-
-
Notifications
You must be signed in to change notification settings - Fork 35.5k
Expand file tree
/
Copy pathnfrule.cpp
More file actions
1694 lines (1566 loc) Β· 66.2 KB
/
nfrule.cpp
File metadata and controls
1694 lines (1566 loc) Β· 66.2 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
// Β© 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
******************************************************************************
* Copyright (C) 1997-2015, International Business Machines
* Corporation and others. All Rights Reserved.
******************************************************************************
* file name: nfrule.cpp
* encoding: UTF-8
* tab size: 8 (not used)
* indentation:4
*
* Modification history
* Date Name Comments
* 10/11/2001 Doug Ported from ICU4J
*/
#include "nfrule.h"
#if U_HAVE_RBNF
#include "unicode/localpointer.h"
#include "unicode/rbnf.h"
#include "unicode/tblcoll.h"
#include "unicode/plurfmt.h"
#include "unicode/upluralrules.h"
#include "unicode/coleitr.h"
#include "unicode/uchar.h"
#include "nfrs.h"
#include "nfrlist.h"
#include "nfsubs.h"
#include "patternprops.h"
#include "putilimp.h"
U_NAMESPACE_BEGIN
NFRule::NFRule(const RuleBasedNumberFormat* _rbnf, const UnicodeString &_ruleText, UErrorCode &status)
: baseValue(static_cast<int32_t>(0))
, radix(10)
, exponent(0)
, decimalPoint(0)
, fRuleText(_ruleText)
, sub1(nullptr)
, sub2(nullptr)
, formatter(_rbnf)
, rulePatternFormat(nullptr)
{
if (!fRuleText.isEmpty()) {
parseRuleDescriptor(fRuleText, status);
}
}
NFRule::~NFRule()
{
if (sub1 != sub2) {
delete sub2;
sub2 = nullptr;
}
delete sub1;
sub1 = nullptr;
delete rulePatternFormat;
rulePatternFormat = nullptr;
}
static const char16_t gLeftBracket = 0x005b;
static const char16_t gRightBracket = 0x005d;
static const char16_t gVerticalLine = 0x007C;
static const char16_t gColon = 0x003a;
static const char16_t gZero = 0x0030;
static const char16_t gNine = 0x0039;
static const char16_t gSpace = 0x0020;
static const char16_t gSlash = 0x002f;
static const char16_t gGreaterThan = 0x003e;
static const char16_t gLessThan = 0x003c;
static const char16_t gComma = 0x002c;
static const char16_t gDot = 0x002e;
static const char16_t gTick = 0x0027;
//static const char16_t gMinus = 0x002d;
static const char16_t gSemicolon = 0x003b;
static const char16_t gX = 0x0078;
static const char16_t gMinusX[] = {0x2D, 0x78, 0}; /* "-x" */
static const char16_t gInf[] = {0x49, 0x6E, 0x66, 0}; /* "Inf" */
static const char16_t gNaN[] = {0x4E, 0x61, 0x4E, 0}; /* "NaN" */
static const char16_t gDollarOpenParenthesis[] = {0x24, 0x28, 0}; /* "$(" */
static const char16_t gClosedParenthesisDollar[] = {0x29, 0x24, 0}; /* ")$" */
static const char16_t gLessLess[] = {0x3C, 0x3C, 0}; /* "<<" */
static const char16_t gLessPercent[] = {0x3C, 0x25, 0}; /* "<%" */
static const char16_t gLessHash[] = {0x3C, 0x23, 0}; /* "<#" */
static const char16_t gLessZero[] = {0x3C, 0x30, 0}; /* "<0" */
static const char16_t gGreaterGreater[] = {0x3E, 0x3E, 0}; /* ">>" */
static const char16_t gGreaterPercent[] = {0x3E, 0x25, 0}; /* ">%" */
static const char16_t gGreaterHash[] = {0x3E, 0x23, 0}; /* ">#" */
static const char16_t gGreaterZero[] = {0x3E, 0x30, 0}; /* ">0" */
static const char16_t gEqualPercent[] = {0x3D, 0x25, 0}; /* "=%" */
static const char16_t gEqualHash[] = {0x3D, 0x23, 0}; /* "=#" */
static const char16_t gEqualZero[] = {0x3D, 0x30, 0}; /* "=0" */
static const char16_t gGreaterGreaterGreater[] = {0x3E, 0x3E, 0x3E, 0}; /* ">>>" */
static const char16_t * const RULE_PREFIXES[] = {
gLessLess, gLessPercent, gLessHash, gLessZero,
