1 /* 2 * Copyright (c) 1994, 2021, Oracle and/or its affiliates. All rights reserved. 3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. 4 * 5 * This code is free software; you can redistribute it and/or modify it 6 * under the terms of the GNU General Public License version 2 only, as 7 * published by the Free Software Foundation. Oracle designates this 8 * particular file as subject to the "Classpath" exception as provided 9 * by Oracle in the LICENSE file that accompanied this code. 10 * 11 * This code is distributed in the hope that it will be useful, but WITHOUT 12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 14 * version 2 for more details (a copy is included in the LICENSE file that 15 * accompanied this code). 16 * 17 * You should have received a copy of the GNU General Public License version 18 * 2 along with this work; if not, write to the Free Software Foundation, 19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 20 * 21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 22 * or visit www.oracle.com if you need additional information or have any 23 * questions. 24 */ 25 26 package java.lang; 27 28 // BEGIN Android-removed: dynamic constants not supported on Android. 29 /* 30 import java.lang.invoke.MethodHandles; 31 import java.lang.constant.Constable; 32 import java.lang.constant.ConstantDesc; 33 import java.util.Optional; 34 */ 35 // END Android-removed: dynamic constants not supported on Android. 36 37 import jdk.internal.math.FloatingDecimal; 38 import jdk.internal.math.DoubleConsts; 39 import jdk.internal.vm.annotation.IntrinsicCandidate; 40 41 /** 42 * The {@code Double} class wraps a value of the primitive type 43 * {@code double} in an object. An object of type 44 * {@code Double} contains a single field whose type is 45 * {@code double}. 46 * 47 * <p>In addition, this class provides several methods for converting a 48 * {@code double} to a {@code String} and a 49 * {@code String} to a {@code double}, as well as other 50 * constants and methods useful when dealing with a 51 * {@code double}. 52 * 53 * <!-- Android-removed: paragraph on ValueBased 54 * <p>This is a <a href="{@docRoot}/java.base/java/lang/doc-files/ValueBased.html">value-based</a> 55 * class; programmers should treat instances that are 56 * {@linkplain #equals(Object) equal} as interchangeable and should not 57 * use instances for synchronization, or unpredictable behavior may 58 * occur. For example, in a future release, synchronization may fail. 59 * --> 60 * 61 * <h2><a id=equivalenceRelation>Floating-point Equality, Equivalence, 62 * and Comparison</a></h2> 63 * 64 * IEEE 754 floating-point values include finite nonzero values, 65 * signed zeros ({@code +0.0} and {@code -0.0}), signed infinities 66 * {@linkplain Double#POSITIVE_INFINITY positive infinity} and 67 * {@linkplain Double#NEGATIVE_INFINITY negative infinity}), and 68 * {@linkplain Double#NaN NaN} (not-a-number). 69 * 70 * <p>An <em>equivalence relation</em> on a set of values is a boolean 71 * relation on pairs of values that is reflexive, symmetric, and 72 * transitive. For more discussion of equivalence relations and object 73 * equality, see the {@link Object#equals Object.equals} 74 * specification. An equivalence relation partitions the values it 75 * operates over into sets called <i>equivalence classes</i>. All the 76 * members of the equivalence class are equal to each other under the 77 * relation. An equivalence class may contain only a single member. At 78 * least for some purposes, all the members of an equivalence class 79 * are substitutable for each other. In particular, in a numeric 80 * expression equivalent values can be <em>substituted</em> for one 81 * another without changing the result of the expression, meaning 82 * changing the equivalence class of the result of the expression. 83 * 84 * <p>Notably, the built-in {@code ==} operation on floating-point 85 * values is <em>not</em> an equivalence relation. Despite not 86 * defining an equivalence relation, the semantics of the IEEE 754 87 * {@code ==} operator were deliberately designed to meet other needs 88 * of numerical computation. There are two exceptions where the 89 * properties of an equivalence relation are not satisfied by {@code 90 * ==} on floating-point values: 91 * 92 * <ul> 93 * 94 * <li>If {@code v1} and {@code v2} are both NaN, then {@code v1 95 * == v2} has the value {@code false}. Therefore, for two NaN 96 * arguments the <em>reflexive</em> property of an equivalence 97 * relation is <em>not</em> satisfied by the {@code ==} operator. 98 * 99 * <li>If {@code v1} represents {@code +0.0} while {@code v2} 100 * represents {@code -0.0}, or vice versa, then {@code v1 == v2} has 101 * the value {@code true} even though {@code +0.0} and {@code -0.0} 102 * are distinguishable under various floating-point operations. For 103 * example, {@code 1.0/+0.0} evaluates to positive infinity while 104 * {@code 1.0/-0.0} evaluates to <em>negative</em> infinity and 105 * positive infinity and negative infinity are neither equal to each 106 * other nor equivalent to each other. Thus, while a signed zero input 107 * most commonly determines the sign of a zero result, because of 108 * dividing by zero, {@code +0.0} and {@code -0.0} may not be 109 * substituted for each other in general. The sign of a zero input 110 * also has a non-substitutable effect on the result of some math 111 * library methods. 112 * 113 * </ul> 114 * 115 * <p>For ordered comparisons using the built-in comparison operators 116 * ({@code <}, {@code <=}, etc.), NaN values have another anomalous 117 * situation: a NaN is neither less than, nor greater than, nor equal 118 * to any value, including itself. This means the <i>trichotomy of 119 * comparison</i> does <em>not</em> hold. 120 * 121 * <p>To provide the appropriate semantics for {@code equals} and 122 * {@code compareTo} methods, those methods cannot simply be wrappers 123 * around {@code ==} or ordered comparison operations. Instead, {@link 124 * Double#equals equals} defines NaN arguments to be equal to each 125 * other and defines {@code +0.0} to <em>not</em> be equal to {@code 126 * -0.0}, restoring reflexivity. For comparisons, {@link 127 * Double#compareTo compareTo} defines a total order where {@code 128 * -0.0} is less than {@code +0.0} and where a NaN is equal to itself 129 * and considered greater than positive infinity. 