integration of PolySpace mods into MLton

[Stephen Weeks]

> > * I didn't understand the reason for the test on line 1151 of gc.c
> >  		    or (s->toSize >= (128ul*1024ul*1024ul))
> >   that forces toSpace to be unmapped if it is too small.
> 
> If is my understanding that this extra code instead forces the
> unmapping of the old space if it is too large, not too small. I
> added this code in order to improve the VM behaviour of the system
> when paging.

Makes sense.  I added it in.

> > * Do you need the ability set RAMslop from the command line?  We have
> >   it at .85, but you have it at .9.
> 
> Preferably, as we tend to have very large memory footprints.

OK.  I added a ram-slop runtime system argument.

> > * Do you still need -stop s when compiling -native false?  We had
> >   eliminated it.
> 
> As long as we can compile and link in other C files, this is OK.

Any additional files on the command line will be added in the final
call to gcc that links to build the main executable.  So you can do
things like "mlton z.sml foo.c".  Hopefully that is good enough.

> > * Do you need the -O option when compiling -native false?  We had 
> >   eliminated it
> 
> We need to be able to set the optimisation level.

OK.  I added back -O.  It is an "expert" option, and is hence not
displayed in the normal help message.  But, you can see it (along with
all the other expert options) with 'mlton -v -z'.