1 /* 2 * Copyright (c) 1997, 2018, Oracle and/or its affiliates. All rights reserved. 3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. 4 * 5 * This code is free software; you can redistribute it and/or modify it 6 * under the terms of the GNU General Public License version 2 only, as 7 * published by the Free Software Foundation. Oracle designates this 8 * particular file as subject to the "Classpath" exception as provided 9 * by Oracle in the LICENSE file that accompanied this code. 10 * 11 * This code is distributed in the hope that it will be useful, but WITHOUT 12 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or 13 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License 14 * version 2 for more details (a copy is included in the LICENSE file that 15 * accompanied this code). 16 * 17 * You should have received a copy of the GNU General Public License version 18 * 2 along with this work; if not, write to the Free Software Foundation, 19 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. 20 * 21 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA 22 * or visit www.oracle.com if you need additional information or have any 23 * questions. 24 */ 25 26 package java.util; 27 28 import java.io.IOException; 29 import java.io.InvalidObjectException; 30 import java.io.Serializable; 31 import java.lang.reflect.ParameterizedType; 32 import java.lang.reflect.Type; 33 import java.util.function.BiConsumer; 34 import java.util.function.BiFunction; 35 import java.util.function.Consumer; 36 import java.util.function.Function; 37 import jdk.internal.access.SharedSecrets; 38 39 /** 40 * Hash table based implementation of the {@code Map} interface. This 41 * implementation provides all of the optional map operations, and permits 42 * {@code null} values and the {@code null} key. (The {@code HashMap} 43 * class is roughly equivalent to {@code Hashtable}, except that it is 44 * unsynchronized and permits nulls.) This class makes no guarantees as to 45 * the order of the map; in particular, it does not guarantee that the order 46 * will remain constant over time. 47 * 48 * <p>This implementation provides constant-time performance for the basic 49 * operations ({@code get} and {@code put}), assuming the hash function 50 * disperses the elements properly among the buckets. Iteration over 51 * collection views requires time proportional to the "capacity" of the 52 * {@code HashMap} instance (the number of buckets) plus its size (the number 53 * of key-value mappings). Thus, it's very important not to set the initial 54 * capacity too high (or the load factor too low) if iteration performance is 55 * important. 56 * 57 * <p>An instance of {@code HashMap} has two parameters that affect its 58 * performance: <i>initial capacity</i> and <i>load factor</i>. The 59 * <i>capacity</i> is the number of buckets in the hash table, and the initial 60 * capacity is simply the capacity at the time the hash table is created. The 61 * <i>load factor</i> is a measure of how full the hash table is allowed to 62 * get before its capacity is automatically increased. When the number of 63 * entries in the hash table exceeds the product of the load factor and the 64 * current capacity, the hash table is <i>rehashed</i> (that is, internal data 65 * structures are rebuilt) so that the hash table has approximately twice the 66 * number of buckets. 67 * 68 * <p>As a general rule, the default load factor (.75) offers a good 69 * tradeoff between time and space costs. Higher values decrease the 70 * space overhead but increase the lookup cost (reflected in most of 71 * the operations of the {@code HashMap} class, including 72 * {@code get} and {@code put}). The expected number of entries in 73 * the map and its load factor should be taken into account when 74 * setting its initial capacity, so as to minimize the number of 75 * rehash operations. If the initial capacity is greater than the 76 * maximum number of entries divided by the load factor, no rehash 77 * operations will ever occur. 78 * 79 * <p>If many mappings are to be stored in a {@code HashMap} 80 * instance, creating it with a sufficiently large capacity will allow 81 * the mappings to be stored more efficiently than letting it perform 82 * automatic rehashing as needed to grow the table. Note that using 83 * many keys with the same {@code hashCode()} is a sure way to slow 84 * down performance of any hash table. To ameliorate impact, when keys 85 * are {@link Comparable}, this class may use comparison order among 86 * keys to help break ties. 87 * 88 * <p><strong>Note that this implementation is not synchronized.</strong> 89 * If multiple threads access a hash map concurrently, and at least one of 90 * the threads modifies the map structurally, it <i>must</i> be 91 * synchronized externally. (A structural modification is any operation 92 * that adds or deletes one or more mappings; merely changing the value 93 * associated with a key that an instance already contains is not a 94 * structural modification.) This is typically accomplished by 95 * synchronizing on some object that naturally encapsulates the map. 96 * 97 * If no such object exists, the map should be "wrapped" using the 98 * {@link Collections#synchronizedMap Collections.synchronizedMap} 99 * method. This is best done at creation time, to prevent accidental 100 * unsynchronized access to the map:<pre> 101 * Map m = Collections.synchronizedMap(new HashMap(...));</pre> 102 * 103 * <p>The iterators returned by all of this class's "collection view methods" 104 * are <i>fail-fast</i>: if the map is structurally modified at any time after 105 * the iterator is created, in any way except through the iterator's own 106 * {@code remove} method, the iterator will throw a 107 * {@link ConcurrentModificationException}. Thus, in the face of concurrent 108 * modification, the iterator fails quickly and cleanly, rather than risking 109 * arbitrary, non-deterministic behavior at an undetermined time in the 110 * future. 111 * 112 * <p>Note that the fail-fast behavior of an iterator cannot be guaranteed 113 * as it is, generally speaking, impossible to make any hard guarantees in the 114 * presence of unsynchronized concurrent modification. Fail-fast iterators 115 * throw {@code ConcurrentModificationException} on a best-effort basis. 116 * Therefore, it would be wrong to write a program that depended on this 117 * exception for its correctness: <i>the fail-fast behavior of iterators 118 * should be used only to detect bugs.</i> 119 * 120 * <p>This class is a member of the 121 * <a href="{@docRoot}/java.base/java/util/package-summary.html#CollectionsFramework"> 122 * Java Collections Framework</a>. 123 * 124 * @param <K> the type of keys maintained by this map 125 * @param <V> the type of mapped values 126 * 127 * @author Doug Lea 128 * @author Josh Bloch 129 * @author Arthur van Hoff 130 * @author Neal Gafter 131 * @see Object#hashCode() 132 * @see Collection 133 * @see Map 134 * @see TreeMap 135 * @see Hashtable 136 * @since 1.2 137 */ 138 public class HashMap<K,V> extends AbstractMap<K,V> 139 implements Map<K,V>, Cloneable, Serializable { 140 141 private static final long serialVersionUID = 362498820763181265L; 142 143 /* 144 * Implementation notes. 145 * 146 * This map usually acts as a binned (bucketed) hash table, but 147 * when bins get too large, they are transformed into bins of 148 * TreeNodes, each structured similarly to those in 149 * java.util.TreeMap. Most methods try to use normal bins, but 150 * relay to TreeNode methods when applicable (simply by checking 151 * instanceof a node). Bins of TreeNodes may be traversed and 152 * used like any others, but additionally support faster lookup 153 * when overpopulated. However, since the vast majority of bins in 154 * normal use are not overpopulated, checking for existence of 155 * tree bins may be delayed in the course of table methods. 156 * 157 * Tree bins (i.e., bins whose elements are all TreeNodes) are 158 * ordered primarily by hashCode, but in the case of ties, if two 159 * elements are of the same "class C implements Comparable<C>", 160 * type then their compareTo method is used for ordering. (We 161 * conservatively check generic types via reflection to validate 162 * this -- see method comparableClassFor). The added complexity 163 * of tree bins is worthwhile in providing worst-case O(log n) 164 * operations when keys either have distinct hashes or are 165 * orderable, Thus, performance degrades gracefully under 166 * accidental or malicious usages in which hashCode() methods 167 * return values that are poorly distributed, as well as those in 168 * which many keys share a hashCode, so long as they are also 169 * Comparable. (If neither of these apply, we may waste about a 170 * factor of two in time and space compared to taking no 171 * precautions. But the only known cases stem from poor user 172 * programming practices that are already so slow that this makes 173 * little difference.) 174 * 175 * Because TreeNodes are about twice the size of regular nodes, we 176 * use them only when bins contain enough nodes to warrant use 177 * (see TREEIFY_THRESHOLD). And when they become too small (due to 178 * removal or resizing) they are converted back to plain bins. In 179 * usages with well-distributed user hashCodes, tree bins are 180 * rarely used. Ideally, under random hashCodes, the frequency of 181 * nodes in bins follows a Poisson distribution 182 * (http://en.wikipedia.org/wiki/Poisson_distribution) with a 183 * parameter of about 0.5 on average for the default resizing 184 * threshold of 0.75, although with a large variance because of 185 * resizing granularity. Ignoring variance, the expected 186 * occurrences of list size k are (exp(-0.5) * pow(0.5, k) / 187 * factorial(k)). The first values are: 188 * 189 * 0: 0.60653066 190 * 1: 0.30326533 191 * 2: 0.07581633 192 * 3: 0.01263606 193 * 4: 0.00157952 194 * 5: 0.00015795 195 * 6: 0.00001316 196 * 7: 0.00000094 197 * 8: 0.00000006 198 * more: less than 1 in ten million 199 * 200 * The root of a tree bin is normally its first node. However, 201 * sometimes (currently only upon Iterator.remove), the root might 202 * be elsewhere, but can be recovered following parent links 203 * (method TreeNode.root()). 