130 * 131 * <p>The operational semantics of {@code equals} and {@code 132 * compareTo} are expressed in terms of {@linkplain #doubleToLongBits 133 * bit-wise converting} the floating-point values to integral values. 134 * 135 * <p>The <em>natural ordering</em> implemented by {@link #compareTo 136 * compareTo} is {@linkplain Comparable consistent with equals}. That 137 * is, two objects are reported as equal by {@code equals} if and only 138 * if {@code compareTo} on those objects returns zero. 139 * 140 * <p>The adjusted behaviors defined for {@code equals} and {@code 141 * compareTo} allow instances of wrapper classes to work properly with 142 * conventional data structures. For example, defining NaN 143 * values to be {@code equals} to one another allows NaN to be used as 144 * an element of a {@link java.util.HashSet HashSet} or as the key of 145 * a {@link java.util.HashMap HashMap}. Similarly, defining {@code 146 * compareTo} as a total ordering, including {@code +0.0}, {@code 147 * -0.0}, and NaN, allows instances of wrapper classes to be used as 148 * elements of a {@link java.util.SortedSet SortedSet} or as keys of a 149 * {@link java.util.SortedMap SortedMap}. 150 * 151 * @jls 4.2.3 Floating-Point Types, Formats, and Values 152 * @jls 4.2.4. Floating-Point Operations 153 * @jls 15.21.1 Numerical Equality Operators == and != 154 * @jls 15.20.1 Numerical Comparison Operators {@code <}, {@code <=}, {@code >}, and {@code >=} 155 * 156 * @author Lee Boynton 157 * @author Arthur van Hoff 158 * @author Joseph D. Darcy 159 * @since 1.0 160 */ 161 @jdk.internal.ValueBased 162 public final class Double extends Number 163 implements Comparable<Double> 164 // Android-removed: no Constable support. 165 // , Constable, ConstantDesc 166 { 167 /** 168 * A constant holding the positive infinity of type 169 * {@code double}. It is equal to the value returned by 170 * {@code Double.longBitsToDouble(0x7ff0000000000000L)}. 171 */ 172 public static final double POSITIVE_INFINITY = 1.0 / 0.0; 173 174 /** 175 * A constant holding the negative infinity of type 176 * {@code double}. It is equal to the value returned by 177 * {@code Double.longBitsToDouble(0xfff0000000000000L)}. 178 */ 179 public static final double NEGATIVE_INFINITY = -1.0 / 0.0; 180 181 /** 182 * A constant holding a Not-a-Number (NaN) value of type 183 * {@code double}. It is equivalent to the value returned by 184 * {@code Double.longBitsToDouble(0x7ff8000000000000L)}. 185 */ 186 public static final double NaN = 0.0d / 0.0; 187 188 /** 189 * A constant holding the largest positive finite value of type 190 * {@code double}, 191 * (2-2<sup>-52</sup>)·2<sup>1023</sup>. It is equal to 192 * the hexadecimal floating-point literal 193 * {@code 0x1.fffffffffffffP+1023} and also equal to 194 * {@code Double.longBitsToDouble(0x7fefffffffffffffL)}. 195 */ 196 public static final double MAX_VALUE = 0x1.fffffffffffffP+1023; // 1.7976931348623157e+308 197 198 /** 199 * A constant holding the smallest positive normal value of type 200 * {@code double}, 2<sup>-1022</sup>. It is equal to the 201 * hexadecimal floating-point literal {@code 0x1.0p-1022} and also 202 * equal to {@code Double.longBitsToDouble(0x0010000000000000L)}. 203 * 204 * @since 1.6 205 */ 206 public static final double MIN_NORMAL = 0x1.0p-1022; // 2.2250738585072014E-308 207 208 /** 209 * A constant holding the smallest positive nonzero value of type 210 * {@code double}, 2<sup>-1074</sup>. It is equal to the 211 * hexadecimal floating-point literal 212 * {@code 0x0.0000000000001P-1022} and also equal to 213 * {@code Double.longBitsToDouble(0x1L)}. 214 */ 215 public static final double MIN_VALUE = 0x0.0000000000001P-1022; // 4.9e-324 216 217 /** 218 * Maximum exponent a finite {@code double} variable may have. 219 * It is equal to the value returned by 220 * {@code Math.getExponent(Double.MAX_VALUE)}. 221 * 222 * @since 1.6 223 */ 224 public static final int MAX_EXPONENT = 1023; 225 226 /** 227 * Minimum exponent a normalized {@code double} variable may 228 * have. It is equal to the value returned by 229 * {@code Math.getExponent(Double.MIN_NORMAL)}. 230 * 231 * @since 1.6 232 */ 233 public static final int MIN_EXPONENT = -1022; 234 235 /** 236 * The number of bits used to represent a {@code double} value. 237 * 238 * @since 1.5 239 */ 240 public static final int SIZE = 64; 241 242 /** 243 * The number of bytes used to represent a {@code double} value. 244 * 245 * @since 1.8 246 */ 247 public static final int BYTES = SIZE / Byte.SIZE; 248 249 /** 250 * The {@code Class} instance representing the primitive type 251 * {@code double}. 252 * 253 * @since 1.1 254 */ 255 @SuppressWarnings("unchecked") 256 public static final Class<Double> TYPE = (Class<Double>) Class.getPrimitiveClass("double"); 257 258 /** 259 * Returns a string representation of the {@code double} 260 * argument. All characters mentioned below are ASCII characters. 261 * <ul> 262 * <li>If the argument is NaN, the result is the string 263 * "{@code NaN}". 264 * <li>Otherwise, the result is a string that represents the sign and 265 * magnitude (absolute value) of the argument. If the sign is negative, 266 * the first character of the result is '{@code -}' 267 * ({@code '\u005Cu002D'}); if the sign is positive, no sign character 268 * appears in the result. As for the magnitude <i>m</i>: 269 * <ul> 270 * <li>If <i>m</i> is infinity, it is represented by the characters 271 * {@code "Infinity"}; thus, positive infinity produces the result 272 * {@code "Infinity"} and negative infinity produces the result 273 * {@code "-Infinity"}. 274 * 275 * <li>If <i>m</i> is zero, it is represented by the characters 276 * {@code "0.0"}; thus, negative zero produces the result 277 * {@code "-0.0"} and positive zero produces the result 278 * {@code "0.0"}. 