FreeTDS Development Group

[brian AT bruns.com]

Jim,

I'll put my non-voting shares up for keeping it in the source tree but
perhaps disabling by default (need to pass a --enable-tds42 to
configure or whatever).

But, automatic detection would be best, I agree.

My reasoning goes like this.  We occasionally get a user from either
Russia or somewhere in the third world using SQL Server 6.5 on NT
3.51/4.0 on low end hardware.   I'd hate to cut these users off just
because we can all afford enough hardware to run SQL 2k.  Granted, I
haven't seen one of these on the list in two years maybe, but I'm sure
they still exist.  I personally beleive they are all running copies of
the same (legally questionable) software.  The old adage that you only
sell 1 copy of a software package to China. ;-)

Brian

On Thu, 17 Jun 2004 11:35:54 -0400, "Lowden, James K" wrote:

> 
> > From: ZIGLIO, Frediano, VF-IT
> > Sent: June 17, 2004 10:33 AM
> > 
> > >From my point of view if server crash it's a server bug :)
> 
> Perhaps, if the TDS protocol were open and Sybase invited all
comers. 
> But Sybase doesn't support other libraries.  If Sybase's own drivers
> don't triggger the crash, then it's a failure of FreeTDS to be
> compatible.  You can define it as a server bug, and I might agree. 
It
> still has to be avoided if it can't be fixed.  
> 
> > If we are able to reproduce the problem we can try to fix it.
> > The problem it's that 4.2 is supported by both vendors. 
> 
> Yes but only needed by Bronze Age servers.  
> 
> > IMHO it would be
> > better to emulate Microsoft drivers and try to detect better
protocol
> > version at startup caching results for forward references.
> 
> Quite so, and orthogonal.  :-)  
> 
> If no one's using 4.2, and we're not determined to fix it so it's
> safe, then it's just a trap for new installations.  
> 
> --jkl