204 * 205 * All applicable internal methods accept a hash code as an 206 * argument (as normally supplied from a public method), allowing 207 * them to call each other without recomputing user hashCodes. 208 * Most internal methods also accept a "tab" argument, that is 209 * normally the current table, but may be a new or old one when 210 * resizing or converting. 211 * 212 * When bin lists are treeified, split, or untreeified, we keep 213 * them in the same relative access/traversal order (i.e., field 214 * Node.next) to better preserve locality, and to slightly 215 * simplify handling of splits and traversals that invoke 216 * iterator.remove. When using comparators on insertion, to keep a 217 * total ordering (or as close as is required here) across 218 * rebalancings, we compare classes and identityHashCodes as 219 * tie-breakers. 220 * 221 * The use and transitions among plain vs tree modes is 222 * complicated by the existence of subclass LinkedHashMap. See 223 * below for hook methods defined to be invoked upon insertion, 224 * removal and access that allow LinkedHashMap internals to 225 * otherwise remain independent of these mechanics. (This also 226 * requires that a map instance be passed to some utility methods 227 * that may create new nodes.) 228 * 229 * The concurrent-programming-like SSA-based coding style helps 230 * avoid aliasing errors amid all of the twisty pointer operations. 231 */ 232 233 /** 234 * The default initial capacity - MUST be a power of two. 235 */ 236 static final int DEFAULT_INITIAL_CAPACITY = 1 << 4; // aka 16 237 238 /** 239 * The maximum capacity, used if a higher value is implicitly specified 240 * by either of the constructors with arguments. 241 * MUST be a power of two <= 1<<30. 242 */ 243 static final int MAXIMUM_CAPACITY = 1 << 30; 244 245 /** 246 * The load factor used when none specified in constructor. 247 */ 248 static final float DEFAULT_LOAD_FACTOR = 0.75f; 249 250 /** 251 * The bin count threshold for using a tree rather than list for a 252 * bin. Bins are converted to trees when adding an element to a 253 * bin with at least this many nodes. The value must be greater 254 * than 2 and should be at least 8 to mesh with assumptions in 255 * tree removal about conversion back to plain bins upon 256 * shrinkage. 257 */ 258 static final int TREEIFY_THRESHOLD = 8; 259 260 /** 261 * The bin count threshold for untreeifying a (split) bin during a 262 * resize operation. Should be less than TREEIFY_THRESHOLD, and at 263 * most 6 to mesh with shrinkage detection under removal. 264 */ 265 static final int UNTREEIFY_THRESHOLD = 6; 266 267 /** 268 * The smallest table capacity for which bins may be treeified. 269 * (Otherwise the table is resized if too many nodes in a bin.) 270 * Should be at least 4 * TREEIFY_THRESHOLD to avoid conflicts 271 * between resizing and treeification thresholds. 272 */ 273 static final int MIN_TREEIFY_CAPACITY = 64; 274 275 /** 276 * Basic hash bin node, used for most entries. (See below for 277 * TreeNode subclass, and in LinkedHashMap for its Entry subclass.) 278 */ 279 static class Node<K,V> implements Map.Entry<K,V> { 280 final int hash; 281 final K key; 282 V value; 283 Node<K,V> next; 284 Node(int hash, K key, V value, Node<K,V> next)285 Node(int hash, K key, V value, Node<K,V> next) { 286 this.hash = hash; 287 this.key = key; 288 this.value = value; 289 this.next = next; 290 } 291 getKey()292 public final K getKey() { return key; } getValue()293 public final V getValue() { return value; } toString()294 public final String toString() { return key + "=" + value; } 295 hashCode()296 public final int hashCode() { 297 return Objects.hashCode(key) ^ Objects.hashCode(value); 298 } 299 setValue(V newValue)300 public final V setValue(V newValue) { 301 V oldValue = value; 302 value = newValue; 303 return oldValue; 304 } 305 equals(Object o)306 public final boolean equals(Object o) { 307 if (o == this) 308 return true; 309 if (o instanceof Map.Entry) { 310 Map.Entry<?,?> e = (Map.Entry<?,?>)o; 311 if (Objects.equals(key, e.getKey()) && 312 Objects.equals(value, e.getValue())) 313 return true; 314 } 315 return false; 316 } 317 } 318 319 /* ---------------- Static utilities -------------- */ 320 321 /** 322 * Computes key.hashCode() and spreads (XORs) higher bits of hash 323 * to lower. Because the table uses power-of-two masking, sets of 324 * hashes that vary only in bits above the current mask will 325 * always collide. (Among known examples are sets of Float keys 326 * holding consecutive whole numbers in small tables.) So we 327 * apply a transform that spreads the impact of higher bits 328 * downward. There is a tradeoff between speed, utility, and 329 * quality of bit-spreading. Because many common sets of hashes 330 * are already reasonably distributed (so don't benefit from 331 * spreading), and because we use trees to handle large sets of 332 * collisions in bins, we just XOR some shifted bits in the 333 * cheapest possible way to reduce systematic lossage, as well as 334 * to incorporate impact of the highest bits that would otherwise 335 * never be used in index calculations because of table bounds. 336 */ hash(Object key)337 static final int hash(Object key) { 338 int h; 339 return (key == null) ? 0 : (h = key.hashCode()) ^ (h >>> 16); 340 } 341 342 /** 343 * Returns x's Class if it is of the form "class C implements 344 * Comparable<C>", else null. 345 */ comparableClassFor(Object x)346 static Class<?> comparableClassFor(Object x) { 347 if (x instanceof Comparable) { 348 Class<?> c; Type[] ts, as; ParameterizedType p; 349 if ((c = x.getClass()) == String.class) // bypass checks 350 return c; 351 if ((ts = c.getGenericInterfaces()) != null) { 352 for (Type t : ts) { 353 if ((t instanceof ParameterizedType) && 354 ((p = (ParameterizedType) t).getRawType() == 355 Comparable.class) && 356 (as = p.getActualTypeArguments()) != null && 357 as.length == 1 && as[0] == c) // type arg is c 358 return c; 359 } 360 } 361 } 362 return null; 363 } 364 365 /** 366 * Returns k.compareTo(x) if x matches kc (k's screened comparable 367 * class), else 0. 368 */ 369 @SuppressWarnings({"rawtypes","unchecked"}) // for cast to Comparable compareComparables(Class<?> kc, Object k, Object x)370 static int compareComparables(Class<?> kc, Object k, Object x) { 371 return (x == null || x.getClass() != kc ? 0 : 372 ((Comparable)k).compareTo(x)); 373 } 374 375 /** 376 * Returns a power of two size for the given target capacity. 377 */ tableSizeFor(int cap)378 static final int tableSizeFor(int cap) { 379 int n = -1 >>> Integer.numberOfLeadingZeros(cap - 1); 380 return (n < 0) ? 1 : (n >= MAXIMUM_CAPACITY) ? MAXIMUM_CAPACITY : n + 1; 381 } 382 383 /* ---------------- Fields -------------- */ 384 385 /** 386 * The table, initialized on first use, and resized as 387 * necessary. When allocated, length is always a power of two. 388 * (We also tolerate length zero in some operations to allow 389 * bootstrapping mechanics that are currently not needed.) 390 */ 391 transient Node<K,V>[] table; 392 393 /** 394 * Holds cached entrySet(). Note that AbstractMap fields are used 395 * for keySet() and values(). 396 */ 397 transient Set<Map.Entry<K,V>> entrySet; 398 399 /** 400 * The number of key-value mappings contained in this map. 401 */ 402 transient int size; 403 404 /** 405 * The number of times this HashMap has been structurally modified 406 * Structural modifications are those that change the number of mappings in 407 * the HashMap or otherwise modify its internal structure (e.g., 408 * rehash). This field is used to make iterators on Collection-views of 409 * the HashMap fail-fast. (See ConcurrentModificationException). 410 */ 411 transient int modCount; 412 413 /** 414 * The next size value at which to resize (capacity * load factor). 415 * 416 * @serial 417 */ 418 // (The javadoc description is true upon serialization. 419 // Additionally, if the table array has not been allocated, this 420 // field holds the initial array capacity, or zero signifying 421 // DEFAULT_INITIAL_CAPACITY.) 422 int threshold; 423 424 /** 425 * The load factor for the hash table. 426 * 427 * @serial 428 */ 429 final float loadFactor; 430 431 /* ---------------- Public operations -------------- */ 432 433 /** 434 * Constructs an empty {@code HashMap} with the specified initial 435 * capacity and load factor. 436 * 437 * @param initialCapacity the initial capacity 438 * @param loadFactor the load factor 439 * @throws IllegalArgumentException if the initial capacity is negative 440 * or the load factor is nonpositive 441 */ HashMap(int initialCapacity, float loadFactor)442 public HashMap(int initialCapacity, float loadFactor) { 443 if (initialCapacity < 0) 444 throw new IllegalArgumentException("Illegal initial capacity: " + 445 initialCapacity); 446 if (initialCapacity > MAXIMUM_CAPACITY) 447 initialCapacity = MAXIMUM_CAPACITY; 448 if (loadFactor <= 0 || Float.isNaN(loadFactor)) 449 throw new IllegalArgumentException("Illegal load factor: " + 450 loadFactor); 451 this.loadFactor = loadFactor; 452 this.threshold = tableSizeFor(initialCapacity); 453 } 454 455 /** 456 * Constructs an empty {@code HashMap} with the specified initial 457 * capacity and the default load factor (0.75). 458 * 459 * @param initialCapacity the initial capacity. 460 * @throws IllegalArgumentException if the initial capacity is negative. 461 */ HashMap(int initialCapacity)462 public HashMap(int initialCapacity) { 463 this(initialCapacity, DEFAULT_LOAD_FACTOR); 464 } 465 466 /** 467 * Constructs an empty {@code HashMap} with the default initial capacity 468 * (16) and the default load factor (0.75). 469 */ HashMap()470 public HashMap() { 471 this.loadFactor = DEFAULT_LOAD_FACTOR; // all other fields defaulted 472 } 473 474 /** 475 * Constructs a new {@code HashMap} with the same mappings as the 476 * specified {@code Map}. The {@code HashMap} is created with 477 * default load factor (0.75) and an initial capacity sufficient to 478 * hold the mappings in the specified {@code Map}. 