gGreaterGreater, gGreaterPercent,gGreaterHash, gGreaterZero,
gEqualPercent, gEqualHash, gEqualZero, nullptr
};
void
NFRule::makeRules(UnicodeString& description,
NFRuleSet *owner,
const NFRule *predecessor,
const RuleBasedNumberFormat *rbnf,
NFRuleList& rules,
UErrorCode& status)
{
if (U_FAILURE(status)) {
return;
}
// we know we're making at least one rule, so go ahead and
// new it up and initialize its basevalue and divisor
// (this also strips the rule descriptor, if any, off the
// description string)
LocalPointer<NFRule> rule1(new NFRule(rbnf, description, status));
if (U_FAILURE(status)) {
return;
}
/* test for nullptr */
if (rule1.isNull()) {
status = U_MEMORY_ALLOCATION_ERROR;
return;
}
description = rule1->fRuleText;
// check the description to see whether there's text enclosed
// in brackets
int32_t brack1 = description.indexOf(gLeftBracket);
int32_t brack2 = brack1 < 0 ? -1 : description.indexOf(gRightBracket);
// if the description doesn't contain a matched pair of brackets,
// or if it's of a type that doesn't recognize bracketed text,
// then leave the description alone, initialize the rule's
// rule text and substitutions, and return that rule
if (brack2 < 0 || brack1 > brack2
|| rule1->getType() == kProperFractionRule
|| rule1->getType() == kNegativeNumberRule
|| rule1->getType() == kInfinityRule
|| rule1->getType() == kNaNRule)
{
rule1->extractSubstitutions(owner, description, predecessor, status);
if (U_FAILURE(status)) {
return;
}
}
else {
// if the description does contain a matched pair of brackets,
// then it's really shorthand for two rules (with one exception)
LocalPointer<NFRule> rule2;
UnicodeString sbuf;
int32_t orElseOp = description.indexOf(gVerticalLine);
uint64_t mod = util64_pow(rule1->radix, rule1->exponent);
// we'll actually only split the rule into two rules if its
// base value is an even multiple of its divisor (or it's one
// of the special rules)
if (rule1->baseValue > 0 && rule1->radix != 0 && mod == 0) {
status = U_NUMBER_ARG_OUTOFBOUNDS_ERROR;
return;
}
if ((rule1->baseValue > 0
&& (rule1->radix != 0) // ICU-23109 Ensure next line won't "% 0"
&& (rule1->baseValue % mod == 0))
|| rule1->getType() == kImproperFractionRule
|| rule1->getType() == kDefaultRule) {
// if it passes that test, new up the second rule. If the
// rule set both rules will belong to is a fraction rule
// set, they both have the same base value; otherwise,
// increment the original rule's base value ("rule1" actually
// goes SECOND in the rule set's rule list)
rule2.adoptInstead(new NFRule(rbnf, UnicodeString(), status));
if (U_FAILURE(status)) {
return;
}
/* test for nullptr */
if (rule2.isNull()) {
status = U_MEMORY_ALLOCATION_ERROR;
return;
}
if (rule1->baseValue >= 0) {
rule2->baseValue = rule1->baseValue;
if (!owner->isFractionRuleSet()) {
++rule1->baseValue;
}
}
// if the description began with "x.x" and contains bracketed
// text, it describes both the improper fraction rule and
// the proper fraction rule
else if (rule1->getType() == kImproperFractionRule) {
rule2->setType(kProperFractionRule);
}
// if the description began with "x.0" and contains bracketed
// text, it describes both the default rule and the
// improper fraction rule
else if (rule1->getType() == kDefaultRule) {
rule2->baseValue = rule1->baseValue;
rule1->setType(kImproperFractionRule);
}
// both rules have the same radix and exponent (i.e., the
// same divisor)
rule2->radix = rule1->radix;
rule2->exponent = rule1->exponent;
// By default, rule2's rule text omits the stuff in brackets,
// unless it contains a | between the brackets.
