FreeTDS Development Group

[Eric Deutsch <edeutsch AT systemsbiology.org>]

Hi, I just grabbed the latest distribution from CVS and checked to see how
it performed against my hacked version of 0.51 release.  I have the
following comments, a few of which may be relevant for the 0.52 release:

(using Red Hat 7.0 + Perl 5.6.0 + DBD::Sybase 0.91 + FreeTDS 0.52CVS
to talk to MS SQL Server 2000)

1) freetds.conf works very nicely, I like it.  One possible snag:
freetds.conf.pl appears to be identical to README.freetds.conf.pl instead
of a perl script.  Might be just TODO or possibly CVS checkin issue.

2) Under both TDS 4.2 and 7.0, the final character of a datatype TEXT
field is truncated.  Last-digit truncation has a long history here, but I
think this is a new problem.  I was able to fix it with the following
patch:
*** convert.c:orig      Mon Mar 19 16:20:37 2001
--- convert.c   Fri Jun 15 16:46:51 2001
***************
*** 98,102 ****
         cplen = srclen > destlen ? destlen : srclen;
           memcpy(dest, src, cplen);
!          dest[cplen-1] = '\0';
           return strlen(dest);
        case SYBCHAR:
--- 98,103 ----
         cplen = srclen > destlen ? destlen : srclen;
           memcpy(dest, src, cplen);
!          /* 2001-06-15 Deutsch changed [cplen-1] to [cplen] */
!          dest[cplen] = '\0';
           return strlen(dest);
        case SYBCHAR:

The line in question appears to be new since 0.51.  This looks like a bug
that needs to be fixed, but maybe this breaks some other client+server
combination?

3) I had to make my usual hack to get TEXT results to work at all under
TDS 7.0 without getting the memory errors (not a problem under 4.2):

diff ct.c:orig ct.c
755a756,762
>
>         /* FIX ME Deutsch hack */
>         if (datafmt->maxlength > 100000) {
>           datafmt->maxlength=100000;
>           tdsdump_log("%L inside ct_describe() datafmt->maxlength = %d
after articial limiting\n", datafmt->maxlength);
>         }
>

I see you have this under "Maybe".  Fair enough, I can live with this hack
for now.  FWIW, when I do "SELECT @@TEXTSIZE" under SQL 2000, I get 64512
as a default.

4) I'd like to see a settable date format in freetds.conf, something like:

  dataformat = "%Y-%m-%d %H:%M:%S"
that gets used in convert.c, again not critical for 0.52 release, but it
would be nice.

5) While I'm tossing out features that would be nice: I miss the
ability to bind variables to a prepared statement.  DBD::Sybase supports
it, but FreeTDS doesn't yet.  I don't have the expertise to do it, sadly.
Obviously, this is not a 0.52 item, but maybe sometime soon?

That's my $0.01 on issues and desired features.
Otherwise, FreeTDS is great, I use it every day!

thanks,
Eric



On Wed, 13 Jun 2001, Brian Bruns wrote:

>
> Hi everyone,
>
> Well tonight I sat down to plow through the domain login stuff and got
> greeted with the unfortunate message "Login failed for user
> '(null)'. Reason: Not associated with a trusted SQL Server connection."
>
> I'm still struggling with the exact meaning of this.  I was expecting to
> get a challenge sequence back.  I'm guessing that the self reported client
> host name is checked against existing connections and the challenge is
> issued if it's ok. second possibility is one of the magic numbers contains
> an identifier of some sort, but I tend to doubt that since they are not in
> the right portion of the packet for it.
>
> In any event domain auth looks unlikely to come together prior to 0.52, so
> I'm going to push ahead with the release.  Here is the current tasks.
>
> Prior to 0.52:
>
> 1. Verify freetds.conf changes are working
> 2. Get the convert program for interfaces -> freetds.conf checked in
> 3. Release notes should detail how to run convert program
> 4. Get RPM builds ready to go
> 5. User Guide updated and checked in
> 6. Check in ODBC fixes
>
> Maybe:
>
> 1. Maximum textsize option in TDS 7.0
>
> Post 0.52: (not necessarily in order)
>
> 1. TDS 7 on big endian platforms
> 2. Domain logins
> 3. Unicode conversion stuff
> 4. TDS 8.0 support
> 5. TDS connection pooling stuff needs more work (tds != 4.2 support,
>     remove hard coded values, et al.)
> 6. tdsping program for testing purposes (jkl's idea)
>
>
> Anyone got others?
>
> Brian
>
>
> ---
> You are currently subscribed to freetds as: [edeutsch AT systemsbiology.org]
> To unsubscribe, forward this message to $subst('Email.Unsub')
>

-- 
----------------------------------------------------------------------
Eric Deutsch                        email: edeutsch AT systemsbiology.org
Institute for Systems Biology                    Voice: (206) 732-1397
4225 Roosevelt Way NE, Suite 200                   FAX: (206) 732-1260
Seattle, WA  98105-6099                                  TJ's Location
http://www.systemsbiology.org/