Re: DBD::PostgreSQL

On Monday, November 18, 2002, at 12:32 AM, Ian Barwick wrote:

> Count me in. I'm still in a slight state of shock after wondering over
> to CPAN to find out how DBD::Pg was coming along ;-). At the
> very least I can do testing and documentation, and quite possibly
> "grunt work". Anything else will depend on how quickly I can
> acquaint myself with the internals of DBI. (Note to self: do not
> believe
> this is impossible or anything). Perl is my main
> development language, and I used to work a lot with C.

Well then, once I finish pasting together dbdimp.c and get everything
to compile, I might ask for your help with a code review and writing
tests.

> For clarification: is DBD::Postgres intended to replace DBD::Pg, and
> are
> any maintenance releases of the latter planned (e.g. in conjunction
> with
> the PostgreSQL 7.3. release)?

And on Monday, November 18, 2002, at 06:47 AM, Jon Jensen wrote:

> I didn't see any indication that David's planning on giving a new name
> to
> his rewritten PostgreSQL DBD driver, other than the subject of his
> email.
> It would cause a lot of pain if the driver's name changed from DBD::Pg,
> since every place people have DSNs in code or config files would have
> to
> be updated ...
>
> Would anyone actually consider using a different name than DBD::Pg? It
> seems it would be a big pain with no benefit, and make it unclear which
> driver users should use.

I expect that the PostgreSQL developers will include whatever DBI
driver is the "official" DBI driver for PostgreSQL. At this point, I've
just changed the name so I can feel free to hack it any way I like,
including breaking backward compatibility where necessary (such as in
the escape() method).

If I finish something that actually works, then I'll request some help
from others comparing it to the behavior of the DBD::Pg driver. If it
doesn't break backwards compatibility too much, then I would suggest
that it become DBD::Pg 1.20 or 2.0 or something. But if its behavior is
different enough (and it would need to be tried with a number of
different applications to see what breaks, I think), then it would
probably have to be released under a different name and people could
switch if/when they could. But we're a long ways from determining that
just yet.

Regards,

David

--
David Wheeler AIM: dwTheory
david(at)wheeler(dot)net ICQ: 15726394
http://david.wheeler.net/ Yahoo!: dew7e
Jabber: Theory(at)jabber(dot)org