summaryrefslogtreecommitdiff
path: root/deps/icu-small/source/i18n/tridpars.cpp
blob: cbfdf03c427850101dfaa8d2f7a3292618427488 (plain)
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
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
**********************************************************************
*   Copyright (c) 2002-2014, International Business Machines Corporation
*   and others.  All Rights Reserved.
**********************************************************************
*   Date        Name        Description
*   01/14/2002  aliu        Creation.
**********************************************************************
*/

#include "unicode/utypes.h"

#if !UCONFIG_NO_TRANSLITERATION

#include "tridpars.h"
#include "hash.h"
#include "mutex.h"
#include "transreg.h"
#include "uassert.h"
#include "ucln_in.h"
#include "unicode/parsepos.h"
#include "unicode/translit.h"
#include "unicode/uchar.h"
#include "unicode/uniset.h"
#include "unicode/unistr.h"
#include "unicode/utrans.h"
#include "util.h"
#include "uvector.h"

U_NAMESPACE_BEGIN

static const UChar ID_DELIM    = 0x003B; // ;
static const UChar TARGET_SEP  = 0x002D; // -
static const UChar VARIANT_SEP = 0x002F; // /
static const UChar OPEN_REV    = 0x0028; // (
static const UChar CLOSE_REV   = 0x0029; // )

//static const UChar EMPTY[]     = {0}; // ""
static const UChar ANY[]       = {65,110,121,0}; // "Any"
static const UChar ANY_NULL[]  = {65,110,121,45,78,117,108,108,0}; // "Any-Null"

static const int32_t FORWARD = UTRANS_FORWARD;
static const int32_t REVERSE = UTRANS_REVERSE;

static Hashtable* SPECIAL_INVERSES = NULL;
static UInitOnce gSpecialInversesInitOnce = U_INITONCE_INITIALIZER;

/**
 * The mutex controlling access to SPECIAL_INVERSES
 */
static UMutex *LOCK() {
    static UMutex m = U_MUTEX_INITIALIZER;
    return &m;
}

TransliteratorIDParser::Specs::Specs(const UnicodeString& s, const UnicodeString& t,
                                     const UnicodeString& v, UBool sawS,
                                     const UnicodeString& f) {
    source = s;
    target = t;
    variant = v;
    sawSource = sawS;
    filter = f;
}

TransliteratorIDParser::SingleID::SingleID(const UnicodeString& c, const UnicodeString& b,
                                           const UnicodeString& f) {
    canonID = c;
    basicID = b;
    filter = f;
}

TransliteratorIDParser::SingleID::SingleID(const UnicodeString& c, const UnicodeString& b) {
    canonID = c;
    basicID = b;
}

Transliterator* TransliteratorIDParser::SingleID::createInstance() {
    Transliterator* t;
    if (basicID.length() == 0) {
        t = createBasicInstance(UnicodeString(TRUE, ANY_NULL, 8), &canonID);
    } else {
        t = createBasicInstance(basicID, &canonID);
    }
    if (t != NULL) {
        if (filter.length() != 0) {
            UErrorCode ec = U_ZERO_ERROR;
            UnicodeSet *set = new UnicodeSet(filter, ec);
            if (U_FAILURE(ec)) {
                delete set;
            } else {
                t->adoptFilter(set);
            }
        }
    }
    return t;
}


/**
 * Parse a single ID, that is, an ID of the general form
 * "[f1] s1-t1/v1 ([f2] s2-t3/v2)", with the parenthesized element
 * optional, the filters optional, and the variants optional.
 * @param id the id to be parsed
 * @param pos INPUT-OUTPUT parameter.  On input, the position of
 * the first character to parse.  On output, the position after
 * the last character parsed.
 * @param dir the direction.  If the direction is REVERSE then the
 * SingleID is constructed for the reverse direction.
 * @return a SingleID object or NULL
 */
TransliteratorIDParser::SingleID*
TransliteratorIDParser::parseSingleID(const UnicodeString& id, int32_t& pos,
                                      int32_t dir, UErrorCode& status) {

    int32_t start = pos;

    // The ID will be of the form A, A(), A(B), or (B), where
    // A and B are filter IDs.
    Specs* specsA = NULL;
    Specs* specsB = NULL;
    UBool sawParen = FALSE;

