[MLton-user] Questions on MLton

[Franck Arnaud]

> P.S. While I'm flaming, it is super-annoying to run carefully hand-tensed code
> on an X86 because there are no frigging registers. I need multiple regs
> available across call *and* across return *and* during loops to hold values.

I guess you could generate x86 asm which is known (to hardware 
engineers) to make use of the numerous registers of the internal 
RISC core of modern x86 implementations, so that you are (nearly) 
as efficient as if you could use the registers explicitely using 
the internal core's instruction set directly. It might be a bit 
painful though.