// Initialize its rule text and substitutions accordingly.
sbuf.append(description, 0, brack1);
if (orElseOp >= 0) {
sbuf.append(description, orElseOp + 1, brack2 - orElseOp - 1);
}
if (brack2 + 1 < description.length()) {
sbuf.append(description, brack2 + 1, description.length() - brack2 - 1);
}
rule2->extractSubstitutions(owner, sbuf, predecessor, status);
if (U_FAILURE(status)) {
return;
}
}
// rule1's text includes the text in the brackets but omits
// the brackets themselves: initialize _its_ rule text and
// substitutions accordingly
sbuf.setTo(description, 0, brack1);
if (orElseOp >= 0) {
sbuf.append(description, brack1 + 1, orElseOp - brack1 - 1);
}
else {
sbuf.append(description, brack1 + 1, brack2 - brack1 - 1);
}
if (brack2 + 1 < description.length()) {
sbuf.append(description, brack2 + 1, description.length() - brack2 - 1);
}
rule1->extractSubstitutions(owner, sbuf, predecessor, status);
if (U_FAILURE(status)) {
return;
}
// if we only have one rule, return it; if we have two, return
// a two-element array containing them (notice that rule2 goes
// BEFORE rule1 in the list: in all cases, rule2 OMITS the
// material in the brackets and rule1 INCLUDES the material
// in the brackets)
if (!rule2.isNull()) {
if (rule2->baseValue >= kNoBase) {
rules.add(rule2.orphan());
}
else {
owner->setNonNumericalRule(rule2.orphan());
}
}
}
if (rule1->baseValue >= kNoBase) {
rules.add(rule1.orphan());
}
else {
owner->setNonNumericalRule(rule1.orphan());
}
}
/**
* This function parses the rule's rule descriptor (i.e., the base
* value and/or other tokens that precede the rule's rule text
* in the description) and sets the rule's base value, radix, and
* exponent according to the descriptor. (If the description doesn't
* include a rule descriptor, then this function sets everything to
* default values and the rule set sets the rule's real base value).
* @param description The rule's description
* @return If "description" included a rule descriptor, this is
* "description" with the descriptor and any trailing whitespace
* stripped off. Otherwise; it's "descriptor" unchangd.
*/
void
NFRule::parseRuleDescriptor(UnicodeString& description, UErrorCode& status)
{
// the description consists of a rule descriptor and a rule body,
// separated by a colon. The rule descriptor is optional. If
// it's omitted, just set the base value to 0.
int32_t p = description.indexOf(gColon);
if (p != -1) {
// copy the descriptor out into its own string and strip it,
// along with any trailing whitespace, out of the original
// description
UnicodeString descriptor;
descriptor.setTo(description, 0, p);
++p;
while (p < description.length() && PatternProps::isWhiteSpace(description.charAt(p))) {
++p;
}
description.removeBetween(0, p);
// check first to see if the rule descriptor matches the token
// for one of the special rules. If it does, set the base
// value to the correct identifier value
int descriptorLength = descriptor.length();
char16_t firstChar = descriptor.charAt(0);
char16_t lastChar = descriptor.charAt(descriptorLength - 1);
if (firstChar >= gZero && firstChar <= gNine && lastChar != gX) {
// if the rule descriptor begins with a digit, it's a descriptor
// for a normal rule
// since we don't have Long.parseLong, and this isn't much work anyway,
// just build up the value as we encounter the digits.
int64_t val = 0;
p = 0;
char16_t c = gSpace;
// begin parsing the descriptor: copy digits
// into "tempValue", skip periods, commas, and spaces,
// stop on a slash or > sign (or at the end of the string),
// and throw an exception on any other character
while (p < descriptorLength) {
c = descriptor.charAt(p);
if (c >= gZero && c <= gNine) {
int64_t digit = static_cast<int64_t>(c - gZero);
if ((val > 0 && val > (INT64_MAX - digit) / 10) ||
(val < 0 && val < (INT64_MIN - digit) / 10)) {
// out of int64_t range
status = U_PARSE_ERROR;
return;
}
val = val * 10 + digit;
}
else if (c == gSlash || c == gGreaterThan) {
break;
}
else if (PatternProps::isWhiteSpace(c) || c == gComma || c == gDot) {
}
else {
// throw new IllegalArgumentException("Illegal character in rule descriptor");
status = U_PARSE_ERROR;
return;
}
++p;
}
// we have the base value, so set it
setBaseValue(val, status);
// if we stopped the previous loop on a slash, we're
// now parsing the rule's radix. Again, accumulate digits
// in tempValue, skip punctuation, stop on a > mark, and
// throw an exception on anything else
if (c == gSlash) {
val = 0;
++p;
while (p < descriptorLength) {
c = descriptor.charAt(p);
if (c >= gZero && c <= gNine) {
int64_t digit = static_cast<int64_t>(c - gZero);
if ((val > 0 && val > (INT64_MAX - digit) / 10) ||
(val < 0 && val < (INT64_MIN - digit) / 10)) {
// out of int64_t range
status = U_PARSE_ERROR;
return;
}
val = val * 10 + digit;
}
else if (c == gGreaterThan) {
break;
}
else if (PatternProps::isWhiteSpace(c) || c == gComma || c == gDot) {
}
else {
// throw new IllegalArgumentException("Illegal character is rule descriptor");
status = U_PARSE_ERROR;
return;
}
++p;
}
// tempValue now contain's the rule's radix. Set it
// accordingly, and recalculate the rule's exponent
radix = static_cast<int32_t>(val);
if (radix == 0) {
// throw new IllegalArgumentException("Rule can't have radix of 0");
status = U_PARSE_ERROR;
}
exponent = expectedExponent();
}
// if we stopped the previous loop on a > sign, then continue
// for as long as we still see > signs. For each one,
// decrement the exponent (unless the exponent is already 0).