279 * 280 * <li>If <i>m</i> is greater than or equal to 10<sup>-3</sup> but less 281 * than 10<sup>7</sup>, then it is represented as the integer part of 282 * <i>m</i>, in decimal form with no leading zeroes, followed by 283 * '{@code .}' ({@code '\u005Cu002E'}), followed by one or 284 * more decimal digits representing the fractional part of <i>m</i>. 285 * 286 * <li>If <i>m</i> is less than 10<sup>-3</sup> or greater than or 287 * equal to 10<sup>7</sup>, then it is represented in so-called 288 * "computerized scientific notation." Let <i>n</i> be the unique 289 * integer such that 10<sup><i>n</i></sup> ≤ <i>m</i> {@literal <} 290 * 10<sup><i>n</i>+1</sup>; then let <i>a</i> be the 291 * mathematically exact quotient of <i>m</i> and 292 * 10<sup><i>n</i></sup> so that 1 ≤ <i>a</i> {@literal <} 10. The 293 * magnitude is then represented as the integer part of <i>a</i>, 294 * as a single decimal digit, followed by '{@code .}' 295 * ({@code '\u005Cu002E'}), followed by decimal digits 296 * representing the fractional part of <i>a</i>, followed by the 297 * letter '{@code E}' ({@code '\u005Cu0045'}), followed 298 * by a representation of <i>n</i> as a decimal integer, as 299 * produced by the method {@link Integer#toString(int)}. 300 * </ul> 301 * </ul> 302 * How many digits must be printed for the fractional part of 303 * <i>m</i> or <i>a</i>? There must be at least one digit to represent 304 * the fractional part, and beyond that as many, but only as many, more 305 * digits as are needed to uniquely distinguish the argument value from 306 * adjacent values of type {@code double}. That is, suppose that 307 * <i>x</i> is the exact mathematical value represented by the decimal 308 * representation produced by this method for a finite nonzero argument 309 * <i>d</i>. Then <i>d</i> must be the {@code double} value nearest 310 * to <i>x</i>; or if two {@code double} values are equally close 311 * to <i>x</i>, then <i>d</i> must be one of them and the least 312 * significant bit of the significand of <i>d</i> must be {@code 0}. 313 * 314 * <p>To create localized string representations of a floating-point 315 * value, use subclasses of {@link java.text.NumberFormat}. 316 * 317 * @param d the {@code double} to be converted. 318 * @return a string representation of the argument. 319 */ toString(double d)320 public static String toString(double d) { 321 return FloatingDecimal.toJavaFormatString(d); 322 } 323 324 /** 325 * Returns a hexadecimal string representation of the 326 * {@code double} argument. All characters mentioned below 327 * are ASCII characters. 328 * 329 * <ul> 330 * <li>If the argument is NaN, the result is the string 331 * "{@code NaN}". 332 * <li>Otherwise, the result is a string that represents the sign 333 * and magnitude of the argument. If the sign is negative, the 334 * first character of the result is '{@code -}' 335 * ({@code '\u005Cu002D'}); if the sign is positive, no sign 336 * character appears in the result. As for the magnitude <i>m</i>: 337 * 338 * <ul> 339 * <li>If <i>m</i> is infinity, it is represented by the string 340 * {@code "Infinity"}; thus, positive infinity produces the 341 * result {@code "Infinity"} and negative infinity produces 342 * the result {@code "-Infinity"}. 343 * 344 * <li>If <i>m</i> is zero, it is represented by the string 345 * {@code "0x0.0p0"}; thus, negative zero produces the result 346 * {@code "-0x0.0p0"} and positive zero produces the result 347 * {@code "0x0.0p0"}. 348 * 349 * <li>If <i>m</i> is a {@code double} value with a 350 * normalized representation, substrings are used to represent the 351 * significand and exponent fields. The significand is 352 * represented by the characters {@code "0x1."} 353 * followed by a lowercase hexadecimal representation of the rest 354 * of the significand as a fraction. Trailing zeros in the 355 * hexadecimal representation are removed unless all the digits 356 * are zero, in which case a single zero is used. Next, the 357 * exponent is represented by {@code "p"} followed 358 * by a decimal string of the unbiased exponent as if produced by 359 * a call to {@link Integer#toString(int) Integer.toString} on the 360 * exponent value. 361 * 362 * <li>If <i>m</i> is a {@code double} value with a subnormal 363 * representation, the significand is represented by the 364 * characters {@code "0x0."} followed by a 365 * hexadecimal representation of the rest of the significand as a 366 * fraction. Trailing zeros in the hexadecimal representation are 367 * removed. Next, the exponent is represented by 368 * {@code "p-1022"}. Note that there must be at 369 * least one nonzero digit in a subnormal significand. 370 * 371 * </ul> 372 * 373 * </ul> 374 * 375 * <table class="striped"> 376 * <caption>Examples</caption> 377 * <thead> 378 * <tr><th scope="col">Floating-point Value</th><th scope="col">Hexadecimal String</th> 379 * </thead> 380 * <tbody style="text-align:right"> 381 * <tr><th scope="row">{@code 1.0}</th> <td>{@code 0x1.0p0}</td> 382 * <tr><th scope="row">{@code -1.0}</th> <td>{@code -0x1.0p0}</td> 383 * <tr><th scope="row">{@code 2.0}</th> <td>{@code 0x1.0p1}</td> 384 * <tr><th scope="row">{@code 3.0}</th> <td>{@code 0x1.8p1}</td> 385 * <tr><th scope="row">{@code 0.5}</th> <td>{@code 0x1.0p-1}</td> 386 * <tr><th scope="row">{@code 0.25}</th> <td>{@code 0x1.0p-2}</td> 387 * <tr><th scope="row">{@code Double.MAX_VALUE}</th> 388 * <td>{@code 0x1.fffffffffffffp1023}</td> 389 * <tr><th scope="row">{@code Minimum Normal Value}</th> 390 * <td>{@code 0x1.0p-1022}</td> 391 * <tr><th scope="row">{@code Maximum Subnormal Value}</th> 392 * <td>{@code 0x0.fffffffffffffp-1022}</td> 393 * <tr><th scope="row">{@code Double.MIN_VALUE}</th> 394 * <td>{@code 0x0.0000000000001p-1022}</td> 395 * </tbody> 396 * </table> 397 * @param d the {@code double} to be converted. 398 * @return a hex string representation of the argument. 