Repository navigation
Expand file tree
/
Copy pathdictobject.c
More file actions
3807 lines (3452 loc) · 115 KB
/
Copy pathdictobject.c
File metadata and controls
3807 lines (3452 loc) · 115 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
/* Dictionary object implementation using a hash table */
/* The distribution includes a separate file, Objects/dictnotes.txt,
describing explorations into dictionary design and optimization.
It covers typical dictionary use patterns, the parameters for
tuning dictionaries, and several ideas for possible optimizations.
*/
/*
There are four kinds of slots in the table:
1. Unused. me_key == me_value == NULL
Does not hold an active (key, value) pair now and never did. Unused can
transition to Active upon key insertion. This is the only case in which
me_key is NULL, and is each slot's initial state.
2. Active. me_key != NULL and me_key != dummy and me_value != NULL
Holds an active (key, value) pair. Active can transition to Dummy or
Pending upon key deletion (for combined and split tables respectively).
This is the only case in which me_value != NULL.
3. Dummy. me_key == dummy and me_value == NULL
Previously held an active (key, value) pair, but that was deleted and an
active pair has not yet overwritten the slot. Dummy can transition to
Active upon key insertion. Dummy slots cannot be made Unused again
(cannot have me_key set to NULL), else the probe sequence in case of
collision would have no way to know they were once active.
4. Pending. Not yet inserted or deleted from a split-table.
key != NULL, key != dummy and value == NULL
The DictObject can be in one of two forms.
Either:
A combined table:
ma_values == NULL, dk_refcnt == 1.
Values are stored in the me_value field of the PyDictKeysObject.
Slot kind 4 is not allowed i.e.
key != NULL, key != dummy and value == NULL is illegal.
Or:
A split table:
ma_values != NULL, dk_refcnt >= 1
Values are stored in the ma_values array.
Only string (unicode) keys are allowed, no <dummy> keys are present.
Note: .popitem() abuses the me_hash field of an Unused or Dummy slot to
hold a search finger. The me_hash field of Unused or Dummy slots has no
meaning otherwise. As a consequence of this popitem always converts the dict
to the combined-table form.
*/
/* PyDict_MINSIZE_SPLIT is the minimum size of a split dictionary.
* It must be a power of 2, and at least 4.
* Resizing of split dictionaries is very rare, so the saving memory is more
* important than the cost of resizing.
*/
#define PyDict_MINSIZE_SPLIT 4
/* PyDict_MINSIZE_COMBINED is the starting size for any new, non-split dict.
* 8 allows dicts with no more than 5 active entries; experiments suggested
* this suffices for the majority of dicts (consisting mostly of usually-small
* dicts created to pass keyword arguments).
* Making this 8, rather than 4 reduces the number of resizes for most
* dictionaries, without any significant extra memory use.
*/
#define PyDict_MINSIZE_COMBINED 8
#include "Python.h"
#include "stringlib/eq.h"
typedef struct {
/* Cached hash code of me_key. */
Py_hash_t me_hash;
PyObject *me_key;
PyObject *me_value; /* This field is only meaningful for combined tables */
} PyDictKeyEntry;
typedef PyDictKeyEntry *(*dict_lookup_func)
(PyDictObject *mp, PyObject *key, Py_hash_t hash, PyObject ***value_addr);
struct _dictkeysobject {
Py_ssize_t dk_refcnt;
Py_ssize_t dk_size;
dict_lookup_func dk_lookup;
Py_ssize_t dk_usable;
PyDictKeyEntry dk_entries[1];
};
/*
To ensure the lookup algorithm terminates, there must be at least one Unused
slot (NULL key) in the table.
To avoid slowing down lookups on a near-full table, we resize the table when
it's USABLE_FRACTION (currently two-thirds) full.
*/
/* Set a key error with the specified argument, wrapping it in a
* tuple automatically so that tuple keys are not unpacked as the
* exception arguments. */
static void
set_key_error(PyObject *arg)
{
PyObject *tup;
tup = PyTuple_Pack(1, arg);
if (!tup)
return; /* caller will expect error to be set anyway */
PyErr_SetObject(PyExc_KeyError, tup);
Py_DECREF(tup);
}
#define PERTURB_SHIFT 5
/*
Major subtleties ahead: Most hash schemes depend on having a "good" hash
function, in the sense of simulating randomness. Python doesn't: its most
important hash functions (for strings and ints) are very regular in common
cases:
>>> map(hash, (0, 1, 2, 3))
[0, 1, 2, 3]
>>> map(hash, ("namea", "nameb", "namec", "named"))
[-1658398457, -1658398460, -1658398459, -1658398462]
>>>
This isn't necessarily bad! To the contrary, in a table of size 2**i, taking
the low-order i bits as the initial table index is extremely fast, and there
are no collisions at all for dicts indexed by a contiguous range of ints.
