[omniORB] update from 2.7.1 to 2.8.0pre2 and tie_

HA Quoc Viet qvha@axlog.fr
Thu, 09 Sep 1999 17:50:23 +0200


following on my previous posting,

the third echo example (eg3_tieimpl/eg3_clt) doesn't work anymore
(from 2.8.0pre2 source tree) though it did run the first time ... !

I have restarted omniNames and I've checked that the CosNaming
service was really up and running, and yes it was.

eg3_tieimpl dies right on the following call, no exception thrown
( catch(...) catches nothing )

testContext = rootContext->bind_new_context(contextName);


... ? ....


Viet

> I've updated my omniorb 2.7.1 to 2.8.0pre2 (linux 2.2.10, i386,
> egcs-2.91.66)
> I have carefully read ReleaseNote_omniORB_280.txt, but couldn't find out
> what's killing my "server"
> 
> the same code compiles fine with both.
> it runs fine on 2.7.1
> 
> however, my server crashes when it's sending
> tie_obj->_obj_is_ready(boa);
> 
> I am heavily using the tie_ feature, could it be that ?
> 
> since _narrow is now deprecated, I have tried _downcast instead, but
> then
> the compiler complains "there is no _downcast member in your class,
> dummy !"
> 
> the typical piece of code that kills my server is
> 
> Variable_i                  *InstanceLocale;
> _tie_Variable<Variable_i,1> *InstanceIDL;
> 
> InstanceLocale  = new Variable_i;
> InstanceIDL     = new _tie_Variable<Variable_i,1>(InstanceLocale);
> InstanceIDL->_obj_is_ready(boa);       // server dies
> 
> Any suggestion welcome ... :o)
> 
> Viet

-- 

In any case, I hope that no one is offended by my cynical questions.
For some reason, these questions blitzed me this morning, so I just core
dumped my brain on this email, partly for therapeutic reasons.  Perhaps
I
need more coffee.

HA Quoc Viet               Rue du 8 Mai 1945            -o)
Axlog Ingenierie           Arceuil 94110                /\\
+33 (0) 01 41 24 31 21     France                      _\_v