FreeTDS Development Group

[Brian Bruns <camber AT umcc.ais.org>]

On Tue, 14 Dec 1999, Michael Peppler wrote:

> Adrian Blakey writes:
>  > Firstly, FreeTDS is a godsend -- thank you.
>  > 
>  > I am testing TDS .5 on RH Linux against a MS SS Ver 7.0.
>  > 
>  > I got and tried to make DBD::FreeTDS 0.02, but it does not seem to be
>  > maintained judging from the compile errors.
> 
> You are correct. I actually think that it should be removed from
> CPAN. 
> 
>  > 
>  > So I have been testing with DBD::Sybase, built using the FreeTDS libs.
>  > 
>  > Things mostly work, however error handling and reporting is not up to the
>  > same level as if DBD::Sybase were linked with the Sybase libs.
> 
> I'd love to get some specifics on this.
>
me too. All the server errors should be propagating up to DBD::Sybase.
 
>  > So I wonder if there would be any point in forking the DBD::Sybase 
> codeline
>  > and doing a better job of providing a DBD::FreeTDS?
> 
> I'm not convinced that DBD::Sybase can't be made to behave correctly
> with both implementations, but I'm open to suggestions. In any case
> the DBD::Sybase code isn't all that complicated :-)
>
I'm not aware of any problems with supporting both either. TDS 7 has some
additional datatypes, but they shouldn't conflict with existing types. It
may require a couple of ifdef's or some clever header file hacks :-) but
it is certainly doable.

>  > 
>  > This makes me wonder about about impedance matches between the libs,
>  > protocol and server.
>  > 
>  > Sybase considers ctLib its annointed interface to TDS 5.0 and Adaptive
>  > Server Enterprise. And they probably do not want to help out MS by 
> extending
>  > ctLib.
>  > 
>  > Microsoft has no ctLib implementation and is interested in furthering 
> dbLib
>  > as the interface to TDS 7.0 and Sql Server.
> 
> Doesn't Microsoft push ODBC over DBlibrary?
> 
This is my understanding.  Once things settle out with TDS 7, I'll
probably attack the ODBC CLI.  TDS 7 isn't that fundamentally different
than 4.2 or 5.0, just more datatypes and unicode support. Unicode is the
interesting problem, but solvable I think.

>  > FreeTDS exports TDS 7.0 functionality through the reimplemented FreeTDS
>  > ctLib -- there is no dblib (rightly so). Therefore, are there ctLib
>  > functions which either behave differently or not at all when directed at 
> a
>  > MS database? So will it be necessary to implement some ctLib functions 
> that
>  > are not part of the Sybase "standard" library to support MS?
> 
> I guess Brian and/or others who are using the TDS 7 stuff will have to 
> step in here...
> 
TDS 7 is exported through both ctlib and dblib in FreeTDS. DBLib has
actually gotten more testing with this stuff. There is also a
configure-time switch to implement MSSQL behaviour in preference to Sybase
where they differ (dbdatecrack() is my single example of this currently)

Brian