The same is approximately true when keys are "consecutive" strings. So this
gives better-than-random behavior in common cases, and that's very desirable.
OTOH, when collisions occur, the tendency to fill contiguous slices of the
hash table makes a good collision resolution strategy crucial. Taking only
the last i bits of the hash code is also vulnerable: for example, consider
the list [i << 16 for i in range(20000)] as a set of keys. Since ints are
their own hash codes, and this fits in a dict of size 2**15, the last 15 bits
of every hash code are all 0: they *all* map to the same table index.
But catering to unusual cases should not slow the usual ones, so we just take
the last i bits anyway. It's up to collision resolution to do the rest. If
we *usually* find the key we're looking for on the first try (and, it turns
out, we usually do -- the table load factor is kept under 2/3, so the odds
are solidly in our favor), then it makes best sense to keep the initial index
computation dirt cheap.
The first half of collision resolution is to visit table indices via this
recurrence:
j = ((5*j) + 1) mod 2**i
For any initial j in range(2**i), repeating that 2**i times generates each
int in range(2**i) exactly once (see any text on random-number generation for
proof). By itself, this doesn't help much: like linear probing (setting
j += 1, or j -= 1, on each loop trip), it scans the table entries in a fixed
order. This would be bad, except that's not the only thing we do, and it's
actually *good* in the common cases where hash keys are consecutive. In an
example that's really too small to make this entirely clear, for a table of
size 2**3 the order of indices is:
0 -> 1 -> 6 -> 7 -> 4 -> 5 -> 2 -> 3 -> 0 [and here it's repeating]
If two things come in at index 5, the first place we look after is index 2,
not 6, so if another comes in at index 6 the collision at 5 didn't hurt it.
Linear probing is deadly in this case because there the fixed probe order
is the *same* as the order consecutive keys are likely to arrive. But it's
extremely unlikely hash codes will follow a 5*j+1 recurrence by accident,
and certain that consecutive hash codes do not.
The other half of the strategy is to get the other bits of the hash code
into play. This is done by initializing a (unsigned) vrbl "perturb" to the
full hash code, and changing the recurrence to:
j = (5*j) + 1 + perturb;
perturb >>= PERTURB_SHIFT;
use j % 2**i as the next table index;
Now the probe sequence depends (eventually) on every bit in the hash code,
and the pseudo-scrambling property of recurring on 5*j+1 is more valuable,
because it quickly magnifies small differences in the bits that didn't affect
the initial index. Note that because perturb is unsigned, if the recurrence
is executed often enough perturb eventually becomes and remains 0. At that
point (very rarely reached) the recurrence is on (just) 5*j+1 again, and
that's certain to find an empty slot eventually (since it generates every int
in range(2**i), and we make sure there's always at least one empty slot).
Selecting a good value for PERTURB_SHIFT is a balancing act. You want it
small so that the high bits of the hash code continue to affect the probe
sequence across iterations; but you want it large so that in really bad cases
the high-order hash bits have an effect on early iterations. 5 was "the
best" in minimizing total collisions across experiments Tim Peters ran (on
both normal and pathological cases), but 4 and 6 weren't significantly worse.
Historical: Reimer Behrends contributed the idea of using a polynomial-based
approach, using repeated multiplication by x in GF(2**n) where an irreducible
polynomial for each table size was chosen such that x was a primitive root.
Christian Tismer later extended that to use division by x instead, as an
efficient way to get the high bits of the hash code into play. This scheme
also gave excellent collision statistics, but was more expensive: two
if-tests were required inside the loop; computing "the next" index took about
the same number of operations but without as much potential parallelism
(e.g., computing 5*j can go on at the same time as computing 1+perturb in the
above, and then shifting perturb can be done while the table index is being
masked); and the PyDictObject struct required a member to hold the table's
polynomial. In Tim's experiments the current scheme ran faster, produced
equally good collision statistics, needed less code & used less memory.
