FreeTDS Development Group

["Ludo Brands" <ludo.brands AT free.fr>]

> > I have no problems sharing the changes I made.
> 
> Thank you for that offer, Ludo.  
> 
> > - since this is a sybase only extension, I assume the 
> > TDS_REQ_LARGEIDENT capability request should be set only for recent 
> > Sybase servers, based on the existence of the SYBECAPTYP 
> error. What 
> > are the possibilities?
> 
> I see no need to worry about capacity negotiation with old 
> Sybase servers.  
> 
> >From the design of the packet and Sybase's track record, I bet old
> servers simply ignore (or acknowledge in the negative) 
> capacity requests they don't recognize. If you're a server 
> that supports 65 capacity bits and I'm a client asking for 
> #83, you can easily and safely say No.  
> 
> If someone reading this message knows of a new Sybase client 
> that cannot connect to an old Sybase server (that supports 
> TDS 5.0), that would certainly be evidence to the contrary.  
> 
> Absent that evidence, we can at least start with no switch.  
> That has the advantage of simplicity and of expoiting new 
> servers, which is, er, the way of the future.  
> 
> Regarding ENABLE_EXTRA_CHECKS, I would remove outdated code 
> from compilation until someone gets around to updating it.  Just put
> 
>       # if CAPACITY_8_BYTES
> 
> around the obsolete validation for now.  If you want to fix 
> the validation instead, that's fine, too.  
> 

Attached the patch. 
 
"# if CAPACITY_8_BYTES" is put around the body of "static void
tds_capability_test(void)" to make mods limited to mem.c. I wasn't able to
test with ENABLE_EXTRA_CHECKS 1 because of missing functions such as
tds_check_tds_extra (windows).

Ludo

Attachment: tds_capacity.diff
Description: Binary data