FreeTDS Development Group

[John Seers <John.Seers AT uk.lionbioscience.com>]

James et al

Hurray. Some success at last. CONNECTION SUCCEEDED from the C++ program. 

I am afraid I have only hacked around the problem for now. I was putting
in the callback code to see if that would solve the problem and at the
same time trying to track down the problem. 

The basic cause of the problem seemed to be in the ct_init call not
initialising the context correctly. The _servermsg_cb is set to 0x10.
This is strange as the _clientmsg_cb looks as if it is set to a
reasonable value.


(gdb) p tdscontext->_clientmsg_cb
$3 = (int (*)(void)) 0x846fa90
(gdb) 
(gdb) p tdscontext->_servermsg_cb
$1 = (int (*)(void)) 0x10
(gdb) 

So I am afraid I just set the _servermsg_cb to 0 for now, as that is
checked for and ignored, and it carries on successfully.

If anyone can cast any light on the cause or how it could be done better
then that would be helpful. It could be that I have something corrupted
somewhere. If so I expect it will rear itself again further down the
track. 

I will carry on developing and looking at the callback that was
recommended.

Thank you everybody for your help and pointing me in the right
directions.

I expect you will hear from me again sometime soon. 


Regards.



John Seers



> Ah.  Good old message 5701, eh?  Sometimes I think MS stuck that in there to
> remind people to write message handlers.

That's the one.

> 
> That variable is was added since 0.53.

Ah. That explains that. I spent quite a lot of time fiddling with this
and in freetds.conf with no luck.
What kept me trying was that running the unittests (dblib I think)
produced a config file that helped me know that the correct freetds.conf
was being picked up. The feature must be half installed in 0.53, but I
do not know how to switch it on.


> > 2. When compiling using g++ the compiler flags as warnings the various
> > freetds routines as "implicitly defined". Do the freetds .h files not
> > define the routines? Or am I missing something here?
> 
> You're not missing anything.  That was the state of affairs then.  Having
> added a pedant or two to the project since, I don't expect the next release
> will yield compiler warnings on most architectures.

Well I discovered the joys of using cproto to generate the definitions.
Are they defined anywhere for download? Or do I have to move to the next
release? 

END






>
begin:vcard 
n:Seers;John
tel;cell:07900 498885
tel;fax:01223 224701
tel;home:01603 660746
tel;work:01223 224729
x-mozilla-html:FALSE
adr:;;;;;;
version:2.1
email;internet:John.Seers AT uk.lionbioscience.com
fn:John Seers
end:vcard