[omniORB] Exception not catched

Attila Pletyak attila.pletyak@anemo.com
Mon, 09 Jul 2001 16:29:34 +0200


One more datum: when we put a delay of 100 nanosecs in the object 
creation process, the error does not come up any more. So it looks like 
that something needs time on omniORB side.

The problem is that the whole program suspends after getting the omniORB 
error message.

Attila Pletyak wrote:

> Hello,
>
>   We are implementing a small test program on Debian Linux 2.2, 
> omniORB 3.03, C++. The test program uses the NameServer quite much.
>
>   On every object creation we check whether or not we have an object 
> registered on the nameserver with the same name. At the moment we do 
> it with the resolve() method of the CosNaming, and try to catch the 
> NamingContext:NotFound exception. For making sure every exception is 
> catched, we catch the ... as well.
>
>   The problem is coming up when this creation process is working too 
> fast: Even tough we try to catch everything, from time to time we get 
> the following:
>
> omniORB: WARNING -- method 'Run' on: boa<0x3b497109b8fb021000010473>
> raised the exception: IDL:omg.org/CosNaming/NamingContext/NotFound:1.0
> omniORB: An exception has occured and was caught by tcpSocketMT Worker 
> thread.
>  
>   As we do not really understand this, we post it here and would like 
> to get any help.
>
>   Sincerely,
>
>   Attila Pletyak
>   Anemo Ltd.
>
>
>