1 /*
   2  * Copyright (c) 1996, 2014, 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 /*
  27  * Portions Copyright (c) 1995  Colin Plumb.  All rights reserved.
  28  */
  29 
  30 package java.math;
  31 
  32 import java.io.IOException;
  33 import java.io.ObjectInputStream;
  34 import java.io.ObjectOutputStream;
  35 import java.io.ObjectStreamField;
  36 import java.util.Arrays;
  37 import java.util.Objects;
  38 import java.util.Random;
  39 import java.util.concurrent.ThreadLocalRandom;
  40 
  41 import sun.misc.DoubleConsts;
  42 import sun.misc.FloatConsts;
  43 import jdk.internal.HotSpotIntrinsicCandidate;
  44 
  45 /**
  46  * Immutable arbitrary-precision integers.  All operations behave as if
  47  * BigIntegers were represented in two's-complement notation (like Java's
  48  * primitive integer types).  BigInteger provides analogues to all of Java's
  49  * primitive integer operators, and all relevant methods from java.lang.Math.
  50  * Additionally, BigInteger provides operations for modular arithmetic, GCD
  51  * calculation, primality testing, prime generation, bit manipulation,
  52  * and a few other miscellaneous operations.
  53  *
  54  * <p>Semantics of arithmetic operations exactly mimic those of Java's integer
  55  * arithmetic operators, as defined in <i>The Java Language Specification</i>.
  56  * For example, division by zero throws an {@code ArithmeticException}, and
  57  * division of a negative by a positive yields a negative (or zero) remainder.
  58  * All of the details in the Spec concerning overflow are ignored, as
  59  * BigIntegers are made as large as necessary to accommodate the results of an
  60  * operation.
  61  *
  62  * <p>Semantics of shift operations extend those of Java's shift operators
  63  * to allow for negative shift distances.  A right-shift with a negative
  64  * shift distance results in a left shift, and vice-versa.  The unsigned
  65  * right shift operator ({@code >>>}) is omitted, as this operation makes
  66  * little sense in combination with the "infinite word size" abstraction
  67  * provided by this class.
  68  *
  69  * <p>Semantics of bitwise logical operations exactly mimic those of Java's
  70  * bitwise integer operators.  The binary operators ({@code and},
  71  * {@code or}, {@code xor}) implicitly perform sign extension on the shorter
  72  * of the two operands prior to performing the operation.
  73  *
  74  * <p>Comparison operations perform signed integer comparisons, analogous to
  75  * those performed by Java's relational and equality operators.
  76  *
  77  * <p>Modular arithmetic operations are provided to compute residues, perform
  78  * exponentiation, and compute multiplicative inverses.  These methods always
  79  * return a non-negative result, between {@code 0} and {@code (modulus - 1)},
  80  * inclusive.
  81  *
  82  * <p>Bit operations operate on a single bit of the two's-complement
  83  * representation of their operand.  If necessary, the operand is sign-
  84  * extended so that it contains the designated bit.  None of the single-bit
  85  * operations can produce a BigInteger with a different sign from the
  86  * BigInteger being operated on, as they affect only a single bit, and the
  87  * "infinite word size" abstraction provided by this class ensures that there
  88  * are infinitely many "virtual sign bits" preceding each BigInteger.
  89  *
  90  * <p>For the sake of brevity and clarity, pseudo-code is used throughout the
  91  * descriptions of BigInteger methods.  The pseudo-code expression
  92  * {@code (i + j)} is shorthand for "a BigInteger whose value is
  93  * that of the BigInteger {@code i} plus that of the BigInteger {@code j}."
  94  * The pseudo-code expression {@code (i == j)} is shorthand for
  95  * "{@code true} if and only if the BigInteger {@code i} represents the same
  96  * value as the BigInteger {@code j}."  Other pseudo-code expressions are
  97  * interpreted similarly.
  98  *
  99  * <p>All methods and constructors in this class throw
 100  * {@code NullPointerException} when passed
 101  * a null object reference for any input parameter.
 102  *
 103  * BigInteger must support values in the range
 104  * -2<sup>{@code Integer.MAX_VALUE}</sup> (exclusive) to
 105  * +2<sup>{@code Integer.MAX_VALUE}</sup> (exclusive)
 106  * and may support values outside of that range.
 107  *
 108  * The range of probable prime values is limited and may be less than
 109  * the full supported positive range of {@code BigInteger}.
 110  * The range must be at least 1 to 2<sup>500000000</sup>.
 111  *
 112  * @implNote
 113  * BigInteger constructors and operations throw {@code ArithmeticException} when
 114  * the result is out of the supported range of
 115  * -2<sup>{@code Integer.MAX_VALUE}</sup> (exclusive) to
 116  * +2<sup>{@code Integer.MAX_VALUE}</sup> (exclusive).
 117  *
 118  * @see     BigDecimal
 119  * @author  Josh Bloch
 120  * @author  Michael McCloskey
 121  * @author  Alan Eliasen
 122  * @author  Timothy Buktu
 123  * @since 1.1
 124  */
 125 
 126 public class BigInteger extends Number implements Comparable<BigInteger> {
 127     /**
 128      * The signum of this BigInteger: -1 for negative, 0 for zero, or
 129      * 1 for positive.  Note that the BigInteger zero <i>must</i> have
 130      * a signum of 0.  This is necessary to ensures that there is exactly one
 131      * representation for each BigInteger value.
 132      */
 133     final int signum;
 134 
 135     /**
 136      * The magnitude of this BigInteger, in <i>big-endian</i> order: the
 137      * zeroth element of this array is the most-significant int of the
 138      * magnitude.  The magnitude must be "minimal" in that the most-significant
 139      * int ({@code mag[0]}) must be non-zero.  This is necessary to
 140      * ensure that there is exactly one representation for each BigInteger
 141      * value.  Note that this implies that the BigInteger zero has a
 142      * zero-length mag array.
 143      */
 144     final int[] mag;
 145 
 146     // The following fields are stable variables. A stable variable's value
 147     // changes at most once from the default zero value to a non-zero stable
 148     // value. A stable value is calculated lazily on demand.
 149 
 150     /**
 151      * One plus the bitCount of this BigInteger. This is a stable variable.
 152      *
 153      * @see #bitCount
 154      */
 155     private int bitCountPlusOne;
 156 
 157     /**
 158      * One plus the bitLength of this BigInteger. This is a stable variable.
 159      * (either value is acceptable).
 160      *
 161      * @see #bitLength()
 162      */
 163     private int bitLengthPlusOne;
 164 
 165     /**
 166      * Two plus the lowest set bit of this BigInteger. This is a stable variable.
 167      *
 168      * @see #getLowestSetBit
 169      */
 170     private int lowestSetBitPlusTwo;
 171 
 172     /**
 173      * Two plus the index of the lowest-order int in the magnitude of this
 174      * BigInteger that contains a nonzero int. This is a stable variable. The
 175      * least significant int has int-number 0, the next int in order of
 176      * increasing significance has int-number 1, and so forth.
 177      *
 178      * <p>Note: never used for a BigInteger with a magnitude of zero.
 179      *
 180      * @see #firstNonzeroIntNum()
 181      */
 182     private int firstNonzeroIntNumPlusTwo;
 183 
 184     /**
 185      * This mask is used to obtain the value of an int as if it were unsigned.
 186      */
 187     static final long LONG_MASK = 0xffffffffL;
 188 
 189     /**
 190      * This constant limits {@code mag.length} of BigIntegers to the supported
 191      * range.
 192      */
 193     private static final int MAX_MAG_LENGTH = Integer.MAX_VALUE / Integer.SIZE + 1; // (1 << 26)
 194 
 195     /**
 196      * Bit lengths larger than this constant can cause overflow in searchLen
 197      * calculation and in BitSieve.singleSearch method.
 198      */
 199     private static final  int PRIME_SEARCH_BIT_LENGTH_LIMIT = 500000000;
 200 
 201     /**
 202      * The threshold value for using Karatsuba multiplication.  If the number
 203      * of ints in both mag arrays are greater than this number, then
 204      * Karatsuba multiplication will be used.   This value is found
 205      * experimentally to work well.
 206      */
 207     private static final int KARATSUBA_THRESHOLD = 80;
 208 
 209     /**
 210      * The threshold value for using 3-way Toom-Cook multiplication.
 211      * If the number of ints in each mag array is greater than the
 212      * Karatsuba threshold, and the number of ints in at least one of
 213      * the mag arrays is greater than this threshold, then Toom-Cook
 214      * multiplication will be used.
 215      */
 216     private static final int TOOM_COOK_THRESHOLD = 240;
 217 
 218     /**
 219      * The threshold value for using Karatsuba squaring.  If the number
 220      * of ints in the number are larger than this value,
 221      * Karatsuba squaring will be used.   This value is found
 222      * experimentally to work well.
 223      */
 224     private static final int KARATSUBA_SQUARE_THRESHOLD = 128;
 225 
 226     /**
 227      * The threshold value for using Toom-Cook squaring.  If the number
 228      * of ints in the number are larger than this value,
 229      * Toom-Cook squaring will be used.   This value is found
 230      * experimentally to work well.
 231      */
 232     private static final int TOOM_COOK_SQUARE_THRESHOLD = 216;
 233 
 234     /**
 235      * The threshold value for using Burnikel-Ziegler division.  If the number
 236      * of ints in the divisor are larger than this value, Burnikel-Ziegler
 237      * division may be used.  This value is found experimentally to work well.
 238      */
 239     static final int BURNIKEL_ZIEGLER_THRESHOLD = 80;
 240 
 241     /**
 242      * The offset value for using Burnikel-Ziegler division.  If the number
 243      * of ints in the divisor exceeds the Burnikel-Ziegler threshold, and the
 244      * number of ints in the dividend is greater than the number of ints in the
 245      * divisor plus this value, Burnikel-Ziegler division will be used.  This
 246      * value is found experimentally to work well.
 247      */
 248     static final int BURNIKEL_ZIEGLER_OFFSET = 40;
 249 
 250     /**
 251      * The threshold value for using Schoenhage recursive base conversion. If
 252      * the number of ints in the number are larger than this value,
 253      * the Schoenhage algorithm will be used.  In practice, it appears that the
 254      * Schoenhage routine is faster for any threshold down to 2, and is
 255      * relatively flat for thresholds between 2-25, so this choice may be
 256      * varied within this range for very small effect.
 257      */
 258     private static final int SCHOENHAGE_BASE_CONVERSION_THRESHOLD = 20;
 259 
 260     /**
 261      * The threshold value for using squaring code to perform multiplication
 262      * of a {@code BigInteger} instance by itself.  If the number of ints in
 263      * the number are larger than this value, {@code multiply(this)} will
 264      * return {@code square()}.
 265      */
 266     private static final int MULTIPLY_SQUARE_THRESHOLD = 20;
 267 
 268     /**
 269      * The threshold for using an intrinsic version of
 270      * implMontgomeryXXX to perform Montgomery multiplication.  If the
 271      * number of ints in the number is more than this value we do not
 272      * use the intrinsic.
 273      */
 274     private static final int MONTGOMERY_INTRINSIC_THRESHOLD = 512;
 275 
 276 
 277     // Constructors
 278 
 279     /**
 280      * Translates a byte sub-array containing the two's-complement binary
 281      * representation of a BigInteger into a BigInteger.  The sub-array is
 282      * specified via an offset into the array and a length.  The sub-array is
 283      * assumed to be in <i>big-endian</i> byte-order: the most significant
 284      * byte is the element at index {@code off}.  The {@code val} array is
 285      * assumed to be unchanged for the duration of the constructor call.
 286      *
 287      * An {@code IndexOutOfBoundsException} is thrown if the length of the array
 288      * {@code val} is non-zero and either {@code off} is negative, {@code len}
 289      * is negative, or {@code off+len} is greater than the length of
 290      * {@code val}.
 291      *
 292      * @param  val byte array containing a sub-array which is the big-endian
 293      *         two's-complement binary representation of a BigInteger.
 294      * @param  off the start offset of the binary representation.
 295      * @param  len the number of bytes to use.
 296      * @throws NumberFormatException {@code val} is zero bytes long.
 297      * @throws IndexOutOfBoundsException if the provided array offset and
 298      *         length would cause an index into the byte array to be
 299      *         negative or greater than or equal to the array length.
 300      * @since 1.9
 301      */
 302     public BigInteger(byte[] val, int off, int len) {
 303         if (val.length == 0) {
 304             throw new NumberFormatException("Zero length BigInteger");
 305         } else if ((off < 0) || (off >= val.length) || (len < 0) ||
 306                    (len > val.length - off)) { // 0 <= off < val.length
 307             throw new IndexOutOfBoundsException();
 308         }
 309 
 310         if (val[off] < 0) {
 311             mag = makePositive(val, off, len);
 312             signum = -1;
 313         } else {
 314             mag = stripLeadingZeroBytes(val, off, len);
 315             signum = (mag.length == 0 ? 0 : 1);
 316         }
 317         if (mag.length >= MAX_MAG_LENGTH) {
 318             checkRange();
 319         }
 320     }
 321 
 322     /**
 323      * Translates a byte array containing the two's-complement binary
 324      * representation of a BigInteger into a BigInteger.  The input array is
 325      * assumed to be in <i>big-endian</i> byte-order: the most significant
 326      * byte is in the zeroth element.  The {@code val} array is assumed to be
 327      * unchanged for the duration of the constructor call.
 328      *
 329      * @param  val big-endian two's-complement binary representation of a
 330      *         BigInteger.
 331      * @throws NumberFormatException {@code val} is zero bytes long.
 332      */
 333     public BigInteger(byte[] val) {
 334         this(val, 0, val.length);
 335     }
 336 
 337     /**
 338      * This private constructor translates an int array containing the
 339      * two's-complement binary representation of a BigInteger into a
 340      * BigInteger. The input array is assumed to be in <i>big-endian</i>
 341      * int-order: the most significant int is in the zeroth element.  The
 342      * {@code val} array is assumed to be unchanged for the duration of
 343      * the constructor call.
 344      */
 345     private BigInteger(int[] val) {
 346         if (val.length == 0)
 347             throw new NumberFormatException("Zero length BigInteger");
 348 
 349         if (val[0] < 0) {
 350             mag = makePositive(val);
 351             signum = -1;
 352         } else {
 353             mag = trustedStripLeadingZeroInts(val);
 354             signum = (mag.length == 0 ? 0 : 1);
 355         }
 356         if (mag.length >= MAX_MAG_LENGTH) {
 357             checkRange();
 358         }
 359     }
 360 
 361     /**
 362      * Translates the sign-magnitude representation of a BigInteger into a
 363      * BigInteger.  The sign is represented as an integer signum value: -1 for
 364      * negative, 0 for zero, or 1 for positive.  The magnitude is a sub-array of
 365      * a byte array in <i>big-endian</i> byte-order: the most significant byte
 366      * is the element at index {@code off}.  A zero value of the length
 367      * {@code len} is permissible, and will result in a BigInteger value of 0,
 368      * whether signum is -1, 0 or 1.  The {@code magnitude} array is assumed to
 369      * be unchanged for the duration of the constructor call.
 370      *
 371      * An {@code IndexOutOfBoundsException} is thrown if the length of the array
 372      * {@code magnitude} is non-zero and either {@code off} is negative,
 373      * {@code len} is negative, or {@code off+len} is greater than the length of
 374      * {@code magnitude}.
 375      *
 376      * @param  signum signum of the number (-1 for negative, 0 for zero, 1
 377      *         for positive).
 378      * @param  magnitude big-endian binary representation of the magnitude of
 379      *         the number.
 380      * @param  off the start offset of the binary representation.
 381      * @param  len the number of bytes to use.
 382      * @throws NumberFormatException {@code signum} is not one of the three
 383      *         legal values (-1, 0, and 1), or {@code signum} is 0 and
 384      *         {@code magnitude} contains one or more non-zero bytes.
 385      * @throws IndexOutOfBoundsException if the provided array offset and
 386      *         length would cause an index into the byte array to be
 387      *         negative or greater than or equal to the array length.
 388      * @since 1.9
 389      */
 390     public BigInteger(int signum, byte[] magnitude, int off, int len) {
 391         if (signum < -1 || signum > 1) {
 392             throw(new NumberFormatException("Invalid signum value"));
 393         } else if ((off < 0) || (len < 0) ||
 394             (len > 0 &&
 395                 ((off >= magnitude.length) ||
 396                  (len > magnitude.length - off)))) { // 0 <= off < magnitude.length
 397             throw new IndexOutOfBoundsException();
 398         }
 399 
 400         // stripLeadingZeroBytes() returns a zero length array if len == 0
 401         this.mag = stripLeadingZeroBytes(magnitude, off, len);
 402 
 403         if (this.mag.length == 0) {
 404             this.signum = 0;
 405         } else {
 406             if (signum == 0)
 407                 throw(new NumberFormatException("signum-magnitude mismatch"));
 408             this.signum = signum;
 409         }
 410         if (mag.length >= MAX_MAG_LENGTH) {
 411             checkRange();
 412         }
 413     }
 414 
 415     /**
 416      * Translates the sign-magnitude representation of a BigInteger into a
 417      * BigInteger.  The sign is represented as an integer signum value: -1 for
 418      * negative, 0 for zero, or 1 for positive.  The magnitude is a byte array
 419      * in <i>big-endian</i> byte-order: the most significant byte is the
 420      * zeroth element.  A zero-length magnitude array is permissible, and will
 421      * result in a BigInteger value of 0, whether signum is -1, 0 or 1.  The
 422      * {@code magnitude} array is assumed to be unchanged for the duration of
 423      * the constructor call.
 424      *
 425      * @param  signum signum of the number (-1 for negative, 0 for zero, 1
 426      *         for positive).
 427      * @param  magnitude big-endian binary representation of the magnitude of
 428      *         the number.
 429      * @throws NumberFormatException {@code signum} is not one of the three
 430      *         legal values (-1, 0, and 1), or {@code signum} is 0 and
 431      *         {@code magnitude} contains one or more non-zero bytes.
 432      */
 433     public BigInteger(int signum, byte[] magnitude) {
 434          this(signum, magnitude, 0, magnitude.length);
 435     }
 436 
 437     /**
 438      * A constructor for internal use that translates the sign-magnitude
 439      * representation of a BigInteger into a BigInteger. It checks the
 440      * arguments and copies the magnitude so this constructor would be
 441      * safe for external use.  The {@code magnitude} array is assumed to be
 442      * unchanged for the duration of the constructor call.
 443      */
 444     private BigInteger(int signum, int[] magnitude) {
 445         this.mag = stripLeadingZeroInts(magnitude);
 446 
 447         if (signum < -1 || signum > 1)
 448             throw(new NumberFormatException("Invalid signum value"));
 449 
 450         if (this.mag.length == 0) {
 451             this.signum = 0;
 452         } else {
 453             if (signum == 0)
 454                 throw(new NumberFormatException("signum-magnitude mismatch"));
 455             this.signum = signum;
 456         }
 457         if (mag.length >= MAX_MAG_LENGTH) {
 458             checkRange();
 459         }
 460     }
 461 
 462     /**
 463      * Translates the String representation of a BigInteger in the
 464      * specified radix into a BigInteger.  The String representation
 465      * consists of an optional minus or plus sign followed by a
 466      * sequence of one or more digits in the specified radix.  The
 467      * character-to-digit mapping is provided by {@code
 468      * Character.digit}.  The String may not contain any extraneous
 469      * characters (whitespace, for example).
 470      *
 471      * @param val String representation of BigInteger.
 472      * @param radix radix to be used in interpreting {@code val}.
 473      * @throws NumberFormatException {@code val} is not a valid representation
 474      *         of a BigInteger in the specified radix, or {@code radix} is
 475      *         outside the range from {@link Character#MIN_RADIX} to
 476      *         {@link Character#MAX_RADIX}, inclusive.
 477      * @see    Character#digit
 478      */
 479     public BigInteger(String val, int radix) {
 480         int cursor = 0, numDigits;
 481         final int len = val.length();
 482 
 483         if (radix < Character.MIN_RADIX || radix > Character.MAX_RADIX)
 484             throw new NumberFormatException("Radix out of range");
 485         if (len == 0)
 486             throw new NumberFormatException("Zero length BigInteger");
 487 
 488         // Check for at most one leading sign
 489         int sign = 1;
 490         int index1 = val.lastIndexOf('-');
 491         int index2 = val.lastIndexOf('+');
 492         if (index1 >= 0) {
 493             if (index1 != 0 || index2 >= 0) {
 494                 throw new NumberFormatException("Illegal embedded sign character");
 495             }
 496             sign = -1;
 497             cursor = 1;
 498         } else if (index2 >= 0) {
 499             if (index2 != 0) {
 500                 throw new NumberFormatException("Illegal embedded sign character");
 501             }
 502             cursor = 1;
 503         }
 504         if (cursor == len)
 505             throw new NumberFormatException("Zero length BigInteger");
 506 
 507         // Skip leading zeros and compute number of digits in magnitude
 508         while (cursor < len &&
 509                Character.digit(val.charAt(cursor), radix) == 0) {
 510             cursor++;
 511         }
 512 
 513         if (cursor == len) {
 514             signum = 0;
 515             mag = ZERO.mag;
 516             return;
 517         }
 518 
 519         numDigits = len - cursor;
 520         signum = sign;
 521 
 522         // Pre-allocate array of expected size. May be too large but can
 523         // never be too small. Typically exact.
 524         long numBits = ((numDigits * bitsPerDigit[radix]) >>> 10) + 1;
 525         if (numBits + 31 >= (1L << 32)) {
 526             reportOverflow();
 527         }
 528         int numWords = (int) (numBits + 31) >>> 5;
 529         int[] magnitude = new int[numWords];
 530 
 531         // Process first (potentially short) digit group
 532         int firstGroupLen = numDigits % digitsPerInt[radix];
 533         if (firstGroupLen == 0)
 534             firstGroupLen = digitsPerInt[radix];
 535         String group = val.substring(cursor, cursor += firstGroupLen);
 536         magnitude[numWords - 1] = Integer.parseInt(group, radix);
 537         if (magnitude[numWords - 1] < 0)
 538             throw new NumberFormatException("Illegal digit");
 539 
 540         // Process remaining digit groups
 541         int superRadix = intRadix[radix];
 542         int groupVal = 0;
 543         while (cursor < len) {
 544             group = val.substring(cursor, cursor += digitsPerInt[radix]);
 545             groupVal = Integer.parseInt(group, radix);
 546             if (groupVal < 0)
 547                 throw new NumberFormatException("Illegal digit");
 548             destructiveMulAdd(magnitude, superRadix, groupVal);
 549         }
 550         // Required for cases where the array was overallocated.
 551         mag = trustedStripLeadingZeroInts(magnitude);
 552         if (mag.length >= MAX_MAG_LENGTH) {
 553             checkRange();
 554         }
 555     }
 556 
 557     /*
 558      * Constructs a new BigInteger using a char array with radix=10.
 559      * Sign is precalculated outside and not allowed in the val. The {@code val}
 560      * array is assumed to be unchanged for the duration of the constructor
 561      * call.
