On Sat, Apr 22, 2006 at 01:23:51PM -0400, Tom Lane wrote:
> "Dave Page" <dpage@vale-housing.co.uk> writes:
> > Err, what? It uses PQsocket & PQgetssl, both of which are official,
> > supported functions that may be being used by countless other apps for
> > all we know.
>
> Dave, weren't you paying attention to the recent discussion? GnuTLS
> support will break anything using PQgetSSL, because it can't return
> an OpenSSL struct if the underlying SSL library is GnuTLS. The current
> thought is to return NULL, so as not to have apps actually crash by
> trying to pass a GnuTLS struct to OpenSSL routines, but that isn't going
> to make psqlODBC *work* in such a situation. You'd need to add code
> that knows how to work with GnuTLS.
It's a messy situation. I hope we're not going require users to include
both GnuTLS and OpenSSL in their source given you can't actually
#include both in the same source easily (namespace issues). Like you
say, backward compatability is important.
> Martijn is being a bit disingenuous by giving the impression that
> acceptance of his patch is a done deal. The compatibility issues
> are serious enough that it's quite possible we'll reject it.
Ouch. I hope I didn't give anyone that impression, I'd prefer to say I
was being optimistic because I don't think the problems are
insurmountable. I only wanted to point out that there's no point
someone doing this as an SoC project before the associated issues have
been sorted out...
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.