[omniORB] omniORBpy Exception Question (and omniidlexception patch).

[Duncan Grisby]

Sorry to all the people on the list who don't care about Python. Maybe
we need an omniORBpy list.

On Tuesday 28 March, Randy Wiser wrote:

> I think I agree with you that CORBA.Exception should inherit from
> exceptions.Exception (as it does now).
[...]

Unfortunately, there is a fundamental conflict between the Python
mapping for CORBA exceptions, and the full use of the
exceptions.Exception class. exceptions.Exception is defined as:

class Exception:
    """Proposed base class for all exceptions."""
    def __init__(self, *args):
        self.args = args

    def __str__(self):
        if not self.args:
            return ''
        elif len(self.args) == 1:
            return str(self.args[0])
        else:
            return str(self.args)

    def __getitem__(self, i):
        return self.args[i]

The big problem is the args data member. Suppose I have an exception
declared in IDL

  interface MyException {
    long args;
  };

An instance of this exception will have a data member called args,
which clashes with args in the exceptions.Exception base class. This
means that the __str__ and __getitem__ methods won't work.
CORBA.Exception already overrides the __str__ method; maybe I need to
provide a __getitem__ method too?  Would that solve your (Randy's)
problem?

Maybe the proper fix is to change the Python mapping to use _args
instead of args in this case. Alternatively, I could try to convince
Guido to change exceptions.Exception to call its data member
__args. Since the args member is documented, I don't imagine he'd be
keen on that...

Cheers,

Duncan.

-- 
 -- Duncan Grisby  \  Research Engineer  --
  -- AT&T Laboratories Cambridge          --
   -- http://www.uk.research.att.com/~dpg1 --