*/
/* Object used as dummy key to fill deleted entries
* This could be any unique object,
* use a custom type in order to minimise coupling.
*/
static PyObject _dummy_struct;
#define dummy (&_dummy_struct)
#ifdef Py_REF_DEBUG
PyObject *
_PyDict_Dummy(void)
{
return dummy;
}
#endif
/* forward declarations */
static PyDictKeyEntry *lookdict(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr);
static PyDictKeyEntry *lookdict_unicode(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr);
static PyDictKeyEntry *
lookdict_unicode_nodummy(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr);
static PyDictKeyEntry *lookdict_split(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr);
static int dictresize(PyDictObject *mp, Py_ssize_t minused);
/* Dictionary reuse scheme to save calls to malloc, free, and memset */
#ifndef PyDict_MAXFREELIST
#define PyDict_MAXFREELIST 80
#endif
static PyDictObject *free_list[PyDict_MAXFREELIST];
static int numfree = 0;
int
PyDict_ClearFreeList(void)
{
PyDictObject *op;
int ret = numfree;
while (numfree) {
op = free_list[--numfree];
assert(PyDict_CheckExact(op));
PyObject_GC_Del(op);
}
return ret;
}
/* Print summary info about the state of the optimized allocator */
void
_PyDict_DebugMallocStats(FILE *out)
{
_PyDebugAllocatorStats(out,
"free PyDictObject", numfree, sizeof(PyDictObject));
}
void
PyDict_Fini(void)
{
PyDict_ClearFreeList();
}
#define DK_DEBUG_INCREF _Py_INC_REFTOTAL _Py_REF_DEBUG_COMMA
#define DK_DEBUG_DECREF _Py_DEC_REFTOTAL _Py_REF_DEBUG_COMMA
#define DK_INCREF(dk) (DK_DEBUG_INCREF ++(dk)->dk_refcnt)
#define DK_DECREF(dk) if (DK_DEBUG_DECREF (--(dk)->dk_refcnt) == 0) free_keys_object(dk)
#define DK_SIZE(dk) ((dk)->dk_size)
#define DK_MASK(dk) (((dk)->dk_size)-1)
#define IS_POWER_OF_2(x) (((x) & (x-1)) == 0)
/* USABLE_FRACTION is the maximum dictionary load.
* Currently set to (2n+1)/3. Increasing this ratio makes dictionaries more
* dense resulting in more collisions. Decreasing it improves sparseness
* at the expense of spreading entries over more cache lines and at the
* cost of total memory consumed.
*
* USABLE_FRACTION must obey the following:
* (0 < USABLE_FRACTION(n) < n) for all n >= 2
*
* USABLE_FRACTION should be very quick to calculate.
* Fractions around 5/8 to 2/3 seem to work well in practice.
*/
/* Use (2n+1)/3 rather than 2n+3 because: it makes no difference for
* combined tables (the two fractions round to the same number n < ),
* but 2*4/3 is 2 whereas (2*4+1)/3 is 3 which potentially saves quite
* a lot of space for small, split tables */
#define USABLE_FRACTION(n) ((((n) << 1)+1)/3)
/* Alternative fraction that is otherwise close enough to (2n+1)/3 to make
* little difference. 8 * 2/3 == 8 * 5/8 == 5. 16 * 2/3 == 16 * 5/8 == 10.
* 32 * 2/3 = 21, 32 * 5/8 = 20.
* Its advantage is that it is faster to compute on machines with slow division.
* #define USABLE_FRACTION(n) (((n) >> 1) + ((n) >> 2) - ((n) >> 3))
*/
/* GROWTH_RATE. Growth rate upon hitting maximum load.
* Currently set to used*2 + capacity/2.
* This means that dicts double in size when growing without deletions,
* but have more head room when the number of deletions is on a par with the
* number of insertions.
* Raising this to used*4 doubles memory consumption depending on the size of
* the dictionary, but results in half the number of resizes, less effort to
* resize.
* GROWTH_RATE was set to used*4 up to version 3.2.
* GROWTH_RATE was set to used*2 in version 3.3.0
*/
#define GROWTH_RATE(d) (((d)->ma_used*2)+((d)->ma_keys->dk_size>>1))
#define ENSURE_ALLOWS_DELETIONS(d) \
if ((d)->ma_keys->dk_lookup == lookdict_unicode_nodummy) { \
(d)->ma_keys->dk_lookup = lookdict_unicode; \
}
/* This immutable, empty PyDictKeysObject is used for PyDict_Clear()
* (which cannot fail and thus can do no allocation).
