Assigning to __class__ : bad form?

[The Pythonista]

I've been wondering for a while about whether assigning to __class__ is 
bad form or not.  Specifically, I mean doing so when some other method of 
implementing the functionality you're after is available (i.e. using an 
adapter, or something like the strategy pattern).

To give an example and a non-example of what I'm talking about, consider 
the following recipes from the online Python Cookbook:

Ring Buffer: http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/68429

In this case, I think the assignment to __class__ just obfuscates things, 
and the example would be better coded as a single class.

On the other hand,

Fast copy of an object having a slow __init__ : http://
aspn.activestate.com/ASPN/Cookbook/Python/Recipe/66507

This seems like a reasonable use case for assigning to __class__ (except 
that it's already implemented in the 'new' module, but, read the Martelli-
bot's comments near the end of the recipe).  I consider this a reasonable 
use case for assigning to __class__, because, short of the 'new' module, 
I don't see any other way to accomplish it.

So, what is the opinion of the broader Python community?  Is code that 
assigns to __class__ just clever trickiness to be avoided, or is it 
sometimes a good thing?

-- 
code.py: A blog about life, the universe, and Python

http://pythonista.wordpress.com
** Posted from http://www.teranews.com **