    // On the first pass, look for (B) or ().  If this fails, then
    // on the second pass, look for A, A(B), or A().
    for (int32_t pass=1; pass<=2; ++pass) {
        if (pass == 2) {
            specsA = parseFilterID(id, pos, TRUE);
            if (specsA == NULL) {
                pos = start;
                return NULL;
            }
        }
        if (ICU_Utility::parseChar(id, pos, OPEN_REV)) {
            sawParen = TRUE;
            if (!ICU_Utility::parseChar(id, pos, CLOSE_REV)) {
                specsB = parseFilterID(id, pos, TRUE);
                // Must close with a ')'
                if (specsB == NULL || !ICU_Utility::parseChar(id, pos, CLOSE_REV)) {
                    delete specsA;
                    pos = start;
                    return NULL;
                }
            }
            break;
        }
    }

    // Assemble return results
    SingleID* single;
    if (sawParen) {
        if (dir == FORWARD) {
            SingleID* b = specsToID(specsB, FORWARD);
            single = specsToID(specsA, FORWARD);
            // Null pointers check
            if (b == NULL || single == NULL) {
		delete b;
		delete single;
		status = U_MEMORY_ALLOCATION_ERROR;
		return NULL;
            }
            single->canonID.append(OPEN_REV)
                .append(b->canonID).append(CLOSE_REV);
            if (specsA != NULL) {
                single->filter = specsA->filter;
            }
            delete b;
        } else {
            SingleID* a = specsToID(specsA, FORWARD);
            single = specsToID(specsB, FORWARD);
            // Check for null pointer.
            if (a == NULL || single == NULL) {
		delete a;
		delete single;
		status = U_MEMORY_ALLOCATION_ERROR;
		return NULL;
            }
            single->canonID.append(OPEN_REV)
                .append(a->canonID).append(CLOSE_REV);
            if (specsB != NULL) {
                single->filter = specsB->filter;
            }
            delete a;
        }
    } else {
        // assert(specsA != NULL);
        if (dir == FORWARD) {
            single = specsToID(specsA, FORWARD);
        } else {
            single = specsToSpecialInverse(*specsA, status);
            if (single == NULL) {
                single = specsToID(specsA, REVERSE);
            }
        }
        // Check for NULL pointer
        if (single == NULL) {
		status = U_MEMORY_ALLOCATION_ERROR;
		return NULL;
        }
        single->filter = specsA->filter;
    }

    delete specsA;
    delete specsB;

    return single;
}

/**
 * Parse a filter ID, that is, an ID of the general form
 * "[f1] s1-t1/v1", with the filters optional, and the variants optional.
 * @param id the id to be parsed
 * @param pos INPUT-OUTPUT parameter.  On input, the position of
 * the first character to parse.  On output, the position after
 * the last character parsed.
 * @return a SingleID object or null if the parse fails
 */
TransliteratorIDParser::SingleID*
TransliteratorIDParser::parseFilterID(const UnicodeString& id, int32_t& pos) {

    int32_t start = pos;

    Specs* specs = parseFilterID(id, pos, TRUE);
    if (specs == NULL) {
        pos = start;
        return NULL;
    }

    // Assemble return results
    SingleID* single = specsToID(specs, FORWARD);
    if (single != NULL) {
        single->filter = specs->filter;
    }
    delete specs;
    return single;
}

/**
 * Parse a global filter of the form "[f]" or "([f])", depending
 * on 'withParens'.
 * @param id the pattern the parse
 * @param pos INPUT-OUTPUT parameter.  On input, the position of
 * the first character to parse.  On output, the position after
 * the last character parsed.
 * @param dir the direction.
 * @param withParens INPUT-OUTPUT parameter.  On entry, if
 * withParens is 0, then parens are disallowed.  If it is 1,
 * then parens are requires.  If it is -1, then parens are
 * optional, and the return result will be set to 0 or 1.
 * @param canonID OUTPUT parameter.  The pattern for the filter
 * added to the canonID, either at the end, if dir is FORWARD, or
 * at the start, if dir is REVERSE.  The pattern will be enclosed
 * in parentheses if appropriate, and will be suffixed with an
 * ID_DELIM character.  May be NULL.
 * @return a UnicodeSet object or NULL.  A non-NULL results
 * indicates a successful parse, regardless of whether the filter
 * applies to the given direction.  The caller should discard it
 * if withParens != (dir == REVERSE).
 */
UnicodeSet* TransliteratorIDParser::parseGlobalFilter(const UnicodeString& id, int32_t& pos,
                                                      int32_t dir,
                                                      int32_t& withParens,
                                                      UnicodeString* canonID) {
    UnicodeSet* filter = NULL;
    int32_t start = pos;

    if (withParens == -1) {
        withParens = ICU_Utility::parseChar(id, pos, OPEN_REV) ? 1 : 0;
    } else if (withParens == 1) {
        if (!ICU_Utility::parseChar(id, pos, OPEN_REV)) {
            pos = start;
            return NULL;
        }
    }