 562      */
 563     BigInteger(char[] val, int sign, int len) {
 564         int cursor = 0, numDigits;
 565 
 566         // Skip leading zeros and compute number of digits in magnitude
 567         while (cursor < len && Character.digit(val[cursor], 10) == 0) {
 568             cursor++;
 569         }
 570         if (cursor == len) {
 571             signum = 0;
 572             mag = ZERO.mag;
 573             return;
 574         }
 575 
 576         numDigits = len - cursor;
 577         signum = sign;
 578         // Pre-allocate array of expected size
 579         int numWords;
 580         if (len < 10) {
 581             numWords = 1;
 582         } else {
 583             long numBits = ((numDigits * bitsPerDigit[10]) >>> 10) + 1;
 584             if (numBits + 31 >= (1L << 32)) {
 585                 reportOverflow();
 586             }
 587             numWords = (int) (numBits + 31) >>> 5;
 588         }
 589         int[] magnitude = new int[numWords];
 590 
 591         // Process first (potentially short) digit group
 592         int firstGroupLen = numDigits % digitsPerInt[10];
 593         if (firstGroupLen == 0)
 594             firstGroupLen = digitsPerInt[10];
 595         magnitude[numWords - 1] = parseInt(val, cursor,  cursor += firstGroupLen);
 596 
 597         // Process remaining digit groups
 598         while (cursor < len) {
 599             int groupVal = parseInt(val, cursor, cursor += digitsPerInt[10]);
 600             destructiveMulAdd(magnitude, intRadix[10], groupVal);
 601         }
 602         mag = trustedStripLeadingZeroInts(magnitude);
 603         if (mag.length >= MAX_MAG_LENGTH) {
 604             checkRange();
 605         }
 606     }
 607 
 608     // Create an integer with the digits between the two indexes
 609     // Assumes start < end. The result may be negative, but it
 610     // is to be treated as an unsigned value.
 611     private int parseInt(char[] source, int start, int end) {
 612         int result = Character.digit(source[start++], 10);
 613         if (result == -1)
 614             throw new NumberFormatException(new String(source));
 615 
 616         for (int index = start; index < end; index++) {
 617             int nextVal = Character.digit(source[index], 10);
 618             if (nextVal == -1)
 619                 throw new NumberFormatException(new String(source));
 620             result = 10*result + nextVal;
 621         }
 622 
 623         return result;
 624     }
 625 
 626     // bitsPerDigit in the given radix times 1024
 627     // Rounded up to avoid underallocation.
 628     private static long bitsPerDigit[] = { 0, 0,
 629         1024, 1624, 2048, 2378, 2648, 2875, 3072, 3247, 3402, 3543, 3672,
 630         3790, 3899, 4001, 4096, 4186, 4271, 4350, 4426, 4498, 4567, 4633,
 631         4696, 4756, 4814, 4870, 4923, 4975, 5025, 5074, 5120, 5166, 5210,
 632                                            5253, 5295};
 633 
 634     // Multiply x array times word y in place, and add word z
 635     private static void destructiveMulAdd(int[] x, int y, int z) {
 636         // Perform the multiplication word by word
 637         long ylong = y & LONG_MASK;
 638         long zlong = z & LONG_MASK;
 639         int len = x.length;
 640 
 641         long product = 0;
 642         long carry = 0;
 643         for (int i = len-1; i >= 0; i--) {
 644             product = ylong * (x[i] & LONG_MASK) + carry;
 645             x[i] = (int)product;
 646             carry = product >>> 32;
 647         }
 648 
 649         // Perform the addition
 650         long sum = (x[len-1] & LONG_MASK) + zlong;
 651         x[len-1] = (int)sum;
 652         carry = sum >>> 32;
 653         for (int i = len-2; i >= 0; i--) {
 654             sum = (x[i] & LONG_MASK) + carry;
 655             x[i] = (int)sum;
 656             carry = sum >>> 32;
 657         }
 658     }
 659 
 660     /**
 661      * Translates the decimal String representation of a BigInteger into a
 662      * BigInteger.  The String representation consists of an optional minus
 663      * sign followed by a sequence of one or more decimal digits.  The
 664      * character-to-digit mapping is provided by {@code Character.digit}.
 665      * The String may not contain any extraneous characters (whitespace, for
 666      * example).
 667      *
 668      * @param val decimal String representation of BigInteger.
 669      * @throws NumberFormatException {@code val} is not a valid representation
 670      *         of a BigInteger.
 671      * @see    Character#digit
 672      */
 673     public BigInteger(String val) {
 674         this(val, 10);
 675     }
 676 
 677     /**
 678      * Constructs a randomly generated BigInteger, uniformly distributed over
 679      * the range 0 to (2<sup>{@code numBits}</sup> - 1), inclusive.
 680      * The uniformity of the distribution assumes that a fair source of random
 681      * bits is provided in {@code rnd}.  Note that this constructor always
 682      * constructs a non-negative BigInteger.
 683      *
 684      * @param  numBits maximum bitLength of the new BigInteger.
 685      * @param  rnd source of randomness to be used in computing the new
 686      *         BigInteger.
 687      * @throws IllegalArgumentException {@code numBits} is negative.
 688      * @see #bitLength()
 689      */
 690     public BigInteger(int numBits, Random rnd) {
 691         this(1, randomBits(numBits, rnd));
 692     }
 693 
 694     private static byte[] randomBits(int numBits, Random rnd) {
 695         if (numBits < 0)
 696             throw new IllegalArgumentException("numBits must be non-negative");
 697         int numBytes = (int)(((long)numBits+7)/8); // avoid overflow
 698         byte[] randomBits = new byte[numBytes];
 699 
 700         // Generate random bytes and mask out any excess bits
 701         if (numBytes > 0) {
 702             rnd.nextBytes(randomBits);
 703             int excessBits = 8*numBytes - numBits;
 704             randomBits[0] &= (1 << (8-excessBits)) - 1;
 705         }
 706         return randomBits;
 707     }
 708 
 709     /**
 710      * Constructs a randomly generated positive BigInteger that is probably
 711      * prime, with the specified bitLength.
 712      *
 713      * <p>It is recommended that the {@link #probablePrime probablePrime}
 714      * method be used in preference to this constructor unless there
 715      * is a compelling need to specify a certainty.
 716      *
 717      * @param  bitLength bitLength of the returned BigInteger.
 718      * @param  certainty a measure of the uncertainty that the caller is
 719      *         willing to tolerate.  The probability that the new BigInteger
 720      *         represents a prime number will exceed
 721      *         (1 - 1/2<sup>{@code certainty}</sup>).  The execution time of
 722      *         this constructor is proportional to the value of this parameter.
 723      * @param  rnd source of random bits used to select candidates to be
 724      *         tested for primality.
 725      * @throws ArithmeticException {@code bitLength < 2} or {@code bitLength} is too large.
 726      * @see    #bitLength()
 727      */
 728     public BigInteger(int bitLength, int certainty, Random rnd) {
 729         BigInteger prime;
 730 
 731         if (bitLength < 2)
 732             throw new ArithmeticException("bitLength < 2");
 733         prime = (bitLength < SMALL_PRIME_THRESHOLD
 734                                 ? smallPrime(bitLength, certainty, rnd)
 735                                 : largePrime(bitLength, certainty, rnd));
 736         signum = 1;
 737         mag = prime.mag;
 738     }
 739 
 740     // Minimum size in bits that the requested prime number has
 741     // before we use the large prime number generating algorithms.
 742     // The cutoff of 95 was chosen empirically for best performance.
 743     private static final int SMALL_PRIME_THRESHOLD = 95;
 744 
 745     // Certainty required to meet the spec of probablePrime
 746     private static final int DEFAULT_PRIME_CERTAINTY = 100;
 747 
 748     /**
 749      * Returns a positive BigInteger that is probably prime, with the
 750      * specified bitLength. The probability that a BigInteger returned
 751      * by this method is composite does not exceed 2<sup>-100</sup>.
 752      *
 753      * @param  bitLength bitLength of the returned BigInteger.
 754      * @param  rnd source of random bits used to select candidates to be
 755      *         tested for primality.
 756      * @return a BigInteger of {@code bitLength} bits that is probably prime
 757      * @throws ArithmeticException {@code bitLength < 2} or {@code bitLength} is too large.
 758      * @see    #bitLength()
 759      * @since 1.4
 760      */
 761     public static BigInteger probablePrime(int bitLength, Random rnd) {
 762         if (bitLength < 2)
 763             throw new ArithmeticException("bitLength < 2");
 764 
 765         return (bitLength < SMALL_PRIME_THRESHOLD ?
 766                 smallPrime(bitLength, DEFAULT_PRIME_CERTAINTY, rnd) :
 767                 largePrime(bitLength, DEFAULT_PRIME_CERTAINTY, rnd));
 768     }
 769 
 770     /**
 771      * Find a random number of the specified bitLength that is probably prime.
 772      * This method is used for smaller primes, its performance degrades on
 773      * larger bitlengths.
 774      *
 775      * This method assumes bitLength > 1.
 776      */
 777     private static BigInteger smallPrime(int bitLength, int certainty, Random rnd) {
 778         int magLen = (bitLength + 31) >>> 5;
 779         int temp[] = new int[magLen];
 780         int highBit = 1 << ((bitLength+31) & 0x1f);  // High bit of high int
 781         int highMask = (highBit << 1) - 1;  // Bits to keep in high int
 782 
 783         while (true) {
 784             // Construct a candidate
 785             for (int i=0; i < magLen; i++)
 786                 temp[i] = rnd.nextInt();
 787             temp[0] = (temp[0] & highMask) | highBit;  // Ensure exact length
 788             if (bitLength > 2)
 789                 temp[magLen-1] |= 1;  // Make odd if bitlen > 2
 790 
 791             BigInteger p = new BigInteger(temp, 1);
 792 
 793             // Do cheap "pre-test" if applicable
 794             if (bitLength > 6) {
 795                 long r = p.remainder(SMALL_PRIME_PRODUCT).longValue();
 796                 if ((r%3==0)  || (r%5==0)  || (r%7==0)  || (r%11==0) ||
 797                     (r%13==0) || (r%17==0) || (r%19==0) || (r%23==0) ||
 798                     (r%29==0) || (r%31==0) || (r%37==0) || (r%41==0))
 799                     continue; // Candidate is composite; try another
 800             }
 801 
 802             // All candidates of bitLength 2 and 3 are prime by this point
 803             if (bitLength < 4)
 804                 return p;
 805 
 806             // Do expensive test if we survive pre-test (or it's inapplicable)
 807             if (p.primeToCertainty(certainty, rnd))
 808                 return p;
 809         }
 810     }
 811 
 812     private static final BigInteger SMALL_PRIME_PRODUCT
 813                        = valueOf(3L*5*7*11*13*17*19*23*29*31*37*41);
 814 
 815     /**
 816      * Find a random number of the specified bitLength that is probably prime.
 817      * This method is more appropriate for larger bitlengths since it uses
 818      * a sieve to eliminate most composites before using a more expensive
 819      * test.
 820      */
 821     private static BigInteger largePrime(int bitLength, int certainty, Random rnd) {
 822         BigInteger p;
 823         p = new BigInteger(bitLength, rnd).setBit(bitLength-1);
 824         p.mag[p.mag.length-1] &= 0xfffffffe;
 825 
 826         // Use a sieve length likely to contain the next prime number
 827         int searchLen = getPrimeSearchLen(bitLength);
 828         BitSieve searchSieve = new BitSieve(p, searchLen);
 829         BigInteger candidate = searchSieve.retrieve(p, certainty, rnd);
 830 
 831         while ((candidate == null) || (candidate.bitLength() != bitLength)) {
 832             p = p.add(BigInteger.valueOf(2*searchLen));
 833             if (p.bitLength() != bitLength)
 834                 p = new BigInteger(bitLength, rnd).setBit(bitLength-1);
 835             p.mag[p.mag.length-1] &= 0xfffffffe;
 836             searchSieve = new BitSieve(p, searchLen);
 837             candidate = searchSieve.retrieve(p, certainty, rnd);
 838         }
 839         return candidate;
 840     }
 841 
 842    /**
 843     * Returns the first integer greater than this {@code BigInteger} that
 844     * is probably prime.  The probability that the number returned by this
 845     * method is composite does not exceed 2<sup>-100</sup>. This method will
 846     * never skip over a prime when searching: if it returns {@code p}, there
 847     * is no prime {@code q} such that {@code this < q < p}.
 848     *
 849     * @return the first integer greater than this {@code BigInteger} that
 850     *         is probably prime.
 851     * @throws ArithmeticException {@code this < 0} or {@code this} is too large.
 852     * @since 1.5
 853     */
 854     public BigInteger nextProbablePrime() {
 855         if (this.signum < 0)
 856             throw new ArithmeticException("start < 0: " + this);
 857 
 858         // Handle trivial cases
 859         if ((this.signum == 0) || this.equals(ONE))
 860             return TWO;
 861 
 862         BigInteger result = this.add(ONE);
 863 
 864         // Fastpath for small numbers
 865         if (result.bitLength() < SMALL_PRIME_THRESHOLD) {
 866 
 867             // Ensure an odd number
 868             if (!result.testBit(0))
 869                 result = result.add(ONE);
 870 
 871             while (true) {
 872                 // Do cheap "pre-test" if applicable
 873                 if (result.bitLength() > 6) {
 874                     long r = result.remainder(SMALL_PRIME_PRODUCT).longValue();
 875                     if ((r%3==0)  || (r%5==0)  || (r%7==0)  || (r%11==0) ||
 876                         (r%13==0) || (r%17==0) || (r%19==0) || (r%23==0) ||
 877                         (r%29==0) || (r%31==0) || (r%37==0) || (r%41==0)) {
 878                         result = result.add(TWO);
 879                         continue; // Candidate is composite; try another
 880                     }
 881                 }
 882 
 883                 // All candidates of bitLength 2 and 3 are prime by this point
 884                 if (result.bitLength() < 4)
 885                     return result;
 886 
 887                 // The expensive test
 888                 if (result.primeToCertainty(DEFAULT_PRIME_CERTAINTY, null))
 889                     return result;
 890 
 891                 result = result.add(TWO);
 892             }
 893         }
 894 
 895         // Start at previous even number
 896         if (result.testBit(0))
 897             result = result.subtract(ONE);
 898 
 899         // Looking for the next large prime
 900         int searchLen = getPrimeSearchLen(result.bitLength());
 901 
 902         while (true) {
 903            BitSieve searchSieve = new BitSieve(result, searchLen);
 904            BigInteger candidate = searchSieve.retrieve(result,
 905                                                  DEFAULT_PRIME_CERTAINTY, null);
 906            if (candidate != null)
 907                return candidate;
 908            result = result.add(BigInteger.valueOf(2 * searchLen));
 909         }
 910     }
 911 
 912     private static int getPrimeSearchLen(int bitLength) {
 913         if (bitLength > PRIME_SEARCH_BIT_LENGTH_LIMIT + 1) {
 914             throw new ArithmeticException("Prime search implementation restriction on bitLength");
 915         }
 916         return bitLength / 20 * 64;
 917     }
 918 
 919     /**
 920      * Returns {@code true} if this BigInteger is probably prime,
 921      * {@code false} if it's definitely composite.
 922      *
 923      * This method assumes bitLength > 2.
 924      *
 925      * @param  certainty a measure of the uncertainty that the caller is
 926      *         willing to tolerate: if the call returns {@code true}
 927      *         the probability that this BigInteger is prime exceeds
 928      *         {@code (1 - 1/2<sup>certainty</sup>)}.  The execution time of
 929      *         this method is proportional to the value of this parameter.
 930      * @return {@code true} if this BigInteger is probably prime,
 931      *         {@code false} if it's definitely composite.
 932      */
 933     boolean primeToCertainty(int certainty, Random random) {
 934         int rounds = 0;
 935         int n = (Math.min(certainty, Integer.MAX_VALUE-1)+1)/2;
 936 
 937         // The relationship between the certainty and the number of rounds
 938         // we perform is given in the draft standard ANSI X9.80, "PRIME
 939         // NUMBER GENERATION, PRIMALITY TESTING, AND PRIMALITY CERTIFICATES".
 940         int sizeInBits = this.bitLength();
 941         if (sizeInBits < 100) {
 942             rounds = 50;
 943             rounds = n < rounds ? n : rounds;
 944             return passesMillerRabin(rounds, random);
 945         }
 946 
 947         if (sizeInBits < 256) {
 948             rounds = 27;
 949         } else if (sizeInBits < 512) {
 950             rounds = 15;
 951         } else if (sizeInBits < 768) {
 952             rounds = 8;
 953         } else if (sizeInBits < 1024) {
 954             rounds = 4;
 955         } else {
 956             rounds = 2;
 957         }
 958         rounds = n < rounds ? n : rounds;
 959 
 960         return passesMillerRabin(rounds, random) && passesLucasLehmer();
 961     }
 962 
 963     /**
 964      * Returns true iff this BigInteger is a Lucas-Lehmer probable prime.
 965      *
 966      * The following assumptions are made:
 967      * This BigInteger is a positive, odd number.
 968      */
 969     private boolean passesLucasLehmer() {
 970         BigInteger thisPlusOne = this.add(ONE);
 971 
 972         // Step 1
 973         int d = 5;
 974         while (jacobiSymbol(d, this) != -1) {
 975             // 5, -7, 9, -11, ...
 976             d = (d < 0) ? Math.abs(d)+2 : -(d+2);
 977         }
 978 
 979         // Step 2
 980         BigInteger u = lucasLehmerSequence(d, thisPlusOne, this);
 981 
 982         // Step 3
 983         return u.mod(this).equals(ZERO);
 984     }
 985 
 986     /**
 987      * Computes Jacobi(p,n).
 988      * Assumes n positive, odd, n>=3.
 989      */
 990     private static int jacobiSymbol(int p, BigInteger n) {
 991         if (p == 0)
 992             return 0;
 993 
 994         // Algorithm and comments adapted from Colin Plumb's C library.
 995         int j = 1;
 996         int u = n.mag[n.mag.length-1];
 997 
 998         // Make p positive
 999         if (p < 0) {
1000             p = -p;
1001             int n8 = u & 7;
1002             if ((n8 == 3) || (n8 == 7))
1003                 j = -j; // 3 (011) or 7 (111) mod 8
1004         }
1005 
1006         // Get rid of factors of 2 in p
1007         while ((p & 3) == 0)
1008             p >>= 2;
1009         if ((p & 1) == 0) {
1010             p >>= 1;
1011             if (((u ^ (u>>1)) & 2) != 0)
1012                 j = -j; // 3 (011) or 5 (101) mod 8
1013         }
1014         if (p == 1)
1015             return j;
1016         // Then, apply quadratic reciprocity
1017         if ((p & u & 2) != 0)   // p = u = 3 (mod 4)?
1018             j = -j;
1019         // And reduce u mod p
1020         u = n.mod(BigInteger.valueOf(p)).intValue();
1021 
1022         // Now compute Jacobi(u,p), u < p
1023         while (u != 0) {
1024             while ((u & 3) == 0)
1025                 u >>= 2;
1026             if ((u & 1) == 0) {
1027                 u >>= 1;
1028                 if (((p ^ (p>>1)) & 2) != 0)
1029                     j = -j;     // 3 (011) or 5 (101) mod 8
1030             }
1031             if (u == 1)
1032                 return j;
1033             // Now both u and p are odd, so use quadratic reciprocity
1034             assert (u < p);
1035             int t = u; u = p; p = t;
1036             if ((u & p & 2) != 0) // u = p = 3 (mod 4)?
1037                 j = -j;
1038             // Now u >= p, so it can be reduced
1039             u %= p;
1040         }
1041         return 0;
1042     }
1043 
1044     private static BigInteger lucasLehmerSequence(int z, BigInteger k, BigInteger n) {
1045         BigInteger d = BigInteger.valueOf(z);
1046         BigInteger u = ONE; BigInteger u2;
1047         BigInteger v = ONE; BigInteger v2;
1048 
1049         for (int i=k.bitLength()-2; i >= 0; i--) {
1050             u2 = u.multiply(v).mod(n);
1051 
1052             v2 = v.square().add(d.multiply(u.square())).mod(n);
1053             if (v2.testBit(0))
1054                 v2 = v2.subtract(n);
1055 
1056             v2 = v2.shiftRight(1);
1057 
1058             u = u2; v = v2;
1059             if (k.testBit(i)) {
1060                 u2 = u.add(v).mod(n);
1061                 if (u2.testBit(0))
1062                     u2 = u2.subtract(n);
1063 
1064                 u2 = u2.shiftRight(1);
1065                 v2 = v.add(d.multiply(u)).mod(n);
1066                 if (v2.testBit(0))
1067                     v2 = v2.subtract(n);
1068                 v2 = v2.shiftRight(1);
1069 
1070                 u = u2; v = v2;
1071             }
1072         }
1073         return u;
1074     }
1075 
1076     /**
1077      * Returns true iff this BigInteger passes the specified number of
1078      * Miller-Rabin tests. This test is taken from the DSA spec (NIST FIPS
1079      * 186-2).
1080      *
1081      * The following assumptions are made:
1082      * This BigInteger is a positive, odd number greater than 2.
1083      * iterations<=50.
1084      */
1085     private boolean passesMillerRabin(int iterations, Random rnd) {
1086         // Find a and m such that m is odd and this == 1 + 2**a * m
1087         BigInteger thisMinusOne = this.subtract(ONE);
1088         BigInteger m = thisMinusOne;
1089         int a = m.getLowestSetBit();
1090         m = m.shiftRight(a);
1091 
1092         // Do the tests
1093         if (rnd == null) {
1094             rnd = ThreadLocalRandom.current();
1095         }
1096         for (int i=0; i < iterations; i++) {
1097             // Generate a uniform random on (1, this)
1098             BigInteger b;
1099             do {
1100                 b = new BigInteger(this.bitLength(), rnd);
1101             } while (b.compareTo(ONE) <= 0 || b.compareTo(this) >= 0);
1102 
1103             int j = 0;
1104             BigInteger z = b.modPow(m, this);
1105             while (!((j == 0 && z.equals(ONE)) || z.equals(thisMinusOne))) {
1106                 if (j > 0 && z.equals(ONE) || ++j == a)
1107                     return false;
1108                 z = z.modPow(TWO, this);
1109             }
1110         }
1111         return true;
1112     }
1113 
1114     /**
1115      * This internal constructor differs from its public cousin
1116      * with the arguments reversed in two ways: it assumes that its
1117      * arguments are correct, and it doesn't copy the magnitude array.
1118      */
1119     BigInteger(int[] magnitude, int signum) {
1120         this.signum = (magnitude.length == 0 ? 0 : signum);
1121         this.mag = magnitude;
1122         if (mag.length >= MAX_MAG_LENGTH) {
1123             checkRange();
1124         }
1125     }
1126 
1127     /**
1128      * This private constructor is for internal use and assumes that its
1129      * arguments are correct.  The {@code magnitude} array is assumed to be
1130      * unchanged for the duration of the constructor call.
1131      */
1132     private BigInteger(byte[] magnitude, int signum) {
1133         this.signum = (magnitude.length == 0 ? 0 : signum);
1134         this.mag = stripLeadingZeroBytes(magnitude, 0, magnitude.length);
1135         if (mag.length >= MAX_MAG_LENGTH) {
1136             checkRange();
1137         }
1138     }
1139 
1140     /**
1141      * Throws an {@code ArithmeticException} if the {@code BigInteger} would be
1142      * out of the supported range.
1143      *
1144      * @throws ArithmeticException if {@code this} exceeds the supported range.