// If we see another character before reaching the end of
// the descriptor, that's also a syntax error.
if (c == gGreaterThan) {
while (p < descriptor.length()) {
c = descriptor.charAt(p);
if (c == gGreaterThan && exponent > 0) {
--exponent;
} else {
// throw new IllegalArgumentException("Illegal character in rule descriptor");
status = U_PARSE_ERROR;
return;
}
++p;
}
}
}
else if (0 == descriptor.compare(gMinusX, 2)) {
setType(kNegativeNumberRule);
}
else if (descriptorLength == 3) {
if (firstChar == gZero && lastChar == gX) {
setBaseValue(kProperFractionRule, status);
decimalPoint = descriptor.charAt(1);
}
else if (firstChar == gX && lastChar == gX) {
setBaseValue(kImproperFractionRule, status);
decimalPoint = descriptor.charAt(1);
}
else if (firstChar == gX && lastChar == gZero) {
setBaseValue(kDefaultRule, status);
decimalPoint = descriptor.charAt(1);
}
else if (descriptor.compare(gNaN, 3) == 0) {
setBaseValue(kNaNRule, status);
}
else if (descriptor.compare(gInf, 3) == 0) {
setBaseValue(kInfinityRule, status);
}
}
}
// else use the default base value for now.
// finally, if the rule body begins with an apostrophe, strip it off
// (this is generally used to put whitespace at the beginning of
// a rule's rule text)
if (!description.isEmpty() && description.charAt(0) == gTick) {
description.removeBetween(0, 1);
}
// return the description with all the stuff we've just waded through
// stripped off the front. It now contains just the rule body.
// return description;
}
/**
* Searches the rule's rule text for the substitution tokens,
* creates the substitutions, and removes the substitution tokens
* from the rule's rule text.
* @param owner The rule set containing this rule
* @param predecessor The rule preseding this one in "owners" rule list
* @param ownersOwner The RuleBasedFormat that owns this rule
*/
void
NFRule::extractSubstitutions(const NFRuleSet* ruleSet,
const UnicodeString &ruleText,
const NFRule* predecessor,
UErrorCode& status)
{
if (U_FAILURE(status)) {
return;
}
fRuleText = ruleText;
sub1 = extractSubstitution(ruleSet, predecessor, status);
if (sub1 == nullptr) {
// Small optimization. There is no need to create a redundant NullSubstitution.
sub2 = nullptr;
}
else {
sub2 = extractSubstitution(ruleSet, predecessor, status);
}
int32_t pluralRuleStart = fRuleText.indexOf(gDollarOpenParenthesis, -1, 0);
int32_t pluralRuleEnd = (pluralRuleStart >= 0 ? fRuleText.indexOf(gClosedParenthesisDollar, -1, pluralRuleStart) : -1);
if (pluralRuleEnd >= 0) {
int32_t endType = fRuleText.indexOf(gComma, pluralRuleStart);
if (endType < 0) {
status = U_PARSE_ERROR;
return;
}
UnicodeString type(fRuleText.tempSubString(pluralRuleStart + 2, endType - pluralRuleStart - 2));
UPluralType pluralType;
if (type.startsWith(UNICODE_STRING_SIMPLE("cardinal"))) {
pluralType = UPLURAL_TYPE_CARDINAL;
}
else if (type.startsWith(UNICODE_STRING_SIMPLE("ordinal"))) {
pluralType = UPLURAL_TYPE_ORDINAL;
}
else {
status = U_ILLEGAL_ARGUMENT_ERROR;
return;
}
rulePatternFormat = formatter->createPluralFormat(pluralType,
fRuleText.tempSubString(endType + 1, pluralRuleEnd - endType - 1), status);
}
}
/**
* Searches the rule's rule text for the first substitution token,
* creates a substitution based on it, and removes the token from
* the rule's rule text.
