Release 110.65

regenerated mdl.lex.sml

re-run lex after update to ml-ulex

Release 110.64

  Added FSQRT instructions.

(again) re-generated mdl.lex.sml after ml-ulex fix

  Fixed a couple bugs in generating floating-point arithmetic.

re-generated mdl.lex.sml after ml-ulex fix

  Added peephole optimization.

  Fixed $ bug.

  New code generator for AMD64, which now uses SSE for floating point.

  Deleted the old amd64 code generator.

added results of running lexer and parser generators to repository;
this is necessary for bootstrapping:
ml-ulex is not yet installed when this gets compiled by the installer

BlockPlacement functor formal parameter needs a where structure definition in order to be consistent with WeightedBlockPlacementFn formal parameters

  Removed stale functor AMD64GenCCallFn.

  Substantially changed the signature and implementation of AMD64 SVID.  It now looks similar to the ia32 SVID, yet uses staged allocation.

Release 110.63

fixed problem with path anchors in AMD64.cm

Fixed C calls to properly kill caller save registers.

Fixed bug where spill code uses bogus sizes for operands.

Added path for c-calls via staged allocation.

  Comment out rewrite modules.

Added path for staged allocation.

Added support for Staged Allocation (http://www.eecs.harvard.edu/~nr/pubs/staged-abstract.html)

    x86MCEmitter crashed when the immediate operand to MOVB was outside
    of the range -128 ... 127.  I've changed the code so that the range
    check is disabled.   Only the low order 8 bits of the immediate operand
    is now significant.

Release 110.62

Release 110.61

Release 110.60

fixed build scripts for nowhere

moved MLRISC to toplevel

created releases for MLRISC

created branches for MLRISC

created toplevel dir for MLRISC