1 /*
2  **********************************************************************
3  *   Copyright (C) 1999-2011, International Business Machines
4  *   Corporation and others.  All Rights Reserved.
5  **********************************************************************
6  *   Date        Name        Description
7  *   11/17/99    aliu        Creation.
8  **********************************************************************
9  */
10 
11 #include "unicode/utypes.h"
12 
13 #if !UCONFIG_NO_TRANSLITERATION
14 
15 #include "unicode/rep.h"
16 #include "unicode/unifilt.h"
17 #include "unicode/uniset.h"
18 #include "unicode/utf16.h"
19 #include "rbt_rule.h"
20 #include "rbt_data.h"
21 #include "cmemory.h"
22 #include "strmatch.h"
23 #include "strrepl.h"
24 #include "util.h"
25 #include "putilimp.h"
26 
27 static const UChar FORWARD_OP[] = {32,62,32,0}; // " > "
28 
29 U_NAMESPACE_BEGIN
30 
31 /**
32  * Construct a new rule with the given input, output text, and other
33  * attributes.  A cursor position may be specified for the output text.
34  * @param input input string, including key and optional ante and
35  * post context
36  * @param anteContextPos offset into input to end of ante context, or -1 if
37  * none.  Must be <= input.length() if not -1.
38  * @param postContextPos offset into input to start of post context, or -1
39  * if none.  Must be <= input.length() if not -1, and must be >=
40  * anteContextPos.
41  * @param output output string
42  * @param cursorPosition offset into output at which cursor is located, or -1 if
43  * none.  If less than zero, then the cursor is placed after the
44  * <code>output</code>; that is, -1 is equivalent to
45  * <code>output.length()</code>.  If greater than
46  * <code>output.length()</code> then an exception is thrown.
47  * @param segs array of UnicodeFunctors corresponding to input pattern
48  * segments, or null if there are none.  The array itself is adopted,
49  * but the pointers within it are not.
50  * @param segsCount number of elements in segs[]
51  * @param anchorStart TRUE if the the rule is anchored on the left to
52  * the context start
53  * @param anchorEnd TRUE if the rule is anchored on the right to the
54  * context limit
55  */
TransliterationRule(const UnicodeString & input,int32_t anteContextPos,int32_t postContextPos,const UnicodeString & outputStr,int32_t cursorPosition,int32_t cursorOffset,UnicodeFunctor ** segs,int32_t segsCount,UBool anchorStart,UBool anchorEnd,const TransliterationRuleData * theData,UErrorCode & status)56 TransliterationRule::TransliterationRule(const UnicodeString& input,
57                                          int32_t anteContextPos, int32_t postContextPos,
58                                          const UnicodeString& outputStr,
59                                          int32_t cursorPosition, int32_t cursorOffset,
60                                          UnicodeFunctor** segs,
61                                          int32_t segsCount,
62                                          UBool anchorStart, UBool anchorEnd,
63                                          const TransliterationRuleData* theData,
64                                          UErrorCode& status) :
65     UMemory(),
66     segments(0),
67     data(theData) {
68 
69     if (U_FAILURE(status)) {
70         return;
71     }
72     // Do range checks only when warranted to save time
73     if (anteContextPos < 0) {
74         anteContextLength = 0;
75     } else {
76         if (anteContextPos > input.length()) {
77             // throw new IllegalArgumentException("Invalid ante context");
78             status = U_ILLEGAL_ARGUMENT_ERROR;
79             return;
80         }
81         anteContextLength = anteContextPos;
82     }
83     if (postContextPos < 0) {
84         keyLength = input.length() - anteContextLength;
85     } else {
86         if (postContextPos < anteContextLength ||
87             postContextPos > input.length()) {
88             // throw new IllegalArgumentException("Invalid post context");
89             status = U_ILLEGAL_ARGUMENT_ERROR;
90             return;
91         }
92         keyLength = postContextPos - anteContextLength;
93     }
94     if (cursorPosition < 0) {
95         cursorPosition = outputStr.length();
96     } else if (cursorPosition > outputStr.length()) {
97         // throw new IllegalArgumentException("Invalid cursor position");
98         status = U_ILLEGAL_ARGUMENT_ERROR;
99         return;
100     }
101     // We don't validate the segments array.  The caller must
102     // guarantee that the segments are well-formed (that is, that
103     // all $n references in the output refer to indices of this
104     // array, and that no array elements are null).
