[plt-scheme] symbols redefined in SRFIs

Danny, thanks for the details -- wish I had something like this before
I started digging :)

The ideea was not to change the module system's behaviour in case of
conflicts, but to keep around a single version of reverse!, assoc etc
-- see Matthew Flatt's earlier analysis.

It seems that mostly map and for-each would be problematic (and
perhaps reverse! due to an extra check in the reference
implementation, which doesn't seem to be mandated by the SRFI though).

-- Dan