    ICU_Utility::skipWhitespace(id, pos, TRUE);

    if (UnicodeSet::resemblesPattern(id, pos)) {
        ParsePosition ppos(pos);
        UErrorCode ec = U_ZERO_ERROR;
        filter = new UnicodeSet(id, ppos, USET_IGNORE_SPACE, NULL, ec);
        /* test for NULL */
        if (filter == 0) {
            pos = start;
            return 0;
        }
        if (U_FAILURE(ec)) {
            delete filter;
            pos = start;
            return NULL;
        }

        UnicodeString pattern;
        id.extractBetween(pos, ppos.getIndex(), pattern);
        pos = ppos.getIndex();

        if (withParens == 1 && !ICU_Utility::parseChar(id, pos, CLOSE_REV)) {
            pos = start;
            return NULL;
        }

        // In the forward direction, append the pattern to the
        // canonID.  In the reverse, insert it at zero, and invert
        // the presence of parens ("A" <-> "(A)").
        if (canonID != NULL) {
            if (dir == FORWARD) {
                if (withParens == 1) {
                    pattern.insert(0, OPEN_REV);
                    pattern.append(CLOSE_REV);
                }
                canonID->append(pattern).append(ID_DELIM);
            } else {
                if (withParens == 0) {
                    pattern.insert(0, OPEN_REV);
                    pattern.append(CLOSE_REV);
                }
                canonID->insert(0, pattern);
                canonID->insert(pattern.length(), ID_DELIM);
            }
        }
    }

    return filter;
}

U_CDECL_BEGIN
static void U_CALLCONV _deleteSingleID(void* obj) {
    delete (TransliteratorIDParser::SingleID*) obj;
}

static void U_CALLCONV _deleteTransliteratorTrIDPars(void* obj) {
    delete (Transliterator*) obj;
}
U_CDECL_END

/**
 * Parse a compound ID, consisting of an optional forward global
 * filter, a separator, one or more single IDs delimited by
 * separators, an an optional reverse global filter.  The
 * separator is a semicolon.  The global filters are UnicodeSet
 * patterns.  The reverse global filter must be enclosed in
 * parentheses.
 * @param id the pattern the parse
 * @param dir the direction.
 * @param canonID OUTPUT parameter that receives the canonical ID,
 * consisting of canonical IDs for all elements, as returned by
 * parseSingleID(), separated by semicolons.  Previous contents
 * are discarded.
 * @param list OUTPUT parameter that receives a list of SingleID
 * objects representing the parsed IDs.  Previous contents are
 * discarded.
 * @param globalFilter OUTPUT parameter that receives a pointer to
 * a newly created global filter for this ID in this direction, or
 * NULL if there is none.
 * @return TRUE if the parse succeeds, that is, if the entire
 * id is consumed without syntax error.
 */
UBool TransliteratorIDParser::parseCompoundID(const UnicodeString& id, int32_t dir,
                                              UnicodeString& canonID,
                                              UVector& list,
                                              UnicodeSet*& globalFilter) {
    UErrorCode ec = U_ZERO_ERROR;
    int32_t i;
    int32_t pos = 0;
    int32_t withParens = 1;
    list.removeAllElements();
    UnicodeSet* filter;
    globalFilter = NULL;
    canonID.truncate(0);

    // Parse leading global filter, if any
    withParens = 0; // parens disallowed
    filter = parseGlobalFilter(id, pos, dir, withParens, &canonID);
    if (filter != NULL) {
        if (!ICU_Utility::parseChar(id, pos, ID_DELIM)) {
            // Not a global filter; backup and resume
            canonID.truncate(0);
            pos = 0;
        }
        if (dir == FORWARD) {
            globalFilter = filter;
        } else {
            delete filter;
        }
        filter = NULL;
    }