399 * @since 1.5 400 * @author Joseph D. Darcy 401 */ toHexString(double d)402 public static String toHexString(double d) { 403 /* 404 * Modeled after the "a" conversion specifier in C99, section 405 * 7.19.6.1; however, the output of this method is more 406 * tightly specified. 407 */ 408 if (!isFinite(d) ) 409 // For infinity and NaN, use the decimal output. 410 return Double.toString(d); 411 else { 412 // Initialized to maximum size of output. 413 StringBuilder answer = new StringBuilder(24); 414 415 if (Math.copySign(1.0, d) == -1.0) // value is negative, 416 answer.append("-"); // so append sign info 417 418 answer.append("0x"); 419 420 d = Math.abs(d); 421 422 if(d == 0.0) { 423 answer.append("0.0p0"); 424 } else { 425 boolean subnormal = (d < Double.MIN_NORMAL); 426 427 // Isolate significand bits and OR in a high-order bit 428 // so that the string representation has a known 429 // length. 430 long signifBits = (Double.doubleToLongBits(d) 431 & DoubleConsts.SIGNIF_BIT_MASK) | 432 0x1000000000000000L; 433 434 // Subnormal values have a 0 implicit bit; normal 435 // values have a 1 implicit bit. 436 answer.append(subnormal ? "0." : "1."); 437 438 // Isolate the low-order 13 digits of the hex 439 // representation. If all the digits are zero, 440 // replace with a single 0; otherwise, remove all 441 // trailing zeros. 442 String signif = Long.toHexString(signifBits).substring(3,16); 443 answer.append(signif.equals("0000000000000") ? // 13 zeros 444 "0": 445 signif.replaceFirst("0{1,12}$", "")); 446 447 answer.append('p'); 448 // If the value is subnormal, use the E_min exponent 449 // value for double; otherwise, extract and report d's 450 // exponent (the representation of a subnormal uses 451 // E_min -1). 452 answer.append(subnormal ? 453 Double.MIN_EXPONENT: 454 Math.getExponent(d)); 455 } 456 return answer.toString(); 457 } 458 } 459 460 /** 461 * Returns a {@code Double} object holding the 462 * {@code double} value represented by the argument string 463 * {@code s}. 464 * 465 * <p>If {@code s} is {@code null}, then a 466 * {@code NullPointerException} is thrown. 467 * 468 * <p>Leading and trailing whitespace characters in {@code s} 469 * are ignored. Whitespace is removed as if by the {@link 470 * String#trim} method; that is, both ASCII space and control 471 * characters are removed. The rest of {@code s} should 472 * constitute a <i>FloatValue</i> as described by the lexical 473 * syntax rules: 474 * 475 * <blockquote> 476 * <dl> 477 * <dt><i>FloatValue:</i> 478 * <dd><i>Sign<sub>opt</sub></i> {@code NaN} 479 * <dd><i>Sign<sub>opt</sub></i> {@code Infinity} 480 * <dd><i>Sign<sub>opt</sub> FloatingPointLiteral</i> 481 * <dd><i>Sign<sub>opt</sub> HexFloatingPointLiteral</i> 482 * <dd><i>SignedInteger</i> 483 * </dl> 484 * 485 * <dl> 486 * <dt><i>HexFloatingPointLiteral</i>: 487 * <dd> <i>HexSignificand BinaryExponent FloatTypeSuffix<sub>opt</sub></i> 488 * </dl> 489 * 490 * <dl> 491 * <dt><i>HexSignificand:</i> 492 * <dd><i>HexNumeral</i> 493 * <dd><i>HexNumeral</i> {@code .} 494 * <dd>{@code 0x} <i>HexDigits<sub>opt</sub> 495 * </i>{@code .}<i> HexDigits</i> 496 * <dd>{@code 0X}<i> HexDigits<sub>opt</sub> 497 * </i>{@code .} <i>HexDigits</i> 498 * </dl> 499 * 500 * <dl> 501 * <dt><i>BinaryExponent:</i> 502 * <dd><i>BinaryExponentIndicator SignedInteger</i> 503 * </dl> 504 * 505 * <dl> 506 * <dt><i>BinaryExponentIndicator:</i> 507 * <dd>{@code p} 508 * <dd>{@code P} 509 * </dl> 510 * 511 * </blockquote> 512 * 513 * where <i>Sign</i>, <i>FloatingPointLiteral</i>, 514 * <i>HexNumeral</i>, <i>HexDigits</i>, <i>SignedInteger</i> and 515 * <i>FloatTypeSuffix</i> are as defined in the lexical structure 516 * sections of 517 * <cite>The Java Language Specification</cite>, 518 * except that underscores are not accepted between digits. 519 * If {@code s} does not have the form of 520 * a <i>FloatValue</i>, then a {@code NumberFormatException} 521 * is thrown. Otherwise, {@code s} is regarded as 522 * representing an exact decimal value in the usual 523 * "computerized scientific notation" or as an exact 524 * hexadecimal value; this exact numerical value is then 525 * conceptually converted to an "infinitely precise" 526 * binary value that is then rounded to type {@code double} 527 * by the usual round-to-nearest rule of IEEE 754 floating-point 528 * arithmetic, which includes preserving the sign of a zero 529 * value. 530 * 531 * Note that the round-to-nearest rule also implies overflow and 532 * underflow behaviour; if the exact value of {@code s} is large 533 * enough in magnitude (greater than or equal to ({@link 534 * #MAX_VALUE} + {@link Math#ulp(double) ulp(MAX_VALUE)}/2), 535 * rounding to {@code double} will result in an infinity and if the 536 * exact value of {@code s} is small enough in magnitude (less 537 * than or equal to {@link #MIN_VALUE}/2), rounding to float will 538 * result in a zero. 539 * 540 * Finally, after rounding a {@code Double} object representing 541 * this {@code double} value is returned. 542 * 543 * <p> To interpret localized string representations of a 544 * floating-point value, use subclasses of {@link 545 * java.text.NumberFormat}. 546 * 547 * <p>Note that trailing format specifiers, specifiers that 548 * determine the type of a floating-point literal 549 * ({@code 1.0f} is a {@code float} value; 550 * {@code 1.0d} is a {@code double} value), do 551 * <em>not</em> influence the results of this method. In other 552 * words, the numerical value of the input string is converted 553 * directly to the target floating-point type. The two-step 554 * sequence of conversions, string to {@code float} followed 555 * by {@code float} to {@code double}, is <em>not</em> 556 * equivalent to converting a string directly to 557 * {@code double}. For example, the {@code float} 558 * literal {@code 0.1f} is equal to the {@code double} 559 * value {@code 0.10000000149011612}; the {@code float} 560 * literal {@code 0.1f} represents a different numerical 561 * value than the {@code double} literal 562 * {@code 0.1}. (The numerical value 0.1 cannot be exactly 563 * represented in a binary floating-point number.) 