1145      */
1146     private void checkRange() {
1147         if (mag.length > MAX_MAG_LENGTH || mag.length == MAX_MAG_LENGTH && mag[0] < 0) {
1148             reportOverflow();
1149         }
1150     }
1151 
1152     private static void reportOverflow() {
1153         throw new ArithmeticException("BigInteger would overflow supported range");
1154     }
1155 
1156     //Static Factory Methods
1157 
1158     /**
1159      * Returns a BigInteger whose value is equal to that of the
1160      * specified {@code long}.  This "static factory method" is
1161      * provided in preference to a ({@code long}) constructor
1162      * because it allows for reuse of frequently used BigIntegers.
1163      *
1164      * @param  val value of the BigInteger to return.
1165      * @return a BigInteger with the specified value.
1166      */
1167     public static BigInteger valueOf(long val) {
1168         // If -MAX_CONSTANT < val < MAX_CONSTANT, return stashed constant
1169         if (val == 0)
1170             return ZERO;
1171         if (val > 0 && val <= MAX_CONSTANT)
1172             return posConst[(int) val];
1173         else if (val < 0 && val >= -MAX_CONSTANT)
1174             return negConst[(int) -val];
1175 
1176         return new BigInteger(val);
1177     }
1178 
1179     /**
1180      * Constructs a BigInteger with the specified value, which may not be zero.
1181      */
1182     private BigInteger(long val) {
1183         if (val < 0) {
1184             val = -val;
1185             signum = -1;
1186         } else {
1187             signum = 1;
1188         }
1189 
1190         int highWord = (int)(val >>> 32);
1191         if (highWord == 0) {
1192             mag = new int[1];
1193             mag[0] = (int)val;
1194         } else {
1195             mag = new int[2];
1196             mag[0] = highWord;
1197             mag[1] = (int)val;
1198         }
1199     }
1200 
1201     /**
1202      * Returns a BigInteger with the given two's complement representation.
1203      * Assumes that the input array will not be modified (the returned
1204      * BigInteger will reference the input array if feasible).
1205      */
1206     private static BigInteger valueOf(int val[]) {
1207         return (val[0] > 0 ? new BigInteger(val, 1) : new BigInteger(val));
1208     }
1209 
1210     // Constants
1211 
1212     /**
1213      * Initialize static constant array when class is loaded.
1214      */
1215     private static final int MAX_CONSTANT = 16;
1216     private static BigInteger posConst[] = new BigInteger[MAX_CONSTANT+1];
1217     private static BigInteger negConst[] = new BigInteger[MAX_CONSTANT+1];
1218 
1219     /**
1220      * The cache of powers of each radix.  This allows us to not have to
1221      * recalculate powers of radix^(2^n) more than once.  This speeds
1222      * Schoenhage recursive base conversion significantly.
1223      */
1224     private static volatile BigInteger[][] powerCache;
1225 
1226     /** The cache of logarithms of radices for base conversion. */
1227     private static final double[] logCache;
1228 
1229     /** The natural log of 2.  This is used in computing cache indices. */
1230     private static final double LOG_TWO = Math.log(2.0);
1231 
1232     static {
1233         for (int i = 1; i <= MAX_CONSTANT; i++) {
1234             int[] magnitude = new int[1];
1235             magnitude[0] = i;
1236             posConst[i] = new BigInteger(magnitude,  1);
1237             negConst[i] = new BigInteger(magnitude, -1);
1238         }
1239 
1240         /*
1241          * Initialize the cache of radix^(2^x) values used for base conversion
1242          * with just the very first value.  Additional values will be created
1243          * on demand.
1244          */
1245         powerCache = new BigInteger[Character.MAX_RADIX+1][];
1246         logCache = new double[Character.MAX_RADIX+1];
1247 
1248         for (int i=Character.MIN_RADIX; i <= Character.MAX_RADIX; i++) {
1249             powerCache[i] = new BigInteger[] { BigInteger.valueOf(i) };
1250             logCache[i] = Math.log(i);
1251         }
1252     }
1253 
1254     /**
1255      * The BigInteger constant zero.
1256      *
1257      * @since   1.2
1258      */
1259     public static final BigInteger ZERO = new BigInteger(new int[0], 0);
1260 
1261     /**
1262      * The BigInteger constant one.
1263      *
1264      * @since   1.2
1265      */
1266     public static final BigInteger ONE = valueOf(1);
1267 
1268     /**
1269      * The BigInteger constant two.  (Not exported.)
1270      */
1271     private static final BigInteger TWO = valueOf(2);
1272 
1273     /**
1274      * The BigInteger constant -1.  (Not exported.)
1275      */
1276     private static final BigInteger NEGATIVE_ONE = valueOf(-1);
1277 
1278     /**
1279      * The BigInteger constant ten.
1280      *
1281      * @since   1.5
1282      */
1283     public static final BigInteger TEN = valueOf(10);
1284 
1285     // Arithmetic Operations
1286 
1287     /**
1288      * Returns a BigInteger whose value is {@code (this + val)}.
1289      *
1290      * @param  val value to be added to this BigInteger.
1291      * @return {@code this + val}
1292      */
1293     public BigInteger add(BigInteger val) {
1294         if (val.signum == 0)
1295             return this;
1296         if (signum == 0)
1297             return val;
1298         if (val.signum == signum)
1299             return new BigInteger(add(mag, val.mag), signum);
1300 
1301         int cmp = compareMagnitude(val);
1302         if (cmp == 0)
1303             return ZERO;
1304         int[] resultMag = (cmp > 0 ? subtract(mag, val.mag)
1305                            : subtract(val.mag, mag));
1306         resultMag = trustedStripLeadingZeroInts(resultMag);
1307 
1308         return new BigInteger(resultMag, cmp == signum ? 1 : -1);
1309     }
1310 
1311     /**
1312      * Package private methods used by BigDecimal code to add a BigInteger
1313      * with a long. Assumes val is not equal to INFLATED.
1314      */
1315     BigInteger add(long val) {
1316         if (val == 0)
1317             return this;
1318         if (signum == 0)
1319             return valueOf(val);
1320         if (Long.signum(val) == signum)
1321             return new BigInteger(add(mag, Math.abs(val)), signum);
1322         int cmp = compareMagnitude(val);
1323         if (cmp == 0)
1324             return ZERO;
1325         int[] resultMag = (cmp > 0 ? subtract(mag, Math.abs(val)) : subtract(Math.abs(val), mag));
1326         resultMag = trustedStripLeadingZeroInts(resultMag);
1327         return new BigInteger(resultMag, cmp == signum ? 1 : -1);
1328     }
1329 
1330     /**
1331      * Adds the contents of the int array x and long value val. This
1332      * method allocates a new int array to hold the answer and returns
1333      * a reference to that array.  Assumes x.length &gt; 0 and val is
1334      * non-negative
1335      */
1336     private static int[] add(int[] x, long val) {
1337         int[] y;
1338         long sum = 0;
1339         int xIndex = x.length;
1340         int[] result;
1341         int highWord = (int)(val >>> 32);
1342         if (highWord == 0) {
1343             result = new int[xIndex];
1344             sum = (x[--xIndex] & LONG_MASK) + val;
1345             result[xIndex] = (int)sum;
1346         } else {
1347             if (xIndex == 1) {
1348                 result = new int[2];
1349                 sum = val  + (x[0] & LONG_MASK);
1350                 result[1] = (int)sum;
1351                 result[0] = (int)(sum >>> 32);
1352                 return result;
1353             } else {
1354                 result = new int[xIndex];
1355                 sum = (x[--xIndex] & LONG_MASK) + (val & LONG_MASK);
1356                 result[xIndex] = (int)sum;
1357                 sum = (x[--xIndex] & LONG_MASK) + (highWord & LONG_MASK) + (sum >>> 32);
1358                 result[xIndex] = (int)sum;
1359             }
1360         }
1361         // Copy remainder of longer number while carry propagation is required
1362         boolean carry = (sum >>> 32 != 0);
1363         while (xIndex > 0 && carry)
1364             carry = ((result[--xIndex] = x[xIndex] + 1) == 0);
1365         // Copy remainder of longer number
1366         while (xIndex > 0)
1367             result[--xIndex] = x[xIndex];
1368         // Grow result if necessary
1369         if (carry) {
1370             int bigger[] = new int[result.length + 1];
1371             System.arraycopy(result, 0, bigger, 1, result.length);
1372             bigger[0] = 0x01;
1373             return bigger;
1374         }
1375         return result;
1376     }
1377 
1378     /**
1379      * Adds the contents of the int arrays x and y. This method allocates
1380      * a new int array to hold the answer and returns a reference to that
1381      * array.
1382      */
1383     private static int[] add(int[] x, int[] y) {
1384         // If x is shorter, swap the two arrays
1385         if (x.length < y.length) {
1386             int[] tmp = x;
1387             x = y;
1388             y = tmp;
1389         }
1390 
1391         int xIndex = x.length;
1392         int yIndex = y.length;
1393         int result[] = new int[xIndex];
1394         long sum = 0;
1395         if (yIndex == 1) {
1396             sum = (x[--xIndex] & LONG_MASK) + (y[0] & LONG_MASK) ;
1397             result[xIndex] = (int)sum;
1398         } else {
1399             // Add common parts of both numbers
1400             while (yIndex > 0) {
1401                 sum = (x[--xIndex] & LONG_MASK) +
1402                       (y[--yIndex] & LONG_MASK) + (sum >>> 32);
1403                 result[xIndex] = (int)sum;
1404             }
1405         }
1406         // Copy remainder of longer number while carry propagation is required
1407         boolean carry = (sum >>> 32 != 0);
1408         while (xIndex > 0 && carry)
1409             carry = ((result[--xIndex] = x[xIndex] + 1) == 0);
1410 
1411         // Copy remainder of longer number
1412         while (xIndex > 0)
1413             result[--xIndex] = x[xIndex];
1414 
1415         // Grow result if necessary
1416         if (carry) {
1417             int bigger[] = new int[result.length + 1];
1418             System.arraycopy(result, 0, bigger, 1, result.length);
1419             bigger[0] = 0x01;
1420             return bigger;
1421         }
1422         return result;
1423     }
1424 
1425     private static int[] subtract(long val, int[] little) {
1426         int highWord = (int)(val >>> 32);
1427         if (highWord == 0) {
1428             int result[] = new int[1];
1429             result[0] = (int)(val - (little[0] & LONG_MASK));
1430             return result;
1431         } else {
1432             int result[] = new int[2];
1433             if (little.length == 1) {
1434                 long difference = ((int)val & LONG_MASK) - (little[0] & LONG_MASK);
1435                 result[1] = (int)difference;
1436                 // Subtract remainder of longer number while borrow propagates
1437                 boolean borrow = (difference >> 32 != 0);
1438                 if (borrow) {
1439                     result[0] = highWord - 1;
1440                 } else {        // Copy remainder of longer number
1441                     result[0] = highWord;
1442                 }
1443                 return result;
1444             } else { // little.length == 2
1445                 long difference = ((int)val & LONG_MASK) - (little[1] & LONG_MASK);
1446                 result[1] = (int)difference;
1447                 difference = (highWord & LONG_MASK) - (little[0] & LONG_MASK) + (difference >> 32);
1448                 result[0] = (int)difference;
1449                 return result;
1450             }
1451         }
1452     }
1453 
1454     /**
1455      * Subtracts the contents of the second argument (val) from the
1456      * first (big).  The first int array (big) must represent a larger number
1457      * than the second.  This method allocates the space necessary to hold the
1458      * answer.
1459      * assumes val &gt;= 0
1460      */
1461     private static int[] subtract(int[] big, long val) {
1462         int highWord = (int)(val >>> 32);
1463         int bigIndex = big.length;
1464         int result[] = new int[bigIndex];
1465         long difference = 0;
1466 
1467         if (highWord == 0) {
1468             difference = (big[--bigIndex] & LONG_MASK) - val;
1469             result[bigIndex] = (int)difference;
1470         } else {
1471             difference = (big[--bigIndex] & LONG_MASK) - (val & LONG_MASK);
1472             result[bigIndex] = (int)difference;
1473             difference = (big[--bigIndex] & LONG_MASK) - (highWord & LONG_MASK) + (difference >> 32);
1474             result[bigIndex] = (int)difference;
1475         }
1476 
1477         // Subtract remainder of longer number while borrow propagates
1478         boolean borrow = (difference >> 32 != 0);
1479         while (bigIndex > 0 && borrow)
1480             borrow = ((result[--bigIndex] = big[bigIndex] - 1) == -1);
1481 
1482         // Copy remainder of longer number
1483         while (bigIndex > 0)
1484             result[--bigIndex] = big[bigIndex];
1485 
1486         return result;
1487     }
1488 
1489     /**
1490      * Returns a BigInteger whose value is {@code (this - val)}.
1491      *
1492      * @param  val value to be subtracted from this BigInteger.
1493      * @return {@code this - val}
1494      */
1495     public BigInteger subtract(BigInteger val) {
1496         if (val.signum == 0)
1497             return this;
1498         if (signum == 0)
1499             return val.negate();
1500         if (val.signum != signum)
1501             return new BigInteger(add(mag, val.mag), signum);
1502 
1503         int cmp = compareMagnitude(val);
1504         if (cmp == 0)
1505             return ZERO;
1506         int[] resultMag = (cmp > 0 ? subtract(mag, val.mag)
1507                            : subtract(val.mag, mag));
1508         resultMag = trustedStripLeadingZeroInts(resultMag);
1509         return new BigInteger(resultMag, cmp == signum ? 1 : -1);
1510     }
1511 
1512     /**
1513      * Subtracts the contents of the second int arrays (little) from the
1514      * first (big).  The first int array (big) must represent a larger number
1515      * than the second.  This method allocates the space necessary to hold the
1516      * answer.
1517      */
1518     private static int[] subtract(int[] big, int[] little) {
1519         int bigIndex = big.length;
1520         int result[] = new int[bigIndex];
1521         int littleIndex = little.length;
1522         long difference = 0;
1523 
1524         // Subtract common parts of both numbers
1525         while (littleIndex > 0) {
1526             difference = (big[--bigIndex] & LONG_MASK) -
1527                          (little[--littleIndex] & LONG_MASK) +
1528                          (difference >> 32);
1529             result[bigIndex] = (int)difference;
1530         }
1531 
1532         // Subtract remainder of longer number while borrow propagates
1533         boolean borrow = (difference >> 32 != 0);
1534         while (bigIndex > 0 && borrow)
1535             borrow = ((result[--bigIndex] = big[bigIndex] - 1) == -1);
1536 
1537         // Copy remainder of longer number
1538         while (bigIndex > 0)
1539             result[--bigIndex] = big[bigIndex];
1540 
1541         return result;
1542     }
1543 
1544     /**
1545      * Returns a BigInteger whose value is {@code (this * val)}.
1546      *
1547      * @implNote An implementation may offer better algorithmic
1548      * performance when {@code val == this}.
1549      *
1550      * @param  val value to be multiplied by this BigInteger.
1551      * @return {@code this * val}
1552      */
1553     public BigInteger multiply(BigInteger val) {
1554         if (val.signum == 0 || signum == 0)
1555             return ZERO;
1556 
1557         int xlen = mag.length;
1558 
1559         if (val == this && xlen > MULTIPLY_SQUARE_THRESHOLD) {
1560             return square();
1561         }
1562 
1563         int ylen = val.mag.length;
1564 
1565         if ((xlen < KARATSUBA_THRESHOLD) || (ylen < KARATSUBA_THRESHOLD)) {
1566             int resultSign = signum == val.signum ? 1 : -1;
1567             if (val.mag.length == 1) {
1568                 return multiplyByInt(mag,val.mag[0], resultSign);
1569             }
1570             if (mag.length == 1) {
1571                 return multiplyByInt(val.mag,mag[0], resultSign);
1572             }
1573             int[] result = multiplyToLen(mag, xlen,
1574                                          val.mag, ylen, null);
1575             result = trustedStripLeadingZeroInts(result);
1576             return new BigInteger(result, resultSign);
1577         } else {
1578             if ((xlen < TOOM_COOK_THRESHOLD) && (ylen < TOOM_COOK_THRESHOLD)) {
1579                 return multiplyKaratsuba(this, val);
1580             } else {
1581                 return multiplyToomCook3(this, val);
1582             }
1583         }
1584     }
1585 
1586     private static BigInteger multiplyByInt(int[] x, int y, int sign) {
1587         if (Integer.bitCount(y) == 1) {
1588             return new BigInteger(shiftLeft(x,Integer.numberOfTrailingZeros(y)), sign);
1589         }
1590         int xlen = x.length;
1591         int[] rmag =  new int[xlen + 1];
1592         long carry = 0;
1593         long yl = y & LONG_MASK;
1594         int rstart = rmag.length - 1;
1595         for (int i = xlen - 1; i >= 0; i--) {
1596             long product = (x[i] & LONG_MASK) * yl + carry;
1597             rmag[rstart--] = (int)product;
1598             carry = product >>> 32;
1599         }
1600         if (carry == 0L) {
1601             rmag = java.util.Arrays.copyOfRange(rmag, 1, rmag.length);
1602         } else {
1603             rmag[rstart] = (int)carry;
1604         }
1605         return new BigInteger(rmag, sign);
1606     }
1607 
1608     /**
1609      * Package private methods used by BigDecimal code to multiply a BigInteger
1610      * with a long. Assumes v is not equal to INFLATED.
1611      */
1612     BigInteger multiply(long v) {
1613         if (v == 0 || signum == 0)
1614           return ZERO;
1615         if (v == BigDecimal.INFLATED)
1616             return multiply(BigInteger.valueOf(v));
1617         int rsign = (v > 0 ? signum : -signum);
1618         if (v < 0)
1619             v = -v;
1620         long dh = v >>> 32;      // higher order bits
1621         long dl = v & LONG_MASK; // lower order bits
1622 
1623         int xlen = mag.length;
1624         int[] value = mag;
1625         int[] rmag = (dh == 0L) ? (new int[xlen + 1]) : (new int[xlen + 2]);
1626         long carry = 0;
1627         int rstart = rmag.length - 1;
1628         for (int i = xlen - 1; i >= 0; i--) {
1629             long product = (value[i] & LONG_MASK) * dl + carry;
1630             rmag[rstart--] = (int)product;
1631             carry = product >>> 32;
1632         }
1633         rmag[rstart] = (int)carry;
1634         if (dh != 0L) {
1635             carry = 0;
1636             rstart = rmag.length - 2;
1637             for (int i = xlen - 1; i >= 0; i--) {
1638                 long product = (value[i] & LONG_MASK) * dh +
1639                     (rmag[rstart] & LONG_MASK) + carry;
1640                 rmag[rstart--] = (int)product;
1641                 carry = product >>> 32;
1642             }
1643             rmag[0] = (int)carry;
1644         }
1645         if (carry == 0L)
1646             rmag = java.util.Arrays.copyOfRange(rmag, 1, rmag.length);
1647         return new BigInteger(rmag, rsign);
1648     }
1649 
1650     /**
1651      * Multiplies int arrays x and y to the specified lengths and places
1652      * the result into z. There will be no leading zeros in the resultant array.
1653      */
1654     private static int[] multiplyToLen(int[] x, int xlen, int[] y, int ylen, int[] z) {
1655         multiplyToLenCheck(x, xlen);
1656         multiplyToLenCheck(y, ylen);
1657         return implMultiplyToLen(x, xlen, y, ylen, z);
1658     }
1659 
1660     @HotSpotIntrinsicCandidate
1661     private static int[] implMultiplyToLen(int[] x, int xlen, int[] y, int ylen, int[] z) {
1662         int xstart = xlen - 1;
1663         int ystart = ylen - 1;
1664 
1665         if (z == null || z.length < (xlen+ ylen))
1666             z = new int[xlen+ylen];
1667 
1668         long carry = 0;
1669         for (int j=ystart, k=ystart+1+xstart; j >= 0; j--, k--) {
1670             long product = (y[j] & LONG_MASK) *
1671                            (x[xstart] & LONG_MASK) + carry;
1672             z[k] = (int)product;
1673             carry = product >>> 32;
1674         }
1675         z[xstart] = (int)carry;
1676 
1677         for (int i = xstart-1; i >= 0; i--) {
1678             carry = 0;
1679             for (int j=ystart, k=ystart+1+i; j >= 0; j--, k--) {
1680                 long product = (y[j] & LONG_MASK) *
1681                                (x[i] & LONG_MASK) +
1682                                (z[k] & LONG_MASK) + carry;
1683                 z[k] = (int)product;
1684                 carry = product >>> 32;
1685             }
1686             z[i] = (int)carry;
1687         }
1688         return z;
1689     }
1690 
1691     private static void multiplyToLenCheck(int[] array, int length) {
1692         if (length <= 0) {
1693             return;  // not an error because multiplyToLen won't execute if len <= 0
1694         }
1695 
1696         Objects.requireNonNull(array);
1697 
1698         if (length > array.length) {
1699             throw new ArrayIndexOutOfBoundsException(length - 1);
1700         }
1701     }
1702 
1703     /**
1704      * Multiplies two BigIntegers using the Karatsuba multiplication
1705      * algorithm.  This is a recursive divide-and-conquer algorithm which is
1706      * more efficient for large numbers than what is commonly called the
1707      * "grade-school" algorithm used in multiplyToLen.  If the numbers to be
1708      * multiplied have length n, the "grade-school" algorithm has an
1709      * asymptotic complexity of O(n^2).  In contrast, the Karatsuba algorithm
1710      * has complexity of O(n^(log2(3))), or O(n^1.585).  It achieves this
1711      * increased performance by doing 3 multiplies instead of 4 when
1712      * evaluating the product.  As it has some overhead, should be used when
1713      * both numbers are larger than a certain threshold (found
1714      * experimentally).
1715      *
1716      * See:  http://en.wikipedia.org/wiki/Karatsuba_algorithm
1717      */
1718     private static BigInteger multiplyKaratsuba(BigInteger x, BigInteger y) {
1719         int xlen = x.mag.length;
1720         int ylen = y.mag.length;
1721 
1722         // The number of ints in each half of the number.
1723         int half = (Math.max(xlen, ylen)+1) / 2;
1724 
1725         // xl and yl are the lower halves of x and y respectively,
1726         // xh and yh are the upper halves.
1727         BigInteger xl = x.getLower(half);
1728         BigInteger xh = x.getUpper(half);
1729         BigInteger yl = y.getLower(half);
1730         BigInteger yh = y.getUpper(half);
1731 
1732         BigInteger p1 = xh.multiply(yh);  // p1 = xh*yh
1733         BigInteger p2 = xl.multiply(yl);  // p2 = xl*yl
1734 
1735         // p3=(xh+xl)*(yh+yl)
1736         BigInteger p3 = xh.add(xl).multiply(yh.add(yl));
1737 
1738         // result = p1 * 2^(32*2*half) + (p3 - p1 - p2) * 2^(32*half) + p2
1739         BigInteger result = p1.shiftLeft(32*half).add(p3.subtract(p1).subtract(p2)).shiftLeft(32*half).add(p2);
1740 
1741         if (x.signum != y.signum) {
1742             return result.negate();
1743         } else {
1744             return result;
1745         }
1746     }
1747 
1748     /**
1749      * Multiplies two BigIntegers using a 3-way Toom-Cook multiplication
1750      * algorithm.  This is a recursive divide-and-conquer algorithm which is
1751      * more efficient for large numbers than what is commonly called the
1752      * "grade-school" algorithm used in multiplyToLen.  If the numbers to be
1753      * multiplied have length n, the "grade-school" algorithm has an
1754      * asymptotic complexity of O(n^2).  In contrast, 3-way Toom-Cook has a
1755      * complexity of about O(n^1.465).  It achieves this increased asymptotic
1756      * performance by breaking each number into three parts and by doing 5
1757      * multiplies instead of 9 when evaluating the product.  Due to overhead
1758      * (additions, shifts, and one division) in the Toom-Cook algorithm, it
1759      * should only be used when both numbers are larger than a certain
1760      * threshold (found experimentally).  This threshold is generally larger
1761      * than that for Karatsuba multiplication, so this algorithm is generally
1762      * only used when numbers become significantly larger.
