floatscan: fix incorrect count of leading nonzero digits
authorRich Felker <dalias@aerifal.cx>
Mon, 16 Apr 2012 05:53:52 +0000 (01:53 -0400)
committerRich Felker <dalias@aerifal.cx>
Mon, 16 Apr 2012 05:53:52 +0000 (01:53 -0400)
commit2ac580fdfe1daf9924f221dc98878ae480306bfa
tree78e6e97b1801310be2bb4ae7085f3808b6bff1c8
parent67b25fe0a8947de93e3eddd36dae24a9fec12ade
floatscan: fix incorrect count of leading nonzero digits

this off-by-one error was causing values with just one digit past the
decimal point to be treated by the integer case. in many cases it
would yield the correct result, but if expressions are evaluated in
excess precision, double rounding may occur.
src/internal/floatscan.c