OpenGrok

1.1  mrg Copyright 1999, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011 Free Software Foundation, Inc.
1.1  mrg Contributed by the Arenaire and Cacao projects, INRIA.
1.1  mrg
1.1  mrg This file is part of the GNU MPFR Library.
1.1  mrg
1.1  mrg The GNU MPFR Library is free software; you can redistribute it and/or modify
1.1  mrg it under the terms of the GNU Lesser General Public License as published by
1.1  mrg the Free Software Foundation; either version 3 of the License, or (at your
1.1  mrg option) any later version.
1.1  mrg
1.1  mrg The GNU MPFR Library is distributed in the hope that it will be useful, but
1.1  mrg WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
1.1  mrg or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU Lesser General Public
1.1  mrg License for more details.
1.1  mrg
1.1  mrg You should have received a copy of the GNU Lesser General Public License
1.1  mrg along with the GNU MPFR Library; see the file COPYING.LESSER.  If not, see
1.1  mrg http://www.gnu.org/licenses/ or write to the Free Software Foundation, Inc.,
1.1  mrg 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA.
1.1  mrg
1.1  mrg ##############################################################################
1.1  mrg
1.1  mrg Probably many bugs.
1.1  mrg
1.1  mrg Known bugs:
1.1  mrg
1.1  mrg * The overflow/underflow exceptions may be badly handled in some functions;
1.1  mrg   specially when the intermediary internal results have exponent which
1.1  mrg   exceeds the hardware limit (2^30 for a 32 bits CPU, and 2^62 for a 64 bits
1.1  mrg   CPU) or the exact result is close to an overflow/underflow threshold.
1.1  mrg
1.1  mrg * Under Linux/x86 with the traditional FPU, some functions do not work
1.1  mrg   if the FPU rounding precision has been changed to single (this is a
1.1  mrg   bad practice and should be useless, but one never knows what other
1.1  mrg   software will do).
1.1  mrg
1.1  mrg * Some functions do not use MPFR_SAVE_EXPO_* macros, thus do not behave
1.1  mrg   correctly in a reduced exponent range.
1.1  mrg
1.1  mrg * Function hypot gives incorrect result when on the one hand the difference
1.1  mrg   between parameters' exponents is near 2*MPFR_EMAX_MAX and on the other hand
1.1  mrg   the output precision or the precision of the parameter with greatest
1.1  mrg   absolute value is greater than 2*MPFR_EMAX_MAX-4.
1.1  mrg
1.1  mrg Potential bugs:
1.1  mrg
1.1  mrg * Possible incorrect results due to internal underflow, which can lead to
1.1  mrg   a huge loss of accuracy while the error analysis doesn't take that into
1.1  mrg   account. If the underflow occurs at the last function call (just before
1.1  mrg   the MPFR_CAN_ROUND), the result should be correct (or MPFR gets into an
1.1  mrg   infinite loop). TODO: check the code and the error analysis.
1.1  mrg
1.1  mrg * Possible integer overflows on some machines.
1.1  mrg
1.1  mrg * Possible bugs with huge precisions (> 2^30).
1.1  mrg
1.1  mrg * Possible bugs if the chosen exponent range does not allow to represent
1.1  mrg   the range [1/16, 16].
1.1  mrg
1.1  mrg * Possible infinite loop in some functions for particular cases: when
1.1  mrg   the exact result is an exactly representable number or the middle of
1.1  mrg   consecutive two such numbers. However for non-algebraic functions, it is
1.1  mrg   believed that no such case exists, except the well-known cases like cos(0)=1,
1.1  mrg   exp(0)=1, and so on, and the x^y function when y is an integer or y=1/2^k.
1.1  mrg
1.1  mrg * The mpfr_set_ld function may be quite slow if the long double type has an
1.1  mrg   exponent of more than 15 bits.
1.1  mrg
1.1  mrg * mpfr_set_d may give wrong results on some non-IEEE architectures.
1.1  mrg
1.1  mrg * Error analysis for some functions may be incorrect (out-of-date due
1.1  mrg   to modifications in the code?).
1.1  mrg
1.1  mrg * Possible use of non-portable feature (pre-C99) of the integer division
1.1  mrg   with negative result.