Logged In: YES 
user_id=3066

I've attached a new version of the patch that simplifies the
code for the new types.

It doesn't look like I'll bet getting to magic structseq
support anytime soon.  I'm also no longer convinced that
this needs to be held up for that.  This should be
considered for Python 2.5 on the basis of the feature; the
implementation can be changed some someone get around to
creation of structseq types from Python.