564 * 565 * <p>To avoid calling this method on an invalid string and having 566 * a {@code NumberFormatException} be thrown, the regular 567 * expression below can be used to screen the input string: 568 * 569 * <pre>{@code 570 * final String Digits = "(\\p{Digit}+)"; 571 * final String HexDigits = "(\\p{XDigit}+)"; 572 * // an exponent is 'e' or 'E' followed by an optionally 573 * // signed decimal integer. 574 * final String Exp = "[eE][+-]?"+Digits; 575 * final String fpRegex = 576 * ("[\\x00-\\x20]*"+ // Optional leading "whitespace" 577 * "[+-]?(" + // Optional sign character 578 * "NaN|" + // "NaN" string 579 * "Infinity|" + // "Infinity" string 580 * 581 * // A decimal floating-point string representing a finite positive 582 * // number without a leading sign has at most five basic pieces: 583 * // Digits . Digits ExponentPart FloatTypeSuffix 584 * // 585 * // Since this method allows integer-only strings as input 586 * // in addition to strings of floating-point literals, the 587 * // two sub-patterns below are simplifications of the grammar 588 * // productions from section 3.10.2 of 589 * // The Java Language Specification. 590 * 591 * // Digits ._opt Digits_opt ExponentPart_opt FloatTypeSuffix_opt 592 * "((("+Digits+"(\\.)?("+Digits+"?)("+Exp+")?)|"+ 593 * 594 * // . Digits ExponentPart_opt FloatTypeSuffix_opt 595 * "(\\.("+Digits+")("+Exp+")?)|"+ 596 * 597 * // Hexadecimal strings 598 * "((" + 599 * // 0[xX] HexDigits ._opt BinaryExponent FloatTypeSuffix_opt 600 * "(0[xX]" + HexDigits + "(\\.)?)|" + 601 * 602 * // 0[xX] HexDigits_opt . HexDigits BinaryExponent FloatTypeSuffix_opt 603 * "(0[xX]" + HexDigits + "?(\\.)" + HexDigits + ")" + 604 * 605 * ")[pP][+-]?" + Digits + "))" + 606 * "[fFdD]?))" + 607 * "[\\x00-\\x20]*");// Optional trailing "whitespace" 608 * 609 * if (Pattern.matches(fpRegex, myString)) 610 * Double.valueOf(myString); // Will not throw NumberFormatException 611 * else { 612 * // Perform suitable alternative action 613 * } 614 * }</pre> 615 * 616 * @param s the string to be parsed. 617 * @return a {@code Double} object holding the value 618 * represented by the {@code String} argument. 619 * @throws NumberFormatException if the string does not contain a 620 * parsable number. 621 */ valueOf(String s)622 public static Double valueOf(String s) throws NumberFormatException { 623 return new Double(parseDouble(s)); 624 } 625 626 /** 627 * Returns a {@code Double} instance representing the specified 628 * {@code double} value. 629 * If a new {@code Double} instance is not required, this method 630 * should generally be used in preference to the constructor 631 * {@link #Double(double)}, as this method is likely to yield 632 * significantly better space and time performance by caching 633 * frequently requested values. 634 * 635 * @param d a double value. 636 * @return a {@code Double} instance representing {@code d}. 637 * @since 1.5 638 */ 639 @IntrinsicCandidate valueOf(double d)640 public static Double valueOf(double d) { 641 return new Double(d); 642 } 643 644 /** 645 * Returns a new {@code double} initialized to the value 646 * represented by the specified {@code String}, as performed 647 * by the {@code valueOf} method of class 648 * {@code Double}. 649 * 650 * @param s the string to be parsed. 651 * @return the {@code double} value represented by the string 652 * argument. 653 * @throws NullPointerException if the string is null 654 * @throws NumberFormatException if the string does not contain 655 * a parsable {@code double}. 656 * @see java.lang.Double#valueOf(String) 657 * @since 1.2 658 */ parseDouble(String s)659 public static double parseDouble(String s) throws NumberFormatException { 660 return FloatingDecimal.parseDouble(s); 661 } 662 663 /** 664 * Returns {@code true} if the specified number is a 665 * Not-a-Number (NaN) value, {@code false} otherwise. 666 * 667 * @param v the value to be tested. 668 * @return {@code true} if the value of the argument is NaN; 669 * {@code false} otherwise. 670 */ isNaN(double v)671 public static boolean isNaN(double v) { 672 return (v != v); 673 } 674 675 /** 676 * Returns {@code true} if the specified number is infinitely 677 * large in magnitude, {@code false} otherwise. 678 * 679 * @param v the value to be tested. 680 * @return {@code true} if the value of the argument is positive 681 * infinity or negative infinity; {@code false} otherwise. 682 */ isInfinite(double v)683 public static boolean isInfinite(double v) { 684 return (v == POSITIVE_INFINITY) || (v == NEGATIVE_INFINITY); 685 } 686 687 /** 688 * Returns {@code true} if the argument is a finite floating-point 689 * value; returns {@code false} otherwise (for NaN and infinity 690 * arguments). 691 * 692 * @param d the {@code double} value to be tested 693 * @return {@code true} if the argument is a finite 694 * floating-point value, {@code false} otherwise. 695 * @since 1.8 696 */ isFinite(double d)697 public static boolean isFinite(double d) { 698 return Math.abs(d) <= Double.MAX_VALUE; 699 } 700 701 /** 702 * The value of the Double. 703 * 704 * @serial 705 */ 706 private final double value; 707 708 /** 709 * Constructs a newly allocated {@code Double} object that 710 * represents the primitive {@code double} argument. 711 * 712 * @param value the value to be represented by the {@code Double}. 713 * 714 * @deprecated 715 * It is rarely appropriate to use this constructor. The static factory 716 * {@link #valueOf(double)} is generally a better choice, as it is 717 * likely to yield significantly better space and time performance. 718 */ 719 // Android-changed: not yet forRemoval on Android. 720 @Deprecated(since="9"/*, forRemoval = true*/) Double(double value)721 public Double(double value) { 722 this.value = value; 723 } 724 725 /** 726 * Constructs a newly allocated {@code Double} object that 727 * represents the floating-point value of type {@code double} 728 * represented by the string. The string is converted to a 729 * {@code double} value as if by the {@code valueOf} method. 