1763      *
1764      * The algorithm used is the "optimal" 3-way Toom-Cook algorithm outlined
1765      * by Marco Bodrato.
1766      *
1767      *  See: http://bodrato.it/toom-cook/
1768      *       http://bodrato.it/papers/#WAIFI2007
1769      *
1770      * "Towards Optimal Toom-Cook Multiplication for Univariate and
1771      * Multivariate Polynomials in Characteristic 2 and 0." by Marco BODRATO;
1772      * In C.Carlet and B.Sunar, Eds., "WAIFI'07 proceedings", p. 116-133,
1773      * LNCS #4547. Springer, Madrid, Spain, June 21-22, 2007.
1774      *
1775      */
1776     private static BigInteger multiplyToomCook3(BigInteger a, BigInteger b) {
1777         int alen = a.mag.length;
1778         int blen = b.mag.length;
1779 
1780         int largest = Math.max(alen, blen);
1781 
1782         // k is the size (in ints) of the lower-order slices.
1783         int k = (largest+2)/3;   // Equal to ceil(largest/3)
1784 
1785         // r is the size (in ints) of the highest-order slice.
1786         int r = largest - 2*k;
1787 
1788         // Obtain slices of the numbers. a2 and b2 are the most significant
1789         // bits of the numbers a and b, and a0 and b0 the least significant.
1790         BigInteger a0, a1, a2, b0, b1, b2;
1791         a2 = a.getToomSlice(k, r, 0, largest);
1792         a1 = a.getToomSlice(k, r, 1, largest);
1793         a0 = a.getToomSlice(k, r, 2, largest);
1794         b2 = b.getToomSlice(k, r, 0, largest);
1795         b1 = b.getToomSlice(k, r, 1, largest);
1796         b0 = b.getToomSlice(k, r, 2, largest);
1797 
1798         BigInteger v0, v1, v2, vm1, vinf, t1, t2, tm1, da1, db1;
1799 
1800         v0 = a0.multiply(b0);
1801         da1 = a2.add(a0);
1802         db1 = b2.add(b0);
1803         vm1 = da1.subtract(a1).multiply(db1.subtract(b1));
1804         da1 = da1.add(a1);
1805         db1 = db1.add(b1);
1806         v1 = da1.multiply(db1);
1807         v2 = da1.add(a2).shiftLeft(1).subtract(a0).multiply(
1808              db1.add(b2).shiftLeft(1).subtract(b0));
1809         vinf = a2.multiply(b2);
1810 
1811         // The algorithm requires two divisions by 2 and one by 3.
1812         // All divisions are known to be exact, that is, they do not produce
1813         // remainders, and all results are positive.  The divisions by 2 are
1814         // implemented as right shifts which are relatively efficient, leaving
1815         // only an exact division by 3, which is done by a specialized
1816         // linear-time algorithm.
1817         t2 = v2.subtract(vm1).exactDivideBy3();
1818         tm1 = v1.subtract(vm1).shiftRight(1);
1819         t1 = v1.subtract(v0);
1820         t2 = t2.subtract(t1).shiftRight(1);
1821         t1 = t1.subtract(tm1).subtract(vinf);
1822         t2 = t2.subtract(vinf.shiftLeft(1));
1823         tm1 = tm1.subtract(t2);
1824 
1825         // Number of bits to shift left.
1826         int ss = k*32;
1827 
1828         BigInteger result = vinf.shiftLeft(ss).add(t2).shiftLeft(ss).add(t1).shiftLeft(ss).add(tm1).shiftLeft(ss).add(v0);
1829 
1830         if (a.signum != b.signum) {
1831             return result.negate();
1832         } else {
1833             return result;
1834         }
1835     }
1836 
1837 
1838     /**
1839      * Returns a slice of a BigInteger for use in Toom-Cook multiplication.
1840      *
1841      * @param lowerSize The size of the lower-order bit slices.
1842      * @param upperSize The size of the higher-order bit slices.
1843      * @param slice The index of which slice is requested, which must be a
1844      * number from 0 to size-1. Slice 0 is the highest-order bits, and slice
1845      * size-1 are the lowest-order bits. Slice 0 may be of different size than
1846      * the other slices.
1847      * @param fullsize The size of the larger integer array, used to align
1848      * slices to the appropriate position when multiplying different-sized
1849      * numbers.
1850      */
1851     private BigInteger getToomSlice(int lowerSize, int upperSize, int slice,
1852                                     int fullsize) {
1853         int start, end, sliceSize, len, offset;
1854 
1855         len = mag.length;
1856         offset = fullsize - len;
1857 
1858         if (slice == 0) {
1859             start = 0 - offset;
1860             end = upperSize - 1 - offset;
1861         } else {
1862             start = upperSize + (slice-1)*lowerSize - offset;
1863             end = start + lowerSize - 1;
1864         }
1865 
1866         if (start < 0) {
1867             start = 0;
1868         }
1869         if (end < 0) {
1870            return ZERO;
1871         }
1872 
1873         sliceSize = (end-start) + 1;
1874 
1875         if (sliceSize <= 0) {
1876             return ZERO;
1877         }
1878 
1879         // While performing Toom-Cook, all slices are positive and
1880         // the sign is adjusted when the final number is composed.
1881         if (start == 0 && sliceSize >= len) {
1882             return this.abs();
1883         }
1884 
1885         int intSlice[] = new int[sliceSize];
1886         System.arraycopy(mag, start, intSlice, 0, sliceSize);
1887 
1888         return new BigInteger(trustedStripLeadingZeroInts(intSlice), 1);
1889     }
1890 
1891     /**
1892      * Does an exact division (that is, the remainder is known to be zero)
1893      * of the specified number by 3.  This is used in Toom-Cook
1894      * multiplication.  This is an efficient algorithm that runs in linear
1895      * time.  If the argument is not exactly divisible by 3, results are
1896      * undefined.  Note that this is expected to be called with positive
1897      * arguments only.
1898      */
1899     private BigInteger exactDivideBy3() {
1900         int len = mag.length;
1901         int[] result = new int[len];
1902         long x, w, q, borrow;
1903         borrow = 0L;
1904         for (int i=len-1; i >= 0; i--) {
1905             x = (mag[i] & LONG_MASK);
1906             w = x - borrow;
1907             if (borrow > x) {      // Did we make the number go negative?
1908                 borrow = 1L;
1909             } else {
1910                 borrow = 0L;
1911             }
1912 
1913             // 0xAAAAAAAB is the modular inverse of 3 (mod 2^32).  Thus,
1914             // the effect of this is to divide by 3 (mod 2^32).
1915             // This is much faster than division on most architectures.
1916             q = (w * 0xAAAAAAABL) & LONG_MASK;
1917             result[i] = (int) q;
1918 
1919             // Now check the borrow. The second check can of course be
1920             // eliminated if the first fails.
1921             if (q >= 0x55555556L) {
1922                 borrow++;
1923                 if (q >= 0xAAAAAAABL)
1924                     borrow++;
1925             }
1926         }
1927         result = trustedStripLeadingZeroInts(result);
1928         return new BigInteger(result, signum);
1929     }
1930 
1931     /**
1932      * Returns a new BigInteger representing n lower ints of the number.
1933      * This is used by Karatsuba multiplication and Karatsuba squaring.
1934      */
1935     private BigInteger getLower(int n) {
1936         int len = mag.length;
1937 
1938         if (len <= n) {
1939             return abs();
1940         }
1941 
1942         int lowerInts[] = new int[n];
1943         System.arraycopy(mag, len-n, lowerInts, 0, n);
1944 
1945         return new BigInteger(trustedStripLeadingZeroInts(lowerInts), 1);
1946     }
1947 
1948     /**
1949      * Returns a new BigInteger representing mag.length-n upper
1950      * ints of the number.  This is used by Karatsuba multiplication and
1951      * Karatsuba squaring.
1952      */
1953     private BigInteger getUpper(int n) {
1954         int len = mag.length;
1955 
1956         if (len <= n) {
1957             return ZERO;
1958         }
1959 
1960         int upperLen = len - n;
1961         int upperInts[] = new int[upperLen];
1962         System.arraycopy(mag, 0, upperInts, 0, upperLen);
1963 
1964         return new BigInteger(trustedStripLeadingZeroInts(upperInts), 1);
1965     }
1966 
1967     // Squaring
1968 
1969     /**
1970      * Returns a BigInteger whose value is {@code (this<sup>2</sup>)}.
1971      *
1972      * @return {@code this<sup>2</sup>}
1973      */
1974     private BigInteger square() {
1975         if (signum == 0) {
1976             return ZERO;
1977         }
1978         int len = mag.length;
1979 
1980         if (len < KARATSUBA_SQUARE_THRESHOLD) {
1981             int[] z = squareToLen(mag, len, null);
1982             return new BigInteger(trustedStripLeadingZeroInts(z), 1);
1983         } else {
1984             if (len < TOOM_COOK_SQUARE_THRESHOLD) {
1985                 return squareKaratsuba();
1986             } else {
1987                 return squareToomCook3();
1988             }
1989         }
1990     }
1991 
1992     /**
1993      * Squares the contents of the int array x. The result is placed into the
1994      * int array z.  The contents of x are not changed.
1995      */
1996     private static final int[] squareToLen(int[] x, int len, int[] z) {
1997          int zlen = len << 1;
1998          if (z == null || z.length < zlen)
1999              z = new int[zlen];
2000 
2001          // Execute checks before calling intrinsified method.
2002          implSquareToLenChecks(x, len, z, zlen);
2003          return implSquareToLen(x, len, z, zlen);
2004      }
2005 
2006      /**
2007       * Parameters validation.
2008       */
2009      private static void implSquareToLenChecks(int[] x, int len, int[] z, int zlen) throws RuntimeException {
2010          if (len < 1) {
2011              throw new IllegalArgumentException("invalid input length: " + len);
2012          }
2013          if (len > x.length) {
2014              throw new IllegalArgumentException("input length out of bound: " +
2015                                         len + " > " + x.length);
2016          }
2017          if (len * 2 > z.length) {
2018              throw new IllegalArgumentException("input length out of bound: " +
2019                                         (len * 2) + " > " + z.length);
2020          }
2021          if (zlen < 1) {
2022              throw new IllegalArgumentException("invalid input length: " + zlen);
2023          }
2024          if (zlen > z.length) {
2025              throw new IllegalArgumentException("input length out of bound: " +
2026                                         len + " > " + z.length);
2027          }
2028      }
2029 
2030      /**
2031       * Java Runtime may use intrinsic for this method.
2032       */
2033      @HotSpotIntrinsicCandidate
2034      private static final int[] implSquareToLen(int[] x, int len, int[] z, int zlen) {
2035         /*
2036          * The algorithm used here is adapted from Colin Plumb's C library.
2037          * Technique: Consider the partial products in the multiplication
2038          * of "abcde" by itself:
2039          *
2040          *               a  b  c  d  e
2041          *            *  a  b  c  d  e
2042          *          ==================
2043          *              ae be ce de ee
2044          *           ad bd cd dd de
2045          *        ac bc cc cd ce
2046          *     ab bb bc bd be
2047          *  aa ab ac ad ae
2048          *
2049          * Note that everything above the main diagonal:
2050          *              ae be ce de = (abcd) * e
2051          *           ad bd cd       = (abc) * d
2052          *        ac bc             = (ab) * c
2053          *     ab                   = (a) * b
2054          *
2055          * is a copy of everything below the main diagonal:
2056          *                       de
2057          *                 cd ce
2058          *           bc bd be
2059          *     ab ac ad ae
2060          *
2061          * Thus, the sum is 2 * (off the diagonal) + diagonal.
2062          *
2063          * This is accumulated beginning with the diagonal (which
2064          * consist of the squares of the digits of the input), which is then
2065          * divided by two, the off-diagonal added, and multiplied by two
2066          * again.  The low bit is simply a copy of the low bit of the
2067          * input, so it doesn't need special care.
2068          */
2069 
2070         // Store the squares, right shifted one bit (i.e., divided by 2)
2071         int lastProductLowWord = 0;
2072         for (int j=0, i=0; j < len; j++) {
2073             long piece = (x[j] & LONG_MASK);
2074             long product = piece * piece;
2075             z[i++] = (lastProductLowWord << 31) | (int)(product >>> 33);
2076             z[i++] = (int)(product >>> 1);
2077             lastProductLowWord = (int)product;
2078         }
2079 
2080         // Add in off-diagonal sums
2081         for (int i=len, offset=1; i > 0; i--, offset+=2) {
2082             int t = x[i-1];
2083             t = mulAdd(z, x, offset, i-1, t);
2084             addOne(z, offset-1, i, t);
2085         }
2086 
2087         // Shift back up and set low bit
2088         primitiveLeftShift(z, zlen, 1);
2089         z[zlen-1] |= x[len-1] & 1;
2090 
2091         return z;
2092     }
2093 
2094     /**
2095      * Squares a BigInteger using the Karatsuba squaring algorithm.  It should
2096      * be used when both numbers are larger than a certain threshold (found
2097      * experimentally).  It is a recursive divide-and-conquer algorithm that
2098      * has better asymptotic performance than the algorithm used in
2099      * squareToLen.
2100      */
2101     private BigInteger squareKaratsuba() {
2102         int half = (mag.length+1) / 2;
2103 
2104         BigInteger xl = getLower(half);
2105         BigInteger xh = getUpper(half);
2106 
2107         BigInteger xhs = xh.square();  // xhs = xh^2
2108         BigInteger xls = xl.square();  // xls = xl^2
2109 
2110         // xh^2 << 64  +  (((xl+xh)^2 - (xh^2 + xl^2)) << 32) + xl^2
2111         return xhs.shiftLeft(half*32).add(xl.add(xh).square().subtract(xhs.add(xls))).shiftLeft(half*32).add(xls);
2112     }
2113 
2114     /**
2115      * Squares a BigInteger using the 3-way Toom-Cook squaring algorithm.  It
2116      * should be used when both numbers are larger than a certain threshold
2117      * (found experimentally).  It is a recursive divide-and-conquer algorithm
2118      * that has better asymptotic performance than the algorithm used in
2119      * squareToLen or squareKaratsuba.
2120      */
2121     private BigInteger squareToomCook3() {
2122         int len = mag.length;
2123 
2124         // k is the size (in ints) of the lower-order slices.
2125         int k = (len+2)/3;   // Equal to ceil(largest/3)
2126 
2127         // r is the size (in ints) of the highest-order slice.
2128         int r = len - 2*k;
2129 
2130         // Obtain slices of the numbers. a2 is the most significant
2131         // bits of the number, and a0 the least significant.
2132         BigInteger a0, a1, a2;
2133         a2 = getToomSlice(k, r, 0, len);
2134         a1 = getToomSlice(k, r, 1, len);
2135         a0 = getToomSlice(k, r, 2, len);
2136         BigInteger v0, v1, v2, vm1, vinf, t1, t2, tm1, da1;
2137 
2138         v0 = a0.square();
2139         da1 = a2.add(a0);
2140         vm1 = da1.subtract(a1).square();
2141         da1 = da1.add(a1);
2142         v1 = da1.square();
2143         vinf = a2.square();
2144         v2 = da1.add(a2).shiftLeft(1).subtract(a0).square();
2145 
2146         // The algorithm requires two divisions by 2 and one by 3.
2147         // All divisions are known to be exact, that is, they do not produce
2148         // remainders, and all results are positive.  The divisions by 2 are
2149         // implemented as right shifts which are relatively efficient, leaving
2150         // only a division by 3.
2151         // The division by 3 is done by an optimized algorithm for this case.
2152         t2 = v2.subtract(vm1).exactDivideBy3();
2153         tm1 = v1.subtract(vm1).shiftRight(1);
2154         t1 = v1.subtract(v0);
2155         t2 = t2.subtract(t1).shiftRight(1);
2156         t1 = t1.subtract(tm1).subtract(vinf);
2157         t2 = t2.subtract(vinf.shiftLeft(1));
2158         tm1 = tm1.subtract(t2);
2159 
2160         // Number of bits to shift left.
2161         int ss = k*32;
2162 
2163         return vinf.shiftLeft(ss).add(t2).shiftLeft(ss).add(t1).shiftLeft(ss).add(tm1).shiftLeft(ss).add(v0);
2164     }
2165 
2166     // Division
2167 
2168     /**
2169      * Returns a BigInteger whose value is {@code (this / val)}.
2170      *
2171      * @param  val value by which this BigInteger is to be divided.
2172      * @return {@code this / val}
2173      * @throws ArithmeticException if {@code val} is zero.
2174      */
2175     public BigInteger divide(BigInteger val) {
2176         if (val.mag.length < BURNIKEL_ZIEGLER_THRESHOLD ||
2177                 mag.length - val.mag.length < BURNIKEL_ZIEGLER_OFFSET) {
2178             return divideKnuth(val);
2179         } else {
2180             return divideBurnikelZiegler(val);
2181         }
2182     }
2183 
2184     /**
2185      * Returns a BigInteger whose value is {@code (this / val)} using an O(n^2) algorithm from Knuth.
2186      *
2187      * @param  val value by which this BigInteger is to be divided.
2188      * @return {@code this / val}
2189      * @throws ArithmeticException if {@code val} is zero.
2190      * @see MutableBigInteger#divideKnuth(MutableBigInteger, MutableBigInteger, boolean)
2191      */
2192     private BigInteger divideKnuth(BigInteger val) {
2193         MutableBigInteger q = new MutableBigInteger(),
2194                           a = new MutableBigInteger(this.mag),
2195                           b = new MutableBigInteger(val.mag);
2196 
2197         a.divideKnuth(b, q, false);
2198         return q.toBigInteger(this.signum * val.signum);
2199     }
2200 
2201     /**
2202      * Returns an array of two BigIntegers containing {@code (this / val)}
2203      * followed by {@code (this % val)}.
2204      *
2205      * @param  val value by which this BigInteger is to be divided, and the
2206      *         remainder computed.
2207      * @return an array of two BigIntegers: the quotient {@code (this / val)}
2208      *         is the initial element, and the remainder {@code (this % val)}
2209      *         is the final element.
2210      * @throws ArithmeticException if {@code val} is zero.
2211      */
2212     public BigInteger[] divideAndRemainder(BigInteger val) {
2213         if (val.mag.length < BURNIKEL_ZIEGLER_THRESHOLD ||
2214                 mag.length - val.mag.length < BURNIKEL_ZIEGLER_OFFSET) {
2215             return divideAndRemainderKnuth(val);
2216         } else {
2217             return divideAndRemainderBurnikelZiegler(val);
2218         }
2219     }
2220 
2221     /** Long division */
2222     private BigInteger[] divideAndRemainderKnuth(BigInteger val) {
2223         BigInteger[] result = new BigInteger[2];
2224         MutableBigInteger q = new MutableBigInteger(),
2225                           a = new MutableBigInteger(this.mag),
2226                           b = new MutableBigInteger(val.mag);
2227         MutableBigInteger r = a.divideKnuth(b, q);
2228         result[0] = q.toBigInteger(this.signum == val.signum ? 1 : -1);
2229         result[1] = r.toBigInteger(this.signum);
2230         return result;
2231     }
2232 
2233     /**
2234      * Returns a BigInteger whose value is {@code (this % val)}.
2235      *
2236      * @param  val value by which this BigInteger is to be divided, and the
2237      *         remainder computed.
2238      * @return {@code this % val}
2239      * @throws ArithmeticException if {@code val} is zero.
2240      */
2241     public BigInteger remainder(BigInteger val) {
2242         if (val.mag.length < BURNIKEL_ZIEGLER_THRESHOLD ||
2243                 mag.length - val.mag.length < BURNIKEL_ZIEGLER_OFFSET) {
2244             return remainderKnuth(val);
2245         } else {
2246             return remainderBurnikelZiegler(val);
2247         }
2248     }
2249 
2250     /** Long division */
2251     private BigInteger remainderKnuth(BigInteger val) {
2252         MutableBigInteger q = new MutableBigInteger(),
2253                           a = new MutableBigInteger(this.mag),
2254                           b = new MutableBigInteger(val.mag);
2255 
2256         return a.divideKnuth(b, q).toBigInteger(this.signum);
2257     }
2258 
2259     /**
2260      * Calculates {@code this / val} using the Burnikel-Ziegler algorithm.
2261      * @param  val the divisor
2262      * @return {@code this / val}
2263      */
2264     private BigInteger divideBurnikelZiegler(BigInteger val) {
2265         return divideAndRemainderBurnikelZiegler(val)[0];
2266     }
2267 
2268     /**
2269      * Calculates {@code this % val} using the Burnikel-Ziegler algorithm.
2270      * @param val the divisor
2271      * @return {@code this % val}
2272      */
2273     private BigInteger remainderBurnikelZiegler(BigInteger val) {
2274         return divideAndRemainderBurnikelZiegler(val)[1];
2275     }
2276 
2277     /**
2278      * Computes {@code this / val} and {@code this % val} using the
2279      * Burnikel-Ziegler algorithm.
2280      * @param val the divisor
2281      * @return an array containing the quotient and remainder
2282      */
2283     private BigInteger[] divideAndRemainderBurnikelZiegler(BigInteger val) {
2284         MutableBigInteger q = new MutableBigInteger();
2285         MutableBigInteger r = new MutableBigInteger(this).divideAndRemainderBurnikelZiegler(new MutableBigInteger(val), q);
2286         BigInteger qBigInt = q.isZero() ? ZERO : q.toBigInteger(signum*val.signum);
2287         BigInteger rBigInt = r.isZero() ? ZERO : r.toBigInteger(signum);
2288         return new BigInteger[] {qBigInt, rBigInt};
2289     }
2290 
2291     /**
2292      * Returns a BigInteger whose value is <code>(this<sup>exponent</sup>)</code>.
2293      * Note that {@code exponent} is an integer rather than a BigInteger.
2294      *
2295      * @param  exponent exponent to which this BigInteger is to be raised.
2296      * @return <code>this<sup>exponent</sup></code>
2297      * @throws ArithmeticException {@code exponent} is negative.  (This would
2298      *         cause the operation to yield a non-integer value.)
2299      */
2300     public BigInteger pow(int exponent) {
2301         if (exponent < 0) {
2302             throw new ArithmeticException("Negative exponent");
2303         }
2304         if (signum == 0) {
2305             return (exponent == 0 ? ONE : this);
2306         }
2307 
2308         BigInteger partToSquare = this.abs();
2309 
2310         // Factor out powers of two from the base, as the exponentiation of
2311         // these can be done by left shifts only.
2312         // The remaining part can then be exponentiated faster.  The
2313         // powers of two will be multiplied back at the end.
2314         int powersOfTwo = partToSquare.getLowestSetBit();
2315         long bitsToShift = (long)powersOfTwo * exponent;
2316         if (bitsToShift > Integer.MAX_VALUE) {
2317             reportOverflow();
2318         }
2319 
2320         int remainingBits;
2321 
2322         // Factor the powers of two out quickly by shifting right, if needed.