    UBool sawDelimiter = TRUE;
    for (;;) {
        SingleID* single = parseSingleID(id, pos, dir, ec);
        if (single == NULL) {
            break;
        }
        if (dir == FORWARD) {
            list.addElement(single, ec);
        } else {
            list.insertElementAt(single, 0, ec);
        }
        if (U_FAILURE(ec)) {
            goto FAIL;
        }
        if (!ICU_Utility::parseChar(id, pos, ID_DELIM)) {
            sawDelimiter = FALSE;
            break;
        }
    }

    if (list.size() == 0) {
        goto FAIL;
    }

    // Construct canonical ID
    for (i=0; i<list.size(); ++i) {
        SingleID* single = (SingleID*) list.elementAt(i);
        canonID.append(single->canonID);
        if (i != (list.size()-1)) {
            canonID.append(ID_DELIM);
        }
    }

    // Parse trailing global filter, if any, and only if we saw
    // a trailing delimiter after the IDs.
    if (sawDelimiter) {
        withParens = 1; // parens required
        filter = parseGlobalFilter(id, pos, dir, withParens, &canonID);
        if (filter != NULL) {
            // Don't require trailing ';', but parse it if present
            ICU_Utility::parseChar(id, pos, ID_DELIM);

            if (dir == REVERSE) {
                globalFilter = filter;
            } else {
                delete filter;
            }
            filter = NULL;
        }
    }

    // Trailing unparsed text is a syntax error
    ICU_Utility::skipWhitespace(id, pos, TRUE);
    if (pos != id.length()) {
        goto FAIL;
    }

    return TRUE;

 FAIL:
    UObjectDeleter *save = list.setDeleter(_deleteSingleID);
    list.removeAllElements();
    list.setDeleter(save);
    delete globalFilter;
    globalFilter = NULL;
    return FALSE;
}

/**
 * Convert the elements of the 'list' vector, which are SingleID
 * objects, into actual Transliterator objects.  In the course of
 * this, some (or all) entries may be removed.  If all entries
 * are removed, the NULL transliterator will be added.
 *
 * Delete entries with empty basicIDs; these are generated by
 * elements like "(A)" in the forward direction, or "A()" in
 * the reverse.  THIS MAY RESULT IN AN EMPTY VECTOR.  Convert
 * SingleID entries to actual transliterators.
 *
 * @param list vector of SingleID objects.  On exit, vector
 * of one or more Transliterators.
 * @return new value of insertIndex.  The index will shift if
 * there are empty items, like "(Lower)", with indices less than
 * insertIndex.
 */
void TransliteratorIDParser::instantiateList(UVector& list,
                                                UErrorCode& ec) {
    UVector tlist(ec);
    if (U_FAILURE(ec)) {
        goto RETURN;
    }
    tlist.setDeleter(_deleteTransliteratorTrIDPars);

    Transliterator* t;
    int32_t i;
    for (i=0; i<=list.size(); ++i) { // [sic]: i<=list.size()
        // We run the loop too long by one, so we can
        // do an insert after the last element
        if (i==list.size()) {
            break;
        }

        SingleID* single = (SingleID*) list.elementAt(i);
        if (single->basicID.length() != 0) {
            t = single->createInstance();
            if (t == NULL) {
                ec = U_INVALID_ID;
                goto RETURN;
            }
            tlist.addElement(t, ec);
            if (U_FAILURE(ec)) {
                delete t;
                goto RETURN;
            }
        }
    }

    // An empty list is equivalent to a NULL transliterator.
    if (tlist.size() == 0) {
        t = createBasicInstance(UnicodeString(TRUE, ANY_NULL, 8), NULL);
        if (t == NULL) {
            // Should never happen
            ec = U_INTERNAL_TRANSLITERATOR_ERROR;
        }
        tlist.addElement(t, ec);
        if (U_FAILURE(ec)) {
            delete t;
        }
    }

 RETURN:

    UObjectDeleter *save = list.setDeleter(_deleteSingleID);
    list.removeAllElements();

    if (U_SUCCESS(ec)) {
        list.setDeleter(_deleteTransliteratorTrIDPars);

        while (tlist.size() > 0) {
            t = (Transliterator*) tlist.orphanElementAt(0);
            list.addElement(t, ec);
            if (U_FAILURE(ec)) {
                delete t;
                list.removeAllElements();
                break;
            }
        }
    }

