Re: Bug in ORBit 0.5.1 with floats/doubles ?

["Povl H. Pedersen" <pope netguide dk>]

> 
> I think there was a bug related to floats/doubles that was fixed in 0.5.3
> - please get the latest & least-worst :)

I got it, but my problem now is, that the IOR string it writes out is
bad.

Using http://www.parc.xerox.com/istl/projects/ILU/parseIOR/
to parse the resulting IOR I get:
_IIOP_ParseCDR:  byte order LittleEndian, repository id <IDL:Cashcow/CashcowCORBA:1.0>, 1 profile
_IIOP_ParseCDR:  profile 1 is 80 bytes, tag 3135221450, LittleEndian byte order
_IIOP_ParseCDR:  error:  No object for IOR.
Error parsing IOR:  inv_objref

And my Java client also gives me an error trying to narrow the object
I got from orb.string_to_object:

org.omg.CORBA.INV_OBJREF:   minor code: 1  completed: No
...

I am not sure what has changed.

I am using the ugly account example from the docs (the main code using
#include to get the implementations). Here is the calls used on the
server to create the object:

orb = CORBA_ORB_init(&argc, argv, "orbit-local-orb", ev);
root_poa = (PortableServer_POA)
  CORBA_ORB_resolve_initial_references(orb, "RootPOA", ev);
cashcow = impl_Cashcow_CashcowCORBA__create(root_poa, ev);
ior = CORBA_ORB_object_to_string(orb, cashcow, ev);

And then I print out ior to a file, and ends with:

  pm = PortableServer_POA__get_the_POAManager(root_poa, ev);
  PortableServer_POAManager_activate(pm,ev);

  CORBA_ORB_run(orb, ev);

My implementation is based on the --skeleton-impl created file
(when is usage ever changed on orbit-idl to describe this flag
correctly ?)