730 * 731 * @param s a string to be converted to a {@code Double}. 732 * @throws NumberFormatException if the string does not contain a 733 * parsable number. 734 * 735 * @deprecated 736 * It is rarely appropriate to use this constructor. 737 * Use {@link #parseDouble(String)} to convert a string to a 738 * {@code double} primitive, or use {@link #valueOf(String)} 739 * to convert a string to a {@code Double} object. 740 */ 741 // Android-changed: not yet forRemoval on Android. 742 @Deprecated(since="9"/*, forRemoval = true */) Double(String s)743 public Double(String s) throws NumberFormatException { 744 value = parseDouble(s); 745 } 746 747 /** 748 * Returns {@code true} if this {@code Double} value is 749 * a Not-a-Number (NaN), {@code false} otherwise. 750 * 751 * @return {@code true} if the value represented by this object is 752 * NaN; {@code false} otherwise. 753 */ isNaN()754 public boolean isNaN() { 755 return isNaN(value); 756 } 757 758 /** 759 * Returns {@code true} if this {@code Double} value is 760 * infinitely large in magnitude, {@code false} otherwise. 761 * 762 * @return {@code true} if the value represented by this object is 763 * positive infinity or negative infinity; 764 * {@code false} otherwise. 765 */ isInfinite()766 public boolean isInfinite() { 767 return isInfinite(value); 768 } 769 770 /** 771 * Returns a string representation of this {@code Double} object. 772 * The primitive {@code double} value represented by this 773 * object is converted to a string exactly as if by the method 774 * {@code toString} of one argument. 775 * 776 * @return a {@code String} representation of this object. 777 * @see java.lang.Double#toString(double) 778 */ toString()779 public String toString() { 780 return toString(value); 781 } 782 783 /** 784 * Returns the value of this {@code Double} as a {@code byte} 785 * after a narrowing primitive conversion. 786 * 787 * @return the {@code double} value represented by this object 788 * converted to type {@code byte} 789 * @jls 5.1.3 Narrowing Primitive Conversion 790 * @since 1.1 791 */ byteValue()792 public byte byteValue() { 793 return (byte)value; 794 } 795 796 /** 797 * Returns the value of this {@code Double} as a {@code short} 798 * after a narrowing primitive conversion. 799 * 800 * @return the {@code double} value represented by this object 801 * converted to type {@code short} 802 * @jls 5.1.3 Narrowing Primitive Conversion 803 * @since 1.1 804 */ shortValue()805 public short shortValue() { 806 return (short)value; 807 } 808 809 /** 810 * Returns the value of this {@code Double} as an {@code int} 811 * after a narrowing primitive conversion. 812 * @jls 5.1.3 Narrowing Primitive Conversion 813 * 814 * @return the {@code double} value represented by this object 815 * converted to type {@code int} 816 */ intValue()817 public int intValue() { 818 return (int)value; 819 } 820 821 /** 822 * Returns the value of this {@code Double} as a {@code long} 823 * after a narrowing primitive conversion. 824 * 825 * @return the {@code double} value represented by this object 826 * converted to type {@code long} 827 * @jls 5.1.3 Narrowing Primitive Conversion 828 */ longValue()829 public long longValue() { 830 return (long)value; 831 } 832 833 /** 834 * Returns the value of this {@code Double} as a {@code float} 835 * after a narrowing primitive conversion. 836 * 837 * @return the {@code double} value represented by this object 838 * converted to type {@code float} 839 * @jls 5.1.3 Narrowing Primitive Conversion 840 * @since 1.0 841 */ floatValue()842 public float floatValue() { 843 return (float)value; 844 } 845 846 /** 847 * Returns the {@code double} value of this {@code Double} object. 848 * 849 * @return the {@code double} value represented by this object 850 */ 851 @IntrinsicCandidate doubleValue()852 public double doubleValue() { 853 return value; 854 } 855 856 /** 857 * Returns a hash code for this {@code Double} object. The 858 * result is the exclusive OR of the two halves of the 859 * {@code long} integer bit representation, exactly as 860 * produced by the method {@link #doubleToLongBits(double)}, of 861 * the primitive {@code double} value represented by this 862 * {@code Double} object. That is, the hash code is the value 863 * of the expression: 864 * 865 * <blockquote> 866 * {@code (int)(v^(v>>>32))} 867 * </blockquote> 868 * 869 * where {@code v} is defined by: 870 * 871 * <blockquote> 872 * {@code long v = Double.doubleToLongBits(this.doubleValue());} 873 * </blockquote> 874 * 875 * @return a {@code hash code} value for this object. 876 */ 877 @Override hashCode()878 public int hashCode() { 879 return Double.hashCode(value); 880 } 881 882 /** 883 * Returns a hash code for a {@code double} value; compatible with 884 * {@code Double.hashCode()}. 885 * 886 * @param value the value to hash 887 * @return a hash code value for a {@code double} value. 888 * @since 1.8 889 */ hashCode(double value)890 public static int hashCode(double value) { 891 long bits = doubleToLongBits(value); 892 return (int)(bits ^ (bits >>> 32)); 893 } 894 895 /** 896 * Compares this object against the specified object. The result 897 * is {@code true} if and only if the argument is not 898 * {@code null} and is a {@code Double} object that 899 * represents a {@code double} that has the same value as the 900 * {@code double} represented by this object. For this 901 * purpose, two {@code double} values are considered to be 902 * the same if and only if the method {@link 903 * #doubleToLongBits(double)} returns the identical 904 * {@code long} value when applied to each. 905 * 906 * @apiNote 907 * This method is defined in terms of {@link 908 * #doubleToLongBits(double)} rather than the {@code ==} operator 909 * on {@code double} values since the {@code ==} operator does 910 * <em>not</em> define an equivalence relation and to satisfy the 911 * {@linkplain Object#equals equals contract} an equivalence 912 * relation must be implemented; see <a 913 * href="#equivalenceRelation">this discussion</a> for details of 914 * floating-point equality and equivalence. 