*/
static PyDictKeysObject empty_keys_struct = {
2, /* dk_refcnt 1 for this struct, 1 for dummy_struct */
1, /* dk_size */
lookdict_split, /* dk_lookup */
0, /* dk_usable (immutable) */
{
{ 0, 0, 0 } /* dk_entries (empty) */
}
};
static PyObject *empty_values[1] = { NULL };
#define Py_EMPTY_KEYS &empty_keys_struct
static PyDictKeysObject *new_keys_object(Py_ssize_t size)
{
PyDictKeysObject *dk;
Py_ssize_t i;
PyDictKeyEntry *ep0;
assert(size >= PyDict_MINSIZE_SPLIT);
assert(IS_POWER_OF_2(size));
dk = PyMem_MALLOC(sizeof(PyDictKeysObject) +
sizeof(PyDictKeyEntry) * (size-1));
if (dk == NULL) {
PyErr_NoMemory();
return NULL;
}
DK_DEBUG_INCREF dk->dk_refcnt = 1;
dk->dk_size = size;
dk->dk_usable = USABLE_FRACTION(size);
ep0 = &dk->dk_entries[0];
/* Hash value of slot 0 is used by popitem, so it must be initialized */
ep0->me_hash = 0;
for (i = 0; i < size; i++) {
ep0[i].me_key = NULL;
ep0[i].me_value = NULL;
}
dk->dk_lookup = lookdict_unicode_nodummy;
return dk;
}
static void
free_keys_object(PyDictKeysObject *keys)
{
PyDictKeyEntry *entries = &keys->dk_entries[0];
Py_ssize_t i, n;
for (i = 0, n = DK_SIZE(keys); i < n; i++) {
Py_XDECREF(entries[i].me_key);
Py_XDECREF(entries[i].me_value);
}
PyMem_FREE(keys);
}
#define new_values(size) PyMem_NEW(PyObject *, size)
#define free_values(values) PyMem_FREE(values)
/* Consumes a reference to the keys object */
static PyObject *
new_dict(PyDictKeysObject *keys, PyObject **values)
{
PyDictObject *mp;
if (numfree) {
mp = free_list[--numfree];
assert (mp != NULL);
assert (Py_TYPE(mp) == &PyDict_Type);
_Py_NewReference((PyObject *)mp);
}
else {
mp = PyObject_GC_New(PyDictObject, &PyDict_Type);
if (mp == NULL) {
DK_DECREF(keys);
free_values(values);
return NULL;
}
}
mp->ma_keys = keys;
mp->ma_values = values;
mp->ma_used = 0;
return (PyObject *)mp;
}
/* Consumes a reference to the keys object */
static PyObject *
new_dict_with_shared_keys(PyDictKeysObject *keys)
{
PyObject **values;
Py_ssize_t i, size;
size = DK_SIZE(keys);
values = new_values(size);
if (values == NULL) {
DK_DECREF(keys);
return PyErr_NoMemory();
}
for (i = 0; i < size; i++) {
values[i] = NULL;
}
return new_dict(keys, values);
}
PyObject *
PyDict_New(void)
{
return new_dict(new_keys_object(PyDict_MINSIZE_COMBINED), NULL);
}
/*
The basic lookup function used by all operations.
This is based on Algorithm D from Knuth Vol. 3, Sec. 6.4.
Open addressing is preferred over chaining since the link overhead for
chaining would be substantial (100% with typical malloc overhead).
The initial probe index is computed as hash mod the table size. Subsequent
probe indices are computed as explained earlier.
All arithmetic on hash should ignore overflow.
The details in this version are due to Tim Peters, building on many past
contributions by Reimer Behrends, Jyrki Alakuijala, Vladimir Marangozov and
Christian Tismer.
lookdict() is general-purpose, and may return NULL if (and only if) a
comparison raises an exception (this was new in Python 2.5).
lookdict_unicode() below is specialized to string keys, comparison of which can
never raise an exception; that function can never return NULL.
lookdict_unicode_nodummy is further specialized for string keys that cannot be
the <dummy> value.
For both, when the key isn't found a PyDictEntry* is returned
where the key would have been found, *value_addr points to the matching value
slot.