    list.setDeleter(save);
}

/**
 * Parse an ID into pieces.  Take IDs of the form T, T/V, S-T,
 * S-T/V, or S/V-T.  If the source is missing, return a source of
 * ANY.
 * @param id the id string, in any of several forms
 * @return an array of 4 strings: source, target, variant, and
 * isSourcePresent.  If the source is not present, ANY will be
 * given as the source, and isSourcePresent will be NULL.  Otherwise
 * isSourcePresent will be non-NULL.  The target may be empty if the
 * id is not well-formed.  The variant may be empty.
 */
void TransliteratorIDParser::IDtoSTV(const UnicodeString& id,
                                     UnicodeString& source,
                                     UnicodeString& target,
                                     UnicodeString& variant,
                                     UBool& isSourcePresent) {
    source.setTo(ANY, 3);
    target.truncate(0);
    variant.truncate(0);

    int32_t sep = id.indexOf(TARGET_SEP);
    int32_t var = id.indexOf(VARIANT_SEP);
    if (var < 0) {
        var = id.length();
    }
    isSourcePresent = FALSE;

    if (sep < 0) {
        // Form: T/V or T (or /V)
        id.extractBetween(0, var, target);
        id.extractBetween(var, id.length(), variant);
    } else if (sep < var) {
        // Form: S-T/V or S-T (or -T/V or -T)
        if (sep > 0) {
            id.extractBetween(0, sep, source);
            isSourcePresent = TRUE;
        }
        id.extractBetween(++sep, var, target);
        id.extractBetween(var, id.length(), variant);
    } else {
        // Form: (S/V-T or /V-T)
        if (var > 0) {
            id.extractBetween(0, var, source);
            isSourcePresent = TRUE;
        }
        id.extractBetween(var, sep++, variant);
        id.extractBetween(sep, id.length(), target);
    }

    if (variant.length() > 0) {
        variant.remove(0, 1);
    }
}

/**
 * Given source, target, and variant strings, concatenate them into a
 * full ID.  If the source is empty, then "Any" will be used for the
 * source, so the ID will always be of the form s-t/v or s-t.
 */
void TransliteratorIDParser::STVtoID(const UnicodeString& source,
                                     const UnicodeString& target,
                                     const UnicodeString& variant,
                                     UnicodeString& id) {
    id = source;
    if (id.length() == 0) {
        id.setTo(ANY, 3);
    }
    id.append(TARGET_SEP).append(target);
    if (variant.length() != 0) {
        id.append(VARIANT_SEP).append(variant);
    }
    // NUL-terminate the ID string for getTerminatedBuffer.
    // This prevents valgrind and Purify warnings.
    id.append((UChar)0);
    id.truncate(id.length()-1);
}

/**
 * Register two targets as being inverses of one another.  For
 * example, calling registerSpecialInverse("NFC", "NFD", TRUE) causes
 * Transliterator to form the following inverse relationships:
 *
 * <pre>NFC => NFD
 * Any-NFC => Any-NFD
 * NFD => NFC
 * Any-NFD => Any-NFC</pre>
 *
 * (Without the special inverse registration, the inverse of NFC
 * would be NFC-Any.)  Note that NFD is shorthand for Any-NFD, but
 * that the presence or absence of "Any-" is preserved.
 *
 * <p>The relationship is symmetrical; registering (a, b) is
 * equivalent to registering (b, a).
 *
 * <p>The relevant IDs must still be registered separately as
 * factories or classes.
 *
 * <p>Only the targets are specified.  Special inverses always
 * have the form Any-Target1 <=> Any-Target2.  The target should
 * have canonical casing (the casing desired to be produced when
 * an inverse is formed) and should contain no whitespace or other
 * extraneous characters.
 *
 * @param target the target against which to register the inverse
 * @param inverseTarget the inverse of target, that is
 * Any-target.getInverse() => Any-inverseTarget
 * @param bidirectional if TRUE, register the reverse relation
 * as well, that is, Any-inverseTarget.getInverse() => Any-target
 */
void TransliteratorIDParser::registerSpecialInverse(const UnicodeString& target,
                                                    const UnicodeString& inverseTarget,
                                                    UBool bidirectional,
                                                    UErrorCode &status) {
    umtx_initOnce(gSpecialInversesInitOnce, init, status);
    if (U_FAILURE(status)) {
        return;
    }

    // If target == inverseTarget then force bidirectional => FALSE
    if (bidirectional && 0==target.caseCompare(inverseTarget, U_FOLD_CASE_DEFAULT)) {
        bidirectional = FALSE;
    }

    Mutex lock(LOCK());

