Re: [certi-dev] Federation startup crash

[Eric Noulard]

2010/11/8 Michael Raab <address@hidden>:
> Hi all,
>
> I have again problems with CERTI.

Sorry about that.

> This time the start up of my federation
> crashes. For testing purposes I have only two federates.

[...]
Would you be able to send us you example federate .
I may try to reproduce the issue here.

> They get started both in sync, each federate waits for my reaction (with a
> MessageBox :-)) after line 7 (queryFederateTime), to be sure that none of
> them advances time before the other has joined and configured. This worked
> well with DMSO RTI for several years.

Yeah, but DMSO RTI stopped being distributed for several years too :-[

> After I permit advancement for both, Federate 1 throws an exception during
> the process of publishment of one of its interaction classes. But not always
> for the same interaction class :-(. Apparently the errors occurs during
> getParamterHandle() function call. The Exception message is: libRTI: Network
> Read Error waiting RTI reply.
> Has someone an idea what may cause this behaviour?

It may be linked to the reported issue:
https://savannah.nongnu.org/bugs/index.php?31551

we seems to have some race condition w.r.t. publishing AFTER
 timeConstrained/RegulationEnabled.

I'll try to track this down.
having your example federate would help.

> FYI:  I'm using 3.3.3cvs

Are you on Windows or Linux?

I have just uploaded a snapshot tarball of 3.4.0cvs
http://download.savannah.gnu.org/releases-noredirect/certi/beta/CERTI-3.4.0cvs-Source-8nov2010.zip
would you be able to test with this version.
(you may do with CVS checkout too)

-- 
Erk
Membre de l'April - « promouvoir et défendre le logiciel libre » -
http://www.april.org