479 * 480 * @param m the map whose mappings are to be placed in this map 481 * @throws NullPointerException if the specified map is null 482 */ HashMap(Map<? extends K, ? extends V> m)483 public HashMap(Map<? extends K, ? extends V> m) { 484 this.loadFactor = DEFAULT_LOAD_FACTOR; 485 putMapEntries(m, false); 486 } 487 488 /** 489 * Implements Map.putAll and Map constructor. 490 * 491 * @param m the map 492 * @param evict false when initially constructing this map, else 493 * true (relayed to method afterNodeInsertion). 494 */ putMapEntries(Map<? extends K, ? extends V> m, boolean evict)495 final void putMapEntries(Map<? extends K, ? extends V> m, boolean evict) { 496 int s = m.size(); 497 if (s > 0) { 498 if (table == null) { // pre-size 499 float ft = ((float)s / loadFactor) + 1.0F; 500 int t = ((ft < (float)MAXIMUM_CAPACITY) ? 501 (int)ft : MAXIMUM_CAPACITY); 502 if (t > threshold) 503 threshold = tableSizeFor(t); 504 } 505 else if (s > threshold) 506 resize(); 507 for (Map.Entry<? extends K, ? extends V> e : m.entrySet()) { 508 K key = e.getKey(); 509 V value = e.getValue(); 510 putVal(hash(key), key, value, false, evict); 511 } 512 } 513 } 514 515 /** 516 * Returns the number of key-value mappings in this map. 517 * 518 * @return the number of key-value mappings in this map 519 */ size()520 public int size() { 521 return size; 522 } 523 524 /** 525 * Returns {@code true} if this map contains no key-value mappings. 526 * 527 * @return {@code true} if this map contains no key-value mappings 528 */ isEmpty()529 public boolean isEmpty() { 530 return size == 0; 531 } 532 533 /** 534 * Returns the value to which the specified key is mapped, 535 * or {@code null} if this map contains no mapping for the key. 536 * 537 * <p>More formally, if this map contains a mapping from a key 538 * {@code k} to a value {@code v} such that {@code (key==null ? k==null : 539 * key.equals(k))}, then this method returns {@code v}; otherwise 540 * it returns {@code null}. (There can be at most one such mapping.) 541 * 542 * <p>A return value of {@code null} does not <i>necessarily</i> 543 * indicate that the map contains no mapping for the key; it's also 544 * possible that the map explicitly maps the key to {@code null}. 545 * The {@link #containsKey containsKey} operation may be used to 546 * distinguish these two cases. 547 * 548 * @see #put(Object, Object) 549 */ get(Object key)550 public V get(Object key) { 551 Node<K,V> e; 552 return (e = getNode(hash(key), key)) == null ? null : e.value; 553 } 554 555 /** 556 * Implements Map.get and related methods. 557 * 558 * @param hash hash for key 559 * @param key the key 560 * @return the node, or null if none 561 */ getNode(int hash, Object key)562 final Node<K,V> getNode(int hash, Object key) { 563 Node<K,V>[] tab; Node<K,V> first, e; int n; K k; 564 if ((tab = table) != null && (n = tab.length) > 0 && 565 (first = tab[(n - 1) & hash]) != null) { 566 if (first.hash == hash && // always check first node 567 ((k = first.key) == key || (key != null && key.equals(k)))) 568 return first; 569 if ((e = first.next) != null) { 570 if (first instanceof TreeNode) 571 return ((TreeNode<K,V>)first).getTreeNode(hash, key); 572 do { 573 if (e.hash == hash && 574 ((k = e.key) == key || (key != null && key.equals(k)))) 575 return e; 576 } while ((e = e.next) != null); 577 } 578 } 579 return null; 580 } 581 582 /** 583 * Returns {@code true} if this map contains a mapping for the 584 * specified key. 585 * 586 * @param key The key whose presence in this map is to be tested 587 * @return {@code true} if this map contains a mapping for the specified 588 * key. 589 */ containsKey(Object key)590 public boolean containsKey(Object key) { 591 return getNode(hash(key), key) != null; 592 } 593 594 /** 595 * Associates the specified value with the specified key in this map. 596 * If the map previously contained a mapping for the key, the old 597 * value is replaced. 598 * 599 * @param key key with which the specified value is to be associated 600 * @param value value to be associated with the specified key 601 * @return the previous value associated with {@code key}, or 602 * {@code null} if there was no mapping for {@code key}. 603 * (A {@code null} return can also indicate that the map 604 * previously associated {@code null} with {@code key}.) 605 */ put(K key, V value)606 public V put(K key, V value) { 607 return putVal(hash(key), key, value, false, true); 608 } 609 610 /** 611 * Implements Map.put and related methods. 612 * 613 * @param hash hash for key 614 * @param key the key 615 * @param value the value to put 616 * @param onlyIfAbsent if true, don't change existing value 617 * @param evict if false, the table is in creation mode. 618 * @return previous value, or null if none 619 */ putVal(int hash, K key, V value, boolean onlyIfAbsent, boolean evict)620 final V putVal(int hash, K key, V value, boolean onlyIfAbsent, 621 boolean evict) { 622 Node<K,V>[] tab; Node<K,V> p; int n, i; 623 if ((tab = table) == null || (n = tab.length) == 0) 624 n = (tab = resize()).length; 625 if ((p = tab[i = (n - 1) & hash]) == null) 626 tab[i] = newNode(hash, key, value, null); 627 else { 628 Node<K,V> e; K k; 629 if (p.hash == hash && 630 ((k = p.key) == key || (key != null && key.equals(k)))) 631 e = p; 632 else if (p instanceof TreeNode) 633 e = ((TreeNode<K,V>)p).putTreeVal(this, tab, hash, key, value); 634 else { 635 for (int binCount = 0; ; ++binCount) { 636 if ((e = p.next) == null) { 637 p.next = newNode(hash, key, value, null); 638 if (binCount >= TREEIFY_THRESHOLD - 1) // -1 for 1st 639 treeifyBin(tab, hash); 640 break; 641 } 642 if (e.hash == hash && 643 ((k = e.key) == key || (key != null && key.equals(k)))) 644 break; 645 p = e; 646 } 647 } 648 if (e != null) { // existing mapping for key 649 V oldValue = e.value; 650 if (!onlyIfAbsent || oldValue == null) 651 e.value = value; 652 afterNodeAccess(e); 653 return oldValue; 654 } 655 } 656 ++modCount; 657 if (++size > threshold) 658 resize(); 659 afterNodeInsertion(evict); 660 return null; 661 } 662 663 /** 664 * Initializes or doubles table size. If null, allocates in 665 * accord with initial capacity target held in field threshold. 666 * Otherwise, because we are using power-of-two expansion, the 667 * elements from each bin must either stay at same index, or move 668 * with a power of two offset in the new table. 669 * 670 * @return the table 671 */ resize()672 final Node<K,V>[] resize() { 673 Node<K,V>[] oldTab = table; 674 int oldCap = (oldTab == null) ? 0 : oldTab.length; 675 int oldThr = threshold; 676 int newCap, newThr = 0; 677 if (oldCap > 0) { 678 if (oldCap >= MAXIMUM_CAPACITY) { 679 threshold = Integer.MAX_VALUE; 680 return oldTab; 681 } 682 else if ((newCap = oldCap << 1) < MAXIMUM_CAPACITY && 683 oldCap >= DEFAULT_INITIAL_CAPACITY) 684 newThr = oldThr << 1; // double threshold 685 } 686 else if (oldThr > 0) // initial capacity was placed in threshold 687 newCap = oldThr; 688 else { // zero initial threshold signifies using defaults 689 newCap = DEFAULT_INITIAL_CAPACITY; 690 newThr = (int)(DEFAULT_LOAD_FACTOR * DEFAULT_INITIAL_CAPACITY); 691 } 692 if (newThr == 0) { 693 float ft = (float)newCap * loadFactor; 694 newThr = (newCap < MAXIMUM_CAPACITY && ft < (float)MAXIMUM_CAPACITY ? 695 (int)ft : Integer.MAX_VALUE); 696 } 697 threshold = newThr; 698 @SuppressWarnings({"rawtypes","unchecked"}) 699 Node<K,V>[] newTab = (Node<K,V>[])new Node[newCap]; 700 table = newTab; 701 if (oldTab != null) { 702 for (int j = 0; j < oldCap; ++j) { 703 Node<K,V> e; 704 if ((e = oldTab[j]) != null) { 705 oldTab[j] = null; 706 if (e.next == null) 707 newTab[e.hash & (newCap - 1)] = e; 708 else if (e instanceof TreeNode) 709 ((TreeNode<K,V>)e).split(this, newTab, j, oldCap); 710 else { // preserve order 711 Node<K,V> loHead = null, loTail = null; 712 Node<K,V> hiHead = null, hiTail = null; 713 Node<K,V> next; 714 do { 715 next = e.next; 716 if ((e.hash & oldCap) == 0) { 717 if (loTail == null) 718 loHead = e; 719 else 720 loTail.next = e; 721 loTail = e; 722 } 723 else { 724 if (hiTail == null) 725 hiHead = e; 726 else 727 hiTail.next = e; 728 hiTail = e; 729 } 730 } while ((e = next) != null); 731 if (loTail != null) { 732 loTail.next = null; 733 newTab[j] = loHead; 734 } 735 if (hiTail != null) { 736 hiTail.next = null; 737 newTab[j + oldCap] = hiHead; 738 } 739 } 740 } 741 } 742 } 743 return newTab; 744 } 745 746 /** 747 * Replaces all linked nodes in bin at index for given hash unless 748 * table is too small, in which case resizes instead. 749 */ treeifyBin(Node<K,V>[] tab, int hash)750 final void treeifyBin(Node<K,V>[] tab, int hash) { 751 int n, index; Node<K,V> e; 752 if (tab == null || (n = tab.length) < MIN_TREEIFY_CAPACITY) 753 resize(); 754 else if ((e = tab[index = (n - 1) & hash]) != null) { 755 TreeNode<K,V> hd = null, tl = null; 756 do { 757 TreeNode<K,V> p = replacementTreeNode(e, null); 758 if (tl == null) 759 hd = p; 760 else { 761 p.prev = tl; 762 tl.next = p; 763 } 764 tl = p; 765 } while ((e = e.next) != null); 766 if ((tab[index] = hd) != null) 767 hd.treeify(tab); 768 } 769 } 770 771 /** 772 * Copies all of the mappings from the specified map to this map. 773 * These mappings will replace any mappings that this map had for 774 * any of the keys currently in the specified map. 775 * 776 * @param m mappings to be stored in this map 777 * @throws NullPointerException if the specified map is null 778 */ putAll(Map<? extends K, ? extends V> m)779 public void putAll(Map<? extends K, ? extends V> m) { 780 putMapEntries(m, true); 781 } 782 783 /** 784 * Removes the mapping for the specified key from this map if present. 785 * 786 * @param key key whose mapping is to be removed from the map 787 * @return the previous value associated with {@code key}, or 788 * {@code null} if there was no mapping for {@code key}. 789 * (A {@code null} return can also indicate that the map 790 * previously associated {@code null} with {@code key}.) 791 */ remove(Object key)792 public V remove(Object key) { 793 Node<K,V> e; 794 return (e = removeNode(hash(key), key, null, false, true)) == null ? 795 null : e.value; 796 } 797 798 /** 799 * Implements Map.remove and related methods. 