105     this->segments = segs;
106     this->segmentsCount = segsCount;
107 
108     pattern = input;
109     flags = 0;
110     if (anchorStart) {
111         flags |= ANCHOR_START;
112     }
113     if (anchorEnd) {
114         flags |= ANCHOR_END;
115     }
116 
117     anteContext = NULL;
118     if (anteContextLength > 0) {
119         anteContext = new StringMatcher(pattern, 0, anteContextLength,
120                                         FALSE, *data);
121         /* test for NULL */
122         if (anteContext == 0) {
123             status = U_MEMORY_ALLOCATION_ERROR;
124             return;
125         }
126     }
127 
128     key = NULL;
129     if (keyLength > 0) {
130         key = new StringMatcher(pattern, anteContextLength, anteContextLength + keyLength,
131                                 FALSE, *data);
132         /* test for NULL */
133         if (key == 0) {
134             status = U_MEMORY_ALLOCATION_ERROR;
135             return;
136         }
137     }
138 
139     int32_t postContextLength = pattern.length() - keyLength - anteContextLength;
140     postContext = NULL;
141     if (postContextLength > 0) {
142         postContext = new StringMatcher(pattern, anteContextLength + keyLength, pattern.length(),
143                                         FALSE, *data);
144         /* test for NULL */
145         if (postContext == 0) {
146             status = U_MEMORY_ALLOCATION_ERROR;
147             return;
148         }
149     }
150 
151     this->output = new StringReplacer(outputStr, cursorPosition + cursorOffset, data);
152     /* test for NULL */
153     if (this->output == 0) {
154         status = U_MEMORY_ALLOCATION_ERROR;
155         return;
156     }
157 }
158 
159 /**
160  * Copy constructor.
161  */
TransliterationRule(TransliterationRule & other)162 TransliterationRule::TransliterationRule(TransliterationRule& other) :
163     UMemory(other),
164     anteContext(NULL),
165     key(NULL),
166     postContext(NULL),
167     pattern(other.pattern),
168     anteContextLength(other.anteContextLength),
169     keyLength(other.keyLength),
170     flags(other.flags),
171     data(other.data) {
172 
173     segments = NULL;
174     segmentsCount = 0;
175     if (other.segmentsCount > 0) {
176         segments = (UnicodeFunctor **)uprv_malloc(other.segmentsCount * sizeof(UnicodeFunctor *));
177         uprv_memcpy(segments, other.segments, other.segmentsCount*sizeof(segments[0]));
178     }
179 
180     if (other.anteContext != NULL) {
181         anteContext = (StringMatcher*) other.anteContext->clone();
182     }
183     if (other.key != NULL) {
184         key = (StringMatcher*) other.key->clone();
185     }
186     if (other.postContext != NULL) {
187         postContext = (StringMatcher*) other.postContext->clone();
188     }
189     output = other.output->clone();
190 }
191 
~TransliterationRule()192 TransliterationRule::~TransliterationRule() {
193     uprv_free(segments);
194     delete anteContext;
195     delete key;
196     delete postContext;
197     delete output;
198 }
199 
200 /**
201  * Return the preceding context length.  This method is needed to
202  * support the <code>Transliterator</code> method
203  * <code>getMaximumContextLength()</code>.  Internally, this is
204  * implemented as the anteContextLength, optionally plus one if
205  * there is a start anchor.  The one character anchor gap is
206  * needed to make repeated incremental transliteration with
207  * anchors work.
208  */
getContextLength(void) const209 int32_t TransliterationRule::getContextLength(void) const {
210     return anteContextLength + ((flags & ANCHOR_START) ? 1 : 0);
211 }
212 
213 /**
214  * Internal method.  Returns 8-bit index value for this rule.
