Confused again
Duncan Smith
buzzard at urubu.freeserve.co.uk
Mon Feb 24 10:16:21 EST 2003
"Steven Taschuk" <staschuk at telusplanet.net> wrote in message
news:mailman.1046060842.23969.python-list at python.org...
> Quoth Duncan Smith:
> > I'm still (again) having a little trouble with operator overloading. I
> > changed the original code I posted yesterday (Confusing problem) so that
it
> > now (apparently) does what I want. I simply changed "def __cmp__(self,
> > other):" to "def __eq__(self, other):". [...]
>
> This has nothing to do with your problem below, but you probably
> also want to provide __ne__. Notice:
> >>> class spam(object):
> ... def __eq__(self, other):
> ... return 1 # all spams are created equal
> ...
> >>> a = spam()
> >>> b = spam()
> >>> a == b
> 1
> >>> a != b
> 1
> Users of spam might be confused by this behaviour.
>
> What's happening is that spam has no __ne__, so != falls back to
> ... an implementation that tests 'is not same object', I think.
> (Certainly a != a yields 0.) In any case, it doesn't do what you
> want. Adding, for example,
> def __ne__(self, other):
> return not self == other
> will suffice (unless you're using NotImplemented tricks in __eq__).
>
Yes. I wondered about that last night, and used "not table1 == table2"
rather than "table1 != table2" to be on the 'safe' side. I'll have to add
the above function to my code to tidy it up.
> > [...] But if I try to raise an exception
> > after calling the function, the exception is apparently ignored. eg.
the
> > following code snippet,
> >
> > print table == numtable2
> > if table == numtable2:
> > to_remove.append(table)
> > else:
> > print 'what?'
> > raise Error
> >
> > class Error:
> > pass
> >
> >
> > >>> c.clusters[(3,)].pruneTables()
> > 0 # evaluates to false as expected
> > what? # prints 'what?' OK
> > >>> # but no error
>
> Puzzling. What's the rest of pruneTables()?
Puzzling until you know how stupid I can be when I'm writing code in the
early hours. I was catching the exception! A complete function is in my
response to Alex's post. Cheers.
Duncan
>
> --
> Steven Taschuk o- @
> staschuk at telusplanet.net 7O )
> " (
>
More information about the Python-list
mailing list