    UnicodeString *tempus = new UnicodeString(inverseTarget);  // Used for null pointer check before usage.
    if (tempus == NULL) {
	status = U_MEMORY_ALLOCATION_ERROR;
	return;
    }
    SPECIAL_INVERSES->put(target, tempus, status);
    if (bidirectional) {
	tempus = new UnicodeString(target);
	if (tempus == NULL) {
		status = U_MEMORY_ALLOCATION_ERROR;
		return;
	}
        SPECIAL_INVERSES->put(inverseTarget, tempus, status);
    }
}

//----------------------------------------------------------------
// Private implementation
//----------------------------------------------------------------

/**
 * Parse an ID into component pieces.  Take IDs of the form T,
 * T/V, S-T, S-T/V, or S/V-T.  If the source is missing, return a
 * source of ANY.
 * @param id the id string, in any of several forms
 * @param pos INPUT-OUTPUT parameter.  On input, pos is the
 * offset of the first character to parse in id.  On output,
 * pos is the offset after the last parsed character.  If the
 * parse failed, pos will be unchanged.
 * @param allowFilter2 if TRUE, a UnicodeSet pattern is allowed
 * at any location between specs or delimiters, and is returned
 * as the fifth string in the array.
 * @return a Specs object, or NULL if the parse failed.  If
 * neither source nor target was seen in the parsed id, then the
 * parse fails.  If allowFilter is TRUE, then the parsed filter
 * pattern is returned in the Specs object, otherwise the returned
 * filter reference is NULL.  If the parse fails for any reason
 * NULL is returned.
 */
TransliteratorIDParser::Specs*
TransliteratorIDParser::parseFilterID(const UnicodeString& id, int32_t& pos,
                                      UBool allowFilter) {
    UnicodeString first;
    UnicodeString source;
    UnicodeString target;
    UnicodeString variant;
    UnicodeString filter;
    UChar delimiter = 0;
    int32_t specCount = 0;
    int32_t start = pos;

    // This loop parses one of the following things with each
    // pass: a filter, a delimiter character (either '-' or '/'),
    // or a spec (source, target, or variant).
    for (;;) {
        ICU_Utility::skipWhitespace(id, pos, TRUE);
        if (pos == id.length()) {
            break;
        }

        // Parse filters
        if (allowFilter && filter.length() == 0 &&
            UnicodeSet::resemblesPattern(id, pos)) {

            ParsePosition ppos(pos);
            UErrorCode ec = U_ZERO_ERROR;
            UnicodeSet set(id, ppos, USET_IGNORE_SPACE, NULL, ec);
            if (U_FAILURE(ec)) {
                pos = start;
                return NULL;
            }
            id.extractBetween(pos, ppos.getIndex(), filter);
            pos = ppos.getIndex();
            continue;
        }

        if (delimiter == 0) {
            UChar c = id.charAt(pos);
            if ((c == TARGET_SEP && target.length() == 0) ||
                (c == VARIANT_SEP && variant.length() == 0)) {
                delimiter = c;
                ++pos;
                continue;
            }
        }

        // We are about to try to parse a spec with no delimiter
        // when we can no longer do so (we can only do so at the
        // start); break.
        if (delimiter == 0 && specCount > 0) {
            break;
        }

        UnicodeString spec = ICU_Utility::parseUnicodeIdentifier(id, pos);
        if (spec.length() == 0) {
            // Note that if there was a trailing delimiter, we
            // consume it.  So Foo-, Foo/, Foo-Bar/, and Foo/Bar-
            // are legal.
            break;
        }

        switch (delimiter) {
        case 0:
            first = spec;
            break;
        case TARGET_SEP:
            target = spec;
            break;
        case VARIANT_SEP:
            variant = spec;
            break;
        }
        ++specCount;
        delimiter = 0;
    }

    // A spec with no prior character is either source or target,
    // depending on whether an explicit "-target" was seen.
    if (first.length() != 0) {
        if (target.length() == 0) {
            target = first;
        } else {
            source = first;
        }
    }

    // Must have either source or target
    if (source.length() == 0 && target.length() == 0) {
        pos = start;
        return NULL;
    }

