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