800 * 801 * @param hash hash for key 802 * @param key the key 803 * @param value the value to match if matchValue, else ignored 804 * @param matchValue if true only remove if value is equal 805 * @param movable if false do not move other nodes while removing 806 * @return the node, or null if none 807 */ removeNode(int hash, Object key, Object value, boolean matchValue, boolean movable)808 final Node<K,V> removeNode(int hash, Object key, Object value, 809 boolean matchValue, boolean movable) { 810 Node<K,V>[] tab; Node<K,V> p; int n, index; 811 if ((tab = table) != null && (n = tab.length) > 0 && 812 (p = tab[index = (n - 1) & hash]) != null) { 813 Node<K,V> node = null, e; K k; V v; 814 if (p.hash == hash && 815 ((k = p.key) == key || (key != null && key.equals(k)))) 816 node = p; 817 else if ((e = p.next) != null) { 818 if (p instanceof TreeNode) 819 node = ((TreeNode<K,V>)p).getTreeNode(hash, key); 820 else { 821 do { 822 if (e.hash == hash && 823 ((k = e.key) == key || 824 (key != null && key.equals(k)))) { 825 node = e; 826 break; 827 } 828 p = e; 829 } while ((e = e.next) != null); 830 } 831 } 832 if (node != null && (!matchValue || (v = node.value) == value || 833 (value != null && value.equals(v)))) { 834 if (node instanceof TreeNode) 835 ((TreeNode<K,V>)node).removeTreeNode(this, tab, movable); 836 else if (node == p) 837 tab[index] = node.next; 838 else 839 p.next = node.next; 840 ++modCount; 841 --size; 842 afterNodeRemoval(node); 843 return node; 844 } 845 } 846 return null; 847 } 848 849 /** 850 * Removes all of the mappings from this map. 851 * The map will be empty after this call returns. 852 */ clear()853 public void clear() { 854 Node<K,V>[] tab; 855 modCount++; 856 if ((tab = table) != null && size > 0) { 857 size = 0; 858 for (int i = 0; i < tab.length; ++i) 859 tab[i] = null; 860 } 861 } 862 863 /** 864 * Returns {@code true} if this map maps one or more keys to the 865 * specified value. 866 * 867 * @param value value whose presence in this map is to be tested 868 * @return {@code true} if this map maps one or more keys to the 869 * specified value 870 */ containsValue(Object value)871 public boolean containsValue(Object value) { 872 Node<K,V>[] tab; V v; 873 if ((tab = table) != null && size > 0) { 874 for (Node<K,V> e : tab) { 875 for (; e != null; e = e.next) { 876 if ((v = e.value) == value || 877 (value != null && value.equals(v))) 878 return true; 879 } 880 } 881 } 882 return false; 883 } 884 885 /** 886 * Returns a {@link Set} view of the keys contained in this map. 887 * The set is backed by the map, so changes to the map are 888 * reflected in the set, and vice-versa. If the map is modified 889 * while an iteration over the set is in progress (except through 890 * the iterator's own {@code remove} operation), the results of 891 * the iteration are undefined. The set supports element removal, 892 * which removes the corresponding mapping from the map, via the 893 * {@code Iterator.remove}, {@code Set.remove}, 894 * {@code removeAll}, {@code retainAll}, and {@code clear} 895 * operations. It does not support the {@code add} or {@code addAll} 896 * operations. 897 * 898 * @return a set view of the keys contained in this map 899 */ keySet()900 public Set<K> keySet() { 901 Set<K> ks = keySet; 902 if (ks == null) { 903 ks = new KeySet(); 904 keySet = ks; 905 } 906 return ks; 907 } 908 909 final class KeySet extends AbstractSet<K> { size()910 public final int size() { return size; } clear()911 public final void clear() { HashMap.this.clear(); } iterator()912 public final Iterator<K> iterator() { return new KeyIterator(); } contains(Object o)913 public final boolean contains(Object o) { return containsKey(o); } remove(Object key)914 public final boolean remove(Object key) { 915 return removeNode(hash(key), key, null, false, true) != null; 916 } spliterator()917 public final Spliterator<K> spliterator() { 918 return new KeySpliterator<>(HashMap.this, 0, -1, 0, 0); 919 } forEach(Consumer<? super K> action)920 public final void forEach(Consumer<? super K> action) { 921 Node<K,V>[] tab; 922 if (action == null) 923 throw new NullPointerException(); 924 if (size > 0 && (tab = table) != null) { 925 int mc = modCount; 926 for (Node<K,V> e : tab) { 927 for (; e != null; e = e.next) 928 action.accept(e.key); 929 } 930 if (modCount != mc) 931 throw new ConcurrentModificationException(); 932 } 933 } 934 } 935 936 /** 937 * Returns a {@link Collection} view of the values contained in this map. 938 * The collection is backed by the map, so changes to the map are 939 * reflected in the collection, and vice-versa. If the map is 940 * modified while an iteration over the collection is in progress 941 * (except through the iterator's own {@code remove} operation), 942 * the results of the iteration are undefined. The collection 943 * supports element removal, which removes the corresponding 944 * mapping from the map, via the {@code Iterator.remove}, 945 * {@code Collection.remove}, {@code removeAll}, 946 * {@code retainAll} and {@code clear} operations. It does not 947 * support the {@code add} or {@code addAll} operations. 948 * 949 * @return a view of the values contained in this map 950 */ values()951 public Collection<V> values() { 952 Collection<V> vs = values; 953 if (vs == null) { 954 vs = new Values(); 955 values = vs; 956 } 957 return vs; 958 } 959 960 final class Values extends AbstractCollection<V> { size()961 public final int size() { return size; } clear()962 public final void clear() { HashMap.this.clear(); } iterator()963 public final Iterator<V> iterator() { return new ValueIterator(); } contains(Object o)964 public final boolean contains(Object o) { return containsValue(o); } spliterator()965 public final Spliterator<V> spliterator() { 966 return new ValueSpliterator<>(HashMap.this, 0, -1, 0, 0); 967 } forEach(Consumer<? super V> action)968 public final void forEach(Consumer<? super V> action) { 969 Node<K,V>[] tab; 970 if (action == null) 971 throw new NullPointerException(); 972 if (size > 0 && (tab = table) != null) { 973 int mc = modCount; 974 for (Node<K,V> e : tab) { 975 for (; e != null; e = e.next) 976 action.accept(e.value); 977 } 978 if (modCount != mc) 979 throw new ConcurrentModificationException(); 980 } 981 } 982 } 983 984 /** 985 * Returns a {@link Set} view of the mappings contained in this map. 986 * The set is backed by the map, so changes to the map are 987 * reflected in the set, and vice-versa. If the map is modified 988 * while an iteration over the set is in progress (except through 989 * the iterator's own {@code remove} operation, or through the 990 * {@code setValue} operation on a map entry returned by the 991 * iterator) the results of the iteration are undefined. The set 992 * supports element removal, which removes the corresponding 993 * mapping from the map, via the {@code Iterator.remove}, 994 * {@code Set.remove}, {@code removeAll}, {@code retainAll} and 995 * {@code clear} operations. It does not support the 996 * {@code add} or {@code addAll} operations. 997 * 998 * @return a set view of the mappings contained in this map 999 */ entrySet()1000 public Set<Map.Entry<K,V>> entrySet() { 1001 Set<Map.Entry<K,V>> es; 1002 return (es = entrySet) == null ? (entrySet = new EntrySet()) : es; 1003 } 1004 1005 final class EntrySet extends AbstractSet<Map.Entry<K,V>> { size()1006 public final int size() { return size; } clear()1007 public final void clear() { HashMap.this.clear(); } iterator()1008 public final Iterator<Map.Entry<K,V>> iterator() { 1009 return new EntryIterator(); 1010 } contains(Object o)1011 public final boolean contains(Object o) { 1012 if (!(o instanceof Map.Entry)) 1013 return false; 1014 Map.Entry<?,?> e = (Map.Entry<?,?>) o; 1015 Object key = e.getKey(); 1016 Node<K,V> candidate = getNode(hash(key), key); 1017 return candidate != null && candidate.equals(e); 1018 } remove(Object o)1019 public final boolean remove(Object o) { 1020 if (o instanceof Map.Entry) { 1021 Map.Entry<?,?> e = (Map.Entry<?,?>) o; 1022 Object key = e.getKey(); 1023 Object value = e.getValue(); 1024 return removeNode(hash(key), key, value, true, true) != null; 1025 } 1026 return false; 1027 } spliterator()1028 public final Spliterator<Map.Entry<K,V>> spliterator() { 1029 return new EntrySpliterator<>(HashMap.this, 0, -1, 0, 0); 1030 } forEach(Consumer<? super Map.Entry<K,V>> action)1031 public final void forEach(Consumer<? super Map.Entry<K,V>> action) { 1032 Node<K,V>[] tab; 1033 if (action == null) 1034 throw new NullPointerException(); 1035 if (size > 0 && (tab = table) != null) { 1036 int mc = modCount; 1037 for (Node<K,V> e : tab) { 1038 for (; e != null; e = e.next) 1039 action.accept(e); 1040 } 1041 if (modCount != mc) 1042 throw new ConcurrentModificationException(); 1043 } 1044 } 1045 } 1046 1047 // Overrides of JDK8 Map extension methods 1048 1049 @Override getOrDefault(Object key, V defaultValue)1050 public V getOrDefault(Object key, V defaultValue) { 1051 Node<K,V> e; 1052 return (e = getNode(hash(key), key)) == null ? defaultValue : e.value; 1053 } 1054 1055 @Override putIfAbsent(K key, V value)1056 public V putIfAbsent(K key, V value) { 1057 return putVal(hash(key), key, value, true, true); 1058 } 1059 1060 @Override remove(Object key, Object value)1061 public boolean remove(Object key, Object value) { 1062 return removeNode(hash(key), key, value, true, true) != null; 1063 } 1064 1065 @Override replace(K key, V oldValue, V newValue)1066 public boolean replace(K key, V oldValue, V newValue) { 1067 Node<K,V> e; V v; 1068 if ((e = getNode(hash(key), key)) != null && 1069 ((v = e.value) == oldValue || (v != null && v.equals(oldValue)))) { 1070 e.value = newValue; 1071 afterNodeAccess(e); 1072 return true; 1073 } 1074 return false; 1075 } 1076 1077 @Override replace(K key, V value)1078 public V replace(K key, V value) { 1079 Node<K,V> e; 1080 if ((e = getNode(hash(key), key)) != null) { 1081 V oldValue = e.value; 1082 e.value = value; 1083 afterNodeAccess(e); 1084 return oldValue; 1085 } 1086 return null; 1087 } 1088 1089 /** 1090 * {@inheritDoc} 1091 * 1092 * <p>This method will, on a best-effort basis, throw a 1093 * {@link ConcurrentModificationException} if it is detected that the 1094 * mapping function modifies this map during computation. 