2323         if (powersOfTwo > 0) {
2324             partToSquare = partToSquare.shiftRight(powersOfTwo);
2325             remainingBits = partToSquare.bitLength();
2326             if (remainingBits == 1) {  // Nothing left but +/- 1?
2327                 if (signum < 0 && (exponent&1) == 1) {
2328                     return NEGATIVE_ONE.shiftLeft(powersOfTwo*exponent);
2329                 } else {
2330                     return ONE.shiftLeft(powersOfTwo*exponent);
2331                 }
2332             }
2333         } else {
2334             remainingBits = partToSquare.bitLength();
2335             if (remainingBits == 1) { // Nothing left but +/- 1?
2336                 if (signum < 0  && (exponent&1) == 1) {
2337                     return NEGATIVE_ONE;
2338                 } else {
2339                     return ONE;
2340                 }
2341             }
2342         }
2343 
2344         // This is a quick way to approximate the size of the result,
2345         // similar to doing log2[n] * exponent.  This will give an upper bound
2346         // of how big the result can be, and which algorithm to use.
2347         long scaleFactor = (long)remainingBits * exponent;
2348 
2349         // Use slightly different algorithms for small and large operands.
2350         // See if the result will safely fit into a long. (Largest 2^63-1)
2351         if (partToSquare.mag.length == 1 && scaleFactor <= 62) {
2352             // Small number algorithm.  Everything fits into a long.
2353             int newSign = (signum <0  && (exponent&1) == 1 ? -1 : 1);
2354             long result = 1;
2355             long baseToPow2 = partToSquare.mag[0] & LONG_MASK;
2356 
2357             int workingExponent = exponent;
2358 
2359             // Perform exponentiation using repeated squaring trick
2360             while (workingExponent != 0) {
2361                 if ((workingExponent & 1) == 1) {
2362                     result = result * baseToPow2;
2363                 }
2364 
2365                 if ((workingExponent >>>= 1) != 0) {
2366                     baseToPow2 = baseToPow2 * baseToPow2;
2367                 }
2368             }
2369 
2370             // Multiply back the powers of two (quickly, by shifting left)
2371             if (powersOfTwo > 0) {
2372                 if (bitsToShift + scaleFactor <= 62) { // Fits in long?
2373                     return valueOf((result << bitsToShift) * newSign);
2374                 } else {
2375                     return valueOf(result*newSign).shiftLeft((int) bitsToShift);
2376                 }
2377             }
2378             else {
2379                 return valueOf(result*newSign);
2380             }
2381         } else {
2382             // Large number algorithm.  This is basically identical to
2383             // the algorithm above, but calls multiply() and square()
2384             // which may use more efficient algorithms for large numbers.
2385             BigInteger answer = ONE;
2386 
2387             int workingExponent = exponent;
2388             // Perform exponentiation using repeated squaring trick
2389             while (workingExponent != 0) {
2390                 if ((workingExponent & 1) == 1) {
2391                     answer = answer.multiply(partToSquare);
2392                 }
2393 
2394                 if ((workingExponent >>>= 1) != 0) {
2395                     partToSquare = partToSquare.square();
2396                 }
2397             }
2398             // Multiply back the (exponentiated) powers of two (quickly,
2399             // by shifting left)
2400             if (powersOfTwo > 0) {
2401                 answer = answer.shiftLeft(powersOfTwo*exponent);
2402             }
2403 
2404             if (signum < 0 && (exponent&1) == 1) {
2405                 return answer.negate();
2406             } else {
2407                 return answer;
2408             }
2409         }
2410     }
2411 
2412     /**
2413      * Returns a BigInteger whose value is the greatest common divisor of
2414      * {@code abs(this)} and {@code abs(val)}.  Returns 0 if
2415      * {@code this == 0 && val == 0}.
2416      *
2417      * @param  val value with which the GCD is to be computed.
2418      * @return {@code GCD(abs(this), abs(val))}
2419      */
2420     public BigInteger gcd(BigInteger val) {
2421         if (val.signum == 0)
2422             return this.abs();
2423         else if (this.signum == 0)
2424             return val.abs();
2425 
2426         MutableBigInteger a = new MutableBigInteger(this);
2427         MutableBigInteger b = new MutableBigInteger(val);
2428 
2429         MutableBigInteger result = a.hybridGCD(b);
2430 
2431         return result.toBigInteger(1);
2432     }
2433 
2434     /**
2435      * Package private method to return bit length for an integer.
2436      */
2437     static int bitLengthForInt(int n) {
2438         return 32 - Integer.numberOfLeadingZeros(n);
2439     }
2440 
2441     /**
2442      * Left shift int array a up to len by n bits. Returns the array that
2443      * results from the shift since space may have to be reallocated.
2444      */
2445     private static int[] leftShift(int[] a, int len, int n) {
2446         int nInts = n >>> 5;
2447         int nBits = n&0x1F;
2448         int bitsInHighWord = bitLengthForInt(a[0]);
2449 
2450         // If shift can be done without recopy, do so
2451         if (n <= (32-bitsInHighWord)) {
2452             primitiveLeftShift(a, len, nBits);
2453             return a;
2454         } else { // Array must be resized
2455             if (nBits <= (32-bitsInHighWord)) {
2456                 int result[] = new int[nInts+len];
2457                 System.arraycopy(a, 0, result, 0, len);
2458                 primitiveLeftShift(result, result.length, nBits);
2459                 return result;
2460             } else {
2461                 int result[] = new int[nInts+len+1];
2462                 System.arraycopy(a, 0, result, 0, len);
2463                 primitiveRightShift(result, result.length, 32 - nBits);
2464                 return result;
2465             }
2466         }
2467     }
2468 
2469     // shifts a up to len right n bits assumes no leading zeros, 0<n<32
2470     static void primitiveRightShift(int[] a, int len, int n) {
2471         int n2 = 32 - n;
2472         for (int i=len-1, c=a[i]; i > 0; i--) {
2473             int b = c;
2474             c = a[i-1];
2475             a[i] = (c << n2) | (b >>> n);
2476         }
2477         a[0] >>>= n;
2478     }
2479 
2480     // shifts a up to len left n bits assumes no leading zeros, 0<=n<32
2481     static void primitiveLeftShift(int[] a, int len, int n) {
2482         if (len == 0 || n == 0)
2483             return;
2484 
2485         int n2 = 32 - n;
2486         for (int i=0, c=a[i], m=i+len-1; i < m; i++) {
2487             int b = c;
2488             c = a[i+1];
2489             a[i] = (b << n) | (c >>> n2);
2490         }
2491         a[len-1] <<= n;
2492     }
2493 
2494     /**
2495      * Calculate bitlength of contents of the first len elements an int array,
2496      * assuming there are no leading zero ints.
2497      */
2498     private static int bitLength(int[] val, int len) {
2499         if (len == 0)
2500             return 0;
2501         return ((len - 1) << 5) + bitLengthForInt(val[0]);
2502     }
2503 
2504     /**
2505      * Returns a BigInteger whose value is the absolute value of this
2506      * BigInteger.
2507      *
2508      * @return {@code abs(this)}
2509      */
2510     public BigInteger abs() {
2511         return (signum >= 0 ? this : this.negate());
2512     }
2513 
2514     /**
2515      * Returns a BigInteger whose value is {@code (-this)}.
2516      *
2517      * @return {@code -this}
2518      */
2519     public BigInteger negate() {
2520         return new BigInteger(this.mag, -this.signum);
2521     }
2522 
2523     /**
2524      * Returns the signum function of this BigInteger.
2525      *
2526      * @return -1, 0 or 1 as the value of this BigInteger is negative, zero or
2527      *         positive.
2528      */
2529     public int signum() {
2530         return this.signum;
2531     }
2532 
2533     // Modular Arithmetic Operations
2534 
2535     /**
2536      * Returns a BigInteger whose value is {@code (this mod m}).  This method
2537      * differs from {@code remainder} in that it always returns a
2538      * <i>non-negative</i> BigInteger.
2539      *
2540      * @param  m the modulus.
2541      * @return {@code this mod m}
2542      * @throws ArithmeticException {@code m} &le; 0
2543      * @see    #remainder
2544      */
2545     public BigInteger mod(BigInteger m) {
2546         if (m.signum <= 0)
2547             throw new ArithmeticException("BigInteger: modulus not positive");
2548 
2549         BigInteger result = this.remainder(m);
2550         return (result.signum >= 0 ? result : result.add(m));
2551     }
2552 
2553     /**
2554      * Returns a BigInteger whose value is
2555      * <code>(this<sup>exponent</sup> mod m)</code>.  (Unlike {@code pow}, this
2556      * method permits negative exponents.)
2557      *
2558      * @param  exponent the exponent.
2559      * @param  m the modulus.
2560      * @return <code>this<sup>exponent</sup> mod m</code>
2561      * @throws ArithmeticException {@code m} &le; 0 or the exponent is
2562      *         negative and this BigInteger is not <i>relatively
2563      *         prime</i> to {@code m}.
2564      * @see    #modInverse
2565      */
2566     public BigInteger modPow(BigInteger exponent, BigInteger m) {
2567         if (m.signum <= 0)
2568             throw new ArithmeticException("BigInteger: modulus not positive");
2569 
2570         // Trivial cases
2571         if (exponent.signum == 0)
2572             return (m.equals(ONE) ? ZERO : ONE);
2573 
2574         if (this.equals(ONE))
2575             return (m.equals(ONE) ? ZERO : ONE);
2576 
2577         if (this.equals(ZERO) && exponent.signum >= 0)
2578             return ZERO;
2579 
2580         if (this.equals(negConst[1]) && (!exponent.testBit(0)))
2581             return (m.equals(ONE) ? ZERO : ONE);
2582 
2583         boolean invertResult;
2584         if ((invertResult = (exponent.signum < 0)))
2585             exponent = exponent.negate();
2586 
2587         BigInteger base = (this.signum < 0 || this.compareTo(m) >= 0
2588                            ? this.mod(m) : this);
2589         BigInteger result;
2590         if (m.testBit(0)) { // odd modulus
2591             result = base.oddModPow(exponent, m);
2592         } else {
2593             /*
2594              * Even modulus.  Tear it into an "odd part" (m1) and power of two
2595              * (m2), exponentiate mod m1, manually exponentiate mod m2, and
2596              * use Chinese Remainder Theorem to combine results.
2597              */
2598 
2599             // Tear m apart into odd part (m1) and power of 2 (m2)
2600             int p = m.getLowestSetBit();   // Max pow of 2 that divides m
2601 
2602             BigInteger m1 = m.shiftRight(p);  // m/2**p
2603             BigInteger m2 = ONE.shiftLeft(p); // 2**p
2604 
2605             // Calculate new base from m1
2606             BigInteger base2 = (this.signum < 0 || this.compareTo(m1) >= 0
2607                                 ? this.mod(m1) : this);
2608 
2609             // Caculate (base ** exponent) mod m1.
2610             BigInteger a1 = (m1.equals(ONE) ? ZERO :
2611                              base2.oddModPow(exponent, m1));
2612 
2613             // Calculate (this ** exponent) mod m2
2614             BigInteger a2 = base.modPow2(exponent, p);
2615 
2616             // Combine results using Chinese Remainder Theorem
2617             BigInteger y1 = m2.modInverse(m1);
2618             BigInteger y2 = m1.modInverse(m2);
2619 
2620             if (m.mag.length < MAX_MAG_LENGTH / 2) {
2621                 result = a1.multiply(m2).multiply(y1).add(a2.multiply(m1).multiply(y2)).mod(m);
2622             } else {
2623                 MutableBigInteger t1 = new MutableBigInteger();
2624                 new MutableBigInteger(a1.multiply(m2)).multiply(new MutableBigInteger(y1), t1);
2625                 MutableBigInteger t2 = new MutableBigInteger();
2626                 new MutableBigInteger(a2.multiply(m1)).multiply(new MutableBigInteger(y2), t2);
2627                 t1.add(t2);
2628                 MutableBigInteger q = new MutableBigInteger();
2629                 result = t1.divide(new MutableBigInteger(m), q).toBigInteger();
2630             }
2631         }
2632 
2633         return (invertResult ? result.modInverse(m) : result);
2634     }
2635 
2636     // Montgomery multiplication.  These are wrappers for
2637     // implMontgomeryXX routines which are expected to be replaced by
2638     // virtual machine intrinsics.  We don't use the intrinsics for
2639     // very large operands: MONTGOMERY_INTRINSIC_THRESHOLD should be
2640     // larger than any reasonable crypto key.
2641     private static int[] montgomeryMultiply(int[] a, int[] b, int[] n, int len, long inv,
2642                                             int[] product) {
2643         implMontgomeryMultiplyChecks(a, b, n, len, product);
2644         if (len > MONTGOMERY_INTRINSIC_THRESHOLD) {
2645             // Very long argument: do not use an intrinsic
2646             product = multiplyToLen(a, len, b, len, product);
2647             return montReduce(product, n, len, (int)inv);
2648         } else {
2649             return implMontgomeryMultiply(a, b, n, len, inv, materialize(product, len));
2650         }
2651     }
2652     private static int[] montgomerySquare(int[] a, int[] n, int len, long inv,
2653                                           int[] product) {
2654         implMontgomeryMultiplyChecks(a, a, n, len, product);
2655         if (len > MONTGOMERY_INTRINSIC_THRESHOLD) {
2656             // Very long argument: do not use an intrinsic
2657             product = squareToLen(a, len, product);
2658             return montReduce(product, n, len, (int)inv);
2659         } else {
2660             return implMontgomerySquare(a, n, len, inv, materialize(product, len));
2661         }
2662     }
2663 
2664     // Range-check everything.
2665     private static void implMontgomeryMultiplyChecks
2666         (int[] a, int[] b, int[] n, int len, int[] product) throws RuntimeException {
2667         if (len % 2 != 0) {
2668             throw new IllegalArgumentException("input array length must be even: " + len);
2669         }
2670 
2671         if (len < 1) {
2672             throw new IllegalArgumentException("invalid input length: " + len);
2673         }
2674 
2675         if (len > a.length ||
2676             len > b.length ||
2677             len > n.length ||
2678             (product != null && len > product.length)) {
2679             throw new IllegalArgumentException("input array length out of bound: " + len);
2680         }
2681     }
2682 
2683     // Make sure that the int array z (which is expected to contain
2684     // the result of a Montgomery multiplication) is present and
2685     // sufficiently large.
2686     private static int[] materialize(int[] z, int len) {
2687          if (z == null || z.length < len)
2688              z = new int[len];
2689          return z;
2690     }
2691 
2692     // These methods are intended to be be replaced by virtual machine
2693     // intrinsics.
2694     @HotSpotIntrinsicCandidate
2695     private static int[] implMontgomeryMultiply(int[] a, int[] b, int[] n, int len,
2696                                          long inv, int[] product) {
2697         product = multiplyToLen(a, len, b, len, product);
2698         return montReduce(product, n, len, (int)inv);
2699     }
2700     @HotSpotIntrinsicCandidate
2701     private static int[] implMontgomerySquare(int[] a, int[] n, int len,
2702                                        long inv, int[] product) {
2703         product = squareToLen(a, len, product);
2704         return montReduce(product, n, len, (int)inv);
2705     }
2706 
2707     static int[] bnExpModThreshTable = {7, 25, 81, 241, 673, 1793,
2708                                                 Integer.MAX_VALUE}; // Sentinel
2709 
2710     /**
2711      * Returns a BigInteger whose value is x to the power of y mod z.
2712      * Assumes: z is odd && x < z.
2713      */
2714     private BigInteger oddModPow(BigInteger y, BigInteger z) {
2715     /*
2716      * The algorithm is adapted from Colin Plumb's C library.
2717      *
2718      * The window algorithm:
2719      * The idea is to keep a running product of b1 = n^(high-order bits of exp)
2720      * and then keep appending exponent bits to it.  The following patterns
2721      * apply to a 3-bit window (k = 3):
2722      * To append   0: square
2723      * To append   1: square, multiply by n^1
2724      * To append  10: square, multiply by n^1, square
2725      * To append  11: square, square, multiply by n^3
2726      * To append 100: square, multiply by n^1, square, square
2727      * To append 101: square, square, square, multiply by n^5
2728      * To append 110: square, square, multiply by n^3, square
2729      * To append 111: square, square, square, multiply by n^7
2730      *
2731      * Since each pattern involves only one multiply, the longer the pattern
2732      * the better, except that a 0 (no multiplies) can be appended directly.
2733      * We precompute a table of odd powers of n, up to 2^k, and can then
2734      * multiply k bits of exponent at a time.  Actually, assuming random
2735      * exponents, there is on average one zero bit between needs to
2736      * multiply (1/2 of the time there's none, 1/4 of the time there's 1,
2737      * 1/8 of the time, there's 2, 1/32 of the time, there's 3, etc.), so
2738      * you have to do one multiply per k+1 bits of exponent.
2739      *
2740      * The loop walks down the exponent, squaring the result buffer as
2741      * it goes.  There is a wbits+1 bit lookahead buffer, buf, that is
2742      * filled with the upcoming exponent bits.  (What is read after the
2743      * end of the exponent is unimportant, but it is filled with zero here.)
2744      * When the most-significant bit of this buffer becomes set, i.e.
2745      * (buf & tblmask) != 0, we have to decide what pattern to multiply
2746      * by, and when to do it.  We decide, remember to do it in future
2747      * after a suitable number of squarings have passed (e.g. a pattern
2748      * of "100" in the buffer requires that we multiply by n^1 immediately;
2749      * a pattern of "110" calls for multiplying by n^3 after one more
2750      * squaring), clear the buffer, and continue.
2751      *
2752      * When we start, there is one more optimization: the result buffer
2753      * is implcitly one, so squaring it or multiplying by it can be
2754      * optimized away.  Further, if we start with a pattern like "100"
2755      * in the lookahead window, rather than placing n into the buffer
2756      * and then starting to square it, we have already computed n^2
2757      * to compute the odd-powers table, so we can place that into
2758      * the buffer and save a squaring.
2759      *
2760      * This means that if you have a k-bit window, to compute n^z,
2761      * where z is the high k bits of the exponent, 1/2 of the time
2762      * it requires no squarings.  1/4 of the time, it requires 1
2763      * squaring, ... 1/2^(k-1) of the time, it reqires k-2 squarings.
2764      * And the remaining 1/2^(k-1) of the time, the top k bits are a
2765      * 1 followed by k-1 0 bits, so it again only requires k-2
2766      * squarings, not k-1.  The average of these is 1.  Add that
2767      * to the one squaring we have to do to compute the table,
2768      * and you'll see that a k-bit window saves k-2 squarings
2769      * as well as reducing the multiplies.  (It actually doesn't
2770      * hurt in the case k = 1, either.)
2771      */
2772         // Special case for exponent of one
2773         if (y.equals(ONE))
2774             return this;
2775 
2776         // Special case for base of zero
2777         if (signum == 0)
2778             return ZERO;
2779 
2780         int[] base = mag.clone();
2781         int[] exp = y.mag;
2782         int[] mod = z.mag;
2783         int modLen = mod.length;
2784 
2785         // Make modLen even. It is conventional to use a cryptographic
2786         // modulus that is 512, 768, 1024, or 2048 bits, so this code
2787         // will not normally be executed. However, it is necessary for
2788         // the correct functioning of the HotSpot intrinsics.
2789         if ((modLen & 1) != 0) {
2790             int[] x = new int[modLen + 1];
2791             System.arraycopy(mod, 0, x, 1, modLen);
2792             mod = x;
2793             modLen++;
2794         }
2795 
2796         // Select an appropriate window size
2797         int wbits = 0;
2798         int ebits = bitLength(exp, exp.length);
2799         // if exponent is 65537 (0x10001), use minimum window size
2800         if ((ebits != 17) || (exp[0] != 65537)) {
2801             while (ebits > bnExpModThreshTable[wbits]) {
2802                 wbits++;
2803             }
2804         }
2805 
2806         // Calculate appropriate table size
2807         int tblmask = 1 << wbits;
2808 
2809         // Allocate table for precomputed odd powers of base in Montgomery form
2810         int[][] table = new int[tblmask][];
2811         for (int i=0; i < tblmask; i++)
2812             table[i] = new int[modLen];
2813 
2814         // Compute the modular inverse of the least significant 64-bit
2815         // digit of the modulus
2816         long n0 = (mod[modLen-1] & LONG_MASK) + ((mod[modLen-2] & LONG_MASK) << 32);
2817         long inv = -MutableBigInteger.inverseMod64(n0);
2818 
2819         // Convert base to Montgomery form
2820         int[] a = leftShift(base, base.length, modLen << 5);
2821 
2822         MutableBigInteger q = new MutableBigInteger(),
2823                           a2 = new MutableBigInteger(a),
2824                           b2 = new MutableBigInteger(mod);
2825         b2.normalize(); // MutableBigInteger.divide() assumes that its
2826                         // divisor is in normal form.
2827 
2828         MutableBigInteger r= a2.divide(b2, q);
2829         table[0] = r.toIntArray();
2830 
2831         // Pad table[0] with leading zeros so its length is at least modLen
2832         if (table[0].length < modLen) {
2833            int offset = modLen - table[0].length;
2834            int[] t2 = new int[modLen];
2835            System.arraycopy(table[0], 0, t2, offset, table[0].length);
2836            table[0] = t2;
2837         }
2838 
2839         // Set b to the square of the base
2840         int[] b = montgomerySquare(table[0], mod, modLen, inv, null);
2841 
2842         // Set t to high half of b
2843         int[] t = Arrays.copyOf(b, modLen);
2844 
2845         // Fill in the table with odd powers of the base
2846         for (int i=1; i < tblmask; i++) {
2847             table[i] = montgomeryMultiply(t, table[i-1], mod, modLen, inv, null);
2848         }
2849 
2850         // Pre load the window that slides over the exponent
2851         int bitpos = 1 << ((ebits-1) & (32-1));
2852 
2853         int buf = 0;
2854         int elen = exp.length;
2855         int eIndex = 0;
2856         for (int i = 0; i <= wbits; i++) {
2857             buf = (buf << 1) | (((exp[eIndex] & bitpos) != 0)?1:0);
2858             bitpos >>>= 1;
2859             if (bitpos == 0) {
2860                 eIndex++;
2861                 bitpos = 1 << (32-1);
2862                 elen--;
2863             }
2864         }
2865 
2866         int multpos = ebits;
2867 
2868         // The first iteration, which is hoisted out of the main loop
2869         ebits--;
2870         boolean isone = true;
2871 
2872         multpos = ebits - wbits;
2873         while ((buf & 1) == 0) {
2874             buf >>>= 1;
2875             multpos++;
2876         }
2877 
2878         int[] mult = table[buf >>> 1];
2879 
2880         buf = 0;
2881         if (multpos == ebits)
2882             isone = false;
2883 
2884         // The main loop
2885         while (true) {
2886             ebits--;
2887             // Advance the window
2888             buf <<= 1;
2889 
2890             if (elen != 0) {
2891                 buf |= ((exp[eIndex] & bitpos) != 0) ? 1 : 0;
2892                 bitpos >>>= 1;
2893                 if (bitpos == 0) {
2894                     eIndex++;
2895                     bitpos = 1 << (32-1);
2896                     elen--;
2897                 }
2898             }
2899 
2900             // Examine the window for pending multiplies
2901             if ((buf & tblmask) != 0) {
2902                 multpos = ebits - wbits;
2903                 while ((buf & 1) == 0) {
2904                     buf >>>= 1;
2905                     multpos++;
2906                 }
2907                 mult = table[buf >>> 1];
2908                 buf = 0;
2909             }
2910 
2911             // Perform multiply
2912             if (ebits == multpos) {
2913                 if (isone) {
2914                     b = mult.clone();
2915                     isone = false;
2916                 } else {
2917                     t = b;
2918                     a = montgomeryMultiply(t, mult, mod, modLen, inv, a);
2919                     t = a; a = b; b = t;
2920                 }
2921             }
2922 
2923             // Check if done
2924             if (ebits == 0)
2925                 break;
2926 
2927             // Square the input
2928             if (!isone) {
2929                 t = b;
2930                 a = montgomerySquare(t, mod, modLen, inv, a);
2931                 t = a; a = b; b = t;
2932             }
2933         }
2934 
2935         // Convert result out of Montgomery form and return
2936         int[] t2 = new int[2*modLen];
2937         System.arraycopy(b, 0, t2, modLen, modLen);
2938 
2939         b = montReduce(t2, mod, modLen, (int)inv);
2940 
2941         t2 = Arrays.copyOf(b, modLen);
2942 
2943         return new BigInteger(1, t2);
2944     }
2945 
2946     /**
2947      * Montgomery reduce n, modulo mod.  This reduces modulo mod and divides
2948      * by 2^(32*mlen). Adapted from Colin Plumb's C library.
