[Python-Dev] PEP 203 Augmented Assignment

[Thomas Wouters]

On Thu, Jul 27, 2000 at 12:03:39PM -0500, Guido van Rossum wrote:

> > So backwards compatibility is out ? I'm not sure howmany Python code
> > uses slice-objects, but I would consider it a (personal ;) failure if
> > slicing has to be *broken* before it can be fixed.

> Huh?  I thought I proposed a b/w compat solution:

>   IF there is a __getslice__ method:
>      IF the slice step is None:
>         call __getslice__(lo, hi)
>      ELSE: # the slice step is not None
>         call __getslice__(lo, hi, step)

> What's wrong with that?

Well, what happens if __getslice__ and __getitem__ both exist, and
__getitem__ is used to handle extended slices, but __getslice__ isn't (yet)? 
Currently, if it's a single item, __getitem__ is called. If it's a basic
slice, __getslice__ is called. If it's an extended slice, __getitem__ is
called.

In the above picture, __getslice__ would be called instead, with the 'wrong'
number of arguments, and the use of extended slices would break.

> > If we can't figure out whether a function supports the 'new syntax'
> > reliably, I don't see how we can transform the error message either. I'll
> > bet inspecting the __getitem__ method to find out whether it supports the
> > one or the other is way too much of a runtime penalty to suffer at each
> > index.

> No, we don't deal with __getitem__; if *it* doesn't support a tuple
> containing slice objects, tough luck.

Sorry, that was my feebleness popping up. I was talking about __getslice__.
How do we find out if __getslice__ accepts 4 arguments, rather than 3 ?

-- 
Thomas Wouters <thomas@xs4all.net>

Hi! I'm a .signature virus! copy me into your .signature file to help me spread!