limit check pushed inside loop

[Matthew Fluet]

> > > Which raises an interesting point -- in that non-allocating loop, there is
> > > a check for 24 additional bytes at each entry; probably for the
> > > continuation where I print out the result, but it looks like it got pushed
> > > too far back into the loop.)
> > 
> > This feels like a bug.  I'll look into it.
> 
> Actually, it's not a bug.  Limit checks are only inserted in a few places, loop
> headers being one of them.  The amount is the max over all control flow paths to 
> the next limit check.  So, even if nothing is allocated inside the loop, as long 
> as one of the exits leads to allocation before a limit check, that will get
> pushed inside the loop.

O.K.  I only mentioned it because from a performance point of view, a
limitCheck with zero bytes can simply compare frontier with limit without
some calculation (whether it be via leal or movl/addl).