Issue number: 36

Submitted by:  kcr
Fixed Issue 36 -- Vector3d hashCode() method distinguishes between -0.0 and 0.0


git-svn-id: https://svn.java.net/svn/vecmath~svn/trunk@20 dd45e54d-f42e-c781-df72-dca083a658b1

1 files changed, 84 insertions, 0 deletions

diff --git a/src/javax/vecmath/VecMathUtil.java b/src/javax/vecmath/VecMathUtil.java
new file mode 100644
index 0000000..cc34d90
--- /dev/null
+++ b/src/javax/vecmath/VecMathUtil.java
@@ -0,0 +1,84 @@
+/*
+ * $RCSfile$
+ *
+ * Copyright (c) 2004 Sun Microsystems, Inc. All rights reserved.
+ *
+ * Use is subject to license terms.
+ *
+ * $Revision$
+ * $Date$
+ * $State$
+ */
+
+package javax.vecmath;
+
+/**
+ * Utility vecmath class used when computing the hash code for vecmath
+ * objects containing float or double values. This fixes Issue 36.
+ */
+class VecMathUtil {
+    /**
+     * Returns the representation of the specified floating-point
+     * value according to the IEEE 754 floating-point "single format"
+     * bit layout, after first mapping -0.0 to 0.0. This method is
+     * identical to Float.floatToIntBits(float) except that an integer
+     * value of 0 is returned for a floating-point value of
+     * -0.0f. This is done for the purpose of computing a hash code
+     * that satisfies the contract of hashCode() and equals(). The
+     * equals() method in each vecmath class does a pair-wise "=="
+     * test on each floating-point field in the class (e.g., x, y, and
+     * z for a Tuple3f). Since 0.0f == -0.0f returns true,
+     * we must also return the same hash code for two objects, one of
+     * which has a field with a value of -0.0f and the other of which
+     * has a cooresponding field with a value of 0.0f.
+     *
+     * @param f an input floating-point number
+     * @return the integer bits representing that floating-point
+     * number, after first mapping -0.0f to 0.0f
+     */
+    static int floatToIntBits(float f) {
+	// Check for +0 or -0
+	if (f == 0.0f) {
+	    return 0;
+	}
+	else {
+	    return Float.floatToIntBits(f);
+	}
+    }
+
+    /**
+     * Returns the representation of the specified floating-point
+     * value according to the IEEE 754 floating-point "double format"
+     * bit layout, after first mapping -0.0 to 0.0. This method is
+     * identical to Double.doubleToLongBits(double) except that an
+     * integer value of 0L is returned for a floating-point value of
+     * -0.0. This is done for the purpose of computing a hash code
+     * that satisfies the contract of hashCode() and equals(). The
+     * equals() method in each vecmath class does a pair-wise "=="
+     * test on each floating-point field in the class (e.g., x, y, and
+     * z for a Tuple3d). Since 0.0 == -0.0 returns true, we
+     * must also return the same hash code for two objects, one of
+     * which has a field with a value of -0.0 and the other of which
+     * has a cooresponding field with a value of 0.0.
+     *
+     * @param d an input double precision floating-point number
+     * @return the integer bits representing that floating-point
+     * number, after first mapping -0.0f to 0.0f
+     */
+    static long doubleToLongBits(double d) {
+	// Check for +0 or -0
+	if (d == 0.0) {
+	    return 0L;
+	}
+	else {
+	    return Double.doubleToLongBits(d);
+	}
+    }
+
+
+    /**
+     * Do not construct an instance of this class.
+     */
+    private VecMathUtil() {
+    }
+}