4 - Are we being careful enough about narrowing those out-of-range CmmInts?
8 - keep track of free stack slots
12 - picking the assignment on entry to a block: better to defer this
13 until we know all the assignments. In a loop, we should pick
14 the assignment from the looping jump (fixpointing?), so that any
15 fixup code ends up *outside* the loop. Otherwise, we should
16 pick the assignment that results in the least fixup code.
20 -- -----------------------------------------------------------------------------
23 - x86 genCCall needs to tack on the @size for stdcalls (might not be in the
26 - x86: should really clean up that IMUL64 stuff, and tell the code gen about
27 Intel imul instructions.
29 - x86: we're not careful enough about making sure that we only use
30 byte-addressable registers in byte instructions. Should we do it this
31 way, or stick to using 32-bit registers everywhere?
33 - Use SSE for floating point, optionally.
35 ------------------------------------------------------------------------------
36 -- Further optimisations:
38 - We might be able to extend the scope of the inlining phase so it can
39 skip over more statements that don't affect the value of the inlined