    // Empty source or target defaults to ANY
    UBool sawSource = TRUE;
    if (source.length() == 0) {
        source.setTo(ANY, 3);
        sawSource = FALSE;
    }
    if (target.length() == 0) {
        target.setTo(ANY, 3);
    }

    return new Specs(source, target, variant, sawSource, filter);
}

/**
 * Givens a Spec object, convert it to a SingleID object.  The
 * Spec object is a more unprocessed parse result.  The SingleID
 * object contains information about canonical and basic IDs.
 * @return a SingleID; never returns NULL.  Returned object always
 * has 'filter' field of NULL.
 */
TransliteratorIDParser::SingleID*
TransliteratorIDParser::specsToID(const Specs* specs, int32_t dir) {
    UnicodeString canonID;
    UnicodeString basicID;
    UnicodeString basicPrefix;
    if (specs != NULL) {
        UnicodeString buf;
        if (dir == FORWARD) {
            if (specs->sawSource) {
                buf.append(specs->source).append(TARGET_SEP);
            } else {
                basicPrefix = specs->source;
                basicPrefix.append(TARGET_SEP);
            }
            buf.append(specs->target);
        } else {
            buf.append(specs->target).append(TARGET_SEP).append(specs->source);
        }
        if (specs->variant.length() != 0) {
            buf.append(VARIANT_SEP).append(specs->variant);
        }
        basicID = basicPrefix;
        basicID.append(buf);
        if (specs->filter.length() != 0) {
            buf.insert(0, specs->filter);
        }
        canonID = buf;
    }
    return new SingleID(canonID, basicID);
}

/**
 * Given a Specs object, return a SingleID representing the
 * special inverse of that ID.  If there is no special inverse
 * then return NULL.
 * @return a SingleID or NULL.  Returned object always has
 * 'filter' field of NULL.
 */
TransliteratorIDParser::SingleID*
TransliteratorIDParser::specsToSpecialInverse(const Specs& specs, UErrorCode &status) {
    if (0!=specs.source.caseCompare(ANY, 3, U_FOLD_CASE_DEFAULT)) {
        return NULL;
    }
    umtx_initOnce(gSpecialInversesInitOnce, init, status);
    if (U_FAILURE(status)) {
        return NULL;
    }

    UnicodeString* inverseTarget;

    umtx_lock(LOCK());
    inverseTarget = (UnicodeString*) SPECIAL_INVERSES->get(specs.target);
    umtx_unlock(LOCK());

    if (inverseTarget != NULL) {
        // If the original ID contained "Any-" then make the
        // special inverse "Any-Foo"; otherwise make it "Foo".
        // So "Any-NFC" => "Any-NFD" but "NFC" => "NFD".
        UnicodeString buf;
        if (specs.filter.length() != 0) {
            buf.append(specs.filter);
        }
        if (specs.sawSource) {
            buf.append(ANY, 3).append(TARGET_SEP);
        }
        buf.append(*inverseTarget);

        UnicodeString basicID(TRUE, ANY, 3);
        basicID.append(TARGET_SEP).append(*inverseTarget);

        if (specs.variant.length() != 0) {
            buf.append(VARIANT_SEP).append(specs.variant);
            basicID.append(VARIANT_SEP).append(specs.variant);
        }
        return new SingleID(buf, basicID);
    }
    return NULL;
}

/**
 * Glue method to get around access problems in C++.  This would
 * ideally be inline but we want to avoid a circular header
 * dependency.
 */
Transliterator* TransliteratorIDParser::createBasicInstance(const UnicodeString& id, const UnicodeString* canonID) {
    return Transliterator::createBasicInstance(id, canonID);
}

/**
 * Initialize static memory. Called through umtx_initOnce only.
 */
void U_CALLCONV TransliteratorIDParser::init(UErrorCode &status) {
    U_ASSERT(SPECIAL_INVERSES == NULL);
    ucln_i18n_registerCleanup(UCLN_I18N_TRANSLITERATOR, utrans_transliterator_cleanup);

    SPECIAL_INVERSES = new Hashtable(TRUE, status);
    if (SPECIAL_INVERSES == NULL) {
	status = U_MEMORY_ALLOCATION_ERROR;
	return;
    }
    SPECIAL_INVERSES->setValueDeleter(uprv_deleteUObject);
}

/**
 * Free static memory.
 */
void TransliteratorIDParser::cleanup() {
    if (SPECIAL_INVERSES) {
        delete SPECIAL_INVERSES;
        SPECIAL_INVERSES = NULL;
    }
    gSpecialInversesInitOnce.reset();
}

U_NAMESPACE_END

#endif /* #if !UCONFIG_NO_TRANSLITERATION */

//eof