2949      */
2950     private static int[] montReduce(int[] n, int[] mod, int mlen, int inv) {
2951         int c=0;
2952         int len = mlen;
2953         int offset=0;
2954 
2955         do {
2956             int nEnd = n[n.length-1-offset];
2957             int carry = mulAdd(n, mod, offset, mlen, inv * nEnd);
2958             c += addOne(n, offset, mlen, carry);
2959             offset++;
2960         } while (--len > 0);
2961 
2962         while (c > 0)
2963             c += subN(n, mod, mlen);
2964 
2965         while (intArrayCmpToLen(n, mod, mlen) >= 0)
2966             subN(n, mod, mlen);
2967 
2968         return n;
2969     }
2970 
2971 
2972     /*
2973      * Returns -1, 0 or +1 as big-endian unsigned int array arg1 is less than,
2974      * equal to, or greater than arg2 up to length len.
2975      */
2976     private static int intArrayCmpToLen(int[] arg1, int[] arg2, int len) {
2977         for (int i=0; i < len; i++) {
2978             long b1 = arg1[i] & LONG_MASK;
2979             long b2 = arg2[i] & LONG_MASK;
2980             if (b1 < b2)
2981                 return -1;
2982             if (b1 > b2)
2983                 return 1;
2984         }
2985         return 0;
2986     }
2987 
2988     /**
2989      * Subtracts two numbers of same length, returning borrow.
2990      */
2991     private static int subN(int[] a, int[] b, int len) {
2992         long sum = 0;
2993 
2994         while (--len >= 0) {
2995             sum = (a[len] & LONG_MASK) -
2996                  (b[len] & LONG_MASK) + (sum >> 32);
2997             a[len] = (int)sum;
2998         }
2999 
3000         return (int)(sum >> 32);
3001     }
3002 
3003     /**
3004      * Multiply an array by one word k and add to result, return the carry
3005      */
3006     static int mulAdd(int[] out, int[] in, int offset, int len, int k) {
3007         implMulAddCheck(out, in, offset, len, k);
3008         return implMulAdd(out, in, offset, len, k);
3009     }
3010 
3011     /**
3012      * Parameters validation.
3013      */
3014     private static void implMulAddCheck(int[] out, int[] in, int offset, int len, int k) {
3015         if (len > in.length) {
3016             throw new IllegalArgumentException("input length is out of bound: " + len + " > " + in.length);
3017         }
3018         if (offset < 0) {
3019             throw new IllegalArgumentException("input offset is invalid: " + offset);
3020         }
3021         if (offset > (out.length - 1)) {
3022             throw new IllegalArgumentException("input offset is out of bound: " + offset + " > " + (out.length - 1));
3023         }
3024         if (len > (out.length - offset)) {
3025             throw new IllegalArgumentException("input len is out of bound: " + len + " > " + (out.length - offset));
3026         }
3027     }
3028 
3029     /**
3030      * Java Runtime may use intrinsic for this method.
3031      */
3032     @HotSpotIntrinsicCandidate
3033     private static int implMulAdd(int[] out, int[] in, int offset, int len, int k) {
3034         long kLong = k & LONG_MASK;
3035         long carry = 0;
3036 
3037         offset = out.length-offset - 1;
3038         for (int j=len-1; j >= 0; j--) {
3039             long product = (in[j] & LONG_MASK) * kLong +
3040                            (out[offset] & LONG_MASK) + carry;
3041             out[offset--] = (int)product;
3042             carry = product >>> 32;
3043         }
3044         return (int)carry;
3045     }
3046 
3047     /**
3048      * Add one word to the number a mlen words into a. Return the resulting
3049      * carry.
3050      */
3051     static int addOne(int[] a, int offset, int mlen, int carry) {
3052         offset = a.length-1-mlen-offset;
3053         long t = (a[offset] & LONG_MASK) + (carry & LONG_MASK);
3054 
3055         a[offset] = (int)t;
3056         if ((t >>> 32) == 0)
3057             return 0;
3058         while (--mlen >= 0) {
3059             if (--offset < 0) { // Carry out of number
3060                 return 1;
3061             } else {
3062                 a[offset]++;
3063                 if (a[offset] != 0)
3064                     return 0;
3065             }
3066         }
3067         return 1;
3068     }
3069 
3070     /**
3071      * Returns a BigInteger whose value is (this ** exponent) mod (2**p)
3072      */
3073     private BigInteger modPow2(BigInteger exponent, int p) {
3074         /*
3075          * Perform exponentiation using repeated squaring trick, chopping off
3076          * high order bits as indicated by modulus.
3077          */
3078         BigInteger result = ONE;
3079         BigInteger baseToPow2 = this.mod2(p);
3080         int expOffset = 0;
3081 
3082         int limit = exponent.bitLength();
3083 
3084         if (this.testBit(0))
3085            limit = (p-1) < limit ? (p-1) : limit;
3086 
3087         while (expOffset < limit) {
3088             if (exponent.testBit(expOffset))
3089                 result = result.multiply(baseToPow2).mod2(p);
3090             expOffset++;
3091             if (expOffset < limit)
3092                 baseToPow2 = baseToPow2.square().mod2(p);
3093         }
3094 
3095         return result;
3096     }
3097 
3098     /**
3099      * Returns a BigInteger whose value is this mod(2**p).
3100      * Assumes that this {@code BigInteger >= 0} and {@code p > 0}.
3101      */
3102     private BigInteger mod2(int p) {
3103         if (bitLength() <= p)
3104             return this;
3105 
3106         // Copy remaining ints of mag
3107         int numInts = (p + 31) >>> 5;
3108         int[] mag = new int[numInts];
3109         System.arraycopy(this.mag, (this.mag.length - numInts), mag, 0, numInts);
3110 
3111         // Mask out any excess bits
3112         int excessBits = (numInts << 5) - p;
3113         mag[0] &= (1L << (32-excessBits)) - 1;
3114 
3115         return (mag[0] == 0 ? new BigInteger(1, mag) : new BigInteger(mag, 1));
3116     }
3117 
3118     /**
3119      * Returns a BigInteger whose value is {@code (this}<sup>-1</sup> {@code mod m)}.
3120      *
3121      * @param  m the modulus.
3122      * @return {@code this}<sup>-1</sup> {@code mod m}.
3123      * @throws ArithmeticException {@code  m} &le; 0, or this BigInteger
3124      *         has no multiplicative inverse mod m (that is, this BigInteger
3125      *         is not <i>relatively prime</i> to m).
3126      */
3127     public BigInteger modInverse(BigInteger m) {
3128         if (m.signum != 1)
3129             throw new ArithmeticException("BigInteger: modulus not positive");
3130 
3131         if (m.equals(ONE))
3132             return ZERO;
3133 
3134         // Calculate (this mod m)
3135         BigInteger modVal = this;
3136         if (signum < 0 || (this.compareMagnitude(m) >= 0))
3137             modVal = this.mod(m);
3138 
3139         if (modVal.equals(ONE))
3140             return ONE;
3141 
3142         MutableBigInteger a = new MutableBigInteger(modVal);
3143         MutableBigInteger b = new MutableBigInteger(m);
3144 
3145         MutableBigInteger result = a.mutableModInverse(b);
3146         return result.toBigInteger(1);
3147     }
3148 
3149     // Shift Operations
3150 
3151     /**
3152      * Returns a BigInteger whose value is {@code (this << n)}.
3153      * The shift distance, {@code n}, may be negative, in which case
3154      * this method performs a right shift.
3155      * (Computes <code>floor(this * 2<sup>n</sup>)</code>.)
3156      *
3157      * @param  n shift distance, in bits.
3158      * @return {@code this << n}
3159      * @see #shiftRight
3160      */
3161     public BigInteger shiftLeft(int n) {
3162         if (signum == 0)
3163             return ZERO;
3164         if (n > 0) {
3165             return new BigInteger(shiftLeft(mag, n), signum);
3166         } else if (n == 0) {
3167             return this;
3168         } else {
3169             // Possible int overflow in (-n) is not a trouble,
3170             // because shiftRightImpl considers its argument unsigned
3171             return shiftRightImpl(-n);
3172         }
3173     }
3174 
3175     /**
3176      * Returns a magnitude array whose value is {@code (mag << n)}.
3177      * The shift distance, {@code n}, is considered unnsigned.
3178      * (Computes <code>this * 2<sup>n</sup></code>.)
3179      *
3180      * @param mag magnitude, the most-significant int ({@code mag[0]}) must be non-zero.
3181      * @param  n unsigned shift distance, in bits.
3182      * @return {@code mag << n}
3183      */
3184     private static int[] shiftLeft(int[] mag, int n) {
3185         int nInts = n >>> 5;
3186         int nBits = n & 0x1f;
3187         int magLen = mag.length;
3188         int newMag[] = null;
3189 
3190         if (nBits == 0) {
3191             newMag = new int[magLen + nInts];
3192             System.arraycopy(mag, 0, newMag, 0, magLen);
3193         } else {
3194             int i = 0;
3195             int nBits2 = 32 - nBits;
3196             int highBits = mag[0] >>> nBits2;
3197             if (highBits != 0) {
3198                 newMag = new int[magLen + nInts + 1];
3199                 newMag[i++] = highBits;
3200             } else {
3201                 newMag = new int[magLen + nInts];
3202             }
3203             int j=0;
3204             while (j < magLen-1)
3205                 newMag[i++] = mag[j++] << nBits | mag[j] >>> nBits2;
3206             newMag[i] = mag[j] << nBits;
3207         }
3208         return newMag;
3209     }
3210 
3211     /**
3212      * Returns a BigInteger whose value is {@code (this >> n)}.  Sign
3213      * extension is performed.  The shift distance, {@code n}, may be
3214      * negative, in which case this method performs a left shift.
3215      * (Computes <code>floor(this / 2<sup>n</sup>)</code>.)
3216      *
3217      * @param  n shift distance, in bits.
3218      * @return {@code this >> n}
3219      * @see #shiftLeft
3220      */
3221     public BigInteger shiftRight(int n) {
3222         if (signum == 0)
3223             return ZERO;
3224         if (n > 0) {
3225             return shiftRightImpl(n);
3226         } else if (n == 0) {
3227             return this;
3228         } else {
3229             // Possible int overflow in {@code -n} is not a trouble,
3230             // because shiftLeft considers its argument unsigned
3231             return new BigInteger(shiftLeft(mag, -n), signum);
3232         }
3233     }
3234 
3235     /**
3236      * Returns a BigInteger whose value is {@code (this >> n)}. The shift
3237      * distance, {@code n}, is considered unsigned.
3238      * (Computes <code>floor(this * 2<sup>-n</sup>)</code>.)
3239      *
3240      * @param  n unsigned shift distance, in bits.
3241      * @return {@code this >> n}
3242      */
3243     private BigInteger shiftRightImpl(int n) {
3244         int nInts = n >>> 5;
3245         int nBits = n & 0x1f;
3246         int magLen = mag.length;
3247         int newMag[] = null;
3248 
3249         // Special case: entire contents shifted off the end
3250         if (nInts >= magLen)
3251             return (signum >= 0 ? ZERO : negConst[1]);
3252 
3253         if (nBits == 0) {
3254             int newMagLen = magLen - nInts;
3255             newMag = Arrays.copyOf(mag, newMagLen);
3256         } else {
3257             int i = 0;
3258             int highBits = mag[0] >>> nBits;
3259             if (highBits != 0) {
3260                 newMag = new int[magLen - nInts];
3261                 newMag[i++] = highBits;
3262             } else {
3263                 newMag = new int[magLen - nInts -1];
3264             }
3265 
3266             int nBits2 = 32 - nBits;
3267             int j=0;
3268             while (j < magLen - nInts - 1)
3269                 newMag[i++] = (mag[j++] << nBits2) | (mag[j] >>> nBits);
3270         }
3271 
3272         if (signum < 0) {
3273             // Find out whether any one-bits were shifted off the end.
3274             boolean onesLost = false;
3275             for (int i=magLen-1, j=magLen-nInts; i >= j && !onesLost; i--)
3276                 onesLost = (mag[i] != 0);
3277             if (!onesLost && nBits != 0)
3278                 onesLost = (mag[magLen - nInts - 1] << (32 - nBits) != 0);
3279 
3280             if (onesLost)
3281                 newMag = javaIncrement(newMag);
3282         }
3283 
3284         return new BigInteger(newMag, signum);
3285     }
3286 
3287     int[] javaIncrement(int[] val) {
3288         int lastSum = 0;
3289         for (int i=val.length-1;  i >= 0 && lastSum == 0; i--)
3290             lastSum = (val[i] += 1);
3291         if (lastSum == 0) {
3292             val = new int[val.length+1];
3293             val[0] = 1;
3294         }
3295         return val;
3296     }
3297 
3298     // Bitwise Operations
3299 
3300     /**
3301      * Returns a BigInteger whose value is {@code (this & val)}.  (This
3302      * method returns a negative BigInteger if and only if this and val are
3303      * both negative.)
3304      *
3305      * @param val value to be AND'ed with this BigInteger.
3306      * @return {@code this & val}
3307      */
3308     public BigInteger and(BigInteger val) {
3309         int[] result = new int[Math.max(intLength(), val.intLength())];
3310         for (int i=0; i < result.length; i++)
3311             result[i] = (getInt(result.length-i-1)
3312                          & val.getInt(result.length-i-1));
3313 
3314         return valueOf(result);
3315     }
3316 
3317     /**
3318      * Returns a BigInteger whose value is {@code (this | val)}.  (This method
3319      * returns a negative BigInteger if and only if either this or val is
3320      * negative.)
3321      *
3322      * @param val value to be OR'ed with this BigInteger.
3323      * @return {@code this | val}
3324      */
3325     public BigInteger or(BigInteger val) {
3326         int[] result = new int[Math.max(intLength(), val.intLength())];
3327         for (int i=0; i < result.length; i++)
3328             result[i] = (getInt(result.length-i-1)
3329                          | val.getInt(result.length-i-1));
3330 
3331         return valueOf(result);
3332     }
3333 
3334     /**
3335      * Returns a BigInteger whose value is {@code (this ^ val)}.  (This method
3336      * returns a negative BigInteger if and only if exactly one of this and
3337      * val are negative.)
3338      *
3339      * @param val value to be XOR'ed with this BigInteger.
3340      * @return {@code this ^ val}
3341      */
3342     public BigInteger xor(BigInteger val) {
3343         int[] result = new int[Math.max(intLength(), val.intLength())];
3344         for (int i=0; i < result.length; i++)
3345             result[i] = (getInt(result.length-i-1)
3346                          ^ val.getInt(result.length-i-1));
3347 
3348         return valueOf(result);
3349     }
3350 
3351     /**
3352      * Returns a BigInteger whose value is {@code (~this)}.  (This method
3353      * returns a negative value if and only if this BigInteger is
3354      * non-negative.)
3355      *
3356      * @return {@code ~this}
3357      */
3358     public BigInteger not() {
3359         int[] result = new int[intLength()];
3360         for (int i=0; i < result.length; i++)
3361             result[i] = ~getInt(result.length-i-1);
3362 
3363         return valueOf(result);
3364     }
3365 
3366     /**
3367      * Returns a BigInteger whose value is {@code (this & ~val)}.  This
3368      * method, which is equivalent to {@code and(val.not())}, is provided as
3369      * a convenience for masking operations.  (This method returns a negative
3370      * BigInteger if and only if {@code this} is negative and {@code val} is
3371      * positive.)
3372      *
3373      * @param val value to be complemented and AND'ed with this BigInteger.
3374      * @return {@code this & ~val}
3375      */
3376     public BigInteger andNot(BigInteger val) {
3377         int[] result = new int[Math.max(intLength(), val.intLength())];
3378         for (int i=0; i < result.length; i++)
3379             result[i] = (getInt(result.length-i-1)
3380                          & ~val.getInt(result.length-i-1));
3381 
3382         return valueOf(result);
3383     }
3384 
3385 
3386     // Single Bit Operations
3387 
3388     /**
3389      * Returns {@code true} if and only if the designated bit is set.
3390      * (Computes {@code ((this & (1<<n)) != 0)}.)
3391      *
3392      * @param  n index of bit to test.
3393      * @return {@code true} if and only if the designated bit is set.
3394      * @throws ArithmeticException {@code n} is negative.
3395      */
3396     public boolean testBit(int n) {
3397         if (n < 0)
3398             throw new ArithmeticException("Negative bit address");
3399 
3400         return (getInt(n >>> 5) & (1 << (n & 31))) != 0;
3401     }
3402 
3403     /**
3404      * Returns a BigInteger whose value is equivalent to this BigInteger
3405      * with the designated bit set.  (Computes {@code (this | (1<<n))}.)
3406      *
3407      * @param  n index of bit to set.
3408      * @return {@code this | (1<<n)}
3409      * @throws ArithmeticException {@code n} is negative.
3410      */
3411     public BigInteger setBit(int n) {
3412         if (n < 0)
3413             throw new ArithmeticException("Negative bit address");
3414 
3415         int intNum = n >>> 5;
3416         int[] result = new int[Math.max(intLength(), intNum+2)];
3417 
3418         for (int i=0; i < result.length; i++)
3419             result[result.length-i-1] = getInt(i);
3420 
3421         result[result.length-intNum-1] |= (1 << (n & 31));
3422 
3423         return valueOf(result);
3424     }
3425 
3426     /**
3427      * Returns a BigInteger whose value is equivalent to this BigInteger
3428      * with the designated bit cleared.
3429      * (Computes {@code (this & ~(1<<n))}.)
3430      *
3431      * @param  n index of bit to clear.
3432      * @return {@code this & ~(1<<n)}
3433      * @throws ArithmeticException {@code n} is negative.
3434      */
3435     public BigInteger clearBit(int n) {
3436         if (n < 0)
3437             throw new ArithmeticException("Negative bit address");
3438 
3439         int intNum = n >>> 5;
3440         int[] result = new int[Math.max(intLength(), ((n + 1) >>> 5) + 1)];
3441 
3442         for (int i=0; i < result.length; i++)
3443             result[result.length-i-1] = getInt(i);
3444 
3445         result[result.length-intNum-1] &= ~(1 << (n & 31));
3446 
3447         return valueOf(result);
3448     }
3449 
3450     /**
3451      * Returns a BigInteger whose value is equivalent to this BigInteger
3452      * with the designated bit flipped.
3453      * (Computes {@code (this ^ (1<<n))}.)
3454      *
3455      * @param  n index of bit to flip.
3456      * @return {@code this ^ (1<<n)}
3457      * @throws ArithmeticException {@code n} is negative.
3458      */
3459     public BigInteger flipBit(int n) {
3460         if (n < 0)
3461             throw new ArithmeticException("Negative bit address");
3462 
3463         int intNum = n >>> 5;
3464         int[] result = new int[Math.max(intLength(), intNum+2)];
3465 
3466         for (int i=0; i < result.length; i++)
3467             result[result.length-i-1] = getInt(i);
3468 
3469         result[result.length-intNum-1] ^= (1 << (n & 31));
3470 
3471         return valueOf(result);
3472     }
3473 
3474     /**
3475      * Returns the index of the rightmost (lowest-order) one bit in this
3476      * BigInteger (the number of zero bits to the right of the rightmost
3477      * one bit).  Returns -1 if this BigInteger contains no one bits.
3478      * (Computes {@code (this == 0? -1 : log2(this & -this))}.)
3479      *
3480      * @return index of the rightmost one bit in this BigInteger.
3481      */
3482     public int getLowestSetBit() {
3483         int lsb = lowestSetBitPlusTwo - 2;
3484         if (lsb == -2) {  // lowestSetBit not initialized yet
3485             lsb = 0;
3486             if (signum == 0) {
3487                 lsb -= 1;
3488             } else {
3489                 // Search for lowest order nonzero int
3490                 int i,b;
3491                 for (i=0; (b = getInt(i)) == 0; i++)
3492                     ;
3493                 lsb += (i << 5) + Integer.numberOfTrailingZeros(b);
3494             }
3495             lowestSetBitPlusTwo = lsb + 2;
3496         }
3497         return lsb;
3498     }
3499 
3500 
3501     // Miscellaneous Bit Operations
3502 
3503     /**
3504      * Returns the number of bits in the minimal two's-complement
3505      * representation of this BigInteger, <i>excluding</i> a sign bit.
3506      * For positive BigIntegers, this is equivalent to the number of bits in
3507      * the ordinary binary representation.  (Computes
3508      * {@code (ceil(log2(this < 0 ? -this : this+1)))}.)
3509      *
3510      * @return number of bits in the minimal two's-complement
3511      *         representation of this BigInteger, <i>excluding</i> a sign bit.
3512      */
3513     public int bitLength() {
3514         int n = bitLengthPlusOne - 1;
3515         if (n == -1) { // bitLength not initialized yet
3516             int[] m = mag;
3517             int len = m.length;
3518             if (len == 0) {
3519                 n = 0; // offset by one to initialize
3520             }  else {
3521                 // Calculate the bit length of the magnitude
3522                 int magBitLength = ((len - 1) << 5) + bitLengthForInt(mag[0]);
3523                  if (signum < 0) {
3524                      // Check if magnitude is a power of two
3525                      boolean pow2 = (Integer.bitCount(mag[0]) == 1);
3526                      for (int i=1; i< len && pow2; i++)
3527                          pow2 = (mag[i] == 0);
3528 
3529                      n = (pow2 ? magBitLength -1 : magBitLength);
3530                  } else {
3531                      n = magBitLength;
3532                  }
3533             }
3534             bitLengthPlusOne = n + 1;
3535         }
3536         return n;
3537     }
3538 
3539     /**
3540      * Returns the number of bits in the two's complement representation
3541      * of this BigInteger that differ from its sign bit.  This method is
3542      * useful when implementing bit-vector style sets atop BigIntegers.
3543      *
3544      * @return number of bits in the two's complement representation
3545      *         of this BigInteger that differ from its sign bit.
