this bug is only for hardware floats, I think.
for bigfloats, fpprintprec apparently works right.
The person who recently made fpprintprec apply to hardware floats
did it wrong, but only by 1 digit. Doesn't seem like a "serious" bug.
As for trailing zeros, the intention, when I wrote this code in 1975 or so,
was to not print them. Otherwise
you get stuff like 0.0000000000000000000000000000000000b0.
RJF