Rounds a float, double, or real value to the nearest (away from zero) int value using lroundf.
Rounds a float, double, or real value to the nearest (away from zero) long value using llroundf.
Boost Software License Version 1.0. See LICENSE_BOOST.txt for details. Alternatively, this file may be distributed under the terms of the Tango 3-Clause BSD License (see LICENSE_BSD.txt for details).
Copyright (c) 2009-2016 dunnhumby Germany GmbH. All rights reserved.
Module contains utility functions to convert floats to integer types. Old conversion functions eg. ocean.math.Math.rndint or ocean.math.Math.rndlong currently round x.5 to the nearest even integer. So rndint(4.5) == 4 and rndint(5.5) == 6. This is undesired behaviour for some situations, so the functions in this module round to the nearest integer. So floatToInt(4.5, output) sets output == 5 and floatToInt(5.5, output) sets output == 6 (this is round to nearest integer away from zero rounding see http://man7.org/linux/man-pages/man3/lround.3.html for details on the stdc lround, lroundf, llround, and llroundf functions).
To check for errors the functions feclearexcept and fetestexcept are used. The feclearexcept(FE_ALL_EXCEPT) method is called before calling the mathematical function and after the mathematical function has been called the fetestexcept method is called to check for errors that occured in the mathematical function (for more details in these functions see http://man7.org/linux/man-pages/man7/math_error.7.html$(RPAREN).