215  * This is the low byte of the first character of the key,
216  * unless the first character of the key is a set.  If it's a
217  * set, or otherwise can match multiple keys, the index value is -1.
218  */
getIndexValue() const219 int16_t TransliterationRule::getIndexValue() const {
220     if (anteContextLength == pattern.length()) {
221         // A pattern with just ante context {such as foo)>bar} can
222         // match any key.
223         return -1;
224     }
225     UChar32 c = pattern.char32At(anteContextLength);
226     return (int16_t)(data->lookupMatcher(c) == NULL ? (c & 0xFF) : -1);
227 }
228 
229 /**
230  * Internal method.  Returns true if this rule matches the given
231  * index value.  The index value is an 8-bit integer, 0..255,
232  * representing the low byte of the first character of the key.
233  * It matches this rule if it matches the first character of the
234  * key, or if the first character of the key is a set, and the set
235  * contains any character with a low byte equal to the index
236  * value.  If the rule contains only ante context, as in foo)>bar,
237  * then it will match any key.
238  */
matchesIndexValue(uint8_t v) const239 UBool TransliterationRule::matchesIndexValue(uint8_t v) const {
240     // Delegate to the key, or if there is none, to the postContext.
241     // If there is neither then we match any key; return true.
242     UnicodeMatcher *m = (key != NULL) ? key : postContext;
243     return (m != NULL) ? m->matchesIndexValue(v) : TRUE;
244 }
245 
246 /**
247  * Return true if this rule masks another rule.  If r1 masks r2 then
248  * r1 matches any input string that r2 matches.  If r1 masks r2 and r2 masks
249  * r1 then r1 == r2.  Examples: "a>x" masks "ab>y".  "a>x" masks "a[b]>y".
250  * "[c]a>x" masks "[dc]a>y".
251  */
masks(const TransliterationRule & r2) const252 UBool TransliterationRule::masks(const TransliterationRule& r2) const {
253     /* Rule r1 masks rule r2 if the string formed of the
254      * antecontext, key, and postcontext overlaps in the following
255      * way:
256      *
257      * r1:      aakkkpppp
258      * r2:     aaakkkkkpppp
259      *            ^
260      *
261      * The strings must be aligned at the first character of the
262      * key.  The length of r1 to the left of the alignment point
263      * must be <= the length of r2 to the left; ditto for the
264      * right.  The characters of r1 must equal (or be a superset
265      * of) the corresponding characters of r2.  The superset
266      * operation should be performed to check for UnicodeSet
267      * masking.
268      *
269      * Anchors:  Two patterns that differ only in anchors only
270      * mask one another if they are exactly equal, and r2 has
271      * all the anchors r1 has (optionally, plus some).  Here Y
272      * means the row masks the column, N means it doesn't.
273      *
274      *         ab   ^ab    ab$  ^ab$
275      *   ab    Y     Y     Y     Y
276      *  ^ab    N     Y     N     Y
277      *   ab$   N     N     Y     Y
278      *  ^ab$   N     N     N     Y
279      *
280      * Post context: {a}b masks ab, but not vice versa, since {a}b
281      * matches everything ab matches, and {a}b matches {|a|}b but ab
282      * does not.  Pre context is different (a{b} does not align with
283      * ab).
284      */
285 
286     /* LIMITATION of the current mask algorithm: Some rule
287      * maskings are currently not detected.  For example,
288      * "{Lu}]a>x" masks "A]a>y".  This can be added later. TODO
289      */
290 
291     int32_t len = pattern.length();
292     int32_t left = anteContextLength;
293     int32_t left2 = r2.anteContextLength;
294     int32_t right = len - left;
295     int32_t right2 = r2.pattern.length() - left2;
296     int32_t cachedCompare = r2.pattern.compare(left2 - left, len, pattern);
297 
298     // TODO Clean this up -- some logic might be combinable with the
299     // next statement.
