikolaidis@home.com>
Sent: Tuesday, October 31, 2000 12:27 PM
Subject: RE: [INTERFACES] RE: new maintainer for the ODBC driver?


> So, so far we have a few people with access to a compiler (including
myself
> - VC++ 6 Enterprise) but not necessarily the required C skills.
>
> Is there any one who knows the driver and can maintain it but doesn't have
a
> compiler? If so perhaps between us (the readers/contributers to this list)
> we can get something sorted out.
>
> On a slightly different note, what about the InstallShield code for the
> setup program - do we have access to that (Byron?), it works so well it
> would be a shame to have to re-write it.
>
> Any ideas? :-)
>
> Regards,
> Dave.

Seems to be in:

.../src/interfaces/odbc/setup.rul

Ross

On Tue, Oct 31, 2000 at 02:23:44PM -0500, Byron Nikolaidis wrote:
> Yes, the installshield code should be in the same directory as the driver
> code in the cvs repository.   It has an odd extension I think.
> 
> Byron
> 

Ross Reedstrom
-- 
Open source code is like a natural resource, it's the result of providing
food and sunshine to programmers, and then staying out of their way.
[...] [It] is not going away because it has utility for both the developers 
and users independent of economic motivations.  Jim Flynn, Sunnyvale, Calif.

In article <001b01c04370$15e98ec0$6401a8c0@byron>, Byron Nikolaidis
<byron.nikolaidis@home.com> writes
>Yes, the installshield code should be in the same directory as the driver
>code in the cvs repository.   It has an odd extension I think.
Could someone package the next release of the driver with a newer
version of InstallShield? The version used for the current driver does
not support the options for silent unattended installation. This would
be a great help for automated updates and installations on LANs.
-- 
Jamie Walker              "While there are no known bugs in it, it might 
jamie@sagaxis.co.uk        destroy your filesystems, eat your data and
http://www.sagaxis.co.uk/  start World War III. You have been warned."

> Could someone package the next release of the driver with a newer
> version of InstallShield? The version used for the current driver does
> not support the options for silent unattended installation. This would
> be a great help for automated updates and installations on LANs.
Maybe VC++ 6 Enterprise contains this new InstallShield (Dave?). But,
until we don't have a maintainer, we are happy to be able to put a new DLL
together at all... :-)

Zoltan

> > Could someone package the next release of the driver with a newer
> > version of InstallShield? The version used for the current 
> driver does
> > not support the options for silent unattended installation. 
> This would
> > be a great help for automated updates and installations on LANs.
> Maybe VC++ 6 Enterprise contains this new InstallShield (Dave?). But,
> until we don't have a maintainer, we are happy to be able to 
> put a new DLL
> together at all... :-)

Visual C++ 6 comes with something known as "InstallShield for Microsoft
Visual C++ 6 (Build-130)". This *appears* to be equivalent to InstallShield
5.10.

Another option would be to look at the Microsoft Visual Studio Installer.
It's free if you have Visual Studio (VC), and it create setup files that use
the Windows Installer. Windows Installer is what you are *supposed* to use
if you write for Windows 2000, and it works on other Windows platforms as
well. There is a version of InstallShield that will produce Windows
Installer files as well, but I it's $$$.
AFAIK, Windows Installer natively supports scripting of installations, so
that should work.

All this is assuming somebody has the time to do more than just recompile
the current setup scripts, of course.

//Magnus

> -----Original Message-----
> From: kovacsz [mailto:zoli@pc10.radnoti-szeged.sulinet.hu]
> Sent: 01 November 2000 16:32
> To: Jamie Walker
> Cc: pgsql-interfaces@postgresql.org
> Subject: Re: [INTERFACES] RE: new maintainer for the ODBC driver?
> 
> 
> > Could someone package the next release of the driver with a newer
> > version of InstallShield? The version used for the current 
> driver does
> > not support the options for silent unattended installation. 
> This would
> > be a great help for automated updates and installations on LANs.
> Maybe VC++ 6 Enterprise contains this new InstallShield (Dave?). But,
> until we don't have a maintainer, we are happy to be able to 
> put a new DLL
> together at all... :-)

There is a demo version with Visual Studio, but as I recall it's a bit
restricted. I tend to use the Visual Studio Installer which creates the new
Installer format files which support installation rollback, advertising and
various silent/nearly silent/not silent modes. The VS Installer is also
limited and can't copy with the driver manager detection that the current
system uses, but I know that the latest InstallShield supports Installer
Files with all the bells and whistles. I might be able to buy a copy if we
get someone to take responsibility for the driver code itself...

BTW: I've had email bounced from your reply address Zoltan... is it correct?

Regards,
Dave.

> BTW: I've had email bounced from your reply address Zoltan... is it correct?
No, sorry. The correct address is:
tip@pc10.radnoti-szeged.sulinet.hu

Zoltan