aboutsummaryrefslogtreecommitdiffstats
path: root/contrib/libs/icu/i18n/rbt_set.cpp
blob: d8d0384dda6f41830d6e9926c5e7564842778d81 (plain) (blame)
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
// © 2016 and later: Unicode, Inc. and others.
// License & terms of use: http://www.unicode.org/copyright.html
/*
 **********************************************************************
 *   Copyright (C) 1999-2011, International Business Machines
 *   Corporation and others.  All Rights Reserved.
 **********************************************************************
 *   Date        Name        Description
 *   11/17/99    aliu        Creation.
 **********************************************************************
 */

#include "unicode/utypes.h"

#if !UCONFIG_NO_TRANSLITERATION

#include "unicode/unistr.h"
#include "unicode/uniset.h"
#include "unicode/utf16.h"
#include "rbt_set.h"
#include "rbt_rule.h"
#include "cmemory.h"
#include "putilimp.h"

U_CDECL_BEGIN
static void U_CALLCONV _deleteRule(void *rule) {
    delete (icu::TransliterationRule *)rule;
}
U_CDECL_END

//----------------------------------------------------------------------
// BEGIN Debugging support
//----------------------------------------------------------------------

// #define DEBUG_RBT

#ifdef DEBUG_RBT
#include <stdio.h>
#include "charstr.h"

/**
 * @param appendTo result is appended to this param.
 * @param input the string being transliterated
 * @param pos the index struct
 */
static UnicodeString& _formatInput(UnicodeString &appendTo,
                                   const UnicodeString& input,
                                   const UTransPosition& pos) {
    // Output a string of the form aaa{bbb|ccc|ddd}eee, where
    // the {} indicate the context start and limit, and the ||
    // indicate the start and limit.
    if (0 <= pos.contextStart &&
        pos.contextStart <= pos.start &&
        pos.start <= pos.limit &&
        pos.limit <= pos.contextLimit &&
        pos.contextLimit <= input.length()) {

        UnicodeString a, b, c, d, e;
        input.extractBetween(0, pos.contextStart, a);
        input.extractBetween(pos.contextStart, pos.start, b);
        input.extractBetween(pos.start, pos.limit, c);
        input.extractBetween(pos.limit, pos.contextLimit, d);
        input.extractBetween(pos.contextLimit, input.length(), e);
        appendTo.append(a).append((UChar)123/*{*/).append(b).
            append((UChar)124/*|*/).append(c).append((UChar)124/*|*/).append(d).
            append((UChar)125/*}*/).append(e);
    } else {
        appendTo.append("INVALID UTransPosition");
        //appendTo.append((UnicodeString)"INVALID UTransPosition {cs=" +
        //                pos.contextStart + ", s=" + pos.start + ", l=" +
        //                pos.limit + ", cl=" + pos.contextLimit + "} on " +
        //                input);
    }
    return appendTo;
}

// Append a hex string to the target
UnicodeString& _appendHex(uint32_t number,
                          int32_t digits,
                          UnicodeString& target) {
    static const UChar digitString[] = {
        0x30, 0x31, 0x32, 0x33, 0x34, 0x35, 0x36, 0x37, 0x38, 0x39,
        0x41, 0x42, 0x43, 0x44, 0x45, 0x46, 0
    };
    while (digits--) {
        target += digitString[(number >> (digits*4)) & 0xF];
    }
    return target;
}

// Replace nonprintable characters with unicode escapes
UnicodeString& _escape(const UnicodeString &source,
                       UnicodeString &target) {
    for (int32_t i = 0; i < source.length(); ) {
        UChar32 ch = source.char32At(i);
        i += U16_LENGTH(ch);
        if (ch < 0x09 || (ch > 0x0A && ch < 0x20)|| ch > 0x7E) {
            if (ch <= 0xFFFF) {
                target += "\\u";
                _appendHex(ch, 4, target);
            } else {
                target += "\\U";
                _appendHex(ch, 8, target);
            }
        } else {
            target += ch;
        }
    }
    return target;
}

