[R] How would you run repeated-measures multifactorial MANCOVA?

[Mitchell Maltenfort]

OK, now that I've worked through Venebales and Ripley and a few other
sources, I can see more than one way of attacking a problem I expect
to be facing soon: a repeated-measures MANCOVA with more than one
X-factor.

But that got me wondering how the people who've been playing with R
longer than I have would be doing it...and thinking it might be an
interesting topic for discussion on the list.

How would you prefer to do a repeated-measures multifactor MANCOVA in
R?   Is there a way that's particularly good --- or particularly bad?