300 
301     // Test for anchor masking
302     if (left == left2 && right == right2 &&
303         keyLength <= r2.keyLength &&
304         0 == cachedCompare) {
305         // The following boolean logic implements the table above
306         return (flags == r2.flags) ||
307             (!(flags & ANCHOR_START) && !(flags & ANCHOR_END)) ||
308             ((r2.flags & ANCHOR_START) && (r2.flags & ANCHOR_END));
309     }
310 
311     return left <= left2 &&
312         (right < right2 ||
313          (right == right2 && keyLength <= r2.keyLength)) &&
314          (0 == cachedCompare);
315 }
316 
posBefore(const Replaceable & str,int32_t pos)317 static inline int32_t posBefore(const Replaceable& str, int32_t pos) {
318     return (pos > 0) ?
319         pos - U16_LENGTH(str.char32At(pos-1)) :
320         pos - 1;
321 }
322 
posAfter(const Replaceable & str,int32_t pos)323 static inline int32_t posAfter(const Replaceable& str, int32_t pos) {
324     return (pos >= 0 && pos < str.length()) ?
325         pos + U16_LENGTH(str.char32At(pos)) :
326         pos + 1;
327 }
328 
329 /**
330  * Attempt a match and replacement at the given position.  Return
331  * the degree of match between this rule and the given text.  The
332  * degree of match may be mismatch, a partial match, or a full
333  * match.  A mismatch means at least one character of the text
334  * does not match the context or key.  A partial match means some
335  * context and key characters match, but the text is not long
336  * enough to match all of them.  A full match means all context
337  * and key characters match.
338  *
339  * If a full match is obtained, perform a replacement, update pos,
340  * and return U_MATCH.  Otherwise both text and pos are unchanged.
341  *
342  * @param text the text
343  * @param pos the position indices
344  * @param incremental if TRUE, test for partial matches that may
345  * be completed by additional text inserted at pos.limit.
346  * @return one of <code>U_MISMATCH</code>,
347  * <code>U_PARTIAL_MATCH</code>, or <code>U_MATCH</code>.  If
348  * incremental is FALSE then U_PARTIAL_MATCH will not be returned.
349  */
matchAndReplace(Replaceable & text,UTransPosition & pos,UBool incremental) const350 UMatchDegree TransliterationRule::matchAndReplace(Replaceable& text,
351                                                   UTransPosition& pos,
352                                                   UBool incremental) const {
353     // Matching and replacing are done in one method because the
354     // replacement operation needs information obtained during the
355     // match.  Another way to do this is to have the match method
356     // create a match result struct with relevant offsets, and to pass
357     // this into the replace method.
358 
359     // ============================ MATCH ===========================
360 
361     // Reset segment match data
362     if (segments != NULL) {
363         for (int32_t i=0; i<segmentsCount; ++i) {
364             ((StringMatcher*) segments[i])->resetMatch();
365         }
366     }
367 
368 //    int32_t lenDelta, keyLimit;
369     int32_t keyLimit;
370 
371     // ------------------------ Ante Context ------------------------
372 
373     // A mismatch in the ante context, or with the start anchor,
374     // is an outright U_MISMATCH regardless of whether we are
375     // incremental or not.
376     int32_t oText; // offset into 'text'
377 //    int32_t newStart = 0;
378     int32_t minOText;
379 
380     // Note (1): We process text in 16-bit code units, rather than
381     // 32-bit code points.  This works because stand-ins are
382     // always in the BMP and because we are doing a literal match
383     // operation, which can be done 16-bits at a time.