1095 * 1096 * @throws ConcurrentModificationException if it is detected that the 1097 * mapping function modified this map 1098 */ 1099 @Override computeIfAbsent(K key, Function<? super K, ? extends V> mappingFunction)1100 public V computeIfAbsent(K key, 1101 Function<? super K, ? extends V> mappingFunction) { 1102 if (mappingFunction == null) 1103 throw new NullPointerException(); 1104 int hash = hash(key); 1105 Node<K,V>[] tab; Node<K,V> first; int n, i; 1106 int binCount = 0; 1107 TreeNode<K,V> t = null; 1108 Node<K,V> old = null; 1109 if (size > threshold || (tab = table) == null || 1110 (n = tab.length) == 0) 1111 n = (tab = resize()).length; 1112 if ((first = tab[i = (n - 1) & hash]) != null) { 1113 if (first instanceof TreeNode) 1114 old = (t = (TreeNode<K,V>)first).getTreeNode(hash, key); 1115 else { 1116 Node<K,V> e = first; K k; 1117 do { 1118 if (e.hash == hash && 1119 ((k = e.key) == key || (key != null && key.equals(k)))) { 1120 old = e; 1121 break; 1122 } 1123 ++binCount; 1124 } while ((e = e.next) != null); 1125 } 1126 V oldValue; 1127 if (old != null && (oldValue = old.value) != null) { 1128 afterNodeAccess(old); 1129 return oldValue; 1130 } 1131 } 1132 int mc = modCount; 1133 V v = mappingFunction.apply(key); 1134 if (mc != modCount) { throw new ConcurrentModificationException(); } 1135 if (v == null) { 1136 return null; 1137 } else if (old != null) { 1138 old.value = v; 1139 afterNodeAccess(old); 1140 return v; 1141 } 1142 else if (t != null) 1143 t.putTreeVal(this, tab, hash, key, v); 1144 else { 1145 tab[i] = newNode(hash, key, v, first); 1146 if (binCount >= TREEIFY_THRESHOLD - 1) 1147 treeifyBin(tab, hash); 1148 } 1149 modCount = mc + 1; 1150 ++size; 1151 afterNodeInsertion(true); 1152 return v; 1153 } 1154 1155 /** 1156 * {@inheritDoc} 1157 * 1158 * <p>This method will, on a best-effort basis, throw a 1159 * {@link ConcurrentModificationException} if it is detected that the 1160 * remapping function modifies this map during computation. 1161 * 1162 * @throws ConcurrentModificationException if it is detected that the 1163 * remapping function modified this map 1164 */ 1165 @Override computeIfPresent(K key, BiFunction<? super K, ? super V, ? extends V> remappingFunction)1166 public V computeIfPresent(K key, 1167 BiFunction<? super K, ? super V, ? extends V> remappingFunction) { 1168 if (remappingFunction == null) 1169 throw new NullPointerException(); 1170 Node<K,V> e; V oldValue; 1171 int hash = hash(key); 1172 if ((e = getNode(hash, key)) != null && 1173 (oldValue = e.value) != null) { 1174 int mc = modCount; 1175 V v = remappingFunction.apply(key, oldValue); 1176 if (mc != modCount) { throw new ConcurrentModificationException(); } 1177 if (v != null) { 1178 e.value = v; 1179 afterNodeAccess(e); 1180 return v; 1181 } 1182 else 1183 removeNode(hash, key, null, false, true); 1184 } 1185 return null; 1186 } 1187 1188 /** 1189 * {@inheritDoc} 1190 * 1191 * <p>This method will, on a best-effort basis, throw a 1192 * {@link ConcurrentModificationException} if it is detected that the 1193 * remapping function modifies this map during computation. 1194 * 1195 * @throws ConcurrentModificationException if it is detected that the 1196 * remapping function modified this map 1197 */ 1198 @Override compute(K key, BiFunction<? super K, ? super V, ? extends V> remappingFunction)1199 public V compute(K key, 1200 BiFunction<? super K, ? super V, ? extends V> remappingFunction) { 1201 if (remappingFunction == null) 1202 throw new NullPointerException(); 1203 int hash = hash(key); 1204 Node<K,V>[] tab; Node<K,V> first; int n, i; 1205 int binCount = 0; 1206 TreeNode<K,V> t = null; 1207 Node<K,V> old = null; 1208 if (size > threshold || (tab = table) == null || 1209 (n = tab.length) == 0) 1210 n = (tab = resize()).length; 1211 if ((first = tab[i = (n - 1) & hash]) != null) { 1212 if (first instanceof TreeNode) 1213 old = (t = (TreeNode<K,V>)first).getTreeNode(hash, key); 1214 else { 1215 Node<K,V> e = first; K k; 1216 do { 1217 if (e.hash == hash && 1218 ((k = e.key) == key || (key != null && key.equals(k)))) { 1219 old = e; 1220 break; 1221 } 1222 ++binCount; 1223 } while ((e = e.next) != null); 1224 } 1225 } 1226 V oldValue = (old == null) ? null : old.value; 1227 int mc = modCount; 1228 V v = remappingFunction.apply(key, oldValue); 1229 if (mc != modCount) { throw new ConcurrentModificationException(); } 1230 if (old != null) { 1231 if (v != null) { 1232 old.value = v; 1233 afterNodeAccess(old); 1234 } 1235 else 1236 removeNode(hash, key, null, false, true); 1237 } 1238 else if (v != null) { 1239 if (t != null) 1240 t.putTreeVal(this, tab, hash, key, v); 1241 else { 1242 tab[i] = newNode(hash, key, v, first); 1243 if (binCount >= TREEIFY_THRESHOLD - 1) 1244 treeifyBin(tab, hash); 1245 } 1246 modCount = mc + 1; 1247 ++size; 1248 afterNodeInsertion(true); 1249 } 1250 return v; 1251 } 1252 1253 /** 1254 * {@inheritDoc} 1255 * 1256 * <p>This method will, on a best-effort basis, throw a 1257 * {@link ConcurrentModificationException} if it is detected that the 1258 * remapping function modifies this map during computation. 1259 * 1260 * @throws ConcurrentModificationException if it is detected that the 1261 * remapping function modified this map 1262 */ 1263 @Override merge(K key, V value, BiFunction<? super V, ? super V, ? extends V> remappingFunction)1264 public V merge(K key, V value, 1265 BiFunction<? super V, ? super V, ? extends V> remappingFunction) { 1266 if (value == null || remappingFunction == null) 1267 throw new NullPointerException(); 1268 int hash = hash(key); 1269 Node<K,V>[] tab; Node<K,V> first; int n, i; 1270 int binCount = 0; 1271 TreeNode<K,V> t = null; 1272 Node<K,V> old = null; 1273 if (size > threshold || (tab = table) == null || 1274 (n = tab.length) == 0) 1275 n = (tab = resize()).length; 1276 if ((first = tab[i = (n - 1) & hash]) != null) { 1277 if (first instanceof TreeNode) 1278 old = (t = (TreeNode<K,V>)first).getTreeNode(hash, key); 1279 else { 1280 Node<K,V> e = first; K k; 1281 do { 1282 if (e.hash == hash && 1283 ((k = e.key) == key || (key != null && key.equals(k)))) { 1284 old = e; 1285 break; 1286 } 1287 ++binCount; 1288 } while ((e = e.next) != null); 1289 } 1290 } 1291 if (old != null) { 1292 V v; 1293 if (old.value != null) { 1294 int mc = modCount; 1295 v = remappingFunction.apply(old.value, value); 1296 if (mc != modCount) { 1297 throw new ConcurrentModificationException(); 1298 } 1299 } else { 1300 v = value; 1301 } 1302 if (v != null) { 1303 old.value = v; 1304 afterNodeAccess(old); 1305 } 1306 else 1307 removeNode(hash, key, null, false, true); 1308 return v; 1309 } else { 1310 if (t != null) 1311 t.putTreeVal(this, tab, hash, key, value); 1312 else { 1313 tab[i] = newNode(hash, key, value, first); 1314 if (binCount >= TREEIFY_THRESHOLD - 1) 1315 treeifyBin(tab, hash); 1316 } 1317 ++modCount; 1318 ++size; 1319 afterNodeInsertion(true); 1320 return value; 1321 } 1322 } 1323 1324 @Override forEach(BiConsumer<? super K, ? super V> action)1325 public void forEach(BiConsumer<? super K, ? super V> action) { 1326 Node<K,V>[] tab; 1327 if (action == null) 1328 throw new NullPointerException(); 1329 if (size > 0 && (tab = table) != null) { 1330 int mc = modCount; 1331 for (Node<K,V> e : tab) { 1332 for (; e != null; e = e.next) 1333 action.accept(e.key, e.value); 1334 } 1335 if (modCount != mc) 1336 throw new ConcurrentModificationException(); 1337 } 1338 } 1339 1340 @Override replaceAll(BiFunction<? super K, ? super V, ? extends V> function)1341 public void replaceAll(BiFunction<? super K, ? super V, ? extends V> function) { 1342 Node<K,V>[] tab; 1343 if (function == null) 1344 throw new NullPointerException(); 1345 if (size > 0 && (tab = table) != null) { 1346 int mc = modCount; 1347 for (Node<K,V> e : tab) { 1348 for (; e != null; e = e.next) { 1349 e.value = function.apply(e.key, e.value); 1350 } 1351 } 1352 if (modCount != mc) 1353 throw new ConcurrentModificationException(); 1354 } 1355 } 1356 1357 /* ------------------------------------------------------------ */ 1358 // Cloning and serialization 1359 1360 /** 1361 * Returns a shallow copy of this {@code HashMap} instance: the keys and 1362 * values themselves are not cloned. 1363 * 1364 * @return a shallow copy of this map 1365 */ 1366 @SuppressWarnings("unchecked") 1367 @Override clone()1368 public Object clone() { 1369 HashMap<K,V> result; 1370 try { 1371 result = (HashMap<K,V>)super.clone(); 1372 } catch (CloneNotSupportedException e) { 1373 // this shouldn't happen, since we are Cloneable 1374 throw new InternalError(e); 1375 } 1376 result.reinitialize(); 1377 result.putMapEntries(this, false); 1378 return result; 1379 } 1380 1381 // These methods are also used when serializing HashSets loadFactor()1382 final float loadFactor() { return loadFactor; } capacity()1383 final int capacity() { 1384 return (table != null) ? table.length : 1385 (threshold > 0) ? threshold : 1386 DEFAULT_INITIAL_CAPACITY; 1387 } 1388 1389 /** 1390 * Saves this map to a stream (that is, serializes it). 1391 * 1392 * @param s the stream 1393 * @throws IOException if an I/O error occurs 1394 * @serialData The <i>capacity</i> of the HashMap (the length of the 1395 * bucket array) is emitted (int), followed by the 1396 * <i>size</i> (an int, the number of key-value 1397 * mappings), followed by the key (Object) and value (Object) 1398 * for each key-value mapping. The key-value mappings are 1399 * emitted in no particular order. 1400 */ writeObject(java.io.ObjectOutputStream s)1401 private void writeObject(java.io.ObjectOutputStream s) 1402 throws IOException { 1403 int buckets = capacity(); 1404 // Write out the threshold, loadfactor, and any hidden stuff 1405 s.defaultWriteObject(); 1406 s.writeInt(buckets); 1407 s.writeInt(size); 1408 internalWriteEntries(s); 1409 } 1410 1411 /** 1412 * Reconstitutes this map from a stream (that is, deserializes it). 