FreeTDS Development Group

[Brian Bruns <camber AT ais.org>]

On Thu, 18 Jul 2002, Lowden, James K wrote:

> > From: Bill Thompson [mailto:thompbil AT exchange.uk.ml.com]
> > Sent: July 18, 2002 11:15 AM
> > 
> > I reckon another couple of days work on this next week and 
> > we'll be all
> > done and dusted. If you could hold off applying any patches 
> > to convert.c
> > and dbconvert and cs_convert, it should make it a lot easier 
> > to integrate my patches when they're ready.
> 
> This is wonderful news, Bill, thanks for the update.  I'll be offline myself
> until August, so your work will probably be done while I'm away. <pout>   
> 
> You had said you were putting off generating error messages.  Where does
> that stand?  Seems to me you needed socket plumbing changes from Brian to do
> it right?  

Yeah, I just have to fixup the ctlib handlers a bit and add the parent 
field to TDSCONTEXT.  I've been wasting^Wspending my time working on 
domain logins now that I think I figured out an ingenious way to test 
them, I compile freetds under cygwin on the box with the SQL Server, and 
therefore the box has already "joined the domain" and hopefully everything 
will work enough for me to debug the thing.
 
> I think the autoconf macro is not hard to add later.  Basically, we can wrap
> your "long long" uses inside a HAVE_LONG_LONG test in the header file, and
> in convert.c (or whatever) we can say, approximately:
> 
> #if ! HAVE_LONG_LONG
>       [call error message handler]
>       return FALSE;
> #endif

I don't mind using native types, but we do really need to do something for 
the platforms without a 64bit type.  In what ways in using the current 
numeric code insufficant for the 8 byte money type?

Brian