FreeTDS Development Group

[Will Saxon <WillS AT housing.ufl.edu>]

Recompiling to use TDS 4.2 _does_ work. I am able to execute the TDS
unittests correctly. Unfortunately, there is now a problem with the next set
of tests (ctlib). About 20 symbols in libct.so are 'undefined,' which is
really odd. 

I built from a clean source directory ('make distclean'ed) and configured
with --with-tdsver=4.2

Thanks for the 4.2 suggestion.

-Will

> -----Original Message-----
> From: Brian Bruns [mailto:camber AT ais.org]
> Sent: Wednesday, May 30, 2001 10:35 PM
> To: TDS Development Group
> Subject: [freetds] Re: problems with make check
> 
> 
> Hi Will,
> 
> There may be a problem with running TDS 7 on big endian 
> machines.  I don't
> recall if it was ever corrected, but if you are willing to give some
> input, we'll see if we can't get this straight.
> 
> Firstly, try configuring using './configure 
> --with-tdsver=4.2' and see if
> that works.
> 
> On Mon, 21 May 2001, Will Saxon wrote:
> 
> > Hello,
> > 
> > I realize this may have been covered in previous posts.. I 
> think this
> > situation is slightly different though.
> > 
> > I cannot get FreeTDS 'make check' to work properly on a Sun Ultra-60
> > running Solaris 2.6. We have to use 2.6 because of another 
> application we
> > run that only works with 2.6.
> > 
> > I have a Microsoft SQL Server (v7.0) running on another 
> machine. I have
> > configured FreeTDS to build for TDS version 7.0. I have 
> specified the SQL
> > server name in /etc/hosts to avoid some of the other 
> problems I have read
> > about. FreeTDS compiles just fine, but when I 'make check' I get the
> > following:
> > 
> > t000#.c: Test stuff
> > tds_connect() failed
> > try_tds_login() failed
> > FAIL: t000#
> > 
> > every single time. I have verified that the FreeTDS machine 
> is actually
> > attempting to connect to the SQL server - using a packet 
> sniffer I can see
> > packets coming in. The database is not in single user mode 
> or anything
> > like that. Note that I do not get the 'cannot assign 
> requested address'
> > thing.
> > 
> > One thing that I have NOT done is set any environment 
> variables. Since I
> > have not actually installed FreeTDS (I run make check 
> before installing) I
> > have not set $SYBASE or anything like that.
> > 
> > I have performed another build on a separate machine (x86 
> based FreeBSD
> > 4.3-STABLE) with no problems. I had everything configured 
> the exact same
> > way, with the same information in PWD and interfaces. It 
> worked the first
> > time, no problems at all. I have a few test scripts that I 
> written already
> > (using DBI and DBD::Sybase) and they all work just great. I 
> don't think I
> > am the problem with the situation.
> > 
> > One thing I just did was I compared logs from a successful 
> test on the
> > FreeBSD machine with an unsuccessful one on the Solaris machine. The
> > initial packet sent is very similar in both tests, but not 
> exactly the
> > same. Several bytes seem to shifted or swapped. Is this 
> just a difference
> > between big and little endian bit ordering? Is this the 
> problem? I can
> > proved copies of the logs to anyone who wants to look.
> > 
> > I am really interested in using this software for this 
> project. It seems
> > to be the best way to do things. If only I could get it to work! :).
> > 
> > -Will
> 
> 
> ---
> You are currently subscribed to freetds as: [wills AT housing.ufl.edu]
> To unsubscribe, forward this message to 
> $subst('Email.Unsub')
>