* @param owner The rule set containing this rule
* @param predecessor The rule preceding this one in the rule set's
* rule list
* @param ownersOwner The RuleBasedNumberFormat that owns this rule
* @return The newly-created substitution. This is never null; if
* the rule text doesn't contain any substitution tokens, this will
* be a NullSubstitution.
*/
NFSubstitution *
NFRule::extractSubstitution(const NFRuleSet* ruleSet,
const NFRule* predecessor,
UErrorCode& status)
{
NFSubstitution* result = nullptr;
// search the rule's rule text for the first two characters of
// a substitution token
int32_t subStart = indexOfAnyRulePrefix();
int32_t subEnd = subStart;
// if we didn't find one, create a null substitution positioned
// at the end of the rule text
if (subStart == -1) {
return nullptr;
}
// special-case the ">>>" token, since searching for the > at the
// end will actually find the > in the middle
if (fRuleText.indexOf(gGreaterGreaterGreater, 3, 0) == subStart) {
subEnd = subStart + 2;
// otherwise the substitution token ends with the same character
// it began with
} else {
char16_t c = fRuleText.charAt(subStart);
subEnd = fRuleText.indexOf(c, subStart + 1);
// special case for '<%foo<<'
if (c == gLessThan && subEnd != -1 && subEnd < fRuleText.length() - 1 && fRuleText.charAt(subEnd+1) == c) {
// ordinals use "=#,##0==%abbrev=" as their rule. Notice that the '==' in the middle
// occurs because of the juxtaposition of two different rules. The check for '<' is a hack
// to get around this. Having the duplicate at the front would cause problems with
// rules like "<<%" to format, say, percents...
++subEnd;
}
}
// if we don't find the end of the token (i.e., if we're on a single,
// unmatched token character), create a null substitution positioned
// at the end of the rule
if (subEnd == -1) {
return nullptr;
}
// if we get here, we have a real substitution token (or at least
// some text bounded by substitution token characters). Use
// makeSubstitution() to create the right kind of substitution
UnicodeString subToken;
subToken.setTo(fRuleText, subStart, subEnd + 1 - subStart);
result = NFSubstitution::makeSubstitution(subStart, this, predecessor, ruleSet,
this->formatter, subToken, status);
// remove the substitution from the rule text
fRuleText.removeBetween(subStart, subEnd+1);
return result;
}
/**
* Sets the rule's base value, and causes the radix and exponent
* to be recalculated. This is used during construction when we
* don't know the rule's base value until after it's been
* constructed. It should be used at any other time.
* @param The new base value for the rule.
*/
void
NFRule::setBaseValue(int64_t newBaseValue, UErrorCode& status)
{
// set the base value
baseValue = newBaseValue;
radix = 10;
// if this isn't a special rule, recalculate the radix and exponent
// (the radix always defaults to 10; if it's supposed to be something
// else, it's cleaned up by the caller and the exponent is
// recalculated again-- the only function that does this is
// NFRule.parseRuleDescriptor() )
if (baseValue >= 1) {
exponent = expectedExponent();
// this function gets called on a fully-constructed rule whose
// description didn't specify a base value. This means it
// has substitutions, and some substitutions hold on to copies
// of the rule's divisor. Fix their copies of the divisor.
if (sub1 != nullptr) {
sub1->setDivisor(radix, exponent, status);
}
if (sub2 != nullptr) {
sub2->setDivisor(radix, exponent, status);
}
// if this is a special rule, its radix and exponent are basically
// ignored. Set them to "safe" default values
} else {
exponent = 0;
}
}
/**
* This calculates the rule's exponent based on its radix and base
* value. This will be the highest power the radix can be raised to
* and still produce a result less than or equal to the base value.
*/
int16_t
NFRule::expectedExponent() const
{
// since the log of 0, or the log base 0 of something, causes an
// error, declare the exponent in these cases to be 0 (we also
// deal with the special-rule identifiers here)
if (radix == 0 || baseValue < 1) {
return 0;
}
// we get rounding error in some cases-- for example, log 1000 / log 10
// gives us 1.9999999996 instead of 2. The extra logic here is to take
// that into account
int16_t tempResult = static_cast<int16_t>(uprv_log(static_cast<double>(baseValue)) /
uprv_log(static_cast<double>(radix)));
int64_t temp = util64_pow(radix, tempResult + 1);
if (temp <= baseValue) {
tempResult += 1;
}
return tempResult;
}
/**
* Searches the rule's rule text for any of the specified strings.
