Re: [INTERFACES] pgsql ODBC sources - Mailing list pgsql-interfaces

From Byron Nikolaidis
Subject Re: [INTERFACES] pgsql ODBC sources
Date
Msg-id 356CB92B.DBA14533@insightdist.com
Whole thread Raw
In response to pgsql ODBC sources  ("Thomas G. Lockhart" <lockhart@alumni.caltech.edu>)
List pgsql-interfaces

Thomas G. Lockhart wrote:

> I was poking through the backend source code working on the "name" data
> type and stumbled on the (apparent) fact that NAMEDATALEN is defined to
> be 32 bytes in the backend and only 16 bytes (the old backend value) in
> interfaces/odbc/connection.h.
>

The define NAMEDATALEN is only used when using the 6.2 protocol.  This is a
per datasource option.  It is not used when using the 6.3 protocol.
SM_USER, which is 32 bytes, is used for the 6.3 protocol.

> Also, some lines in the odbc sources seem to have a trailing ^M. Should
> we add more, or take them out? :)
>
>                      - Tom

Yeah, I hate those things.  When I transfer the odbc source files to our
unix box, I use ftp in ascii mode, which filters that crap out.  It works
either way on the Microsoft side so I'd say take em out.

By the way, just to clear things up, as long as people are on the subject
lately, I hate Microsoft just as much as the next guy!  I'm a unix systems
programmer first.  I dont really consider myself a windows programmer, even
though I work on the odbc driver.  We just had to have the driver work well
for our customers and it forced me to work with the windows programming
model.

Sorry about the divergent theme here.  I know you weren't asking about
Microsoft, but I wanted to get that out!

Byron


pgsql-interfaces by date:

Previous
From: Byron Nikolaidis
Date:
Subject: Re: Thank you! But how to get it working with VC++5?
Next
From: "Jose' Soares Da Silva"
Date:
Subject: ODBC is slow with M$-Access Report