changes in sbcl-0.8.4 relative to sbcl-0.8.3:
* fixed compiler performance when processing loops with a step >1;
+ * optimization: restored some effective method precomputation
+ (turned off by an ANSI fix in sbcl-0.8.3); the amount of
+ precomputation is now tunable.
planned incompatible changes in 0.8.x:
* (not done yet, but planned:) When the profiling interface settles