[omniORB] Alpha/Digital UX link errors linking example 1

[agonne - Andrew Gonnet]

Hello.

I'm getting unresolved errors for methods that are called in the 
object generated from the "echoSK.cc" using the supplied make file in 
the "echo" example directory. Using "nm" I can find these methods in 
the "echoSK.o" object, but the version in the supplied object 
"libomniORB2.so" is in the format "__vtbl_Q15CORBA7UNKNOWN" etc - i.e 
there's an extra "Q1" in the middle.

Has anybody come across this problem?



I'm using Digital UX v4.0e, Digital cxx 6.0 (for compiling), gcc 2.8.1 
(just for omniidl2, it complained when it wasn't there) and omniORB 
2.7.1.

The errors during linking are:

ld:
Unresolved:
CORBA::__vtbl_5CORBA7UNKNOWN  Make sure that the first noninlined 
virtual member function of 'CORBA::UNKNOWN' is defined.
CORBA::__vtbl_5CORBA9NO_MEMORY  Make sure that the first noninlined 
virtual member function of 'CORBA::NO_MEMORY' is defined.
CORBA::__vtbl_5CORBA12COMM_FAILURE  Make sure that the first 
noninlined virtual member function of 'CORBA::COMM_FAILURE' is 
defined.
CORBA::__vtbl_5CORBA7MARSHAL  Make sure that the first noninlined 
virtual member function of 'CORBA::MARSHAL' is defined.
CORBA::__vtbl_5CORBA13BAD_OPERATION  Make sure that the first 
noninlined virtual member function of 'CORBA::BAD_OPERATION' is 
defined.
CORBA::__vtbl_5CORBA9TRANSIENT  Make sure that the first noninlined 
virtual member function of 'CORBA::TRANSIENT' is defined.

The lines from nm of echoSK.o greped for "UNKNOWN" are:

CORBA::__vtbl_5CORBA7UNKNOWN     | 0000000000000016 | U | 
0000000000000000
__cxxexsigQ15CORBA7UNKNOWN       | 0000000000000008 | C | 
0000000000000000

...and the lines from libomniORB.so greped for "UNKNOWN" are:

CORBA::UNKNOWN::NP_RepositoryId(void) const  | 0004396972227136 | T | 
0000000000000008
GIOP_Basetypes::SysExceptRepoID::UNKNOWN | 0004398046479696 | B | 
0000000000000000
__T_Q15CORBA7UNKNOWN             | 0004398046332880 | R | 
0000000000000000
__cxxexsigQ15CORBA7UNKNOWN       | 0004398046478632 | S | 
0000000000000000
__vtbl_Q15CORBA7UNKNOWN          | 0004398046331784 | R | 
0000000000000000

Does anybody have any thoughts about this?

Thanks,

Andrew Gonnet.
Acxiom UK.