summaryrefslogtreecommitdiff
path: root/Build/source/libs/mpfr/mpfr-src/src/eq.c
diff options
context:
space:
mode:
Diffstat (limited to 'Build/source/libs/mpfr/mpfr-src/src/eq.c')
-rw-r--r--Build/source/libs/mpfr/mpfr-src/src/eq.c141
1 files changed, 0 insertions, 141 deletions
diff --git a/Build/source/libs/mpfr/mpfr-src/src/eq.c b/Build/source/libs/mpfr/mpfr-src/src/eq.c
deleted file mode 100644
index 68f69464bd9..00000000000
--- a/Build/source/libs/mpfr/mpfr-src/src/eq.c
+++ /dev/null
@@ -1,141 +0,0 @@
-/* mpfr_eq -- Compare two floats up to a specified bit #.
-
-Copyright 1999, 2001, 2003-2004, 2006-2020 Free Software Foundation, Inc.
-Contributed by the AriC and Caramba projects, INRIA.
-
-This file is part of the GNU MPFR Library.
-
-The GNU MPFR Library is free software; you can redistribute it and/or modify
-it under the terms of the GNU Lesser General Public License as published by
-the Free Software Foundation; either version 3 of the License, or (at your
-option) any later version.
-
-The GNU MPFR Library is distributed in the hope that it will be useful, but
-WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY
-or FITNESS FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public
-License for more details.
-
-You should have received a copy of the GNU Lesser General Public License
-along with the GNU MPFR Library; see the file COPYING.LESSER. If not, see
-https://www.gnu.org/licenses/ or write to the Free Software Foundation, Inc.,
-51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA. */
-
-
-#include "mpfr-impl.h"
-
-/* return non-zero if the first n_bits bits of u, v are equal,
- 0 otherwise */
-int
-mpfr_eq (mpfr_srcptr u, mpfr_srcptr v, unsigned long int n_bits)
-{
- mpfr_limb_srcptr up, vp;
- mp_size_t usize, vsize, size, i;
- mpfr_exp_t uexp, vexp;
- int k;
-
- if (MPFR_ARE_SINGULAR(u, v))
- {
- if (MPFR_IS_NAN(u) || MPFR_IS_NAN(v))
- return 0; /* non equal */
- else if (MPFR_IS_INF(u) && MPFR_IS_INF(v))
- return (MPFR_SIGN(u) == MPFR_SIGN(v));
- else if (MPFR_IS_ZERO(u) && MPFR_IS_ZERO(v))
- return 1;
- else
- return 0;
- }
-
- /* 1. Are the signs different? */
- if (MPFR_SIGN(u) != MPFR_SIGN(v))
- return 0;
-
- uexp = MPFR_GET_EXP (u);
- vexp = MPFR_GET_EXP (v);
-
- /* 2. Are the exponents different? */
- if (uexp != vexp)
- return 0; /* no bit agree */
-
- usize = MPFR_LIMB_SIZE (u);
- vsize = MPFR_LIMB_SIZE (v);
-
- if (vsize > usize) /* exchange u and v */
- {
- up = MPFR_MANT(v);
- vp = MPFR_MANT(u);
- size = vsize;
- vsize = usize;
- usize = size;
- }
- else
- {
- up = MPFR_MANT(u);
- vp = MPFR_MANT(v);
- }
-
- /* now usize >= vsize */
- MPFR_ASSERTD(usize >= vsize);
-
- if (usize > vsize)
- {
- if ((unsigned long) vsize * GMP_NUMB_BITS < n_bits)
- {
- /* check if low min(PREC(u), n_bits) - (vsize * GMP_NUMB_BITS)
- bits from u are non-zero */
- unsigned long remains = n_bits - (vsize * GMP_NUMB_BITS);
- k = usize - vsize - 1;
- while (k >= 0 && remains >= GMP_NUMB_BITS && !up[k])
- {
- k--;
- remains -= GMP_NUMB_BITS;
- }
- /* now either k < 0: all low bits from u are zero
- or remains < GMP_NUMB_BITS: check high bits from up[k]
- or up[k] <> 0: different */
- if (k >= 0 && (((remains < GMP_NUMB_BITS) &&
- (up[k] >> (GMP_NUMB_BITS - remains))) ||
- (remains >= GMP_NUMB_BITS && up[k])))
- return 0; /* surely too different */
- }
- size = vsize;
- }
- else
- {
- size = usize;
- }
-
- /* now size = min (usize, vsize) */
-
- /* If size is too large wrt n_bits, reduce it to look only at the
- high n_bits bits.
- Otherwise, if n_bits > size * GMP_NUMB_BITS, reduce n_bits to
- size * GMP_NUMB_BITS, since the extra low bits of one of the
- operands have already been check above. */
- if ((unsigned long) size > 1 + (n_bits - 1) / GMP_NUMB_BITS)
- size = 1 + (n_bits - 1) / GMP_NUMB_BITS;
- else if (n_bits > (unsigned long) size * GMP_NUMB_BITS)
- n_bits = size * GMP_NUMB_BITS;
-
- up += usize - size;
- vp += vsize - size;
-
- for (i = size - 1; i > 0 && n_bits >= GMP_NUMB_BITS; i--)
- {
- if (up[i] != vp[i])
- return 0;
- n_bits -= GMP_NUMB_BITS;
- }
-
- /* now either i=0 or n_bits<GMP_NUMB_BITS */
-
- /* since n_bits <= size * GMP_NUMB_BITS before the above for-loop,
- we have the invariant n_bits <= (i+1) * GMP_NUMB_BITS, thus
- we always have n_bits <= GMP_NUMB_BITS here */
- MPFR_ASSERTD(n_bits <= GMP_NUMB_BITS);
-
- if (n_bits & (GMP_NUMB_BITS - 1))
- return (up[i] >> (GMP_NUMB_BITS - (n_bits & (GMP_NUMB_BITS - 1))) ==
- vp[i] >> (GMP_NUMB_BITS - (n_bits & (GMP_NUMB_BITS - 1))));
- else
- return (up[i] == vp[i]);
-}