*/
static PyDictKeyEntry *
lookdict(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr)
{
register size_t i;
register size_t perturb;
register PyDictKeyEntry *freeslot;
register size_t mask;
PyDictKeyEntry *ep0;
register PyDictKeyEntry *ep;
register int cmp;
PyObject *startkey;
top:
mask = DK_MASK(mp->ma_keys);
ep0 = &mp->ma_keys->dk_entries[0];
i = (size_t)hash & mask;
ep = &ep0[i];
if (ep->me_key == NULL || ep->me_key == key) {
*value_addr = &ep->me_value;
return ep;
}
if (ep->me_key == dummy)
freeslot = ep;
else {
if (ep->me_hash == hash) {
startkey = ep->me_key;
Py_INCREF(startkey);
cmp = PyObject_RichCompareBool(startkey, key, Py_EQ);
Py_DECREF(startkey);
if (cmp < 0)
return NULL;
if (ep0 == mp->ma_keys->dk_entries && ep->me_key == startkey) {
if (cmp > 0) {
*value_addr = &ep->me_value;
return ep;
}
}
else {
/* The dict was mutated, restart */
goto top;
}
}
freeslot = NULL;
}
/* In the loop, me_key == dummy is by far (factor of 100s) the
least likely outcome, so test for that last. */
for (perturb = hash; ; perturb >>= PERTURB_SHIFT) {
i = (i << 2) + i + perturb + 1;
ep = &ep0[i & mask];
if (ep->me_key == NULL) {
if (freeslot == NULL) {
*value_addr = &ep->me_value;
return ep;
} else {
*value_addr = &freeslot->me_value;
return freeslot;
}
}
if (ep->me_key == key) {
*value_addr = &ep->me_value;
return ep;
}
if (ep->me_hash == hash && ep->me_key != dummy) {
startkey = ep->me_key;
Py_INCREF(startkey);
cmp = PyObject_RichCompareBool(startkey, key, Py_EQ);
Py_DECREF(startkey);
if (cmp < 0) {
*value_addr = NULL;
return NULL;
}
if (ep0 == mp->ma_keys->dk_entries && ep->me_key == startkey) {
if (cmp > 0) {
*value_addr = &ep->me_value;
return ep;
}
}
else {
/* The dict was mutated, restart */
goto top;
}
}
else if (ep->me_key == dummy && freeslot == NULL)
freeslot = ep;
}
assert(0); /* NOT REACHED */
return 0;
}
/* Specialized version for string-only keys */
static PyDictKeyEntry *
lookdict_unicode(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr)
{
register size_t i;
register size_t perturb;
register PyDictKeyEntry *freeslot;
register size_t mask = DK_MASK(mp->ma_keys);
PyDictKeyEntry *ep0 = &mp->ma_keys->dk_entries[0];
register PyDictKeyEntry *ep;
/* Make sure this function doesn't have to handle non-unicode keys,
including subclasses of str; e.g., one reason to subclass
unicodes is to override __eq__, and for speed we don't cater to
that here. */
if (!PyUnicode_CheckExact(key)) {
mp->ma_keys->dk_lookup = lookdict;
return lookdict(mp, key, hash, value_addr);
}
i = (size_t)hash & mask;
ep = &ep0[i];
if (ep->me_key == NULL || ep->me_key == key) {
*value_addr = &ep->me_value;
return ep;
}
if (ep->me_key == dummy)
freeslot = ep;
else {
if (ep->me_hash == hash && unicode_eq(ep->me_key, key)) {
*value_addr = &ep->me_value;
return ep;
}
freeslot = NULL;
}
/* In the loop, me_key == dummy is by far (factor of 100s) the
least likely outcome, so test for that last. */
for (perturb = hash; ; perturb >>= PERTURB_SHIFT) {
i = (i << 2) + i + perturb + 1;
ep = &ep0[i & mask];
if (ep->me_key == NULL) {
if (freeslot == NULL) {
*value_addr = &ep->me_value;
return ep;
} else {
*value_addr = &freeslot->me_value;
return freeslot;
}
}
if (ep->me_key == key
|| (ep->me_hash == hash
&& ep->me_key != dummy
&& unicode_eq(ep->me_key, key))) {
*value_addr = &ep->me_value;
return ep;
}
if (ep->me_key == dummy && freeslot == NULL)
freeslot = ep;
}
assert(0); /* NOT REACHED */
return 0;
}
/* Faster version of lookdict_unicode when it is known that no <dummy> keys
* will be present. */
static PyDictKeyEntry *
lookdict_unicode_nodummy(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr)
{
register size_t i;
register size_t perturb;
register size_t mask = DK_MASK(mp->ma_keys);
