subclassing list

[John Machin]

spike wrote:
> I've googled like crazy and can't seem to find an answer to why this
> isn't working.
> 
> I want to create a custom list class that acts as a circular list.
> 
> ie: my_list = (0, 1, 2)

Perhaps you mean [0, 1, 2]
> 
> how I want it to behave:
> 
> my_list[0] -> 0
> my_list[1] -> 1
> my_list[2] -> 2
> my_list[3] -> 0
> my_list[4] -> 1
> ...etc
> 
> so, what I've tried to do is:
> 
> def circular_list(list):
>     def __getitem__(self, i):
>         if (i >= len(self)):
>             return self[i % len(self)]
>         else:
>             return self[i]
> 
> items = circular_list(range(8))
> 
> however, when I want to iterate over it with a for statement, I get:
> 
> TypeError: iteration over non-sequence
> 
> what am I missing?

0.1 Elementary debugging, like a "print" statement after "items = cir......"

0.2 pychecker -- These warning messages might have given you a clue:

spike.py:2: Local variable (__getitem__) not used
spike.py:2: Parameter (list) not used

1. A "class" statement. Your *function* circular_list returns None, 
which is a non-sequence.

2. A definition of what "in" a circular_list really means: in your 
example, 7 is obviously "in", but is 15 "in" or "not in"? Do you want 
"if x in a_circ_list" to be compatible with "for x in a_circ_list"?

3. An abhorrence of superfluous parentheses.

4. Battle-scars: you haven't been hit with "RuntimeError: maximum 
recursion depth exceeded" before, have you? After you've done 
s/def/class/ and put in some elementary test statements like "print 
items[3]", run it again and seen what happens, Google search for 
"__getitem__ infinite recursion" in this newsgroup.

HTH,
John