source: mainline/uspace/lib/math/generic/trunc.c@ ed9043f7

/*
 * Copyright (c) 2015 Jiri Svoboda
 * Copyright (c) 2014 Martin Decky
 * All rights reserved.
 *
 * Redistribution and use in source and binary forms, with or without
 * modification, are permitted provided that the following conditions
 * are met:
 *
 * - Redistributions of source code must retain the above copyright
 *   notice, this list of conditions and the following disclaimer.
 * - Redistributions in binary form must reproduce the above copyright
 *   notice, this list of conditions and the following disclaimer in the
 *   documentation and/or other materials provided with the distribution.
 * - The name of the author may not be used to endorse or promote products
 *   derived from this software without specific prior written permission.
 *
 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
 * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 */

/** @addtogroup libmath
 * @{
 */
/** @file
 */

#include <math.h>
#include <float.h>
#include <stdint.h>

/** Truncate fractional part (round towards zero)
 *
 * Truncate the fractional part of IEEE 754 single
 * precision floating point number by zeroing fraction
 * bits, effectively rounding the number towards zero
 * to the nearest whole number.
 *
 * If the argument is infinity or NaN, an exception
 * should be indicated. This is not implemented yet.
 *
 * @param val Floating point number.
 *
 * @return Number rounded towards zero.
 *
 */
float truncf(float val)
{
        /* If the input is a nan, return a canonical nan. */
        if (isnan(val))
                return __builtin_nanf("");

        const int exp_bias = FLT_MAX_EXP - 1;
        const int mant_bits = FLT_MANT_DIG - 1;
        const uint32_t mant_mask = (UINT32_C(1) << mant_bits) - 1;

        union {
                float f;
                uint32_t i;
        } u = { .f = fabsf(val) };

        int exp = (u.i >> mant_bits) - exp_bias;

        /* If value is less than one, return zero with appropriate sign. */
        if (exp < 0)
                return copysignf(0.0f, val);

        if (exp >= mant_bits)
                return val;

        /* Truncate irrelevant fraction bits */
        u.i &= ~(mant_mask >> exp);
        return copysignf(u.f, val);
}

/** Truncate fractional part (round towards zero)
 *
 * Truncate the fractional part of IEEE 754 double
 * precision floating point number by zeroing fraction
 * bits, effectively rounding the number towards zero
 * to the nearest whole number.
 *
 * If the argument is infinity or NaN, an exception
 * should be indicated. This is not implemented yet.
 *
 * @param val Floating point number.
 *
 * @return Number rounded towards zero.
 *
 */
double trunc(double val)
{
        /* If the input is a nan, return a canonical nan. */
        if (isnan(val))
                return __builtin_nan("");

        const int exp_bias = DBL_MAX_EXP - 1;
        const int mant_bits = DBL_MANT_DIG - 1;
        const uint64_t mant_mask = (UINT64_C(1) << mant_bits) - 1;

        union {
                double f;
                uint64_t i;
        } u = { .f = fabs(val) };

        int exp = ((int)(u.i >> mant_bits)) - exp_bias;

        /* If value is less than one, return zero with appropriate sign. */
        if (exp < 0)
                return copysign(0.0, val);

        if (exp >= mant_bits)
                return val;

        /* Truncate irrelevant fraction bits */
        u.i &= ~(mant_mask >> exp);
        return copysign(u.f, val);
}

/** @}
 */