use cleaner code for handling float rounding in vfprintf
authorSzabolcs Nagy <nsz@port70.net>
Fri, 30 May 2014 15:09:53 +0000 (17:09 +0200)
committerRich Felker <dalias@aerifal.cx>
Fri, 30 May 2014 17:06:51 +0000 (13:06 -0400)
commitbff6095d915f3e41206e47ea2a570ecb937ef926
tree87fe4f5a4541909018baa9e9495ab8b6b29354be
parentc3504686580bd93c3937e3b6d6cfa223c4cd7b74
use cleaner code for handling float rounding in vfprintf

CONCAT(0x1p,LDBL_MANT_DIG) is not safe outside of libc,
use 2/LDBL_EPSILON instead.

fix was proposed by Morten Welinder.
src/stdio/vfprintf.c