915 * 916 * @see java.lang.Double#doubleToLongBits(double) 917 * @jls 15.21.1 Numerical Equality Operators == and != 918 */ equals(Object obj)919 public boolean equals(Object obj) { 920 return (obj instanceof Double) 921 && (doubleToLongBits(((Double)obj).value) == 922 doubleToLongBits(value)); 923 } 924 925 /** 926 * Returns a representation of the specified floating-point value 927 * according to the IEEE 754 floating-point "double 928 * format" bit layout. 929 * 930 * <p>Bit 63 (the bit that is selected by the mask 931 * {@code 0x8000000000000000L}) represents the sign of the 932 * floating-point number. Bits 933 * 62-52 (the bits that are selected by the mask 934 * {@code 0x7ff0000000000000L}) represent the exponent. Bits 51-0 935 * (the bits that are selected by the mask 936 * {@code 0x000fffffffffffffL}) represent the significand 937 * (sometimes called the mantissa) of the floating-point number. 938 * 939 * <p>If the argument is positive infinity, the result is 940 * {@code 0x7ff0000000000000L}. 941 * 942 * <p>If the argument is negative infinity, the result is 943 * {@code 0xfff0000000000000L}. 944 * 945 * <p>If the argument is NaN, the result is 946 * {@code 0x7ff8000000000000L}. 947 * 948 * <p>In all cases, the result is a {@code long} integer that, when 949 * given to the {@link #longBitsToDouble(long)} method, will produce a 950 * floating-point value the same as the argument to 951 * {@code doubleToLongBits} (except all NaN values are 952 * collapsed to a single "canonical" NaN value). 953 * 954 * @param value a {@code double} precision floating-point number. 955 * @return the bits that represent the floating-point number. 956 */ 957 @IntrinsicCandidate doubleToLongBits(double value)958 public static long doubleToLongBits(double value) { 959 if (!isNaN(value)) { 960 return doubleToRawLongBits(value); 961 } 962 return 0x7ff8000000000000L; 963 } 964 965 /** 966 * Returns a representation of the specified floating-point value 967 * according to the IEEE 754 floating-point "double 968 * format" bit layout, preserving Not-a-Number (NaN) values. 969 * 970 * <p>Bit 63 (the bit that is selected by the mask 971 * {@code 0x8000000000000000L}) represents the sign of the 972 * floating-point number. Bits 973 * 62-52 (the bits that are selected by the mask 974 * {@code 0x7ff0000000000000L}) represent the exponent. Bits 51-0 975 * (the bits that are selected by the mask 976 * {@code 0x000fffffffffffffL}) represent the significand 977 * (sometimes called the mantissa) of the floating-point number. 978 * 979 * <p>If the argument is positive infinity, the result is 980 * {@code 0x7ff0000000000000L}. 981 * 982 * <p>If the argument is negative infinity, the result is 983 * {@code 0xfff0000000000000L}. 984 * 985 * <p>If the argument is NaN, the result is the {@code long} 986 * integer representing the actual NaN value. Unlike the 987 * {@code doubleToLongBits} method, 988 * {@code doubleToRawLongBits} does not collapse all the bit 989 * patterns encoding a NaN to a single "canonical" NaN 990 * value. 991 * 992 * <p>In all cases, the result is a {@code long} integer that, 993 * when given to the {@link #longBitsToDouble(long)} method, will 994 * produce a floating-point value the same as the argument to 995 * {@code doubleToRawLongBits}. 996 * 997 * @param value a {@code double} precision floating-point number. 998 * @return the bits that represent the floating-point number. 999 * @since 1.3 1000 */ 1001 @IntrinsicCandidate doubleToRawLongBits(double value)1002 public static native long doubleToRawLongBits(double value); 1003 1004 /** 1005 * Returns the {@code double} value corresponding to a given 1006 * bit representation. 1007 * The argument is considered to be a representation of a 1008 * floating-point value according to the IEEE 754 floating-point 1009 * "double format" bit layout. 1010 * 1011 * <p>If the argument is {@code 0x7ff0000000000000L}, the result 1012 * is positive infinity. 1013 * 1014 * <p>If the argument is {@code 0xfff0000000000000L}, the result 1015 * is negative infinity. 1016 * 1017 * <p>If the argument is any value in the range 1018 * {@code 0x7ff0000000000001L} through 1019 * {@code 0x7fffffffffffffffL} or in the range 1020 * {@code 0xfff0000000000001L} through 1021 * {@code 0xffffffffffffffffL}, the result is a NaN. No IEEE 1022 * 754 floating-point operation provided by Java can distinguish 1023 * between two NaN values of the same type with different bit 1024 * patterns. Distinct values of NaN are only distinguishable by 1025 * use of the {@code Double.doubleToRawLongBits} method. 1026 * 1027 * <p>In all other cases, let <i>s</i>, <i>e</i>, and <i>m</i> be three 1028 * values that can be computed from the argument: 1029 * 1030 * <blockquote><pre>{@code 1031 * int s = ((bits >> 63) == 0) ? 1 : -1; 1032 * int e = (int)((bits >> 52) & 0x7ffL); 1033 * long m = (e == 0) ? 1034 * (bits & 0xfffffffffffffL) << 1 : 1035 * (bits & 0xfffffffffffffL) | 0x10000000000000L; 1036 * }</pre></blockquote> 1037 * 1038 * Then the floating-point result equals the value of the mathematical 1039 * expression <i>s</i>·<i>m</i>·2<sup><i>e</i>-1075</sup>. 1040 * 1041 * <p>Note that this method may not be able to return a 1042 * {@code double} NaN with exactly same bit pattern as the 1043 * {@code long} argument. IEEE 754 distinguishes between two 1044 * kinds of NaNs, quiet NaNs and <i>signaling NaNs</i>. The 1045 * differences between the two kinds of NaN are generally not 1046 * visible in Java. Arithmetic operations on signaling NaNs turn 1047 * them into quiet NaNs with a different, but often similar, bit 1048 * pattern. However, on some processors merely copying a 1049 * signaling NaN also performs that conversion. In particular, 1050 * copying a signaling NaN to return it to the calling method 1051 * may perform this conversion. So {@code longBitsToDouble} 1052 * may not be able to return a {@code double} with a 1053 * signaling NaN bit pattern. Consequently, for some 1054 * {@code long} values, 1055 * {@code doubleToRawLongBits(longBitsToDouble(start))} may 1056 * <i>not</i> equal {@code start}. Moreover, which 1057 * particular bit patterns represent signaling NaNs is platform 1058 * dependent; although all NaN bit patterns, quiet or signaling, 1059 * must be in the NaN range identified above. 