[src/strtofr.c] Updated FIXME, due to new tests added to

tests/tstrtofr.c in r13278-13279,13286-13288.

git-svn-id: svn://scm.gforge.inria.fr/svn/mpfr/trunk@13290 280ebfd0-de03-0410-8827-d642c229c3f4

1 files changed, 1 insertions, 6 deletions

diff --git a/src/strtofr.c b/src/strtofr.c
index dee46eeb7..510719775 100644
--- a/src/strtofr.c
+++ b/src/strtofr.c
@@ -621,12 +621,7 @@ parsed_string_to_mpfr (mpfr_t x, struct parsed_string *pstr, mpfr_rnd_t rnd)
             }
           /* exp = shift count */
           /* FIXME: Why not taking offset (when count != 0) into account?
-             The value of exp does not seem to matter (changing it does
-             not make tstrtofr fail). The reason is that the rounding test
-             does no depend on the exponent and one always loops in this
-             case (except in case of overflow or underflow, but this is
-             not tested). -> This is actually due to the current tstrtofr.c
-             tests, which seem to always be hard-to-round cases!!!
+             Indeed, in r13289, a test fails for diff_ysize = -2.
              Additionally, the lost bits are not taken into account
              in the error analysis below! */
           exp = GMP_NUMB_BITS - count;