[omniORB] System Exception BAD_PARAMETER::InvalidServiceCtxtId
(26)
Wernke zur Borg
wernke.zur.borg at vega.de
Tue Sep 18 17:45:28 BST 2007
>
>
> Using OmniOrb 4.0.3 (SLeS) we get a system exception, with
> minor code 26
> (InvalidServiceCtxtId).
>
> The IDL is shown below;
> interface TCOhpServerClient
> {
> TCOhpStatus getTcoCoeffUTC(in TCOhpTDSsunTime referenceTime,
> out TCOhpCorbaCoefficient coefficient);
> };
>
> The implementation is shown below;
>
> TCOhpStatus TCOhpServer::getTcoCoeffUTC(const
> TCOhpTDSsunTime& refTime,
> TCOhpCorbaCoefficient& coeff)
> throw (CORBA::SystemException)
> {
> // Local variables
> TCOhpStatus retVal = TCO_STATUS_OK;
>
> // Protect the Coefficient cache before reading any data
> pthread_mutex_lock(&m_mutex);
> if (true == m_coefCacheIf->isInCache(startTime))
> {
> // The data is already available in the cache.
> m_coefCacheIf->getCoeffs(startTime, coeffs, nrOfCoefficients);
> }
> else
> {
> retVal = TCO_STATUS_NO_COEFFS_AVAILABLE;
> }
>
> return retVal;
> }
>
>
> Following the return the exception is thrown.
>
> We catch the exception in the client and print the minor
> code. The value is
> 1096024090 (Decimal). Looking in the minorCode.h file, this
> maps to minor
> code 26 (InvalidServiceCtxtId). The code should be a standard
> OMG minor code
> for BAD_PARAM, but looking in the CORBA 3 specification we
> dont find this.
>
> Questions are many;
> 1. Why do we get this error?
> 2. Do we interpret the error code correctly, i.e. is it
> really minor code 26
> we see?
> 3. What is a service context id?
>
> Been wasting almost a week on this problem. We would be very
> thankful for
> any help.
Hi,
the code maps to OMNIORBMinorCode_26, which is MARSHAL_InvalidEnumValue.
Do not mix up the OMNIORB minor codes with the OMG ones...
Perhaps you can also post the complete definition of your return code
enum. Have you made sure that both sides use the same IDL? You may have
added the value TCO_STATUS_NO_COEFFS_AVAILABLE to the enum but not
recompiled the client side with the modified IDL.
Regards, Wernke
More information about the omniORB-list
mailing list