[TransWarp] Any objections to requiring 2.2 from here on out?

[Phillip J. Eby]

If anyone has any objections to TW requiring Python 2.2, please speak up 
*now*.  I hope to merge the Simplification-Branch back onto the trunk this 
weekend and tag it with a "last chance for 2.1" tag, then start some 
serious 2.2 work.

After doing some experimentation with TransWarp and Python 2.2, I've 
determined that a lot of what TransWarp does can be done *much* easier 
using 2.2 metaclass features.  Both in the sense that there are now things 
you can do without using TransWarp, and in the sense that TransWarp itself 
can become less complex, for those things you still need it for in 2.2.

So far, I've determined that with 2.2, ExtensionClasses and Acquisition 
aren't needed to do anything that TW wants to do.  Further, some of the 
"classic" metaclasses in TW can become "new style" metaclasses, which 
should increase their range of abilities.  Also, many things that required 
Postprocessors and other Advisors in 2.1 can now be done more easily with 
metaclasses.  Aspect-oriented programming will still require a "build" 
phase, but many other sorts of generative programming and metaprogramming 
may not.

So, the benefits are too substantial for me to ignore.  The only 
significant drawback that I see is that Jython may be excluded as a TW 
platform for an unknown period of time.  However, for the work Ty and I 
have lined up at present, this is not likely to be an issue, and any Java 
interactions that come up will likely be either 1) do-able without 
TransWarp, or 2) do-able with JPE.

Anyway, if you have objections, please let me know *now*.