3546      */
3547     public int bitCount() {
3548         int bc = bitCountPlusOne - 1;
3549         if (bc == -1) {  // bitCount not initialized yet
3550             bc = 0;      // offset by one to initialize
3551             // Count the bits in the magnitude
3552             for (int i=0; i < mag.length; i++)
3553                 bc += Integer.bitCount(mag[i]);
3554             if (signum < 0) {
3555                 // Count the trailing zeros in the magnitude
3556                 int magTrailingZeroCount = 0, j;
3557                 for (j=mag.length-1; mag[j] == 0; j--)
3558                     magTrailingZeroCount += 32;
3559                 magTrailingZeroCount += Integer.numberOfTrailingZeros(mag[j]);
3560                 bc += magTrailingZeroCount - 1;
3561             }
3562             bitCountPlusOne = bc + 1;
3563         }
3564         return bc;
3565     }
3566 
3567     // Primality Testing
3568 
3569     /**
3570      * Returns {@code true} if this BigInteger is probably prime,
3571      * {@code false} if it's definitely composite.  If
3572      * {@code certainty} is &le; 0, {@code true} is
3573      * returned.
3574      *
3575      * @param  certainty a measure of the uncertainty that the caller is
3576      *         willing to tolerate: if the call returns {@code true}
3577      *         the probability that this BigInteger is prime exceeds
3578      *         (1 - 1/2<sup>{@code certainty}</sup>).  The execution time of
3579      *         this method is proportional to the value of this parameter.
3580      * @return {@code true} if this BigInteger is probably prime,
3581      *         {@code false} if it's definitely composite.
3582      */
3583     public boolean isProbablePrime(int certainty) {
3584         if (certainty <= 0)
3585             return true;
3586         BigInteger w = this.abs();
3587         if (w.equals(TWO))
3588             return true;
3589         if (!w.testBit(0) || w.equals(ONE))
3590             return false;
3591 
3592         return w.primeToCertainty(certainty, null);
3593     }
3594 
3595     // Comparison Operations
3596 
3597     /**
3598      * Compares this BigInteger with the specified BigInteger.  This
3599      * method is provided in preference to individual methods for each
3600      * of the six boolean comparison operators ({@literal <}, ==,
3601      * {@literal >}, {@literal >=}, !=, {@literal <=}).  The suggested
3602      * idiom for performing these comparisons is: {@code
3603      * (x.compareTo(y)} &lt;<i>op</i>&gt; {@code 0)}, where
3604      * &lt;<i>op</i>&gt; is one of the six comparison operators.
3605      *
3606      * @param  val BigInteger to which this BigInteger is to be compared.
3607      * @return -1, 0 or 1 as this BigInteger is numerically less than, equal
3608      *         to, or greater than {@code val}.
3609      */
3610     public int compareTo(BigInteger val) {
3611         if (signum == val.signum) {
3612             switch (signum) {
3613             case 1:
3614                 return compareMagnitude(val);
3615             case -1:
3616                 return val.compareMagnitude(this);
3617             default:
3618                 return 0;
3619             }
3620         }
3621         return signum > val.signum ? 1 : -1;
3622     }
3623 
3624     /**
3625      * Compares the magnitude array of this BigInteger with the specified
3626      * BigInteger's. This is the version of compareTo ignoring sign.
3627      *
3628      * @param val BigInteger whose magnitude array to be compared.
3629      * @return -1, 0 or 1 as this magnitude array is less than, equal to or
3630      *         greater than the magnitude aray for the specified BigInteger's.
3631      */
3632     final int compareMagnitude(BigInteger val) {
3633         int[] m1 = mag;
3634         int len1 = m1.length;
3635         int[] m2 = val.mag;
3636         int len2 = m2.length;
3637         if (len1 < len2)
3638             return -1;
3639         if (len1 > len2)
3640             return 1;
3641         for (int i = 0; i < len1; i++) {
3642             int a = m1[i];
3643             int b = m2[i];
3644             if (a != b)
3645                 return ((a & LONG_MASK) < (b & LONG_MASK)) ? -1 : 1;
3646         }
3647         return 0;
3648     }
3649 
3650     /**
3651      * Version of compareMagnitude that compares magnitude with long value.
3652      * val can't be Long.MIN_VALUE.
3653      */
3654     final int compareMagnitude(long val) {
3655         assert val != Long.MIN_VALUE;
3656         int[] m1 = mag;
3657         int len = m1.length;
3658         if (len > 2) {
3659             return 1;
3660         }
3661         if (val < 0) {
3662             val = -val;
3663         }
3664         int highWord = (int)(val >>> 32);
3665         if (highWord == 0) {
3666             if (len < 1)
3667                 return -1;
3668             if (len > 1)
3669                 return 1;
3670             int a = m1[0];
3671             int b = (int)val;
3672             if (a != b) {
3673                 return ((a & LONG_MASK) < (b & LONG_MASK))? -1 : 1;
3674             }
3675             return 0;
3676         } else {
3677             if (len < 2)
3678                 return -1;
3679             int a = m1[0];
3680             int b = highWord;
3681             if (a != b) {
3682                 return ((a & LONG_MASK) < (b & LONG_MASK))? -1 : 1;
3683             }
3684             a = m1[1];
3685             b = (int)val;
3686             if (a != b) {
3687                 return ((a & LONG_MASK) < (b & LONG_MASK))? -1 : 1;
3688             }
3689             return 0;
3690         }
3691     }
3692 
3693     /**
3694      * Compares this BigInteger with the specified Object for equality.
3695      *
3696      * @param  x Object to which this BigInteger is to be compared.
3697      * @return {@code true} if and only if the specified Object is a
3698      *         BigInteger whose value is numerically equal to this BigInteger.
3699      */
3700     public boolean equals(Object x) {
3701         // This test is just an optimization, which may or may not help
3702         if (x == this)
3703             return true;
3704 
3705         if (!(x instanceof BigInteger))
3706             return false;
3707 
3708         BigInteger xInt = (BigInteger) x;
3709         if (xInt.signum != signum)
3710             return false;
3711 
3712         int[] m = mag;
3713         int len = m.length;
3714         int[] xm = xInt.mag;
3715         if (len != xm.length)
3716             return false;
3717 
3718         for (int i = 0; i < len; i++)
3719             if (xm[i] != m[i])
3720                 return false;
3721 
3722         return true;
3723     }
3724 
3725     /**
3726      * Returns the minimum of this BigInteger and {@code val}.
3727      *
3728      * @param  val value with which the minimum is to be computed.
3729      * @return the BigInteger whose value is the lesser of this BigInteger and
3730      *         {@code val}.  If they are equal, either may be returned.
3731      */
3732     public BigInteger min(BigInteger val) {
3733         return (compareTo(val) < 0 ? this : val);
3734     }
3735 
3736     /**
3737      * Returns the maximum of this BigInteger and {@code val}.
3738      *
3739      * @param  val value with which the maximum is to be computed.
3740      * @return the BigInteger whose value is the greater of this and
3741      *         {@code val}.  If they are equal, either may be returned.
3742      */
3743     public BigInteger max(BigInteger val) {
3744         return (compareTo(val) > 0 ? this : val);
3745     }
3746 
3747 
3748     // Hash Function
3749 
3750     /**
3751      * Returns the hash code for this BigInteger.
3752      *
3753      * @return hash code for this BigInteger.
3754      */
3755     public int hashCode() {
3756         int hashCode = 0;
3757 
3758         for (int i=0; i < mag.length; i++)
3759             hashCode = (int)(31*hashCode + (mag[i] & LONG_MASK));
3760 
3761         return hashCode * signum;
3762     }
3763 
3764     /**
3765      * Returns the String representation of this BigInteger in the
3766      * given radix.  If the radix is outside the range from {@link
3767      * Character#MIN_RADIX} to {@link Character#MAX_RADIX} inclusive,
3768      * it will default to 10 (as is the case for
3769      * {@code Integer.toString}).  The digit-to-character mapping
3770      * provided by {@code Character.forDigit} is used, and a minus
3771      * sign is prepended if appropriate.  (This representation is
3772      * compatible with the {@link #BigInteger(String, int) (String,
3773      * int)} constructor.)
3774      *
3775      * @param  radix  radix of the String representation.
3776      * @return String representation of this BigInteger in the given radix.
3777      * @see    Integer#toString
3778      * @see    Character#forDigit
3779      * @see    #BigInteger(java.lang.String, int)
3780      */
3781     public String toString(int radix) {
3782         if (signum == 0)
3783             return "0";
3784         if (radix < Character.MIN_RADIX || radix > Character.MAX_RADIX)
3785             radix = 10;
3786 
3787         // If it's small enough, use smallToString.
3788         if (mag.length <= SCHOENHAGE_BASE_CONVERSION_THRESHOLD)
3789            return smallToString(radix);
3790 
3791         // Otherwise use recursive toString, which requires positive arguments.
3792         // The results will be concatenated into this StringBuilder
3793         StringBuilder sb = new StringBuilder();
3794         if (signum < 0) {
3795             toString(this.negate(), sb, radix, 0);
3796             sb.insert(0, '-');
3797         }
3798         else
3799             toString(this, sb, radix, 0);
3800 
3801         return sb.toString();
3802     }
3803 
3804     /** This method is used to perform toString when arguments are small. */
3805     private String smallToString(int radix) {
3806         if (signum == 0) {
3807             return "0";
3808         }
3809 
3810         // Compute upper bound on number of digit groups and allocate space
3811         int maxNumDigitGroups = (4*mag.length + 6)/7;
3812         String digitGroup[] = new String[maxNumDigitGroups];
3813 
3814         // Translate number to string, a digit group at a time
3815         BigInteger tmp = this.abs();
3816         int numGroups = 0;
3817         while (tmp.signum != 0) {
3818             BigInteger d = longRadix[radix];
3819 
3820             MutableBigInteger q = new MutableBigInteger(),
3821                               a = new MutableBigInteger(tmp.mag),
3822                               b = new MutableBigInteger(d.mag);
3823             MutableBigInteger r = a.divide(b, q);
3824             BigInteger q2 = q.toBigInteger(tmp.signum * d.signum);
3825             BigInteger r2 = r.toBigInteger(tmp.signum * d.signum);
3826 
3827             digitGroup[numGroups++] = Long.toString(r2.longValue(), radix);
3828             tmp = q2;
3829         }
3830 
3831         // Put sign (if any) and first digit group into result buffer
3832         StringBuilder buf = new StringBuilder(numGroups*digitsPerLong[radix]+1);
3833         if (signum < 0) {
3834             buf.append('-');
3835         }
3836         buf.append(digitGroup[numGroups-1]);
3837 
3838         // Append remaining digit groups padded with leading zeros
3839         for (int i=numGroups-2; i >= 0; i--) {
3840             // Prepend (any) leading zeros for this digit group
3841             int numLeadingZeros = digitsPerLong[radix]-digitGroup[i].length();
3842             if (numLeadingZeros != 0) {
3843                 buf.append(zeros[numLeadingZeros]);
3844             }
3845             buf.append(digitGroup[i]);
3846         }
3847         return buf.toString();
3848     }
3849 
3850     /**
3851      * Converts the specified BigInteger to a string and appends to
3852      * {@code sb}.  This implements the recursive Schoenhage algorithm
3853      * for base conversions.
3854      * <p>
3855      * See Knuth, Donald,  _The Art of Computer Programming_, Vol. 2,
3856      * Answers to Exercises (4.4) Question 14.
3857      *
3858      * @param u      The number to convert to a string.
3859      * @param sb     The StringBuilder that will be appended to in place.
3860      * @param radix  The base to convert to.
3861      * @param digits The minimum number of digits to pad to.
3862      */
3863     private static void toString(BigInteger u, StringBuilder sb, int radix,
3864                                  int digits) {
3865         // If we're smaller than a certain threshold, use the smallToString
3866         // method, padding with leading zeroes when necessary.
3867         if (u.mag.length <= SCHOENHAGE_BASE_CONVERSION_THRESHOLD) {
3868             String s = u.smallToString(radix);
3869 
3870             // Pad with internal zeros if necessary.
3871             // Don't pad if we're at the beginning of the string.
3872             if ((s.length() < digits) && (sb.length() > 0)) {
3873                 for (int i=s.length(); i < digits; i++) {
3874                     sb.append('0');
3875                 }
3876             }
3877 
3878             sb.append(s);
3879             return;
3880         }
3881 
3882         int b, n;
3883         b = u.bitLength();
3884 
3885         // Calculate a value for n in the equation radix^(2^n) = u
3886         // and subtract 1 from that value.  This is used to find the
3887         // cache index that contains the best value to divide u.
3888         n = (int) Math.round(Math.log(b * LOG_TWO / logCache[radix]) / LOG_TWO - 1.0);
3889         BigInteger v = getRadixConversionCache(radix, n);
3890         BigInteger[] results;
3891         results = u.divideAndRemainder(v);
3892 
3893         int expectedDigits = 1 << n;
3894 
3895         // Now recursively build the two halves of each number.
3896         toString(results[0], sb, radix, digits-expectedDigits);
3897         toString(results[1], sb, radix, expectedDigits);
3898     }
3899 
3900     /**
3901      * Returns the value radix^(2^exponent) from the cache.
3902      * If this value doesn't already exist in the cache, it is added.
3903      * <p>
3904      * This could be changed to a more complicated caching method using
3905      * {@code Future}.
3906      */
3907     private static BigInteger getRadixConversionCache(int radix, int exponent) {
3908         BigInteger[] cacheLine = powerCache[radix]; // volatile read
3909         if (exponent < cacheLine.length) {
3910             return cacheLine[exponent];
3911         }
3912 
3913         int oldLength = cacheLine.length;
3914         cacheLine = Arrays.copyOf(cacheLine, exponent + 1);
3915         for (int i = oldLength; i <= exponent; i++) {
3916             cacheLine[i] = cacheLine[i - 1].pow(2);
3917         }
3918 
3919         BigInteger[][] pc = powerCache; // volatile read again
3920         if (exponent >= pc[radix].length) {
3921             pc = pc.clone();
3922             pc[radix] = cacheLine;
3923             powerCache = pc; // volatile write, publish
3924         }
3925         return cacheLine[exponent];
3926     }
3927 
3928     /* zero[i] is a string of i consecutive zeros. */
3929     private static String zeros[] = new String[64];
3930     static {
3931         zeros[63] =
3932             "000000000000000000000000000000000000000000000000000000000000000";
3933         for (int i=0; i < 63; i++)
3934             zeros[i] = zeros[63].substring(0, i);
3935     }
3936 
3937     /**
3938      * Returns the decimal String representation of this BigInteger.
3939      * The digit-to-character mapping provided by
3940      * {@code Character.forDigit} is used, and a minus sign is
3941      * prepended if appropriate.  (This representation is compatible
3942      * with the {@link #BigInteger(String) (String)} constructor, and
3943      * allows for String concatenation with Java's + operator.)
3944      *
3945      * @return decimal String representation of this BigInteger.
3946      * @see    Character#forDigit
3947      * @see    #BigInteger(java.lang.String)
3948      */
3949     public String toString() {
3950         return toString(10);
3951     }
3952 
3953     /**
3954      * Returns a byte array containing the two's-complement
3955      * representation of this BigInteger.  The byte array will be in
3956      * <i>big-endian</i> byte-order: the most significant byte is in
3957      * the zeroth element.  The array will contain the minimum number
3958      * of bytes required to represent this BigInteger, including at
3959      * least one sign bit, which is {@code (ceil((this.bitLength() +
3960      * 1)/8))}.  (This representation is compatible with the
3961      * {@link #BigInteger(byte[]) (byte[])} constructor.)
3962      *
3963      * @return a byte array containing the two's-complement representation of
3964      *         this BigInteger.
3965      * @see    #BigInteger(byte[])
3966      */
3967     public byte[] toByteArray() {
3968         int byteLen = bitLength()/8 + 1;
3969         byte[] byteArray = new byte[byteLen];
3970 
3971         for (int i=byteLen-1, bytesCopied=4, nextInt=0, intIndex=0; i >= 0; i--) {
3972             if (bytesCopied == 4) {
3973                 nextInt = getInt(intIndex++);
3974                 bytesCopied = 1;
3975             } else {
3976                 nextInt >>>= 8;
3977                 bytesCopied++;
3978             }
3979             byteArray[i] = (byte)nextInt;
3980         }
3981         return byteArray;
3982     }
3983 
3984     /**
3985      * Converts this BigInteger to an {@code int}.  This
3986      * conversion is analogous to a
3987      * <i>narrowing primitive conversion</i> from {@code long} to
3988      * {@code int} as defined in section 5.1.3 of
3989      * <cite>The Java&trade; Language Specification</cite>:
3990      * if this BigInteger is too big to fit in an
3991      * {@code int}, only the low-order 32 bits are returned.
3992      * Note that this conversion can lose information about the
3993      * overall magnitude of the BigInteger value as well as return a
3994      * result with the opposite sign.
3995      *
3996      * @return this BigInteger converted to an {@code int}.
3997      * @see #intValueExact()
3998      */
3999     public int intValue() {
4000         int result = 0;
4001         result = getInt(0);
4002         return result;
4003     }
4004 
4005     /**
4006      * Converts this BigInteger to a {@code long}.  This
4007      * conversion is analogous to a
4008      * <i>narrowing primitive conversion</i> from {@code long} to
4009      * {@code int} as defined in section 5.1.3 of
4010      * <cite>The Java&trade; Language Specification</cite>:
4011      * if this BigInteger is too big to fit in a
4012      * {@code long}, only the low-order 64 bits are returned.
4013      * Note that this conversion can lose information about the
4014      * overall magnitude of the BigInteger value as well as return a
4015      * result with the opposite sign.
4016      *
4017      * @return this BigInteger converted to a {@code long}.
4018      * @see #longValueExact()
4019      */
4020     public long longValue() {
4021         long result = 0;
4022 
4023         for (int i=1; i >= 0; i--)
4024             result = (result << 32) + (getInt(i) & LONG_MASK);
4025         return result;
4026     }
4027 
4028     /**
4029      * Converts this BigInteger to a {@code float}.  This
4030      * conversion is similar to the
4031      * <i>narrowing primitive conversion</i> from {@code double} to
4032      * {@code float} as defined in section 5.1.3 of
4033      * <cite>The Java&trade; Language Specification</cite>:
4034      * if this BigInteger has too great a magnitude
4035      * to represent as a {@code float}, it will be converted to
4036      * {@link Float#NEGATIVE_INFINITY} or {@link
4037      * Float#POSITIVE_INFINITY} as appropriate.  Note that even when
4038      * the return value is finite, this conversion can lose
4039      * information about the precision of the BigInteger value.
4040      *
4041      * @return this BigInteger converted to a {@code float}.
4042      */
4043     public float floatValue() {
4044         if (signum == 0) {
4045             return 0.0f;
4046         }
4047 
4048         int exponent = ((mag.length - 1) << 5) + bitLengthForInt(mag[0]) - 1;
4049 
4050         // exponent == floor(log2(abs(this)))
4051         if (exponent < Long.SIZE - 1) {
4052             return longValue();
4053         } else if (exponent > Float.MAX_EXPONENT) {
4054             return signum > 0 ? Float.POSITIVE_INFINITY : Float.NEGATIVE_INFINITY;
4055         }
4056 
4057         /*
4058          * We need the top SIGNIFICAND_WIDTH bits, including the "implicit"
4059          * one bit. To make rounding easier, we pick out the top
4060          * SIGNIFICAND_WIDTH + 1 bits, so we have one to help us round up or
4061          * down. twiceSignifFloor will contain the top SIGNIFICAND_WIDTH + 1
4062          * bits, and signifFloor the top SIGNIFICAND_WIDTH.
4063          *
4064          * It helps to consider the real number signif = abs(this) *
4065          * 2^(SIGNIFICAND_WIDTH - 1 - exponent).
4066          */
4067         int shift = exponent - FloatConsts.SIGNIFICAND_WIDTH;
4068 
4069         int twiceSignifFloor;
4070         // twiceSignifFloor will be == abs().shiftRight(shift).intValue()
4071         // We do the shift into an int directly to improve performance.
4072 
4073         int nBits = shift & 0x1f;
4074         int nBits2 = 32 - nBits;
4075 
4076         if (nBits == 0) {
4077             twiceSignifFloor = mag[0];
4078         } else {
4079             twiceSignifFloor = mag[0] >>> nBits;
4080             if (twiceSignifFloor == 0) {
4081                 twiceSignifFloor = (mag[0] << nBits2) | (mag[1] >>> nBits);
4082             }
4083         }
4084 
4085         int signifFloor = twiceSignifFloor >> 1;
4086         signifFloor &= FloatConsts.SIGNIF_BIT_MASK; // remove the implied bit
4087 
4088         /*
4089          * We round up if either the fractional part of signif is strictly
4090          * greater than 0.5 (which is true if the 0.5 bit is set and any lower
4091          * bit is set), or if the fractional part of signif is >= 0.5 and
4092          * signifFloor is odd (which is true if both the 0.5 bit and the 1 bit
4093          * are set). This is equivalent to the desired HALF_EVEN rounding.
4094          */
4095         boolean increment = (twiceSignifFloor & 1) != 0
4096                 && ((signifFloor & 1) != 0 || abs().getLowestSetBit() < shift);
4097         int signifRounded = increment ? signifFloor + 1 : signifFloor;
4098         int bits = ((exponent + FloatConsts.EXP_BIAS))
4099                 << (FloatConsts.SIGNIFICAND_WIDTH - 1);
4100         bits += signifRounded;
4101         /*
4102          * If signifRounded == 2^24, we'd need to set all of the significand
4103          * bits to zero and add 1 to the exponent. This is exactly the behavior
4104          * we get from just adding signifRounded to bits directly. If the
4105          * exponent is Float.MAX_EXPONENT, we round up (correctly) to
4106          * Float.POSITIVE_INFINITY.
4107          */
4108         bits |= signum & FloatConsts.SIGN_BIT_MASK;
4109         return Float.intBitsToFloat(bits);
4110     }
4111 
4112     /**
4113      * Converts this BigInteger to a {@code double}.  This
4114      * conversion is similar to the
4115      * <i>narrowing primitive conversion</i> from {@code double} to
4116      * {@code float} as defined in section 5.1.3 of
4117      * <cite>The Java&trade; Language Specification</cite>:
4118      * if this BigInteger has too great a magnitude
4119      * to represent as a {@code double}, it will be converted to
4120      * {@link Double#NEGATIVE_INFINITY} or {@link
4121      * Double#POSITIVE_INFINITY} as appropriate.  Note that even when
4122      * the return value is finite, this conversion can lose
4123      * information about the precision of the BigInteger value.
4124      *
4125      * @return this BigInteger converted to a {@code double}.
4126      */
4127     public double doubleValue() {
4128         if (signum == 0) {
4129             return 0.0;
4130         }
4131 
4132         int exponent = ((mag.length - 1) << 5) + bitLengthForInt(mag[0]) - 1;
4133 
4134         // exponent == floor(log2(abs(this))Double)
4135         if (exponent < Long.SIZE - 1) {
4136             return longValue();
4137         } else if (exponent > Double.MAX_EXPONENT) {
4138             return signum > 0 ? Double.POSITIVE_INFINITY : Double.NEGATIVE_INFINITY;
4139         }
4140 
4141         /*
4142          * We need the top SIGNIFICAND_WIDTH bits, including the "implicit"
4143          * one bit. To make rounding easier, we pick out the top
4144          * SIGNIFICAND_WIDTH + 1 bits, so we have one to help us round up or
4145          * down. twiceSignifFloor will contain the top SIGNIFICAND_WIDTH + 1
4146          * bits, and signifFloor the top SIGNIFICAND_WIDTH.
