[Python-Dev] extended slicing for lists

[Michael Hudson]

The patch is here, btw:

http://sourceforge.net/patch/?func=detailpatch&patch_id=100998&group_id=5470

Ka-Ping Yee <ping@lfw.org> writes:

> On 28 Jul 2000, Michael Hudson wrote:
> >   >>> l = range(10)
> >   >>> l[2:10:2] = [0]*5
> >   >>> l
> >   [0, 1, 0, 3, 0, 5, 0, 7, 0, 9]
> >   >>>
> 
> Hmph.  I'd be perfectly comfortable with
> 
>     >>> l[2:10:2] = [0]*5
>     TypeError: lists cannot assign to stepped slices
> 
> I can't see this being a commonly desired feature.  If it
> were perfectly clear what should happen in all cases, i'd
> be okay with it -- but the ambiguity you get when the length
> of the right side doesn't match the length of the indicated
> slice is sufficient for me to suggest just dropping it:
> better to avoid confusion altogether.

Well, the BDFL spoke.

> > (and presumably del l[a:b:c])
> 
> That seems reasonable and well-defined.

Yes.

> A separate question: does l[::-1] return l reversed?  

Yes.  All that cleverness is in PySlice_GetIndices, so this patch gets
it for free.

> That is, are the defaults for omitted start/end exchanged if step is
> negative?  This would seem to be quite useful -- having to say
> l[len(l)-1:-1:-1] would be a royal pain.

Well, you could probably always say 

l[sys.maxint:0:-1]

(your example returns the empty list!) but yes, I think l[::-1] is one
of the neater bits of new behaviour.

Cheers,
M.

-- 
  "declare"?  my bogometer indicates that you're really programming
  in some other language and trying  to force Common Lisp into your
  mindset.  this won't work.            -- Erik Naggum, comp.lang.lisp