No subject

Henry Cejtin henry@sourcelight.com
Tue, 9 Oct 2001 16:31:16 -0500


I  realized  that  things like common subexpression hoisting and the like are
really much more valuable in MLton that I would have guessed.  In Scheme  the
point  behind  such  optimizations (as opposed to the user having to do it by
hand) was largely macros (which are going to insert lots  of  opportunities).
In MLton the correspnding opportunities come from cross-module inlining.