1413 * @param s the stream 1414 * @throws ClassNotFoundException if the class of a serialized object 1415 * could not be found 1416 * @throws IOException if an I/O error occurs 1417 */ readObject(java.io.ObjectInputStream s)1418 private void readObject(java.io.ObjectInputStream s) 1419 throws IOException, ClassNotFoundException { 1420 // Read in the threshold (ignored), loadfactor, and any hidden stuff 1421 s.defaultReadObject(); 1422 reinitialize(); 1423 if (loadFactor <= 0 || Float.isNaN(loadFactor)) 1424 throw new InvalidObjectException("Illegal load factor: " + 1425 loadFactor); 1426 s.readInt(); // Read and ignore number of buckets 1427 int mappings = s.readInt(); // Read number of mappings (size) 1428 if (mappings < 0) 1429 throw new InvalidObjectException("Illegal mappings count: " + 1430 mappings); 1431 else if (mappings > 0) { // (if zero, use defaults) 1432 // Size the table using given load factor only if within 1433 // range of 0.25...4.0 1434 float lf = Math.min(Math.max(0.25f, loadFactor), 4.0f); 1435 float fc = (float)mappings / lf + 1.0f; 1436 int cap = ((fc < DEFAULT_INITIAL_CAPACITY) ? 1437 DEFAULT_INITIAL_CAPACITY : 1438 (fc >= MAXIMUM_CAPACITY) ? 1439 MAXIMUM_CAPACITY : 1440 tableSizeFor((int)fc)); 1441 float ft = (float)cap * lf; 1442 threshold = ((cap < MAXIMUM_CAPACITY && ft < MAXIMUM_CAPACITY) ? 1443 (int)ft : Integer.MAX_VALUE); 1444 1445 // Check Map.Entry[].class since it's the nearest public type to 1446 // what we're actually creating. 1447 SharedSecrets.getJavaObjectInputStreamAccess().checkArray(s, Map.Entry[].class, cap); 1448 @SuppressWarnings({"rawtypes","unchecked"}) 1449 Node<K,V>[] tab = (Node<K,V>[])new Node[cap]; 1450 table = tab; 1451 1452 // Read the keys and values, and put the mappings in the HashMap 1453 for (int i = 0; i < mappings; i++) { 1454 @SuppressWarnings("unchecked") 1455 K key = (K) s.readObject(); 1456 @SuppressWarnings("unchecked") 1457 V value = (V) s.readObject(); 1458 putVal(hash(key), key, value, false, false); 1459 } 1460 } 1461 } 1462 1463 /* ------------------------------------------------------------ */ 1464 // iterators 1465 1466 abstract class HashIterator { 1467 Node<K,V> next; // next entry to return 1468 Node<K,V> current; // current entry 1469 int expectedModCount; // for fast-fail 1470 int index; // current slot 1471 HashIterator()1472 HashIterator() { 1473 expectedModCount = modCount; 1474 Node<K,V>[] t = table; 1475 current = next = null; 1476 index = 0; 1477 if (t != null && size > 0) { // advance to first entry 1478 do {} while (index < t.length && (next = t[index++]) == null); 1479 } 1480 } 1481 hasNext()1482 public final boolean hasNext() { 1483 return next != null; 1484 } 1485 nextNode()1486 final Node<K,V> nextNode() { 1487 Node<K,V>[] t; 1488 Node<K,V> e = next; 1489 if (modCount != expectedModCount) 1490 throw new ConcurrentModificationException(); 1491 if (e == null) 1492 throw new NoSuchElementException(); 1493 if ((next = (current = e).next) == null && (t = table) != null) { 1494 do {} while (index < t.length && (next = t[index++]) == null); 1495 } 1496 return e; 1497 } 1498 remove()1499 public final void remove() { 1500 Node<K,V> p = current; 1501 if (p == null) 1502 throw new IllegalStateException(); 1503 if (modCount != expectedModCount) 1504 throw new ConcurrentModificationException(); 1505 current = null; 1506 removeNode(p.hash, p.key, null, false, false); 1507 expectedModCount = modCount; 1508 } 1509 } 1510 1511 final class KeyIterator extends HashIterator 1512 implements Iterator<K> { next()1513 public final K next() { return nextNode().key; } 1514 } 1515 1516 final class ValueIterator extends HashIterator 1517 implements Iterator<V> { next()1518 public final V next() { return nextNode().value; } 1519 } 1520 1521 final class EntryIterator extends HashIterator 1522 implements Iterator<Map.Entry<K,V>> { next()1523 public final Map.Entry<K,V> next() { return nextNode(); } 1524 } 1525 1526 /* ------------------------------------------------------------ */ 1527 // spliterators 1528 1529 static class HashMapSpliterator<K,V> { 1530 final HashMap<K,V> map; 1531 Node<K,V> current; // current node 1532 int index; // current index, modified on advance/split 1533 int fence; // one past last index 1534 int est; // size estimate 1535 int expectedModCount; // for comodification checks 1536 HashMapSpliterator(HashMap<K,V> m, int origin, int fence, int est, int expectedModCount)1537 HashMapSpliterator(HashMap<K,V> m, int origin, 1538 int fence, int est, 1539 int expectedModCount) { 1540 this.map = m; 1541 this.index = origin; 1542 this.fence = fence; 1543 this.est = est; 1544 this.expectedModCount = expectedModCount; 1545 } 1546 getFence()1547 final int getFence() { // initialize fence and size on first use 1548 int hi; 1549 if ((hi = fence) < 0) { 1550 HashMap<K,V> m = map; 1551 est = m.size; 1552 expectedModCount = m.modCount; 1553 Node<K,V>[] tab = m.table; 1554 hi = fence = (tab == null) ? 0 : tab.length; 1555 } 1556 return hi; 1557 } 1558 estimateSize()1559 public final long estimateSize() { 1560 getFence(); // force init 1561 return (long) est; 1562 } 1563 } 1564 1565 static final class KeySpliterator<K,V> 1566 extends HashMapSpliterator<K,V> 1567 implements Spliterator<K> { KeySpliterator(HashMap<K,V> m, int origin, int fence, int est, int expectedModCount)1568 KeySpliterator(HashMap<K,V> m, int origin, int fence, int est, 1569 int expectedModCount) { 1570 super(m, origin, fence, est, expectedModCount); 1571 } 1572 trySplit()1573 public KeySpliterator<K,V> trySplit() { 1574 int hi = getFence(), lo = index, mid = (lo + hi) >>> 1; 1575 return (lo >= mid || current != null) ? null : 1576 new KeySpliterator<>(map, lo, index = mid, est >>>= 1, 1577 expectedModCount); 1578 } 1579 forEachRemaining(Consumer<? super K> action)1580 public void forEachRemaining(Consumer<? super K> action) { 1581 int i, hi, mc; 1582 if (action == null) 1583 throw new NullPointerException(); 1584 HashMap<K,V> m = map; 1585 Node<K,V>[] tab = m.table; 1586 if ((hi = fence) < 0) { 1587 mc = expectedModCount = m.modCount; 1588 hi = fence = (tab == null) ? 0 : tab.length; 1589 } 1590 else 1591 mc = expectedModCount; 1592 if (tab != null && tab.length >= hi && 1593 (i = index) >= 0 && (i < (index = hi) || current != null)) { 1594 Node<K,V> p = current; 1595 current = null; 1596 do { 1597 if (p == null) 1598 p = tab[i++]; 1599 else { 1600 action.accept(p.key); 1601 p = p.next; 1602 } 1603 } while (p != null || i < hi); 1604 if (m.modCount != mc) 1605 throw new ConcurrentModificationException(); 1606 } 1607 } 1608 tryAdvance(Consumer<? super K> action)1609 public boolean tryAdvance(Consumer<? super K> action) { 1610 int hi; 1611 if (action == null) 1612 throw new NullPointerException(); 1613 Node<K,V>[] tab = map.table; 1614 if (tab != null && tab.length >= (hi = getFence()) && index >= 0) { 1615 while (current != null || index < hi) { 1616 if (current == null) 1617 current = tab[index++]; 1618 else { 1619 K k = current.key; 1620 current = current.next; 1621 action.accept(k); 1622 if (map.modCount != expectedModCount) 1623 throw new ConcurrentModificationException(); 1624 return true; 1625 } 1626 } 1627 } 1628 return false; 1629 } 1630 characteristics()1631 public int characteristics() { 1632 return (fence < 0 || est == map.size ? Spliterator.SIZED : 0) | 1633 Spliterator.DISTINCT; 1634 } 1635 } 1636 1637 static final class ValueSpliterator<K,V> 1638 extends HashMapSpliterator<K,V> 1639 implements Spliterator<V> { ValueSpliterator(HashMap<K,V> m, int origin, int fence, int est, int expectedModCount)1640 ValueSpliterator(HashMap<K,V> m, int origin, int fence, int est, 1641 int expectedModCount) { 1642 super(m, origin, fence, est, expectedModCount); 1643 } 1644 trySplit()1645 public ValueSpliterator<K,V> trySplit() { 1646 int hi = getFence(), lo = index, mid = (lo + hi) >>> 1; 1647 return (lo >= mid || current != null) ? null : 1648 new ValueSpliterator<>(map, lo, index = mid, est >>>= 1, 1649 expectedModCount); 1650 } 1651 forEachRemaining(Consumer<? super V> action)1652 public void forEachRemaining(Consumer<? super V> action) { 1653 int i, hi, mc; 1654 if (action == null) 1655 throw new NullPointerException(); 1656 HashMap<K,V> m = map; 1657 Node<K,V>[] tab = m.table; 1658 if ((hi = fence) < 0) { 1659 mc = expectedModCount = m.modCount; 1660 hi = fence = (tab == null) ? 0 : tab.length; 1661 } 1662 else 1663 mc = expectedModCount; 1664 if (tab != null && tab.length >= hi && 1665 (i = index) >= 0 && (i < (index = hi) || current != null)) { 1666 Node<K,V> p = current; 1667 current = null; 1668 do { 1669 if (p == null) 1670 p = tab[i++]; 1671 else { 1672 action.accept(p.value); 1673 p = p.next; 1674 } 1675 } while (p != null || i < hi); 1676 if (m.modCount != mc) 1677 throw new ConcurrentModificationException(); 1678 } 1679 } 1680 tryAdvance(Consumer<? super V> action)1681 public boolean tryAdvance(Consumer<? super V> action) { 1682 int hi; 1683 if (action == null) 1684 throw new NullPointerException(); 1685 Node<K,V>[] tab = map.table; 1686 if (tab != null && tab.length >= (hi = getFence()) && index >= 0) { 1687 while (current != null || index < hi) { 1688 if (current == null) 1689 current = tab[index++]; 1690 else { 1691 V v = current.value; 1692 current = current.next; 1693 action.accept(v); 1694 if (map.modCount != expectedModCount) 1695 throw new ConcurrentModificationException(); 1696 return true; 1697 } 1698 } 1699 } 1700 return false; 1701 } 1702 characteristics()1703 public int characteristics() { 1704 return (fence < 0 || est == map.size ? Spliterator.SIZED : 0); 1705 } 1706 } 1707 1708 static final class EntrySpliterator<K,V> 1709 extends HashMapSpliterator<K,V> 1710 implements Spliterator<Map.Entry<K,V>> { EntrySpliterator(HashMap<K,V> m, int origin, int fence, int est, int expectedModCount)1711 EntrySpliterator(HashMap<K,V> m, int origin, int fence, int est, 1712 int expectedModCount) { 1713 super(m, origin, fence, est, expectedModCount); 1714 } 1715 trySplit()1716 public EntrySpliterator<K,V> trySplit() { 1717 int hi = getFence(), lo = index, mid = (lo + hi) >>> 1; 1718 return (lo >= mid || current != null) ? null : 1719 new EntrySpliterator<>(map, lo, index = mid, est >>>= 1, 1720 expectedModCount); 1721 } 1722 forEachRemaining(Consumer<? super Map.Entry<K,V>> action)1723 public void forEachRemaining(Consumer<? super Map.Entry<K,V>> action) { 1724 int i, hi, mc; 1725 if (action == null) 1726 throw new NullPointerException(); 1727 HashMap<K,V> m = map; 1728 Node<K,V>[] tab = m.table; 1729 if ((hi = fence) < 0) { 1730 mc = expectedModCount = m.modCount; 1731 hi = fence = (tab == null) ? 