apr-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Greg Stein <gst...@lyra.org>
Subject databases in APRUTIL (was: apr_xml)
Date Sat, 19 May 2001 00:39:06 GMT
On Sat, May 19, 2001 at 12:19:28AM +0100, David Reid wrote:
> Seems like a good idea.

Yes, but somebody just has to do it. That is: somebody with an itch.

> Did we ever do anything about defining/writing code
> for a common apr interface to databases.

No, although it would be a nice addition. After designing the DB-API for
Python, I have to say, though: you *MUST* have it properly scoped before
beginning the API design work. The simple fact is that every database is
different in any number of unimaginable ways. To create an API to make them
look the *SAME* is impossible. Given that, you have to set yourself a bar
that says "we'll make them look the same to <THIS> level and no more."

The Python DB-API set a bar that some people think is just right, and some
people think is too low. The Perl DBI bar is a bit higher. Last I looked,
the PHP folks have separate function sets for each database (e.g. there is
no attempted commonality).

I can go on, but the point is moot until somebody wants to spend time
implementing such a beast.

> I'd think MySQL and Postgresql
> would make the most sense, but we should probably try and get people who
> develop using Oracle and MS SQL server involved...  I know it's been
> discussed, just wondering how we're doing :)

Zero work on design or implementation. We have "DBM style" support for most
of the DBMs out there (be nice to add the Samba's TDB(?) support, tho). DBM
is just key/value pairs; nothing at all like SQL.


Greg Stein, http://www.lyra.org/

View raw message