[R6RS] Distinct S-expression in `case'

[dyb at cs.indiana.edu]

> Do we require a check and exception with condition type &syntax for
> duplicate case constants at expansion time?

No.  The restriction preventing duplicate case constants is fine for
hand-written code, but it's undesirable for machine-generated code.  In
fact, I prefer we adopt a left-to-right processing order for case and
allow duplicate datums.  This change wouldn't break any R5RS-compliant
code.

Kent