0 : tab.length; 1732 } 1733 else 1734 mc = expectedModCount; 1735 if (tab != null && tab.length >= hi && 1736 (i = index) >= 0 && (i < (index = hi) || current != null)) { 1737 Node<K,V> p = current; 1738 current = null; 1739 do { 1740 if (p == null) 1741 p = tab[i++]; 1742 else { 1743 action.accept(p); 1744 p = p.next; 1745 } 1746 } while (p != null || i < hi); 1747 if (m.modCount != mc) 1748 throw new ConcurrentModificationException(); 1749 } 1750 } 1751 tryAdvance(Consumer<? super Map.Entry<K,V>> action)1752 public boolean tryAdvance(Consumer<? super Map.Entry<K,V>> action) { 1753 int hi; 1754 if (action == null) 1755 throw new NullPointerException(); 1756 Node<K,V>[] tab = map.table; 1757 if (tab != null && tab.length >= (hi = getFence()) && index >= 0) { 1758 while (current != null || index < hi) { 1759 if (current == null) 1760 current = tab[index++]; 1761 else { 1762 Node<K,V> e = current; 1763 current = current.next; 1764 action.accept(e); 1765 if (map.modCount != expectedModCount) 1766 throw new ConcurrentModificationException(); 1767 return true; 1768 } 1769 } 1770 } 1771 return false; 1772 } 1773 characteristics()1774 public int characteristics() { 1775 return (fence < 0 || est == map.size ? Spliterator.SIZED : 0) | 1776 Spliterator.DISTINCT; 1777 } 1778 } 1779 1780 /* ------------------------------------------------------------ */ 1781 // LinkedHashMap support 1782 1783 1784 /* 1785 * The following package-protected methods are designed to be 1786 * overridden by LinkedHashMap, but not by any other subclass. 1787 * Nearly all other internal methods are also package-protected 1788 * but are declared final, so can be used by LinkedHashMap, view 1789 * classes, and HashSet. 1790 */ 1791 1792 // Create a regular (non-tree) node newNode(int hash, K key, V value, Node<K,V> next)1793 Node<K,V> newNode(int hash, K key, V value, Node<K,V> next) { 1794 return new Node<>(hash, key, value, next); 1795 } 1796 1797 // For conversion from TreeNodes to plain nodes replacementNode(Node<K,V> p, Node<K,V> next)1798 Node<K,V> replacementNode(Node<K,V> p, Node<K,V> next) { 1799 return new Node<>(p.hash, p.key, p.value, next); 1800 } 1801 1802 // Create a tree bin node newTreeNode(int hash, K key, V value, Node<K,V> next)1803 TreeNode<K,V> newTreeNode(int hash, K key, V value, Node<K,V> next) { 1804 return new TreeNode<>(hash, key, value, next); 1805 } 1806 1807 // For treeifyBin replacementTreeNode(Node<K,V> p, Node<K,V> next)1808 TreeNode<K,V> replacementTreeNode(Node<K,V> p, Node<K,V> next) { 1809 return new TreeNode<>(p.hash, p.key, p.value, next); 1810 } 1811 1812 /** 1813 * Reset to initial default state. Called by clone and readObject. 1814 */ reinitialize()1815 void reinitialize() { 1816 table = null; 1817 entrySet = null; 1818 keySet = null; 1819 values = null; 1820 modCount = 0; 1821 threshold = 0; 1822 size = 0; 1823 } 1824 1825 // Callbacks to allow LinkedHashMap post-actions afterNodeAccess(Node<K,V> p)1826 void afterNodeAccess(Node<K,V> p) { } afterNodeInsertion(boolean evict)1827 void afterNodeInsertion(boolean evict) { } afterNodeRemoval(Node<K,V> p)1828 void afterNodeRemoval(Node<K,V> p) { } 1829 1830 // Called only from writeObject, to ensure compatible ordering. internalWriteEntries(java.io.ObjectOutputStream s)1831 void internalWriteEntries(java.io.ObjectOutputStream s) throws IOException { 1832 Node<K,V>[] tab; 1833 if (size > 0 && (tab = table) != null) { 1834 for (Node<K,V> e : tab) { 1835 for (; e != null; e = e.next) { 1836 s.writeObject(e.key); 1837 s.writeObject(e.value); 1838 } 1839 } 1840 } 1841 } 1842 1843 /* ------------------------------------------------------------ */ 1844 // Tree bins 1845 1846 /** 1847 * Entry for Tree bins. Extends LinkedHashMap.Entry (which in turn 1848 * extends Node) so can be used as extension of either regular or 1849 * linked node. 1850 */ 1851 static final class TreeNode<K,V> extends LinkedHashMap.Entry<K,V> { 1852 TreeNode<K,V> parent; // red-black tree links 1853 TreeNode<K,V> left; 1854 TreeNode<K,V> right; 1855 TreeNode<K,V> prev; // needed to unlink next upon deletion 1856 boolean red; TreeNode(int hash, K key, V val, Node<K,V> next)1857 TreeNode(int hash, K key, V val, Node<K,V> next) { 1858 super(hash, key, val, next); 1859 } 1860 1861 /** 1862 * Returns root of tree containing this node. 1863 */ root()1864 final TreeNode<K,V> root() { 1865 for (TreeNode<K,V> r = this, p;;) { 1866 if ((p = r.parent) == null) 1867 return r; 1868 r = p; 1869 } 1870 } 1871 1872 /** 1873 * Ensures that the given root is the first node of its bin. 1874 */ moveRootToFront(Node<K,V>[] tab, TreeNode<K,V> root)1875 static <K,V> void moveRootToFront(Node<K,V>[] tab, TreeNode<K,V> root) { 1876 int n; 1877 if (root != null && tab != null && (n = tab.length) > 0) { 1878 int index = (n - 1) & root.hash; 1879 TreeNode<K,V> first = (TreeNode<K,V>)tab[index]; 1880 if (root != first) { 1881 Node<K,V> rn; 1882 tab[index] = root; 1883 TreeNode<K,V> rp = root.prev; 1884 if ((rn = root.next) != null) 1885 ((TreeNode<K,V>)rn).prev = rp; 1886 if (rp != null) 1887 rp.next = rn; 1888 if (first != null) 1889 first.prev = root; 1890 root.next = first; 1891 root.prev = null; 1892 } 1893 assert checkInvariants(root); 1894 } 1895 } 1896 1897 /** 1898 * Finds the node starting at root p with the given hash and key. 1899 * The kc argument caches comparableClassFor(key) upon first use 1900 * comparing keys. 1901 */ find(int h, Object k, Class<?> kc)1902 final TreeNode<K,V> find(int h, Object k, Class<?> kc) { 1903 TreeNode<K,V> p = this; 1904 do { 1905 int ph, dir; K pk; 1906 TreeNode<K,V> pl = p.left, pr = p.right, q; 1907 if ((ph = p.hash) > h) 1908 p = pl; 1909 else if (ph < h) 1910 p = pr; 1911 else if ((pk = p.key) == k || (k != null && k.equals(pk))) 1912 return p; 1913 else if (pl == null) 1914 p = pr; 1915 else if (pr == null) 1916 p = pl; 1917 else if ((kc != null || 1918 (kc = comparableClassFor(k)) != null) && 1919 (dir = compareComparables(kc, k, pk)) != 0) 1920 p = (dir < 0) ? pl : pr; 1921 else if ((q = pr.find(h, k, kc)) != null) 1922 return q; 1923 else 1924 p = pl; 1925 } while (p != null); 1926 return null; 1927 } 1928 1929 /** 1930 * Calls find for root node. 1931 */ getTreeNode(int h, Object k)1932 final TreeNode<K,V> getTreeNode(int h, Object k) { 1933 return ((parent != null) ? root() : this).find(h, k, null); 1934 } 1935 1936 /** 1937 * Tie-breaking utility for ordering insertions when equal 1938 * hashCodes and non-comparable. We don't require a total 1939 * order, just a consistent insertion rule to maintain 1940 * equivalence across rebalancings. Tie-breaking further than 1941 * necessary simplifies testing a bit. 1942 */ tieBreakOrder(Object a, Object b)1943 static int tieBreakOrder(Object a, Object b) { 1944 int d; 1945 if (a == null || b == null || 1946 (d = a.getClass().getName(). 1947 compareTo(b.getClass().getName())) == 0) 1948 d = (System.identityHashCode(a) <= System.identityHashCode(b) ? 1949 -1 : 1); 1950 return d; 1951 } 1952 1953 /** 1954 * Forms tree of the nodes linked from this node. 1955 */ treeify(Node<K,V>[] tab)1956 final void treeify(Node<K,V>[] tab) { 1957 TreeNode<K,V> root = null; 1958 for (TreeNode<K,V> x = this, next; x != null; x = next) { 1959 next = (TreeNode<K,V>)x.next; 1960 x.left = x.right = null; 1961 if (root == null) { 1962 x.parent = null; 1963 x.red = false; 1964 root = x; 1965 } 1966 else { 1967 K k = x.key; 1968 int h = x.hash; 1969 Class<?> kc = null; 1970 for (TreeNode<K,V> p = root;;) { 1971 int dir, ph; 1972 K pk = p.key; 1973 if ((ph = p.hash) > h) 1974 dir = -1; 1975 else if (ph < h) 1976 dir = 1; 1977 else if ((kc == null && 1978 (kc = comparableClassFor(k)) == null) || 1979 (dir = compareComparables(kc, k, pk)) == 0) 1980 dir = tieBreakOrder(k, pk); 1981 1982 TreeNode<K,V> xp = p; 1983 if ((p = (dir <= 0) ? p.left : p.right) == null) { 1984 x.parent = xp; 1985 if (dir <= 0) 1986 xp.left = x; 1987 else 1988 xp.right = x; 1989 root = balanceInsertion(root, x); 1990 break; 1991 } 1992 } 1993 } 1994 } 1995 moveRootToFront(tab, root); 1996 } 1997 1998 /** 1999 * Returns a list of non-TreeNodes replacing those linked from 2000 * this node. 2001 */ untreeify(HashMap<K,V> map)2002 final Node<K,V> untreeify(HashMap<K,V> map) { 2003 Node<K,V> hd = null, tl = null; 2004 for (Node<K,V> q = this; q != null; q = q.next) { 2005 Node<K,V> p = map.replacementNode(q, null); 2006 if (tl == null) 2007 hd = p; 2008 else 2009 tl.next = p; 2010 tl = p; 2011 } 2012 return hd; 2013 } 2014 2015 /** 2016 * Tree version of putVal. 2017 */ putTreeVal(HashMap<K,V> map, Node<K,V>[] tab, int h, K k, V v)2018 final TreeNode<K,V> putTreeVal(HashMap<K,V> map, Node<K,V>[] tab, 2019 int h, K k, V v) { 2020 Class<?> kc = null; 2021 boolean searched = false; 2022 TreeNode<K,V> root = (parent != null) ? root() : this; 2023 for (TreeNode<K,V> p = root;;) { 2024 int dir, ph; K pk; 2025 if ((ph = p.hash) > h) 2026 dir = -1; 2027 else if (ph < h) 2028 dir = 1; 2029 else if ((pk = p.key) == k || (k != null && k.equals(pk))) 2030 return p; 2031 else if ((kc == null && 2032 (kc = comparableClassFor(k)) == null) || 2033 (dir = compareComparables(kc, k, pk)) == 0) { 2034 if (!searched) { 2035 TreeNode<K,V> q, ch; 2036 searched = true; 2037 if (((ch = p.left) != null && 2038 (q = ch.find(h, k, kc)) != null) || 2039 ((ch = p.right) != null && 2040 (q = ch.find(h, k, kc)) != null)) 2041 return q; 2042 } 2043 dir = tieBreakOrder(k, pk); 2044 } 2045 2046 TreeNode<K,V> xp = p; 2047 if ((p = (dir <= 0) ? p.left : p.right) == null) { 2048 Node<K,V> xpn = xp.next; 2049 TreeNode<K,V> x = map.newTreeNode(h, k, v, xpn); 2050 if (dir <= 0) 2051 xp.left = x; 2052 else 2053 xp.right = x; 2054 xp.next = x; 2055 x.parent = x.prev = xp; 2056 if (xpn != null) 2057 ((TreeNode<K,V>)xpn).prev = x; 2058 moveRootToFront(tab, balanceInsertion(root, x)); 2059 return null; 2060 } 2061 } 2062 } 2063 2064 /** 2065 * Removes the given node, that must be present before this call. 2066 * This is messier than typical red-black deletion code because we 2067 * cannot swap the contents of an interior node with a leaf 2068 * successor that is pinned by "next" pointers that are accessible 2069 * independently during traversal. So instead we swap the tree 2070 * linkages. If the current tree appears to have too few nodes, 2071 * the bin is converted back to a plain bin. (The test triggers 2072 * somewhere between 2 and 6 nodes, depending on tree structure). 