FreeTDS Development Group

[Steve Langasek <vorlon AT netexpress.net>]

On Mon, Jan 06, 2003 at 12:15:25PM -0500, Lowden, James K wrote:
> > From: Andrew Stanley [mailto:astanley AT cs.uml.edu]
> > Sent: January 6, 2003 11:34 AM
> > 
> > (I hacked DBD::Sybase to look like DBD::FreeTDS, so I could have
> > 2 different DBD::Sybase's co-existing)

> [musing]

> Several people have mentioned the need to use FreeTDS and Sybase's libraries
> at the same time in Perl.  At present, that's only possible by using
> DBD::Sybase for Sybase and DBD::ODBC or sybperl for FreeTDS.  In limited
> cases, it may be possible to get away with manipulating @INC to manage two
> DBD::Sybase modules, but almost no one wants the administration headaches
> that are sure to follow.  

> I wonder if there isn't/couldn't be a better way.  For example, could the
> namespace and module name of DBD::Sybase be altered at configuration time?
> Could we use Perl namespaces another way, to convert DBD::Sybase into, say,
> DBD::FreeTDS::Sybase?  

> Andrew used hackery.  Is no mechanism available to make that unnecessary?  

Even if you managed to keep the two modules clear of each other within
the Perl namespace, there would still be (unavoidable) namespace
collisions at the C level.  If "at the same time in Perl" means "in the
same process", anything that would make this work would *always* be a
kludge.

I think efforts are better directed towards answering the question, "why
is one set of libraries not sufficient?", and fixing the corresponding
deficiencies in FreeTDS.

-- 
Steve Langasek
postmodern programmer

Attachment: pgphR9l9EAA_e.pgp
Description: PGP signature