Merge pull request #92 from paulpach/ceil

Simplify ceil division

Author: GabrielTofvesson

MLAPI.Tests/NetworkingManagerComponents/Binary/ArithmeticTest.cs

@@ -15,8 +15,14 @@ public void testCeil()
         {
             Assert.That(Arithmetic.CeilingExact(10, 5), Is.EqualTo(2));
             Assert.That(Arithmetic.CeilingExact(11, 5), Is.EqualTo(3));
-            Assert.That(Arithmetic.CeilingExact(-5, 5), Is.EqualTo(-1));
-            Assert.That(Arithmetic.CeilingExact(-4, 5), Is.EqualTo(-1));
+            Assert.That(Arithmetic.CeilingExact(0, 5), Is.EqualTo(0));
+            Assert.That(Arithmetic.CeilingExact(1, 5), Is.EqualTo(1));
+            Assert.That(Arithmetic.CeilingExact(2, 5), Is.EqualTo(1));
+            Assert.That(Arithmetic.CeilingExact(3, 5), Is.EqualTo(1));
+            Assert.That(Arithmetic.CeilingExact(4, 5), Is.EqualTo(1));
+            Assert.That(Arithmetic.CeilingExact(5, 5), Is.EqualTo(1));
+            Assert.That(Arithmetic.CeilingExact(6, 5), Is.EqualTo(2));
+
         }
 
         [Test]

MLAPI/NetworkingManagerComponents/Binary/Arithmetic.cs

@@ -10,14 +10,15 @@ public static class Arithmetic
         public const sbyte SIGN_BIT_8 = -128;
 
         // Ceiling function that doesn't deal with floating point values
-        public static ulong CeilingExact(ulong u1, ulong u2) => (u1 / u2) + (u1 % u2 == 0 ? 0UL : 1UL);
-        public static long CeilingExact(long u1, long u2) => (u1 / u2) + (u1 % u2 == 0 ? 0L : 1L | (((u1 & SIGN_BIT_64) ^ (u2 & SIGN_BIT_64))>>62));
-        public static uint CeilingExact(uint u1, uint u2) => (u1 / u2) + (u1 % u2 == 0 ? 0U : 1U);
-        public static int CeilingExact(int u1, int u2) => (u1 / u2) + (u1 % u2 == 0 ? 0 : 1 | (((u1 & SIGN_BIT_32) ^ (u2 & SIGN_BIT_32)) >> 30));
-        public static ushort CeilingExact(ushort u1, ushort u2) => (ushort)((u1 / u2) + (u1 % u2 == 0 ? 0 : 1));
-        public static short CeilingExact(short u1, short u2) => (short)((u1 / u2) + (u1 % u2 == 0 ? 0 : 1 | (((u1 & SIGN_BIT_32) ^ (u2 & SIGN_BIT_32)) >> 30)));
-        public static byte CeilingExact(byte u1, byte u2) => (byte)((u1 / u2) + (u1 % u2 == 0 ? 0 : 1));
-        public static sbyte CeilingExact(sbyte u1, sbyte u2) => (sbyte)((u1 / u2) + (u1 % u2 == 0 ? 0 : 1 | (((u1 & SIGN_BIT_32) ^ (u2 & SIGN_BIT_32)) >> 30)));
+        // these only work correctly with possitive numbers
+        public static ulong CeilingExact(ulong u1, ulong u2) => (u1 + u2 - 1) / u2;
+        public static long CeilingExact(long u1, long u2) => (u1 + u2 - 1) / u2;
+        public static uint CeilingExact(uint u1, uint u2) => (u1 + u2 - 1) / u2;
+        public static int CeilingExact(int u1, int u2) => (u1 + u2 - 1) / u2;
+        public static ushort CeilingExact(ushort u1, ushort u2) => (ushort)((u1 + u2 - 1) / u2);
+        public static short CeilingExact(short u1, short u2) => (short)((u1 + u2 - 1) / u2);
+        public static byte CeilingExact(byte u1, byte u2) => (byte)((u1 + u2 - 1) / u2);
+        public static sbyte CeilingExact(sbyte u1, sbyte u2) => (sbyte)((u1 + u2 - 1) / u2);
 
         /// <summary>
         /// ZigZag encodes a signed integer and maps it to a unsigned integer