[FIX] WorldState>>runStepMethodsIn: ( [er][et] still relevant in 3.7a, consider for base )

[cdshaffer at acm.org]

I've been using this CS in several of my images since the original post.
 I think that the change is worthwhile for consideration for base
inclusion.  Basically the deferedUI queue is dispatched for at most
200ms and then morphs are stepped.  Thus if you are building large
numbers of defered events, this CS keeps you from killing the UI
completely.

There could be problems if assumptions are made about execution order of
a deferred request relative to the next stepping pass but anyone who
assumes such an ordering should be...well...discouraged from depending
on such things.  Anyway, I don't think I've ever had problems related to
this.

I guess the only remaining question is "Does it fix any problems which
are occurring in the wild?"  I'm not sure what the answer is but I do
think that it's a good piece of defensive programming :-)













<This post brought to you by BFAV2>