diff options
author | vlefevre <vlefevre@280ebfd0-de03-0410-8827-d642c229c3f4> | 2007-05-22 11:50:06 +0000 |
---|---|---|
committer | vlefevre <vlefevre@280ebfd0-de03-0410-8827-d642c229c3f4> | 2007-05-22 11:50:06 +0000 |
commit | ecefc3453f1d38f4f71b25ee36061e396be3f554 (patch) | |
tree | 0ae8e2152620cf6df9c571a918795838a1062b29 | |
parent | 1524be6e95fdf4fb8d0995de59d4830899bd3fe6 (diff) | |
download | mpfr-ecefc3453f1d38f4f71b25ee36061e396be3f554.tar.gz |
Updated BUGS and TODO (mpfr_lgamma has been implemented, but see BUGS,
and gamma(-integer) with integer >= 1 is defined as NaN).
git-svn-id: svn://scm.gforge.inria.fr/svn/mpfr/trunk@4460 280ebfd0-de03-0410-8827-d642c229c3f4
-rw-r--r-- | BUGS | 11 | ||||
-rw-r--r-- | TODO | 6 |
2 files changed, 11 insertions, 6 deletions
@@ -34,6 +34,17 @@ Known bugs: bad practice and should be useless, but one never knows what other software will do). +* The implementation of mpfr_lgamma is incomplete (case x negative with + small exponent). + +* Incorrect behavior (possible infinite loop, e.g. in mpfr_exp2) in some + functions on tiny arguments, e.g. +/- 2^(emin-1), due to an integer + overflow in MPFR_FAST_COMPUTE_IF_SMALL_INPUT. + +* The mpfr_fma function behaves incorrectly if the multiplication overflows + or underflows. The overflow case has been fixed except in some corner + cases. + Potential bugs: * Possible integer overflows on some machines. @@ -80,10 +80,6 @@ New functions to implement: ############################################################################## - document MPFR_SIGN and/or provide mpfr_get_sign / mpfr_set_sign. -- function mpfr_lgamma, with the prototype - int mpfr_lgamma (mpfr_ptr, int *, mpfr_srcptr, mpfr_rnd_t) - that also returns the value of signgam; see thread - http://sympa.loria.fr/wwsympa/arc/mpfr/2006-10/msg00029.html - functions operating on mpfr_t and double: mpfr_add_d, mpfr_sub_d, mpfr_d_sub, mpfr_mul_d, mpfr_div_d, mpfr_d_div [suggested by Keith Briggs, 3 Jan 2006] @@ -355,8 +351,6 @@ Miscellaneous: z = round(x, p, TO_NEAREST_EVEN) so it avoids the double-rounding problem. -- check/define the sign of infinity for gamma(-integer) - - add tests of the ternary value for constants - When doing Extensive Check (--enable-assert=full), since all the |