* @return The index of the first match in the rule's rule text
* (i.e., the first substring in the rule's rule text that matches
* _any_ of the strings in "strings"). If none of the strings in
* "strings" is found in the rule's rule text, returns -1.
*/
int32_t
NFRule::indexOfAnyRulePrefix() const
{
int result = -1;
for (int i = 0; RULE_PREFIXES[i]; i++) {
int32_t pos = fRuleText.indexOf(*RULE_PREFIXES[i]);
if (pos != -1 && (result == -1 || pos < result)) {
result = pos;
}
}
return result;
}
//-----------------------------------------------------------------------
// boilerplate
//-----------------------------------------------------------------------
static UBool
util_equalSubstitutions(const NFSubstitution* sub1, const NFSubstitution* sub2)
{
if (sub1) {
if (sub2) {
return *sub1 == *sub2;
}
} else if (!sub2) {
return true;
}
return false;
}
/**
* Tests two rules for equality.
* @param that The rule to compare this one against
* @return True is the two rules are functionally equivalent
*/
bool
NFRule::operator==(const NFRule& rhs) const
{
return baseValue == rhs.baseValue
&& radix == rhs.radix
&& exponent == rhs.exponent
&& fRuleText == rhs.fRuleText
&& util_equalSubstitutions(sub1, rhs.sub1)
&& util_equalSubstitutions(sub2, rhs.sub2);
}
/**
* Returns a textual representation of the rule. This won't
* necessarily be the same as the description that this rule
* was created with, but it will produce the same result.
* @return A textual description of the rule
*/
static void util_append64(UnicodeString& result, int64_t n)
{
char16_t buffer[256];
int32_t len = util64_tou(n, buffer, sizeof(buffer));
UnicodeString temp(buffer, len);
result.append(temp);
}
void
NFRule::_appendRuleText(UnicodeString& result) const
{
switch (getType()) {
case kNegativeNumberRule: result.append(gMinusX, 2); break;
case kImproperFractionRule: result.append(gX).append(decimalPoint == 0 ? gDot : decimalPoint).append(gX); break;
case kProperFractionRule: result.append(gZero).append(decimalPoint == 0 ? gDot : decimalPoint).append(gX); break;
case kDefaultRule: result.append(gX).append(decimalPoint == 0 ? gDot : decimalPoint).append(gZero); break;
case kInfinityRule: result.append(gInf, 3); break;
case kNaNRule: result.append(gNaN, 3); break;
default:
// for a normal rule, write out its base value, and if the radix is
// something other than 10, write out the radix (with the preceding
// slash, of course). Then calculate the expected exponent and if
// if isn't the same as the actual exponent, write an appropriate
// number of > signs. Finally, terminate the whole thing with
// a colon.
util_append64(result, baseValue);
if (radix != 10) {
result.append(gSlash);
util_append64(result, radix);
}
int numCarets = expectedExponent() - exponent;
for (int i = 0; i < numCarets; i++) {
result.append(gGreaterThan);
}
break;
}
result.append(gColon);
result.append(gSpace);
// if the rule text begins with a space, write an apostrophe
// (whitespace after the rule descriptor is ignored; the
// apostrophe is used to make the whitespace significant)
if (fRuleText.charAt(0) == gSpace && (sub1 == nullptr || sub1->getPos() != 0)) {
result.append(gTick);
}
// now, write the rule's rule text, inserting appropriate
// substitution tokens in the appropriate places
UnicodeString ruleTextCopy;
ruleTextCopy.setTo(fRuleText);
UnicodeString temp;
if (sub2 != nullptr) {
sub2->toString(temp);
ruleTextCopy.insert(sub2->getPos(), temp);
}
if (sub1 != nullptr) {
sub1->toString(temp);
ruleTextCopy.insert(sub1->getPos(), temp);
}
result.append(ruleTextCopy);
// and finally, top the whole thing off with a semicolon and
// return the result
result.append(gSemicolon);
}
int64_t NFRule::getDivisor() const
{
return util64_pow(radix, exponent);
}
/**
* Internal function to facilitate numerical rounding. See the explanation in MultiplierSubstitution::transformNumber().
