1 1.1.1.6 mrg Copyright 1999, 2001-2023 Free Software Foundation, Inc. 2 1.1.1.3 mrg Contributed by the AriC and Caramba projects, INRIA. 3 1.1 mrg 4 1.1 mrg This file is part of the GNU MPFR Library. 5 1.1 mrg 6 1.1 mrg The GNU MPFR Library is free software; you can redistribute it and/or modify 7 1.1 mrg it under the terms of the GNU Lesser General Public License as published by 8 1.1 mrg the Free Software Foundation; either version 3 of the License, or (at your 9 1.1 mrg option) any later version. 10 1.1 mrg 11 1.1 mrg The GNU MPFR Library is distributed in the hope that it will be useful, but 12 1.1 mrg WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY 13 1.1 mrg or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public 14 1.1 mrg License for more details. 15 1.1 mrg 16 1.1 mrg You should have received a copy of the GNU Lesser General Public License 17 1.1 mrg along with the GNU MPFR Library; see the file COPYING.LESSER. If not, see 18 1.1.1.5 mrg https://www.gnu.org/licenses/ or write to the Free Software Foundation, Inc., 19 1.1 mrg 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. 20 1.1 mrg 21 1.1 mrg ############################################################################## 22 1.1 mrg 23 1.1 mrg Known bugs: 24 1.1 mrg 25 1.1 mrg * The overflow/underflow exceptions may be badly handled in some functions; 26 1.1 mrg specially when the intermediary internal results have exponent which 27 1.1 mrg exceeds the hardware limit (2^30 for a 32 bits CPU, and 2^62 for a 64 bits 28 1.1 mrg CPU) or the exact result is close to an overflow/underflow threshold. 29 1.1 mrg 30 1.1 mrg * Under Linux/x86 with the traditional FPU, some functions do not work 31 1.1 mrg if the FPU rounding precision has been changed to single (this is a 32 1.1 mrg bad practice and should be useless, but one never knows what other 33 1.1 mrg software will do). 34 1.1 mrg 35 1.1 mrg * Some functions do not use MPFR_SAVE_EXPO_* macros, thus do not behave 36 1.1 mrg correctly in a reduced exponent range. 37 1.1 mrg 38 1.1 mrg * Function hypot gives incorrect result when on the one hand the difference 39 1.1 mrg between parameters' exponents is near 2*MPFR_EMAX_MAX and on the other hand 40 1.1 mrg the output precision or the precision of the parameter with greatest 41 1.1 mrg absolute value is greater than 2*MPFR_EMAX_MAX-4. 42 1.1.1.7 mrg Note: Such huge precisions are not possible as they would be larger than 43 1.1.1.7 mrg MPFR_PREC_MAX, unless the types for mpfr_exp_t and/or mpfr_prec_t are 44 1.1.1.7 mrg changed (only for developers or expert users, not officially supported). 45 1.1 mrg 46 1.1 mrg Potential bugs: 47 1.1 mrg 48 1.1 mrg * Possible incorrect results due to internal underflow, which can lead to 49 1.1 mrg a huge loss of accuracy while the error analysis doesn't take that into 50 1.1 mrg account. If the underflow occurs at the last function call (just before 51 1.1 mrg the MPFR_CAN_ROUND), the result should be correct (or MPFR gets into an 52 1.1 mrg infinite loop). TODO: check the code and the error analysis. 53 1.1 mrg 54 1.1.1.4 mrg * Possible bugs with huge precisions (> 2^30) and a 32-bit ABI, in particular 55 1.1.1.4 mrg undetected integer overflows. TODO: use the MPFR_ADD_PREC macro. 56 1.1 mrg 57 1.1 mrg * Possible bugs if the chosen exponent range does not allow to represent 58 1.1 mrg the range [1/16, 16]. 59 1.1 mrg 60 1.1 mrg * Possible infinite loop in some functions for particular cases: when 61 1.1 mrg the exact result is an exactly representable number or the middle of 62 1.1.1.6 mrg consecutive two such numbers. However, for non-algebraic functions, it is 63 1.1 mrg believed that no such case exists, except the well-known cases like cos(0)=1, 64 1.1 mrg exp(0)=1, and so on, and the x^y function when y is an integer or y=1/2^k. 65 1.1 mrg 66 1.1 mrg * The mpfr_set_ld function may be quite slow if the long double type has an 67 1.1 mrg exponent of more than 15 bits. 68 1.1 mrg 69 1.1 mrg * mpfr_set_d may give wrong results on some non-IEEE architectures. 70 1.1 mrg 71 1.1 mrg * Error analysis for some functions may be incorrect (out-of-date due 72 1.1 mrg to modifications in the code?). 73 1.1 mrg 74 1.1 mrg * Possible use of non-portable feature (pre-C99) of the integer division 75 1.1 mrg with negative result. 76
Indexes created Tue Apr 21 00:23:25 UTC 2026