PyDictKeyEntry *ep0 = &mp->ma_keys->dk_entries[0];
register PyDictKeyEntry *ep;
/* Make sure this function doesn't have to handle non-unicode keys,
including subclasses of str; e.g., one reason to subclass
unicodes is to override __eq__, and for speed we don't cater to
that here. */
if (!PyUnicode_CheckExact(key)) {
mp->ma_keys->dk_lookup = lookdict;
return lookdict(mp, key, hash, value_addr);
}
i = (size_t)hash & mask;
ep = &ep0[i];
assert(ep->me_key == NULL || PyUnicode_CheckExact(ep->me_key));
if (ep->me_key == NULL || ep->me_key == key ||
(ep->me_hash == hash && unicode_eq(ep->me_key, key))) {
*value_addr = &ep->me_value;
return ep;
}
for (perturb = hash; ; perturb >>= PERTURB_SHIFT) {
i = (i << 2) + i + perturb + 1;
ep = &ep0[i & mask];
assert(ep->me_key == NULL || PyUnicode_CheckExact(ep->me_key));
if (ep->me_key == NULL || ep->me_key == key ||
(ep->me_hash == hash && unicode_eq(ep->me_key, key))) {
*value_addr = &ep->me_value;
return ep;
}
}
assert(0); /* NOT REACHED */
return 0;
}
/* Version of lookdict for split tables.
* All split tables and only split tables use this lookup function.
* Split tables only contain unicode keys and no dummy keys,
* so algorithm is the same as lookdict_unicode_nodummy.
*/
static PyDictKeyEntry *
lookdict_split(PyDictObject *mp, PyObject *key,
Py_hash_t hash, PyObject ***value_addr)
{
register size_t i;
register size_t perturb;
register size_t mask = DK_MASK(mp->ma_keys);
PyDictKeyEntry *ep0 = &mp->ma_keys->dk_entries[0];
register PyDictKeyEntry *ep;
if (!PyUnicode_CheckExact(key)) {
ep = lookdict(mp, key, hash, value_addr);
/* lookdict expects a combined-table, so fix value_addr */
i = ep - ep0;
*value_addr = &mp->ma_values[i];
return ep;
}
i = (size_t)hash & mask;
ep = &ep0[i];
assert(ep->me_key == NULL || PyUnicode_CheckExact(ep->me_key));
if (ep->me_key == NULL || ep->me_key == key ||
(ep->me_hash == hash && unicode_eq(ep->me_key, key))) {
*value_addr = &mp->ma_values[i];
return ep;
}
for (perturb = hash; ; perturb >>= PERTURB_SHIFT) {
i = (i << 2) + i + perturb + 1;
ep = &ep0[i & mask];
assert(ep->me_key == NULL || PyUnicode_CheckExact(ep->me_key));
if (ep->me_key == NULL || ep->me_key == key ||
(ep->me_hash == hash && unicode_eq(ep->me_key, key))) {
*value_addr = &mp->ma_values[i & mask];
return ep;
}
}
assert(0); /* NOT REACHED */
return 0;
}
int
_PyDict_HasOnlyStringKeys(PyObject *dict)
{
Py_ssize_t pos = 0;
PyObject *key, *value;
assert(PyDict_Check(dict));
/* Shortcut */
if (((PyDictObject *)dict)->ma_keys->dk_lookup != lookdict)
return 1;
while (PyDict_Next(dict, &pos, &key, &value))
if (!PyUnicode_Check(key))
return 0;
return 1;
}
#define MAINTAIN_TRACKING(mp, key, value) \
do { \
if (!_PyObject_GC_IS_TRACKED(mp)) { \
if (_PyObject_GC_MAY_BE_TRACKED(key) || \
_PyObject_GC_MAY_BE_TRACKED(value)) { \
_PyObject_GC_TRACK(mp); \
} \
} \
} while(0)
void
_PyDict_MaybeUntrack(PyObject *op)
{
PyDictObject *mp;
PyObject *value;
Py_ssize_t i, size;
if (!PyDict_CheckExact(op) || !_PyObject_GC_IS_TRACKED(op))
return;
mp = (PyDictObject *) op;
size = DK_SIZE(mp->ma_keys);
if (_PyDict_HasSplitTable(mp)) {
for (i = 0; i < size; i++) {
if ((value = mp->ma_values[i]) == NULL)
continue;
if (_PyObject_GC_MAY_BE_TRACKED(value)) {
assert(!_PyObject_GC_MAY_BE_TRACKED(
mp->ma_keys->dk_entries[i].me_key));
return;
}
}
}
else {
PyDictKeyEntry *ep0 = &mp->ma_keys->dk_entries[0];
for (i = 0; i < size; i++) {
if ((value = ep0[i].me_value) == NULL)
continue;
if (_PyObject_GC_MAY_BE_TRACKED(value) ||
_PyObject_GC_MAY_BE_TRACKED(ep0[i].me_key))
return;
}
}
_PyObject_GC_UNTRACK(op);
}
/* Internal function to find slot for an item from its hash
* when it is known that the key is not present in the dict.
