> -----Original Message-----
> From: Bruce Momjian [mailto:pgman@candle.pha.pa.us]
>
> > Yes, that could happen, though in reality I suspect it proably
> wouldn't as
> > (in my experience - I may be wrong) those technologies do
> little more than
> > putting SELECT * FROM in front of a given table name and
> generating query
> > based updates.
> >
> > I suspect that more problems may be caused for apps like MS
> Query or Seagate
> > Crystal Reports which might attempt to use outer joins on an
> older database
> > (actually Query just enables/disables the relevant join
> options). As far as
> > I can tell, the driver doesn't check the backend version
> (despite returning
> > the DBMS version which is currently hardcoded), so we'd either
> need to add a
> > 7.1 protocol option (not nice as it isn't really a different
> protocol) or do
> > a SELECT VERSION() on connect to get the exact version string
> and figure out
> > what is/isn't possible from that.
> >
> > I'll look into it but I'm on the wrong end of the learning curve here!
>
> We usually are more concerned about old libraries talking to new
> servers. New libraries talking to old servers are not needed as much.
>
I strongly object.
Postgres ODBC driver has supported all the way back to version 6.2(
or 6.0 according to a person's opinion). Why could we easily break the
backward compatibility without any consensus ?
I also object to centralize the version info. It is preferable that client
apps and servers are relatively indepedent as much as possible.
It seems very natural that client apps and the server have separate
versioning policies. Do we always have to wait for the all-in-one
release to get improved client apps ?
Regrads,
Hiroshi Inoue
