Updated bug in utils prtnum.c

When changing the utilities that generate code for floating point numbers a bug was introduced into prtnum.c that impacted single precision numbers. This had no effect on the math library because the tables had been built prior th the bugs introduction.