2073 */ removeTreeNode(HashMap<K,V> map, Node<K,V>[] tab, boolean movable)2074 final void removeTreeNode(HashMap<K,V> map, Node<K,V>[] tab, 2075 boolean movable) { 2076 int n; 2077 if (tab == null || (n = tab.length) == 0) 2078 return; 2079 int index = (n - 1) & hash; 2080 TreeNode<K,V> first = (TreeNode<K,V>)tab[index], root = first, rl; 2081 TreeNode<K,V> succ = (TreeNode<K,V>)next, pred = prev; 2082 if (pred == null) 2083 tab[index] = first = succ; 2084 else 2085 pred.next = succ; 2086 if (succ != null) 2087 succ.prev = pred; 2088 if (first == null) 2089 return; 2090 if (root.parent != null) 2091 root = root.root(); 2092 if (root == null 2093 || (movable 2094 && (root.right == null 2095 || (rl = root.left) == null 2096 || rl.left == null))) { 2097 tab[index] = first.untreeify(map); // too small 2098 return; 2099 } 2100 TreeNode<K,V> p = this, pl = left, pr = right, replacement; 2101 if (pl != null && pr != null) { 2102 TreeNode<K,V> s = pr, sl; 2103 while ((sl = s.left) != null) // find successor 2104 s = sl; 2105 boolean c = s.red; s.red = p.red; p.red = c; // swap colors 2106 TreeNode<K,V> sr = s.right; 2107 TreeNode<K,V> pp = p.parent; 2108 if (s == pr) { // p was s's direct parent 2109 p.parent = s; 2110 s.right = p; 2111 } 2112 else { 2113 TreeNode<K,V> sp = s.parent; 2114 if ((p.parent = sp) != null) { 2115 if (s == sp.left) 2116 sp.left = p; 2117 else 2118 sp.right = p; 2119 } 2120 if ((s.right = pr) != null) 2121 pr.parent = s; 2122 } 2123 p.left = null; 2124 if ((p.right = sr) != null) 2125 sr.parent = p; 2126 if ((s.left = pl) != null) 2127 pl.parent = s; 2128 if ((s.parent = pp) == null) 2129 root = s; 2130 else if (p == pp.left) 2131 pp.left = s; 2132 else 2133 pp.right = s; 2134 if (sr != null) 2135 replacement = sr; 2136 else 2137 replacement = p; 2138 } 2139 else if (pl != null) 2140 replacement = pl; 2141 else if (pr != null) 2142 replacement = pr; 2143 else 2144 replacement = p; 2145 if (replacement != p) { 2146 TreeNode<K,V> pp = replacement.parent = p.parent; 2147 if (pp == null) 2148 (root = replacement).red = false; 2149 else if (p == pp.left) 2150 pp.left = replacement; 2151 else 2152 pp.right = replacement; 2153 p.left = p.right = p.parent = null; 2154 } 2155 2156 TreeNode<K,V> r = p.red ? root : balanceDeletion(root, replacement); 2157 2158 if (replacement == p) { // detach 2159 TreeNode<K,V> pp = p.parent; 2160 p.parent = null; 2161 if (pp != null) { 2162 if (p == pp.left) 2163 pp.left = null; 2164 else if (p == pp.right) 2165 pp.right = null; 2166 } 2167 } 2168 if (movable) 2169 moveRootToFront(tab, r); 2170 } 2171 2172 /** 2173 * Splits nodes in a tree bin into lower and upper tree bins, 2174 * or untreeifies if now too small. Called only from resize; 2175 * see above discussion about split bits and indices. 2176 * 2177 * @param map the map 2178 * @param tab the table for recording bin heads 2179 * @param index the index of the table being split 2180 * @param bit the bit of hash to split on 2181 */ split(HashMap<K,V> map, Node<K,V>[] tab, int index, int bit)2182 final void split(HashMap<K,V> map, Node<K,V>[] tab, int index, int bit) { 2183 TreeNode<K,V> b = this; 2184 // Relink into lo and hi lists, preserving order 2185 TreeNode<K,V> loHead = null, loTail = null; 2186 TreeNode<K,V> hiHead = null, hiTail = null; 2187 int lc = 0, hc = 0; 2188 for (TreeNode<K,V> e = b, next; e != null; e = next) { 2189 next = (TreeNode<K,V>)e.next; 2190 e.next = null; 2191 if ((e.hash & bit) == 0) { 2192 if ((e.prev = loTail) == null) 2193 loHead = e; 2194 else 2195 loTail.next = e; 2196 loTail = e; 2197 ++lc; 2198 } 2199 else { 2200 if ((e.prev = hiTail) == null) 2201 hiHead = e; 2202 else 2203 hiTail.next = e; 2204 hiTail = e; 2205 ++hc; 2206 } 2207 } 2208 2209 if (loHead != null) { 2210 if (lc <= UNTREEIFY_THRESHOLD) 2211 tab[index] = loHead.untreeify(map); 2212 else { 2213 tab[index] = loHead; 2214 if (hiHead != null) // (else is already treeified) 2215 loHead.treeify(tab); 2216 } 2217 } 2218 if (hiHead != null) { 2219 if (hc <= UNTREEIFY_THRESHOLD) 2220 tab[index + bit] = hiHead.untreeify(map); 2221 else { 2222 tab[index + bit] = hiHead; 2223 if (loHead != null) 2224 hiHead.treeify(tab); 2225 } 2226 } 2227 } 2228 2229 /* ------------------------------------------------------------ */ 2230 // Red-black tree methods, all adapted from CLR 2231 rotateLeft(TreeNode<K,V> root, TreeNode<K,V> p)2232 static <K,V> TreeNode<K,V> rotateLeft(TreeNode<K,V> root, 2233 TreeNode<K,V> p) { 2234 TreeNode<K,V> r, pp, rl; 2235 if (p != null && (r = p.right) != null) { 2236 if ((rl = p.right = r.left) != null) 2237 rl.parent = p; 2238 if ((pp = r.parent = p.parent) == null) 2239 (root = r).red = false; 2240 else if (pp.left == p) 2241 pp.left = r; 2242 else 2243 pp.right = r; 2244 r.left = p; 2245 p.parent = r; 2246 } 2247 return root; 2248 } 2249 rotateRight(TreeNode<K,V> root, TreeNode<K,V> p)2250 static <K,V> TreeNode<K,V> rotateRight(TreeNode<K,V> root, 2251 TreeNode<K,V> p) { 2252 TreeNode<K,V> l, pp, lr; 2253 if (p != null && (l = p.left) != null) { 2254 if ((lr = p.left = l.right) != null) 2255 lr.parent = p; 2256 if ((pp = l.parent = p.parent) == null) 2257 (root = l).red = false; 2258 else if (pp.right == p) 2259 pp.right = l; 2260 else 2261 pp.left = l; 2262 l.right = p; 2263 p.parent = l; 2264 } 2265 return root; 2266 } 2267 balanceInsertion(TreeNode<K,V> root, TreeNode<K,V> x)2268 static <K,V> TreeNode<K,V> balanceInsertion(TreeNode<K,V> root, 2269 TreeNode<K,V> x) { 2270 x.red = true; 2271 for (TreeNode<K,V> xp, xpp, xppl, xppr;;) { 2272 if ((xp = x.parent) == null) { 2273 x.red = false; 2274 return x; 2275 } 2276 else if (!xp.red || (xpp = xp.parent) == null) 2277 return root; 2278 if (xp == (xppl = xpp.left)) { 2279 if ((xppr = xpp.right) != null && xppr.red) { 2280 xppr.red = false; 2281 xp.red = false; 2282 xpp.red = true; 2283 x = xpp; 2284 } 2285 else { 2286 if (x == xp.right) { 2287 root = rotateLeft(root, x = xp); 2288 xpp = (xp = x.parent) == null ? null : xp.parent; 2289 } 2290 if (xp != null) { 2291 xp.red = false; 2292 if (xpp != null) { 2293 xpp.red = true; 2294 root = rotateRight(root, xpp); 2295 } 2296 } 2297 } 2298 } 2299 else { 2300 if (xppl != null && xppl.red) { 2301 xppl.red = false; 2302 xp.red = false; 2303 xpp.red = true; 2304 x = xpp; 2305 } 2306 else { 2307 if (x == xp.left) { 2308 root = rotateRight(root, x = xp); 2309 xpp = (xp = x.parent) == null ? null : xp.parent; 2310 } 2311 if (xp != null) { 2312 xp.red = false; 2313 if (xpp != null) { 2314 xpp.red = true; 2315 root = rotateLeft(root, xpp); 2316 } 2317 } 2318 } 2319 } 2320 } 2321 } 2322 balanceDeletion(TreeNode<K,V> root, TreeNode<K,V> x)2323 static <K,V> TreeNode<K,V> balanceDeletion(TreeNode<K,V> root, 2324 TreeNode<K,V> x) { 2325 for (TreeNode<K,V> xp, xpl, xpr;;) { 2326 if (x == null || x == root) 2327 return root; 2328 else if ((xp = x.parent) == null) { 2329 x.red = false; 2330 return x; 2331 } 2332 else if (x.red) { 2333 x.red = false; 2334 return root; 2335 } 2336 else if ((xpl = xp.left) == x) { 2337 if ((xpr = xp.right) != null && xpr.red) { 2338 xpr.red = false; 2339 xp.red = true; 2340 root = rotateLeft(root, xp); 2341 xpr = (xp = x.parent) == null ? null : xp.right; 2342 } 2343 if (xpr == null) 2344 x = xp; 2345 else { 2346 TreeNode<K,V> sl = xpr.left, sr = xpr.right; 2347 if ((sr == null || !sr.red) && 2348 (sl == null || !sl.red)) { 2349 xpr.red = true; 2350 x = xp; 2351 } 2352 else { 2353 if (sr == null || !sr.red) { 2354 if (sl != null) 2355 sl.red = false; 2356 xpr.red = true; 2357 root = rotateRight(root, xpr); 2358 xpr = (xp = x.parent) == null ? 2359 null : xp.right; 2360 } 2361 if (xpr != null) { 2362 xpr.red = (xp == null) ? false : xp.red; 2363 if ((sr = xpr.right) != null) 2364 sr.red = false; 2365 } 2366 if (xp != null) { 2367 xp.red = false; 2368 root = rotateLeft(root, xp); 2369 } 2370 x = root; 2371 } 2372 } 2373 } 2374 else { // symmetric 2375 if (xpl != null && xpl.red) { 2376 xpl.red = false; 2377 xp.red = true; 2378 root = rotateRight(root, xp); 2379 xpl = (xp = x.parent) == null ? null : xp.left; 2380 } 2381 if (xpl == null) 2382 x = xp; 2383 else { 2384 TreeNode<K,V> sl = xpl.left, sr = xpl.right; 2385 if ((sl == null || !sl.red) && 2386 (sr == null || !sr.red)) { 2387 xpl.red = true; 2388 x = xp; 2389 } 2390 else { 2391 if (sl == null || !sl.red) { 2392 if (sr != null) 2393 sr.red = false; 2394 xpl.red = true; 2395 root = rotateLeft(root, xpl); 2396 xpl = (xp = x.parent) == null ? 2397 null : xp.left; 2398 } 2399 if (xpl != null) { 2400 xpl.red = (xp == null) ? false : xp.red; 2401 if ((sl = xpl.left) != null) 2402 sl.red = false; 2403 } 2404 if (xp != null) { 2405 xp.red = false; 2406 root = rotateRight(root, xp); 2407 } 2408 x = root; 2409 } 2410 } 2411 } 2412 } 2413 } 2414 2415 /** 2416 * Recursive invariant check 2417 */ checkInvariants(TreeNode<K,V> t)2418 static <K,V> boolean checkInvariants(TreeNode<K,V> t) { 2419 TreeNode<K,V> tp = t.parent, tl = t.left, tr = t.right, 2420 tb = t.prev, tn = (TreeNode<K,V>)t.next; 2421 if (tb != null && tb.next != t) 2422 return false; 2423 if (tn != null && tn.prev != t) 2424 return false; 2425 if (tp != null && t != tp.left && t != tp.right) 2426 return false; 2427 if (tl != null && (tl.parent != t || tl.hash > t.hash)) 2428 return false; 2429 if (tr != null && (tr.parent != t || tr.hash < t.hash)) 2430 return false; 2431 if (t.red && tl != null && tl.red && tr != null && tr.red) 2432 return false; 2433 if (tl != null && !checkInvariants(tl)) 2434 return false; 2435 if (tr != null && !checkInvariants(tr)) 2436 return false; 2437 return true; 2438 } 2439 } 2440 2441 } 2442