make it clear that we don't propagate the sign bit in mpfr_get_d and mpfr_set_d

and similar functions for binary32, decimal64 and binary128


git-svn-id: https://scm.gforge.inria.fr/anonscm/svn/mpfr/trunk@14296 280ebfd0-de03-0410-8827-d642c229c3f4

1 files changed, 3 insertions, 3 deletions

diff --git a/src/set_flt.c b/src/set_flt.c
index 7a0a7d9b2..6f60b7e5f 100644
--- a/src/set_flt.c
+++ b/src/set_flt.c
@@ -26,9 +26,9 @@ int
 mpfr_set_flt (mpfr_ptr r, float f, mpfr_rnd_t rnd_mode)
 {
   /* we convert f to double precision and use mpfr_set_d;
-     NaN and infinities should be preserved, and all single precision
-     numbers are exactly representable in the double format, thus the
-     conversion is always exact */
+     NaN and infinities should be preserved (except the sign bit for NaN),
+     and all single precision numbers are exactly representable in the double
+     format, thus the conversion is always exact */
   return mpfr_set_d (r, (double) f, rnd_mode);
 }