inline void _debugOut(const char* msg, TransliterationRule* rule,
                      const Replaceable& theText, UTransPosition& pos) {
    UnicodeString buf(msg, "");
    if (rule) {
        UnicodeString r;
        rule->toRule(r, TRUE);
        buf.append((UChar)32).append(r);
    }
    buf.append(UnicodeString(" => ", ""));
    UnicodeString* text = (UnicodeString*)&theText;
    _formatInput(buf, *text, pos);
    UnicodeString esc;
    _escape(buf, esc);
    CharString cbuf(esc);
    printf("%s\n", (const char*) cbuf);
}

#else
#define _debugOut(msg, rule, theText, pos)
#endif

//----------------------------------------------------------------------
// END Debugging support
//----------------------------------------------------------------------

// Fill the precontext and postcontext with the patterns of the rules
// that are masking one another.
static void maskingError(const icu::TransliterationRule& rule1,
                         const icu::TransliterationRule& rule2,
                         UParseError& parseError) {
    icu::UnicodeString r;
    int32_t len;

    parseError.line = parseError.offset = -1;
    
    // for pre-context
    rule1.toRule(r, FALSE);
    len = uprv_min(r.length(), U_PARSE_CONTEXT_LEN-1);
    r.extract(0, len, parseError.preContext);
    parseError.preContext[len] = 0;   
    
    //for post-context
    r.truncate(0);
    rule2.toRule(r, FALSE);
    len = uprv_min(r.length(), U_PARSE_CONTEXT_LEN-1);
    r.extract(0, len, parseError.postContext);
    parseError.postContext[len] = 0;   
}

U_NAMESPACE_BEGIN

/**
 * Construct a new empty rule set.
 */
TransliterationRuleSet::TransliterationRuleSet(UErrorCode& status) : UMemory() {
    ruleVector = new UVector(&_deleteRule, NULL, status);
    if (U_FAILURE(status)) {
        return;
    }
    if (ruleVector == NULL) {
        status = U_MEMORY_ALLOCATION_ERROR;
    }
    rules = NULL;
    maxContextLength = 0;
}

/**
 * Copy constructor.
 */
TransliterationRuleSet::TransliterationRuleSet(const TransliterationRuleSet& other) :
    UMemory(other),
    ruleVector(0),
    rules(0),
    maxContextLength(other.maxContextLength) {

    int32_t i, len;
    uprv_memcpy(index, other.index, sizeof(index));
    UErrorCode status = U_ZERO_ERROR;
    ruleVector = new UVector(&_deleteRule, NULL, status);
    if (other.ruleVector != 0 && ruleVector != 0 && U_SUCCESS(status)) {
        len = other.ruleVector->size();
        for (i=0; i<len && U_SUCCESS(status); ++i) {
            TransliterationRule *tempTranslitRule = new TransliterationRule(*(TransliterationRule*)other.ruleVector->elementAt(i));
            // Null pointer test
            if (tempTranslitRule == NULL) {
                status = U_MEMORY_ALLOCATION_ERROR;
                break;
            }
            ruleVector->addElement(tempTranslitRule, status);
            if (U_FAILURE(status)) {
                break;
            }
        }
    }
    if (other.rules != 0 && U_SUCCESS(status)) {
        UParseError p;
        freeze(p, status);
    }
}

/**
 * Destructor.
 */
TransliterationRuleSet::~TransliterationRuleSet() {
    delete ruleVector; // This deletes the contained rules
    uprv_free(rules);
}

void TransliterationRuleSet::setData(const TransliterationRuleData* d) {
    /**
     * We assume that the ruleset has already been frozen.
     */
    int32_t len = index[256]; // see freeze()
    for (int32_t i=0; i<len; ++i) {
        rules[i]->setData(d);
    }
}

/**
 * Return the maximum context length.
 * @return the length of the longest preceding context.
 */
int32_t TransliterationRuleSet::getMaximumContextLength(void) const {
    return maxContextLength;
}

