[MLton] C codegen and world* regressions

[Matthew Fluet]

The "may_alias" attribute isn't understood (and consequently ignored) by 
gcc 3.2.2, such as exists on Red Hat Enterprise Server AS version 4; this 
is used by runtime/util/pointer.h; I ran the regression suite with 
-codegen c just to verify that it wasn't causing any issues (though, the 
pointer typedef is used most heavily in the runtime/gc, not in C-codegen 
emitted code).  In any case, I'm getting errors on all of the world* 
regressions:

unhandled exception: SysErr: Success [<UNKNOWN>]
unhandled exception: Fail: child failed
Nonzero exit status.

I'm guessing this is due to the reorganization of world save/restore in 
r4642 and r4648; it'll cause problems with bootstraps on non-x86 systems.