[MLton] building mlton on Solaris 10

[Matthew Fluet]

> testing mlton.share
> 1c1
> < size of a is 1600
> ---
> > size of a is 2000

The mlton.share regression checks the behavior of the MLton.share and 
MLton.shareAll functions (http://www.mlton.org/MLtonStructure).  It 
maximizes sharing in the heap of immutable objects.

Because of alignment considerations (and, in the future, because of the 
native sizes of system types), object representations and sizes can vary 
from platform to platform.  In the regression suite, there are 
mlton.share.sparc-solaris.ok and mlton.share.hppa-hpux.ok files, which 
contain the expected output on those platforms.

What's odd is that you seem to be producing the expected sizes for 
sparc-solaris, but the regression script appears to be comparing against 
mlton.share.ok.  What is the output of running './bin/platform'?