/**
 * Add a rule to this set.  Rules are added in order, and order is
 * significant.  The last call to this method must be followed by
 * a call to <code>freeze()</code> before the rule set is used.
 *
 * <p>If freeze() has already been called, calling addRule()
 * unfreezes the rules, and freeze() must be called again.
 *
 * @param adoptedRule the rule to add
 */
void TransliterationRuleSet::addRule(TransliterationRule* adoptedRule,
                                     UErrorCode& status) {
    if (U_FAILURE(status)) {
        delete adoptedRule;
        return;
    }
    ruleVector->addElement(adoptedRule, status);

    int32_t len;
    if ((len = adoptedRule->getContextLength()) > maxContextLength) {
        maxContextLength = len;
    }

    uprv_free(rules);
    rules = 0;
}

/**
 * Check this for masked rules and index it to optimize performance.
 * The sequence of operations is: (1) add rules to a set using
 * <code>addRule()</code>; (2) freeze the set using
 * <code>freeze()</code>; (3) use the rule set.  If
 * <code>addRule()</code> is called after calling this method, it
 * invalidates this object, and this method must be called again.
 * That is, <code>freeze()</code> may be called multiple times,
 * although for optimal performance it shouldn't be.
 */
void TransliterationRuleSet::freeze(UParseError& parseError,UErrorCode& status) {
    /* Construct the rule array and index table.  We reorder the
     * rules by sorting them into 256 bins.  Each bin contains all
     * rules matching the index value for that bin.  A rule
     * matches an index value if string whose first key character
     * has a low byte equal to the index value can match the rule.
     *
     * Each bin contains zero or more rules, in the same order
     * they were found originally.  However, the total rules in
     * the bins may exceed the number in the original vector,
     * since rules that have a variable as their first key
     * character will generally fall into more than one bin.
     *
     * That is, each bin contains all rules that either have that
     * first index value as their first key character, or have
     * a set containing the index value as their first character.
     */
    int32_t n = ruleVector->size();
    int32_t j;
    int16_t x;
    UVector v(2*n, status); // heuristic; adjust as needed

    if (U_FAILURE(status)) {
        return;
    }

    /* Precompute the index values.  This saves a LOT of time.
     * Be careful not to call malloc(0).
     */
    int16_t* indexValue = (int16_t*) uprv_malloc( sizeof(int16_t) * (n > 0 ? n : 1) );
    /* test for NULL */
    if (indexValue == 0) {
        status = U_MEMORY_ALLOCATION_ERROR;
        return;
    }
    for (j=0; j<n; ++j) {
        TransliterationRule* r = (TransliterationRule*) ruleVector->elementAt(j);
        indexValue[j] = r->getIndexValue();
    }
    for (x=0; x<256; ++x) {
        index[x] = v.size();
        for (j=0; j<n; ++j) {
            if (indexValue[j] >= 0) {
                if (indexValue[j] == x) {
                    v.addElement(ruleVector->elementAt(j), status);
                }
            } else {
                // If the indexValue is < 0, then the first key character is
                // a set, and we must use the more time-consuming
                // matchesIndexValue check.  In practice this happens
                // rarely, so we seldom tread this code path.
                TransliterationRule* r = (TransliterationRule*) ruleVector->elementAt(j);
                if (r->matchesIndexValue((uint8_t)x)) {
                    v.addElement(r, status);
                }
            }
        }
    }
    uprv_free(indexValue);
    index[256] = v.size();

    /* Freeze things into an array.
     */
    uprv_free(rules); // Contains alias pointers

    /* You can't do malloc(0)! */
    if (v.size() == 0) {
        rules = NULL;
        return;
    }
    rules = (TransliterationRule **)uprv_malloc(v.size() * sizeof(TransliterationRule *));
    /* test for NULL */
    if (rules == 0) {
        status = U_MEMORY_ALLOCATION_ERROR;
        return;
    }
    for (j=0; j<v.size(); ++j) {
        rules[j] = (TransliterationRule*) v.elementAt(j);
    }