4147          *
4148          * It helps to consider the real number signif = abs(this) *
4149          * 2^(SIGNIFICAND_WIDTH - 1 - exponent).
4150          */
4151         int shift = exponent - DoubleConsts.SIGNIFICAND_WIDTH;
4152 
4153         long twiceSignifFloor;
4154         // twiceSignifFloor will be == abs().shiftRight(shift).longValue()
4155         // We do the shift into a long directly to improve performance.
4156 
4157         int nBits = shift & 0x1f;
4158         int nBits2 = 32 - nBits;
4159 
4160         int highBits;
4161         int lowBits;
4162         if (nBits == 0) {
4163             highBits = mag[0];
4164             lowBits = mag[1];
4165         } else {
4166             highBits = mag[0] >>> nBits;
4167             lowBits = (mag[0] << nBits2) | (mag[1] >>> nBits);
4168             if (highBits == 0) {
4169                 highBits = lowBits;
4170                 lowBits = (mag[1] << nBits2) | (mag[2] >>> nBits);
4171             }
4172         }
4173 
4174         twiceSignifFloor = ((highBits & LONG_MASK) << 32)
4175                 | (lowBits & LONG_MASK);
4176 
4177         long signifFloor = twiceSignifFloor >> 1;
4178         signifFloor &= DoubleConsts.SIGNIF_BIT_MASK; // remove the implied bit
4179 
4180         /*
4181          * We round up if either the fractional part of signif is strictly
4182          * greater than 0.5 (which is true if the 0.5 bit is set and any lower
4183          * bit is set), or if the fractional part of signif is >= 0.5 and
4184          * signifFloor is odd (which is true if both the 0.5 bit and the 1 bit
4185          * are set). This is equivalent to the desired HALF_EVEN rounding.
4186          */
4187         boolean increment = (twiceSignifFloor & 1) != 0
4188                 && ((signifFloor & 1) != 0 || abs().getLowestSetBit() < shift);
4189         long signifRounded = increment ? signifFloor + 1 : signifFloor;
4190         long bits = (long) ((exponent + DoubleConsts.EXP_BIAS))
4191                 << (DoubleConsts.SIGNIFICAND_WIDTH - 1);
4192         bits += signifRounded;
4193         /*
4194          * If signifRounded == 2^53, we'd need to set all of the significand
4195          * bits to zero and add 1 to the exponent. This is exactly the behavior
4196          * we get from just adding signifRounded to bits directly. If the
4197          * exponent is Double.MAX_EXPONENT, we round up (correctly) to
4198          * Double.POSITIVE_INFINITY.
4199          */
4200         bits |= signum & DoubleConsts.SIGN_BIT_MASK;
4201         return Double.longBitsToDouble(bits);
4202     }
4203 
4204     /**
4205      * Returns a copy of the input array stripped of any leading zero bytes.
4206      */
4207     private static int[] stripLeadingZeroInts(int val[]) {
4208         int vlen = val.length;
4209         int keep;
4210 
4211         // Find first nonzero byte
4212         for (keep = 0; keep < vlen && val[keep] == 0; keep++)
4213             ;
4214         return java.util.Arrays.copyOfRange(val, keep, vlen);
4215     }
4216 
4217     /**
4218      * Returns the input array stripped of any leading zero bytes.
4219      * Since the source is trusted the copying may be skipped.
4220      */
4221     private static int[] trustedStripLeadingZeroInts(int val[]) {
4222         int vlen = val.length;
4223         int keep;
4224 
4225         // Find first nonzero byte
4226         for (keep = 0; keep < vlen && val[keep] == 0; keep++)
4227             ;
4228         return keep == 0 ? val : java.util.Arrays.copyOfRange(val, keep, vlen);
4229     }
4230 
4231     /**
4232      * Returns a copy of the input array stripped of any leading zero bytes.
4233      */
4234     private static int[] stripLeadingZeroBytes(byte a[], int off, int len) {
4235         int indexBound = off + len;
4236         int keep;
4237 
4238         // Find first nonzero byte
4239         for (keep = off; keep < indexBound && a[keep] == 0; keep++)
4240             ;
4241 
4242         // Allocate new array and copy relevant part of input array
4243         int intLength = ((indexBound - keep) + 3) >>> 2;
4244         int[] result = new int[intLength];
4245         int b = indexBound - 1;
4246         for (int i = intLength-1; i >= 0; i--) {
4247             result[i] = a[b--] & 0xff;
4248             int bytesRemaining = b - keep + 1;
4249             int bytesToTransfer = Math.min(3, bytesRemaining);
4250             for (int j=8; j <= (bytesToTransfer << 3); j += 8)
4251                 result[i] |= ((a[b--] & 0xff) << j);
4252         }
4253         return result;
4254     }
4255 
4256     /**
4257      * Takes an array a representing a negative 2's-complement number and
4258      * returns the minimal (no leading zero bytes) unsigned whose value is -a.
4259      */
4260     private static int[] makePositive(byte a[], int off, int len) {
4261         int keep, k;
4262         int indexBound = off + len;
4263 
4264         // Find first non-sign (0xff) byte of input
4265         for (keep=off; keep < indexBound && a[keep] == -1; keep++)
4266             ;
4267 
4268 
4269         /* Allocate output array.  If all non-sign bytes are 0x00, we must
4270          * allocate space for one extra output byte. */
4271         for (k=keep; k < indexBound && a[k] == 0; k++)
4272             ;
4273 
4274         int extraByte = (k == indexBound) ? 1 : 0;
4275         int intLength = ((indexBound - keep + extraByte) + 3) >>> 2;
4276         int result[] = new int[intLength];
4277 
4278         /* Copy one's complement of input into output, leaving extra
4279          * byte (if it exists) == 0x00 */
4280         int b = indexBound - 1;
4281         for (int i = intLength-1; i >= 0; i--) {
4282             result[i] = a[b--] & 0xff;
4283             int numBytesToTransfer = Math.min(3, b-keep+1);
4284             if (numBytesToTransfer < 0)
4285                 numBytesToTransfer = 0;
4286             for (int j=8; j <= 8*numBytesToTransfer; j += 8)
4287                 result[i] |= ((a[b--] & 0xff) << j);
4288 
4289             // Mask indicates which bits must be complemented
4290             int mask = -1 >>> (8*(3-numBytesToTransfer));
4291             result[i] = ~result[i] & mask;
4292         }
4293 
4294         // Add one to one's complement to generate two's complement
4295         for (int i=result.length-1; i >= 0; i--) {
4296             result[i] = (int)((result[i] & LONG_MASK) + 1);
4297             if (result[i] != 0)
4298                 break;
4299         }
4300 
4301         return result;
4302     }
4303 
4304     /**
4305      * Takes an array a representing a negative 2's-complement number and
4306      * returns the minimal (no leading zero ints) unsigned whose value is -a.
4307      */
4308     private static int[] makePositive(int a[]) {
4309         int keep, j;
4310 
4311         // Find first non-sign (0xffffffff) int of input
4312         for (keep=0; keep < a.length && a[keep] == -1; keep++)
4313             ;
4314 
4315         /* Allocate output array.  If all non-sign ints are 0x00, we must
4316          * allocate space for one extra output int. */
4317         for (j=keep; j < a.length && a[j] == 0; j++)
4318             ;
4319         int extraInt = (j == a.length ? 1 : 0);
4320         int result[] = new int[a.length - keep + extraInt];
4321 
4322         /* Copy one's complement of input into output, leaving extra
4323          * int (if it exists) == 0x00 */
4324         for (int i = keep; i < a.length; i++)
4325             result[i - keep + extraInt] = ~a[i];
4326 
4327         // Add one to one's complement to generate two's complement
4328         for (int i=result.length-1; ++result[i] == 0; i--)
4329             ;
4330 
4331         return result;
4332     }
4333 
4334     /*
4335      * The following two arrays are used for fast String conversions.  Both
4336      * are indexed by radix.  The first is the number of digits of the given
4337      * radix that can fit in a Java long without "going negative", i.e., the
4338      * highest integer n such that radix**n < 2**63.  The second is the
4339      * "long radix" that tears each number into "long digits", each of which
4340      * consists of the number of digits in the corresponding element in
4341      * digitsPerLong (longRadix[i] = i**digitPerLong[i]).  Both arrays have
4342      * nonsense values in their 0 and 1 elements, as radixes 0 and 1 are not
4343      * used.
4344      */
4345     private static int digitsPerLong[] = {0, 0,
4346         62, 39, 31, 27, 24, 22, 20, 19, 18, 18, 17, 17, 16, 16, 15, 15, 15, 14,
4347         14, 14, 14, 13, 13, 13, 13, 13, 13, 12, 12, 12, 12, 12, 12, 12, 12};
4348 
4349     private static BigInteger longRadix[] = {null, null,
4350         valueOf(0x4000000000000000L), valueOf(0x383d9170b85ff80bL),
4351         valueOf(0x4000000000000000L), valueOf(0x6765c793fa10079dL),
4352         valueOf(0x41c21cb8e1000000L), valueOf(0x3642798750226111L),
4353         valueOf(0x1000000000000000L), valueOf(0x12bf307ae81ffd59L),
4354         valueOf( 0xde0b6b3a7640000L), valueOf(0x4d28cb56c33fa539L),
4355         valueOf(0x1eca170c00000000L), valueOf(0x780c7372621bd74dL),
4356         valueOf(0x1e39a5057d810000L), valueOf(0x5b27ac993df97701L),
4357         valueOf(0x1000000000000000L), valueOf(0x27b95e997e21d9f1L),
4358         valueOf(0x5da0e1e53c5c8000L), valueOf( 0xb16a458ef403f19L),
4359         valueOf(0x16bcc41e90000000L), valueOf(0x2d04b7fdd9c0ef49L),
4360         valueOf(0x5658597bcaa24000L), valueOf( 0x6feb266931a75b7L),
4361         valueOf( 0xc29e98000000000L), valueOf(0x14adf4b7320334b9L),
4362         valueOf(0x226ed36478bfa000L), valueOf(0x383d9170b85ff80bL),
4363         valueOf(0x5a3c23e39c000000L), valueOf( 0x4e900abb53e6b71L),
4364         valueOf( 0x7600ec618141000L), valueOf( 0xaee5720ee830681L),
4365         valueOf(0x1000000000000000L), valueOf(0x172588ad4f5f0981L),
4366         valueOf(0x211e44f7d02c1000L), valueOf(0x2ee56725f06e5c71L),
4367         valueOf(0x41c21cb8e1000000L)};
4368 
4369     /*
4370      * These two arrays are the integer analogue of above.
4371      */
4372     private static int digitsPerInt[] = {0, 0, 30, 19, 15, 13, 11,
4373         11, 10, 9, 9, 8, 8, 8, 8, 7, 7, 7, 7, 7, 7, 7, 6, 6, 6, 6,
4374         6, 6, 6, 6, 6, 6, 6, 6, 6, 6, 5};
4375 
4376     private static int intRadix[] = {0, 0,
4377         0x40000000, 0x4546b3db, 0x40000000, 0x48c27395, 0x159fd800,
4378         0x75db9c97, 0x40000000, 0x17179149, 0x3b9aca00, 0xcc6db61,
4379         0x19a10000, 0x309f1021, 0x57f6c100, 0xa2f1b6f,  0x10000000,
4380         0x18754571, 0x247dbc80, 0x3547667b, 0x4c4b4000, 0x6b5a6e1d,
4381         0x6c20a40,  0x8d2d931,  0xb640000,  0xe8d4a51,  0x1269ae40,
4382         0x17179149, 0x1cb91000, 0x23744899, 0x2b73a840, 0x34e63b41,
4383         0x40000000, 0x4cfa3cc1, 0x5c13d840, 0x6d91b519, 0x39aa400
4384     };
4385 
4386     /**
4387      * These routines provide access to the two's complement representation
4388      * of BigIntegers.
4389      */
4390 
4391     /**
4392      * Returns the length of the two's complement representation in ints,
4393      * including space for at least one sign bit.
4394      */
4395     private int intLength() {
4396         return (bitLength() >>> 5) + 1;
4397     }
4398 
4399     /* Returns sign bit */
4400     private int signBit() {
4401         return signum < 0 ? 1 : 0;
4402     }
4403 
4404     /* Returns an int of sign bits */
4405     private int signInt() {
4406         return signum < 0 ? -1 : 0;
4407     }
4408 
4409     /**
4410      * Returns the specified int of the little-endian two's complement
4411      * representation (int 0 is the least significant).  The int number can
4412      * be arbitrarily high (values are logically preceded by infinitely many
4413      * sign ints).
4414      */
4415     private int getInt(int n) {
4416         if (n < 0)
4417             return 0;
4418         if (n >= mag.length)
4419             return signInt();
4420 
4421         int magInt = mag[mag.length-n-1];
4422 
4423         return (signum >= 0 ? magInt :
4424                 (n <= firstNonzeroIntNum() ? -magInt : ~magInt));
4425     }
4426 
4427     /**
4428     * Returns the index of the int that contains the first nonzero int in the
4429     * little-endian binary representation of the magnitude (int 0 is the
4430     * least significant). If the magnitude is zero, return value is undefined.
4431     *
4432     * <p>Note: never used for a BigInteger with a magnitude of zero.
4433     * @see #getInt.
4434     */
4435     private int firstNonzeroIntNum() {
4436         int fn = firstNonzeroIntNumPlusTwo - 2;
4437         if (fn == -2) { // firstNonzeroIntNum not initialized yet
4438             // Search for the first nonzero int
4439             int i;
4440             int mlen = mag.length;
4441             for (i = mlen - 1; i >= 0 && mag[i] == 0; i--)
4442                 ;
4443             fn = mlen - i - 1;
4444             firstNonzeroIntNumPlusTwo = fn + 2; // offset by two to initialize
4445         }
4446         return fn;
4447     }
4448 
4449     /** use serialVersionUID from JDK 1.1. for interoperability */
4450     private static final long serialVersionUID = -8287574255936472291L;
4451 
4452     /**
4453      * Serializable fields for BigInteger.
4454      *
4455      * @serialField signum  int
4456      *              signum of this BigInteger
4457      * @serialField magnitude byte[]
4458      *              magnitude array of this BigInteger
4459      * @serialField bitCount  int
4460      *              appears in the serialized form for backward compatibility
4461      * @serialField bitLength int
4462      *              appears in the serialized form for backward compatibility
4463      * @serialField firstNonzeroByteNum int
4464      *              appears in the serialized form for backward compatibility
4465      * @serialField lowestSetBit int
4466      *              appears in the serialized form for backward compatibility
4467      */
4468     private static final ObjectStreamField[] serialPersistentFields = {
4469         new ObjectStreamField("signum", Integer.TYPE),
4470         new ObjectStreamField("magnitude", byte[].class),
4471         new ObjectStreamField("bitCount", Integer.TYPE),
4472         new ObjectStreamField("bitLength", Integer.TYPE),
4473         new ObjectStreamField("firstNonzeroByteNum", Integer.TYPE),
4474         new ObjectStreamField("lowestSetBit", Integer.TYPE)
4475         };
4476 
4477     /**
4478      * Reconstitute the {@code BigInteger} instance from a stream (that is,
4479      * deserialize it). The magnitude is read in as an array of bytes
4480      * for historical reasons, but it is converted to an array of ints
4481      * and the byte array is discarded.
4482      * Note:
4483      * The current convention is to initialize the cache fields, bitCountPlusOne,
4484      * bitLengthPlusOne and lowestSetBitPlusTwo, to 0 rather than some other
4485      * marker value. Therefore, no explicit action to set these fields needs to
4486      * be taken in readObject because those fields already have a 0 value by
4487      * default since defaultReadObject is not being used.
4488      */
4489     private void readObject(java.io.ObjectInputStream s)
4490         throws java.io.IOException, ClassNotFoundException {
4491         // prepare to read the alternate persistent fields
4492         ObjectInputStream.GetField fields = s.readFields();
4493 
4494         // Read the alternate persistent fields that we care about
4495         int sign = fields.get("signum", -2);
4496         byte[] magnitude = (byte[])fields.get("magnitude", null);
4497 
4498         // Validate signum
4499         if (sign < -1 || sign > 1) {
4500             String message = "BigInteger: Invalid signum value";
4501             if (fields.defaulted("signum"))
4502                 message = "BigInteger: Signum not present in stream";
4503             throw new java.io.StreamCorruptedException(message);
4504         }
4505         int[] mag = stripLeadingZeroBytes(magnitude, 0, magnitude.length);
4506         if ((mag.length == 0) != (sign == 0)) {
4507             String message = "BigInteger: signum-magnitude mismatch";
4508             if (fields.defaulted("magnitude"))
4509                 message = "BigInteger: Magnitude not present in stream";
4510             throw new java.io.StreamCorruptedException(message);
4511         }
4512 
4513         // Commit final fields via Unsafe
4514         UnsafeHolder.putSign(this, sign);
4515 
4516         // Calculate mag field from magnitude and discard magnitude
4517         UnsafeHolder.putMag(this, mag);
4518         if (mag.length >= MAX_MAG_LENGTH) {
4519             try {
4520                 checkRange();
4521             } catch (ArithmeticException e) {
4522                 throw new java.io.StreamCorruptedException("BigInteger: Out of the supported range");
4523             }
4524         }
4525     }
4526 
4527     // Support for resetting final fields while deserializing
4528     private static class UnsafeHolder {
4529         private static final jdk.internal.misc.Unsafe unsafe;
4530         private static final long signumOffset;
4531         private static final long magOffset;
4532         static {
4533             try {
4534                 unsafe = jdk.internal.misc.Unsafe.getUnsafe();
4535                 signumOffset = unsafe.objectFieldOffset
4536                     (BigInteger.class.getDeclaredField("signum"));
4537                 magOffset = unsafe.objectFieldOffset
4538                     (BigInteger.class.getDeclaredField("mag"));
4539             } catch (Exception ex) {
4540                 throw new ExceptionInInitializerError(ex);
4541             }
4542         }
4543 
4544         static void putSign(BigInteger bi, int sign) {
4545             unsafe.putInt(bi, signumOffset, sign);
4546         }
4547 
4548         static void putMag(BigInteger bi, int[] magnitude) {
4549             unsafe.putObject(bi, magOffset, magnitude);
4550         }
4551     }
4552 
4553     /**
4554      * Save the {@code BigInteger} instance to a stream.  The magnitude of a
4555      * {@code BigInteger} is serialized as a byte array for historical reasons.
4556      * To maintain compatibility with older implementations, the integers
4557      * -1, -1, -2, and -2 are written as the values of the obsolete fields
4558      * {@code bitCount}, {@code bitLength}, {@code lowestSetBit}, and
4559      * {@code firstNonzeroByteNum}, respectively.  These values are compatible
4560      * with older implementations, but will be ignored by current
4561      * implementations.
4562      */
4563     private void writeObject(ObjectOutputStream s) throws IOException {
4564         // set the values of the Serializable fields
4565         ObjectOutputStream.PutField fields = s.putFields();
4566         fields.put("signum", signum);
4567         fields.put("magnitude", magSerializedForm());
4568         // The values written for cached fields are compatible with older
4569         // versions, but are ignored in readObject so don't otherwise matter.
4570         fields.put("bitCount", -1);
4571         fields.put("bitLength", -1);
4572         fields.put("lowestSetBit", -2);
4573         fields.put("firstNonzeroByteNum", -2);
4574 
4575         // save them
4576         s.writeFields();
4577     }
4578 
4579     /**
4580      * Returns the mag array as an array of bytes.
4581      */
4582     private byte[] magSerializedForm() {
4583         int len = mag.length;
4584 
4585         int bitLen = (len == 0 ? 0 : ((len - 1) << 5) + bitLengthForInt(mag[0]));
4586         int byteLen = (bitLen + 7) >>> 3;
4587         byte[] result = new byte[byteLen];
4588 
4589         for (int i = byteLen - 1, bytesCopied = 4, intIndex = len - 1, nextInt = 0;
4590              i >= 0; i--) {
4591             if (bytesCopied == 4) {
4592                 nextInt = mag[intIndex--];
4593                 bytesCopied = 1;
4594             } else {
4595                 nextInt >>>= 8;
4596                 bytesCopied++;
4597             }
4598             result[i] = (byte)nextInt;
4599         }
4600         return result;
4601     }
4602 
4603     /**
4604      * Converts this {@code BigInteger} to a {@code long}, checking
4605      * for lost information.  If the value of this {@code BigInteger}
4606      * is out of the range of the {@code long} type, then an
4607      * {@code ArithmeticException} is thrown.
4608      *
4609      * @return this {@code BigInteger} converted to a {@code long}.
4610      * @throws ArithmeticException if the value of {@code this} will
4611      * not exactly fit in a {@code long}.
4612      * @see BigInteger#longValue
4613      * @since  1.8
4614      */
4615     public long longValueExact() {
4616         if (mag.length <= 2 && bitLength() <= 63)
4617             return longValue();
4618         else
4619             throw new ArithmeticException("BigInteger out of long range");
4620     }
4621 
4622     /**
4623      * Converts this {@code BigInteger} to an {@code int}, checking
4624      * for lost information.  If the value of this {@code BigInteger}
4625      * is out of the range of the {@code int} type, then an
4626      * {@code ArithmeticException} is thrown.
4627      *
4628      * @return this {@code BigInteger} converted to an {@code int}.
4629      * @throws ArithmeticException if the value of {@code this} will
4630      * not exactly fit in a {@code int}.
4631      * @see BigInteger#intValue
4632      * @since  1.8
4633      */
4634     public int intValueExact() {
4635         if (mag.length <= 1 && bitLength() <= 31)
4636             return intValue();
4637         else
4638             throw new ArithmeticException("BigInteger out of int range");
4639     }
4640 
4641     /**
4642      * Converts this {@code BigInteger} to a {@code short}, checking
4643      * for lost information.  If the value of this {@code BigInteger}
4644      * is out of the range of the {@code short} type, then an
4645      * {@code ArithmeticException} is thrown.
4646      *
4647      * @return this {@code BigInteger} converted to a {@code short}.
4648      * @throws ArithmeticException if the value of {@code this} will
4649      * not exactly fit in a {@code short}.
4650      * @see BigInteger#shortValue
4651      * @since  1.8
4652      */
4653     public short shortValueExact() {
4654         if (mag.length <= 1 && bitLength() <= 31) {
4655             int value = intValue();
4656             if (value >= Short.MIN_VALUE && value <= Short.MAX_VALUE)
4657                 return shortValue();
4658         }
4659         throw new ArithmeticException("BigInteger out of short range");
4660     }
4661 
4662     /**
4663      * Converts this {@code BigInteger} to a {@code byte}, checking
4664      * for lost information.  If the value of this {@code BigInteger}
4665      * is out of the range of the {@code byte} type, then an
4666      * {@code ArithmeticException} is thrown.
4667      *
4668      * @return this {@code BigInteger} converted to a {@code byte}.
4669      * @throws ArithmeticException if the value of {@code this} will
4670      * not exactly fit in a {@code byte}.
4671      * @see BigInteger#byteValue
4672      * @since  1.8
4673      */
4674     public byte byteValueExact() {
4675         if (mag.length <= 1 && bitLength() <= 31) {
4676             int value = intValue();
4677             if (value >= Byte.MIN_VALUE && value <= Byte.MAX_VALUE)
4678                 return byteValue();
4679         }
4680         throw new ArithmeticException("BigInteger out of byte range");
4681     }
4682 }