*/
bool NFRule::hasModulusSubstitution() const
{
return (sub1 != nullptr && sub1->isModulusSubstitution()) || (sub2 != nullptr && sub2->isModulusSubstitution());
}
//-----------------------------------------------------------------------
// formatting
//-----------------------------------------------------------------------
/**
* Formats the number, and inserts the resulting text into
* toInsertInto.
* @param number The number being formatted
* @param toInsertInto The string where the resultant text should
* be inserted
* @param pos The position in toInsertInto where the resultant text
* should be inserted
*/
void
NFRule::doFormat(int64_t number, UnicodeString& toInsertInto, int32_t pos, int32_t recursionCount, UErrorCode& status) const
{
// first, insert the rule's rule text into toInsertInto at the
// specified position, then insert the results of the substitutions
// into the right places in toInsertInto (notice we do the
// substitutions in reverse order so that the offsets don't get
// messed up)
int32_t pluralRuleStart = fRuleText.length();
int32_t lengthOffset = 0;
if (!rulePatternFormat) {
toInsertInto.insert(pos, fRuleText);
}
else {
pluralRuleStart = fRuleText.indexOf(gDollarOpenParenthesis, -1, 0);
int pluralRuleEnd = fRuleText.indexOf(gClosedParenthesisDollar, -1, pluralRuleStart);
int initialLength = toInsertInto.length();
if (pluralRuleEnd < fRuleText.length() - 1) {
toInsertInto.insert(pos, fRuleText.tempSubString(pluralRuleEnd + 2));
}
toInsertInto.insert(pos,
rulePatternFormat->format(static_cast<int32_t>(number / util64_pow(radix, exponent)), status));
if (pluralRuleStart > 0) {
toInsertInto.insert(pos, fRuleText.tempSubString(0, pluralRuleStart));
}
lengthOffset = fRuleText.length() - (toInsertInto.length() - initialLength);
}
if (sub2 != nullptr) {
sub2->doSubstitution(number, toInsertInto, pos - (sub2->getPos() > pluralRuleStart ? lengthOffset : 0), recursionCount, status);
}
if (sub1 != nullptr) {
sub1->doSubstitution(number, toInsertInto, pos - (sub1->getPos() > pluralRuleStart ? lengthOffset : 0), recursionCount, status);
}
}
/**
* Formats the number, and inserts the resulting text into
* toInsertInto.
* @param number The number being formatted
* @param toInsertInto The string where the resultant text should
* be inserted
* @param pos The position in toInsertInto where the resultant text
* should be inserted
*/
void
NFRule::doFormat(double number, UnicodeString& toInsertInto, int32_t pos, int32_t recursionCount, UErrorCode& status) const
{
// first, insert the rule's rule text into toInsertInto at the
// specified position, then insert the results of the substitutions
// into the right places in toInsertInto
// [again, we have two copies of this routine that do the same thing
// so that we don't sacrifice precision in a long by casting it
// to a double]
int32_t pluralRuleStart = fRuleText.length();
int32_t lengthOffset = 0;
if (!rulePatternFormat) {
toInsertInto.insert(pos, fRuleText);
}
else {
pluralRuleStart = fRuleText.indexOf(gDollarOpenParenthesis, -1, 0);
int pluralRuleEnd = fRuleText.indexOf(gClosedParenthesisDollar, -1, pluralRuleStart);
int initialLength = toInsertInto.length();
if (pluralRuleEnd < fRuleText.length() - 1) {
toInsertInto.insert(pos, fRuleText.tempSubString(pluralRuleEnd + 2));
}
double pluralVal = number;
if (0 <= pluralVal && pluralVal < 1) {
// We're in a fractional rule, and we have to match the NumeratorSubstitution behavior.
// 2.3 can become 0.2999999999999998 for the fraction due to rounding errors.