    // TODO Add error reporting that indicates the rules that
    //      are being masked.
    //UnicodeString errors;

    /* Check for masking.  This is MUCH faster than our old check,
     * which was each rule against each following rule, since we
     * only have to check for masking within each bin now.  It's
     * 256*O(n2^2) instead of O(n1^2), where n1 is the total rule
     * count, and n2 is the per-bin rule count.  But n2<<n1, so
     * it's a big win.
     */
    for (x=0; x<256; ++x) {
        for (j=index[x]; j<index[x+1]-1; ++j) {
            TransliterationRule* r1 = rules[j];
            for (int32_t k=j+1; k<index[x+1]; ++k) {
                TransliterationRule* r2 = rules[k];
                if (r1->masks(*r2)) {
//|                 if (errors == null) {
//|                     errors = new StringBuffer();
//|                 } else {
//|                     errors.append("\n");
//|                 }
//|                 errors.append("Rule " + r1 + " masks " + r2);
                    status = U_RULE_MASK_ERROR;
                    maskingError(*r1, *r2, parseError);
                    return;
                }
            }
        }
    }

    //if (errors != null) {
    //    throw new IllegalArgumentException(errors.toString());
    //}
}

/**
 * Transliterate the given text with the given UTransPosition
 * indices.  Return TRUE if the transliteration should continue
 * or FALSE if it should halt (because of a U_PARTIAL_MATCH match).
 * Note that FALSE is only ever returned if isIncremental is TRUE.
 * @param text the text to be transliterated
 * @param pos the position indices, which will be updated
 * @param incremental if TRUE, assume new text may be inserted
 * at index.limit, and return FALSE if thre is a partial match.
 * @return TRUE unless a U_PARTIAL_MATCH has been obtained,
 * indicating that transliteration should stop until more text
 * arrives.
 */
UBool TransliterationRuleSet::transliterate(Replaceable& text,
                                            UTransPosition& pos,
                                            UBool incremental) {
    int16_t indexByte = (int16_t) (text.char32At(pos.start) & 0xFF);
    for (int32_t i=index[indexByte]; i<index[indexByte+1]; ++i) {
        UMatchDegree m = rules[i]->matchAndReplace(text, pos, incremental);
        switch (m) {
        case U_MATCH:
            _debugOut("match", rules[i], text, pos);
            return TRUE;
        case U_PARTIAL_MATCH:
            _debugOut("partial match", rules[i], text, pos);
            return FALSE;
        default: /* Ram: added default to make GCC happy */
            break;
        }
    }
    // No match or partial match from any rule
    pos.start += U16_LENGTH(text.char32At(pos.start));
    _debugOut("no match", NULL, text, pos);
    return TRUE;
}

/**
 * Create rule strings that represents this rule set.
 */
UnicodeString& TransliterationRuleSet::toRules(UnicodeString& ruleSource,
                                               UBool escapeUnprintable) const {
    int32_t i;
    int32_t count = ruleVector->size();
    ruleSource.truncate(0);
    for (i=0; i<count; ++i) {
        if (i != 0) {
            ruleSource.append((UChar) 0x000A /*\n*/);
        }
        TransliterationRule *r =
            (TransliterationRule*) ruleVector->elementAt(i);
        r->toRule(ruleSource, escapeUnprintable);
    }
    return ruleSource;
}

/**
 * Return the set of all characters that may be modified
 * (getTarget=false) or emitted (getTarget=true) by this set.
 */
UnicodeSet& TransliterationRuleSet::getSourceTargetSet(UnicodeSet& result,
                               UBool getTarget) const
{
    result.clear();
    int32_t count = ruleVector->size();
    for (int32_t i=0; i<count; ++i) {
        TransliterationRule* r =
            (TransliterationRule*) ruleVector->elementAt(i);
        if (getTarget) {
            r->addTargetSetTo(result);
        } else {
            r->addSourceSetTo(result);
        }
    }
    return result;
}

U_NAMESPACE_END

#endif /* #if !UCONFIG_NO_TRANSLITERATION */