Bug in general floating point format

[John M. Canning <jc@cvl.umd.edu>]

  Here is a bug with the general floating point output format.  I'm posting
to both bug-macl and info-macl so everyone sees this.

         - John Canning (jc@cvl.umd.edu)

? (format t "~&~,2G" 0.08)
8.00E-2
NIL
? (format t "~&~,2G" -0.08)
> Error: NIL is not a valid argument to - .
> While executing: #<An Anonymous Compiled-function>
> Type Command-/ to continue, Command-. to abort.
1 >