Example code for COPE corba/perl binding
I. Reid Aug 7/98
Modified to track COPE changes 11/99.

This example shows four different ways of creating CORBA objects in
COPE.  There is a little trickery involved in letting all these co-exist
in the same server.  See comments on calls to cleanupObjectOnExit() in
Supervisor_impl.pm for the details.

The different methods are:

Temp:	object ids are automatically generated and assigned by the server;
	the object references (IORs) corresponding to these objects are
	only valid for the lifetime of the client connection.

Persistent: object IDs are assigned by the server side of the object
	implementation; because the server side always assigns the same
	ID for the same backend object, the IORs remain valid even if
	the server exits and restarts.

OneServant: object IDs and therefore IORs are persistent, but there is
	not actually a separate Perl object in the server for each CORBA
	object.  Instead, a single "default servant" handles all
	requests for objects of that type.  This is appropriate for
	applications that put an object front end onto a database with
	lots of records.

OnDemand: Like OneServant, except that a real Perl object is created the
	first time a method is invoked on the CORBA object.  This is a
	hybrid model for cases where there are many database entries,
	only a few will be referred to, and you want fast access.  The
	CORBA server only needs to hit the database the first time; from
	then on, the object data is in memory.

The server writes its IOR at startup and the client reads it.
So the programs can be run with these Unix shell commands:

ce_server > /tmp/IOR &
# Pause for startup
ce_client `cat /tmp/IOR`

# $Id: README 1206 2006-05-31 07:11:53Z rvosa $