pluralVal = uprv_round(pluralVal * util64_pow(radix, exponent));
}
else {
pluralVal = pluralVal / util64_pow(radix, exponent);
}
toInsertInto.insert(pos, rulePatternFormat->format(static_cast<int32_t>(pluralVal), status));
if (pluralRuleStart > 0) {
toInsertInto.insert(pos, fRuleText.tempSubString(0, pluralRuleStart));
}
lengthOffset = fRuleText.length() - (toInsertInto.length() - initialLength);
}
if (sub2 != nullptr) {
sub2->doSubstitution(number, toInsertInto, pos - (sub2->getPos() > pluralRuleStart ? lengthOffset : 0), recursionCount, status);
}
if (sub1 != nullptr) {
sub1->doSubstitution(number, toInsertInto, pos - (sub1->getPos() > pluralRuleStart ? lengthOffset : 0), recursionCount, status);
}
}
/**
* Used by the owning rule set to determine whether to invoke the
* rollback rule (i.e., whether this rule or the one that precedes
* it in the rule set's list should be used to format the number)
* @param The number being formatted
* @return True if the rule set should use the rule that precedes
* this one in its list; false if it should use this rule
*/
UBool
NFRule::shouldRollBack(int64_t number) const
{
// we roll back if the rule contains a modulus substitution,
// the number being formatted is an even multiple of the rule's
// divisor, and the rule's base value is NOT an even multiple
// of its divisor
// In other words, if the original description had
// 100: << hundred[ >>];
// that expands into
// 100: << hundred;
// 101: << hundred >>;
// internally. But when we're formatting 200, if we use the rule
// at 101, which would normally apply, we get "two hundred zero".
// To prevent this, we roll back and use the rule at 100 instead.
// This is the logic that makes this happen: the rule at 101 has
// a modulus substitution, its base value isn't an even multiple
// of 100, and the value we're trying to format _is_ an even
// multiple of 100. This is called the "rollback rule."
if ((sub1 != nullptr && sub1->isModulusSubstitution()) || (sub2 != nullptr && sub2->isModulusSubstitution())) {
int64_t re = util64_pow(radix, exponent);
return (number % re) == 0 && (baseValue % re) != 0;
}
return false;
}
//-----------------------------------------------------------------------
// parsing
//-----------------------------------------------------------------------
/**
* Attempts to parse the string with this rule.
* @param text The string being parsed
* @param parsePosition On entry, the value is ignored and assumed to
* be 0. On exit, this has been updated with the position of the first
* character not consumed by matching the text against this rule
* (if this rule doesn't match the text at all, the parse position
* if left unchanged (presumably at 0) and the function returns
* new Long(0)).
* @param isFractionRule True if this rule is contained within a
* fraction rule set. This is only used if the rule has no
* substitutions.
* @return If this rule matched the text, this is the rule's base value
* combined appropriately with the results of parsing the substitutions.
* If nothing matched, this is new Long(0) and the parse position is
* left unchanged. The result will be an instance of Long if the
* result is an integer and Double otherwise. The result is never null.
*/
#ifdef RBNF_DEBUG
#include <stdio.h>
static void dumpUS(FILE* f, const UnicodeString& us) {
int len = us.length();
char* buf = (char *)uprv_malloc((len+1)*sizeof(char)); //new char[len+1];
if (buf != nullptr) {
us.extract(0, len, buf);
buf[len] = 0;
fprintf(f, "%s", buf);
uprv_free(buf); //delete[] buf;
}
}
#endif
UBool
NFRule::doParse(const UnicodeString& text,
ParsePosition& parsePosition,
UBool isFractionRule,
double upperBound,
uint32_t nonNumericalExecutedRuleMask,
int32_t recursionCount,
Formattable& resVal) const
{
// internally we operate on a copy of the string being parsed
// (because we're going to change it) and use our own ParsePosition
ParsePosition pp;
UnicodeString workText(text);
int32_t sub1Pos = sub1 != nullptr ? sub1->getPos() : fRuleText.length();
int32_t sub2Pos = sub2 != nullptr ? sub2->getPos() : fRuleText.length();
// check to see whether the text before the first substitution
// matches the text at the beginning of the string being
// parsed. If it does, strip that off the front of workText;
// otherwise, dump out with a mismatch
UnicodeString prefix;
prefix.setTo(fRuleText, 0, sub1Pos);
#ifdef RBNF_DEBUG
fprintf(stderr, "doParse %p ", this);
{
UnicodeString rt;
_appendRuleText(rt);
dumpUS(stderr, rt);
}
fprintf(stderr, " text: '");
dumpUS(stderr, text);
fprintf(stderr, "' prefix: '");
dumpUS(stderr, prefix);
#endif
stripPrefix(workText, prefix, pp);
int32_t prefixLength = text.length() - workText.length();
#ifdef RBNF_DEBUG
fprintf(stderr, "' pl: %d ppi: %d s1p: %d\n", prefixLength, pp.getIndex(), sub1Pos);
#endif
if (pp.getIndex() == 0 && sub1Pos != 0) {
// commented out because ParsePosition doesn't have error index in 1.1.x
// restored for ICU4C port