Changes by: Brian Emre Aydemir (baydemir at cis.upenn.edu)
Date: 2002-04-03 00:37:31 -0800 (Wed, 03 Apr 2002)
Revision: 3562
Log message:
So, I've defined the compile function in Mp_mc_compile to actually
take an Fir.prog, convert the function definitions to terms,
and then back again, in one big identity operation. The current
term representation of the FIR functions is a bit of a hack that
will need to be cleaned up if anything non-trivial is to be done.
(Each individual function is fine I hope. It's the program as a whole
that's represented rather poorly. It's essentially a (term SymbolTable.t),
one term for each fundef in the original Fir.prog.prog_funs.)
Though, this should be sufficient for basic testing I think.