*/
static PyDictKeyEntry *
find_empty_slot(PyDictObject *mp, PyObject *key, Py_hash_t hash,
PyObject ***value_addr)
{
size_t i;
size_t perturb;
size_t mask = DK_MASK(mp->ma_keys);
PyDictKeyEntry *ep0 = &mp->ma_keys->dk_entries[0];
PyDictKeyEntry *ep;
assert(key != NULL);
if (!PyUnicode_CheckExact(key))
mp->ma_keys->dk_lookup = lookdict;
i = hash & mask;
ep = &ep0[i];
for (perturb = hash; ep->me_key != NULL; perturb >>= PERTURB_SHIFT) {
i = (i << 2) + i + perturb + 1;
ep = &ep0[i & mask];
}
assert(ep->me_value == NULL);
if (mp->ma_values)
*value_addr = &mp->ma_values[i & mask];
else
*value_addr = &ep->me_value;
return ep;
}
static int
insertion_resize(PyDictObject *mp)
{
return dictresize(mp, GROWTH_RATE(mp));
}
/*
Internal routine to insert a new item into the table.
Used both by the internal resize routine and by the public insert routine.
Returns -1 if an error occurred, or 0 on success.
*/
static int
insertdict(PyDictObject *mp, PyObject *key, Py_hash_t hash, PyObject *value)
{
PyObject *old_value;
PyObject **value_addr;
PyDictKeyEntry *ep;
assert(key != dummy);
if (mp->ma_values != NULL && !PyUnicode_CheckExact(key)) {
if (insertion_resize(mp) < 0)
return -1;
}
ep = mp->ma_keys->dk_lookup(mp, key, hash, &value_addr);
if (ep == NULL) {
return -1;
}
Py_INCREF(value);
MAINTAIN_TRACKING(mp, key, value);
old_value = *value_addr;
if (old_value != NULL) {
assert(ep->me_key != NULL && ep->me_key != dummy);
*value_addr = value;
Py_DECREF(old_value); /* which **CAN** re-enter */
}
else {
if (ep->me_key == NULL) {
Py_INCREF(key);
if (mp->ma_keys->dk_usable <= 0) {
/* Need to resize. */
if (insertion_resize(mp) < 0) {
Py_DECREF(key);
Py_DECREF(value);
return -1;
}
ep = find_empty_slot(mp, key, hash, &value_addr);
}
mp->ma_keys->dk_usable--;
assert(mp->ma_keys->dk_usable >= 0);
ep->me_key = key;
ep->me_hash = hash;
}
else {
if (ep->me_key == dummy) {
Py_INCREF(key);
ep->me_key = key;
ep->me_hash = hash;
Py_DECREF(dummy);
} else {
assert(_PyDict_HasSplitTable(mp));
}
}
mp->ma_used++;
*value_addr = value;
}
assert(ep->me_key != NULL && ep->me_key != dummy);
assert(PyUnicode_CheckExact(key) || mp->ma_keys->dk_lookup == lookdict);
return 0;
}
/*
Internal routine used by dictresize() to insert an item which is
known to be absent from the dict. This routine also assumes that
the dict contains no deleted entries. Besides the performance benefit,
using insertdict() in dictresize() is dangerous (SF bug #1456209).
Note that no refcounts are changed by this routine; if needed, the caller
is responsible for incref'ing `key` and `value`.
