[Bf-python] Update: Python Ipocurve/CurNurb module changes

[Stephen Swaney]

The dual pointer comparison stuff strikes me as overly complicated
and as you point out, it still has some holes.

The only place where we know both the location of the control point
array *and* its length is in the Nurb struct.  The way I see it, bpy
control points need to access their point data thru their parent Nurb
struct.  The metaphor here is an iterator initialized to the Nth
element of an array.

The other alternative is to blow off bpy control point types
completely and have the accessors simply return python list copies of
the point data.

As for deleting objects and then accessing them, managing object
lifetime is a fundamental part of doing OO programming.  Language
features can help you, but not protect you completely.  Without
implementing a signaling mechanism, this is the best we can do in this
case.

I have the tp_getset stuff done for CurNurb.  My goal was to finish
the new accessors and get it checked in this morning before the
meeting, but I seem to be caught up in that Heisenbergian project
management state of writing about code instead of actually coding.

-- 
Stephen Swaney			
sswaney at centurytel.net