/****************************************************************************** * Spine Runtimes License Agreement * Last updated January 1, 2020. Replaces all prior versions. * * Copyright (c) 2013-2020, Esoteric Software LLC * * Integration of the Spine Runtimes into software or otherwise creating * derivative works of the Spine Runtimes is permitted under the terms and * conditions of Section 2 of the Spine Editor License Agreement: * http://esotericsoftware.com/spine-editor-license * * Otherwise, it is permitted to integrate the Spine Runtimes into software * or otherwise create derivative works of the Spine Runtimes (collectively, * "Products"), provided that each user of the Products must obtain their own * Spine Editor license and redistribution of the Products in any form must * include this license and copyright notice. * * THE SPINE RUNTIMES ARE PROVIDED BY ESOTERIC SOFTWARE LLC "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 ESOTERIC SOFTWARE LLC BE LIABLE FOR ANY * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, * BUSINESS INTERRUPTION, OR LOSS OF USE, DATA, OR PROFITS) 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 * THE SPINE RUNTIMES, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. *****************************************************************************/ #ifdef SPINE_UE4 #include "SpinePluginPrivatePCH.h" #endif #include #include #include // Required for division by 0 in _isNaN on MSVC #ifdef _MSC_VER #pragma warning(disable:4723) #endif using namespace spine; const float MathUtil::Pi = 3.1415926535897932385f; const float MathUtil::Pi_2 = 3.1415926535897932385f * 2; const float MathUtil::Deg_Rad = (3.1415926535897932385f / 180.0f); const float MathUtil::Rad_Deg = (180.0f / 3.1415926535897932385f); float MathUtil::abs(float v) { return ((v) < 0 ? -(v) : (v)); } float MathUtil::sign(float v) { return ((v) < 0 ? -1.0f : (v) > 0 ? 1.0f : 0.0f); } float MathUtil::clamp(float x, float min, float max) { return ((x) < (min) ? (min) : ((x) > (max) ? (max) : (x))); } float MathUtil::fmod(float a, float b) { return (float)::fmod(a, b); } /// Returns atan2 in radians, faster but less accurate than Math.Atan2. Average error of 0.00231 radians (0.1323 /// degrees), largest error of 0.00488 radians (0.2796 degrees). float MathUtil::atan2(float y, float x) { return (float)::atan2(y, x); } /// Returns the cosine in radians from a lookup table. float MathUtil::cos(float radians) { return (float)::cos(radians); } /// Returns the sine in radians from a lookup table. float MathUtil::sin(float radians) { return (float)::sin(radians); } float MathUtil::sqrt(float v) { return (float)::sqrt(v); } float MathUtil::acos(float v) { return (float)::acos(v); } /// Returns the sine in radians from a lookup table. float MathUtil::sinDeg(float degrees) { return (float)::sin(degrees * MathUtil::Deg_Rad); } /// Returns the cosine in radians from a lookup table. float MathUtil::cosDeg(float degrees) { return (float)::cos(degrees * MathUtil::Deg_Rad); } /* Need to pass 0 as an argument, so VC++ doesn't error with C2124 */ static bool _isNan(float value, float zero) { float _nan = (float) 0.0 / zero; return 0 == memcmp((void *) &value, (void *) &_nan, sizeof(value)); } bool MathUtil::isNan(float v) { return _isNan(v, 0); } float MathUtil::random() { return ::rand() / (float)RAND_MAX; } float MathUtil::randomTriangular(float min, float max) { return randomTriangular(min, max, (min + max) * 0.5f); } float MathUtil::randomTriangular(float min, float max, float mode) { float u = random(); float d = max - min; if (u <= (mode - min) / d) return min + sqrt(u * d * (mode - min)); return max - sqrt((1 - u) * d * (max - mode)); } float MathUtil::pow(float a, float b) { return (float)::pow(a, b); }