1060 * 1061 * @param bits any {@code long} integer. 1062 * @return the {@code double} floating-point value with the same 1063 * bit pattern. 1064 */ 1065 @IntrinsicCandidate longBitsToDouble(long bits)1066 public static native double longBitsToDouble(long bits); 1067 1068 /** 1069 * Compares two {@code Double} objects numerically. 1070 * 1071 * This method imposes a total order on {@code Double} objects 1072 * with two differences compared to the incomplete order defined by 1073 * the Java language numerical comparison operators ({@code <, <=, 1074 * ==, >=, >}) on {@code double} values. 1075 * 1076 * <ul><li> A NaN is <em>unordered</em> with respect to other 1077 * values and unequal to itself under the comparison 1078 * operators. This method chooses to define {@code 1079 * Double.NaN} to be equal to itself and greater than all 1080 * other {@code double} values (including {@code 1081 * Double.POSITIVE_INFINITY}). 1082 * 1083 * <li> Positive zero and negative zero compare equal 1084 * numerically, but are distinct and distinguishable values. 1085 * This method chooses to define positive zero ({@code +0.0d}), 1086 * to be greater than negative zero ({@code -0.0d}). 1087 * </ul> 1088 1089 * This ensures that the <i>natural ordering</i> of {@code Double} 1090 * objects imposed by this method is <i>consistent with 1091 * equals</i>; see <a href="#equivalenceRelation">this 1092 * discussion</a> for details of floating-point comparison and 1093 * ordering. 1094 * 1095 * @param anotherDouble the {@code Double} to be compared. 1096 * @return the value {@code 0} if {@code anotherDouble} is 1097 * numerically equal to this {@code Double}; a value 1098 * less than {@code 0} if this {@code Double} 1099 * is numerically less than {@code anotherDouble}; 1100 * and a value greater than {@code 0} if this 1101 * {@code Double} is numerically greater than 1102 * {@code anotherDouble}. 1103 * 1104 * @jls 15.20.1 Numerical Comparison Operators {@code <}, {@code <=}, {@code >}, and {@code >=} 1105 * @since 1.2 1106 */ compareTo(Double anotherDouble)1107 public int compareTo(Double anotherDouble) { 1108 return Double.compare(value, anotherDouble.value); 1109 } 1110 1111 /** 1112 * Compares the two specified {@code double} values. The sign 1113 * of the integer value returned is the same as that of the 1114 * integer that would be returned by the call: 1115 * <pre> 1116 * new Double(d1).compareTo(new Double(d2)) 1117 * </pre> 1118 * 1119 * @param d1 the first {@code double} to compare 1120 * @param d2 the second {@code double} to compare 1121 * @return the value {@code 0} if {@code d1} is 1122 * numerically equal to {@code d2}; a value less than 1123 * {@code 0} if {@code d1} is numerically less than 1124 * {@code d2}; and a value greater than {@code 0} 1125 * if {@code d1} is numerically greater than 1126 * {@code d2}. 1127 * @since 1.4 1128 */ compare(double d1, double d2)1129 public static int compare(double d1, double d2) { 1130 if (d1 < d2) 1131 return -1; // Neither val is NaN, thisVal is smaller 1132 if (d1 > d2) 1133 return 1; // Neither val is NaN, thisVal is larger 1134 1135 // Cannot use doubleToRawLongBits because of possibility of NaNs. 1136 long thisBits = Double.doubleToLongBits(d1); 1137 long anotherBits = Double.doubleToLongBits(d2); 1138 1139 return (thisBits == anotherBits ? 0 : // Values are equal 1140 (thisBits < anotherBits ? -1 : // (-0.0, 0.0) or (!NaN, NaN) 1141 1)); // (0.0, -0.0) or (NaN, !NaN) 1142 } 1143 1144 /** 1145 * Adds two {@code double} values together as per the + operator. 1146 * 1147 * @param a the first operand 1148 * @param b the second operand 1149 * @return the sum of {@code a} and {@code b} 1150 * @jls 4.2.4 Floating-Point Operations 1151 * @see java.util.function.BinaryOperator 1152 * @since 1.8 1153 */ sum(double a, double b)1154 public static double sum(double a, double b) { 1155 return a + b; 1156 } 1157 1158 /** 1159 * Returns the greater of two {@code double} values 1160 * as if by calling {@link Math#max(double, double) Math.max}. 1161 * 1162 * @param a the first operand 1163 * @param b the second operand 1164 * @return the greater of {@code a} and {@code b} 1165 * @see java.util.function.BinaryOperator 1166 * @since 1.8 1167 */ max(double a, double b)1168 public static double max(double a, double b) { 1169 return Math.max(a, b); 1170 } 1171 1172 /** 1173 * Returns the smaller of two {@code double} values 1174 * as if by calling {@link Math#min(double, double) Math.min}. 1175 * 1176 * @param a the first operand 1177 * @param b the second operand 1178 * @return the smaller of {@code a} and {@code b}. 1179 * @see java.util.function.BinaryOperator 1180 * @since 1.8 1181 */ min(double a, double b)1182 public static double min(double a, double b) { 1183 return Math.min(a, b); 1184 } 1185 1186 // BEGIN Android-removed: dynamic constants not supported on Android. 1187 /** 1188 * Returns an {@link Optional} containing the nominal descriptor for this 1189 * instance, which is the instance itself. 1190 * 1191 * @return an {@link Optional} describing the {@linkplain Double} instance 1192 * @since 12 1193 * 1194 @Override 1195 public Optional<Double> describeConstable() { 1196 return Optional.of(this); 1197 } 1198 1199 /** 1200 * Resolves this instance as a {@link ConstantDesc}, the result of which is 1201 * the instance itself. 1202 * 1203 * @param lookup ignored 1204 * @return the {@linkplain Double} instance 1205 * @since 12 1206 * 1207 @Override 1208 public Double resolveConstantDesc(MethodHandles.Lookup lookup) { 1209 return this; 1210 } 1211 // END Android-removed: dynamic constants not supported on Android. 1212 1213 /** use serialVersionUID from JDK 1.0.2 for interoperability */ 1214 @java.io.Serial 1215 private static final long serialVersionUID = -9172774392245257468L; 1216 } 1217