Neither mp->ma_used nor k->dk_usable are modified by this routine; the caller
must set them correctly
*/
static void
insertdict_clean(PyDictObject *mp, PyObject *key, Py_hash_t hash,
PyObject *value)
{
size_t i;
size_t perturb;
PyDictKeysObject *k = mp->ma_keys;
size_t mask = (size_t)DK_SIZE(k)-1;
PyDictKeyEntry *ep0 = &k->dk_entries[0];
PyDictKeyEntry *ep;
assert(k->dk_lookup != NULL);
assert(value != NULL);
assert(key != NULL);
assert(key != dummy);
assert(PyUnicode_CheckExact(key) || k->dk_lookup == lookdict);
i = hash & mask;
ep = &ep0[i];
for (perturb = hash; ep->me_key != NULL; perturb >>= PERTURB_SHIFT) {
i = (i << 2) + i + perturb + 1;
ep = &ep0[i & mask];
}
assert(ep->me_value == NULL);
ep->me_key = key;
ep->me_hash = hash;
ep->me_value = value;
}
/*
Restructure the table by allocating a new table and reinserting all
items again. When entries have been deleted, the new table may
actually be smaller than the old one.
If a table is split (its keys and hashes are shared, its values are not),
then the values are temporarily copied into the table, it is resized as
a combined table, then the me_value slots in the old table are NULLed out.
After resizing a table is always combined,
but can be resplit by make_keys_shared().
*/
static int
dictresize(PyDictObject *mp, Py_ssize_t minused)
{
Py_ssize_t newsize;
PyDictKeysObject *oldkeys;
PyObject **oldvalues;
Py_ssize_t i, oldsize;
/* Find the smallest table size > minused. */
for (newsize = PyDict_MINSIZE_COMBINED;
newsize <= minused && newsize > 0;
newsize <<= 1)
;
if (newsize <= 0) {
PyErr_NoMemory();
return -1;
}
oldkeys = mp->ma_keys;
oldvalues = mp->ma_values;
/* Allocate a new table. */
mp->ma_keys = new_keys_object(newsize);
if (mp->ma_keys == NULL) {
mp->ma_keys = oldkeys;
return -1;
}
if (oldkeys->dk_lookup == lookdict)
mp->ma_keys->dk_lookup = lookdict;
oldsize = DK_SIZE(oldkeys);
mp->ma_values = NULL;
/* If empty then nothing to copy so just return */
if (oldsize == 1) {
assert(oldkeys == Py_EMPTY_KEYS);
DK_DECREF(oldkeys);
return 0;
}
/* Main loop below assumes we can transfer refcount to new keys
* and that value is stored in me_value.
* Increment ref-counts and copy values here to compensate
* This (resizing a split table) should be relatively rare */
if (oldvalues != NULL) {
for (i = 0; i < oldsize; i++) {
if (oldvalues[i] != NULL) {
Py_INCREF(oldkeys->dk_entries[i].me_key);
oldkeys->dk_entries[i].me_value = oldvalues[i];
}
}
}
/* Main loop */
for (i = 0; i < oldsize; i++) {
PyDictKeyEntry *ep = &oldkeys->dk_entries[i];
if (ep->me_value != NULL) {
assert(ep->me_key != dummy);
insertdict_clean(mp, ep->me_key, ep->me_hash, ep->me_value);
}
}
mp->ma_keys->dk_usable -= mp->ma_used;
if (oldvalues != NULL) {
/* NULL out me_value slot in oldkeys, in case it was shared */
for (i = 0; i < oldsize; i++)
oldkeys->dk_entries[i].me_value = NULL;
assert(oldvalues != empty_values);
free_values(oldvalues);
DK_DECREF(oldkeys);
}
else {
assert(oldkeys->dk_lookup != lookdict_split);
if (oldkeys->dk_lookup != lookdict_unicode_nodummy) {
PyDictKeyEntry *ep0 = &oldkeys->dk_entries[0];
for (i = 0; i < oldsize; i++) {
if (ep0[i].me_key == dummy)
Py_DECREF(dummy);
}
}
assert(oldkeys->dk_refcnt == 1);
DK_DEBUG_DECREF PyMem_FREE(oldkeys);
}
return 0;
}
/* Returns NULL if unable to split table.
* A NULL return does not necessarily indicate an error */
static PyDictKeysObject *
make_keys_shared(PyObject *op)
{
Py_ssize_t i;
Py_ssize_t size;
PyDictObject *mp = (PyDictObject *)op;