The problem is also evident in #// and #\\ -- if the receiver is either
a LargePositiveInteger or a LargeNegativeInteger and the argument is a
negative Smallinteger you see the same "highBit is not defined for
negative integers" walkback in 2.1.

-- Dwight

Travis or Kerrin Griggs wrote:
> 
> Folks,
> 
> Try the following (in 2.1)...
> 
> -4 / 1000000000000000
> 
> It blows up when it tries to find the gcd: of -4 and 1000000000000000.
> Something about negative numbers having no high bit which breaks the
> quo: implementation, which all looks suspiciously like something related
> to this big thread a couple of months back about twos complements of
> negative numbers and related bitwise operations.
> 
> Did we "fix" something a little too much here? :)
> 
> Travis Griggs