384 
385     int32_t anteLimit = posBefore(text, pos.contextStart);
386 
387     UMatchDegree match;
388 
389     // Start reverse match at char before pos.start
390     oText = posBefore(text, pos.start);
391 
392     if (anteContext != NULL) {
393         match = anteContext->matches(text, oText, anteLimit, FALSE);
394         if (match != U_MATCH) {
395             return U_MISMATCH;
396         }
397     }
398 
399     minOText = posAfter(text, oText);
400 
401     // ------------------------ Start Anchor ------------------------
402 
403     if (((flags & ANCHOR_START) != 0) && oText != anteLimit) {
404         return U_MISMATCH;
405     }
406 
407     // -------------------- Key and Post Context --------------------
408 
409     oText = pos.start;
410 
411     if (key != NULL) {
412         match = key->matches(text, oText, pos.limit, incremental);
413         if (match != U_MATCH) {
414             return match;
415         }
416     }
417 
418     keyLimit = oText;
419 
420     if (postContext != NULL) {
421         if (incremental && keyLimit == pos.limit) {
422             // The key matches just before pos.limit, and there is
423             // a postContext.  Since we are in incremental mode,
424             // we must assume more characters may be inserted at
425             // pos.limit -- this is a partial match.
426             return U_PARTIAL_MATCH;
427         }
428 
429         match = postContext->matches(text, oText, pos.contextLimit, incremental);
430         if (match != U_MATCH) {
431             return match;
432         }
433     }
434 
435     // ------------------------- Stop Anchor ------------------------
436 
437     if (((flags & ANCHOR_END)) != 0) {
438         if (oText != pos.contextLimit) {
439             return U_MISMATCH;
440         }
441         if (incremental) {
442             return U_PARTIAL_MATCH;
443         }
444     }
445 
446     // =========================== REPLACE ==========================
447 
448     // We have a full match.  The key is between pos.start and
449     // keyLimit.
450 
451     int32_t newStart;
452     int32_t newLength = output->toReplacer()->replace(text, pos.start, keyLimit, newStart);
453     int32_t lenDelta = newLength - (keyLimit - pos.start);
454 
455     oText += lenDelta;
456     pos.limit += lenDelta;
457     pos.contextLimit += lenDelta;
458     // Restrict new value of start to [minOText, min(oText, pos.limit)].
459     pos.start = uprv_max(minOText, uprv_min(uprv_min(oText, pos.limit), newStart));
460     return U_MATCH;
461 }
462 
463 /**
464  * Create a source string that represents this rule.  Append it to the
465  * given string.
466  */
toRule(UnicodeString & rule,UBool escapeUnprintable) const467 UnicodeString& TransliterationRule::toRule(UnicodeString& rule,
468                                            UBool escapeUnprintable) const {
469 
470     // Accumulate special characters (and non-specials following them)
471     // into quoteBuf.  Append quoteBuf, within single quotes, when
472     // a non-quoted element must be inserted.
473     UnicodeString str, quoteBuf;
474 
475     // Do not emit the braces '{' '}' around the pattern if there
476     // is neither anteContext nor postContext.
477     UBool emitBraces =
478         (anteContext != NULL) || (postContext != NULL);
479 
480     // Emit start anchor
481     if ((flags & ANCHOR_START) != 0) {
482         rule.append((UChar)94/*^*/);
483     }
484 
485     // Emit the input pattern
486     ICU_Utility::appendToRule(rule, anteContext, escapeUnprintable, quoteBuf);
487 
488     if (emitBraces) {
489         ICU_Utility::appendToRule(rule, (UChar) 0x007B /*{*/, TRUE, escapeUnprintable, quoteBuf);
490     }
491 
492     ICU_Utility::appendToRule(rule, key, escapeUnprintable, quoteBuf);
493 
494     if (emitBraces) {
495         ICU_Utility::appendToRule(rule, (UChar) 0x007D /*}*/, TRUE, escapeUnprintable, quoteBuf);
496     }
497 
498     ICU_Utility::appendToRule(rule, postContext, escapeUnprintable, quoteBuf);
499 
500     // Emit end anchor
501     if ((flags & ANCHOR_END) != 0) {
502         rule.append((UChar)36/*$*/);
503     }
504 
505     ICU_Utility::appendToRule(rule, UnicodeString(TRUE, FORWARD_OP, 3), TRUE, escapeUnprintable, quoteBuf);
506 
507     // Emit the output pattern
508 
509     ICU_Utility::appendToRule(rule, output->toReplacer()->toReplacerPattern(str, escapeUnprintable),
510                               TRUE, escapeUnprintable, quoteBuf);
511 
512     ICU_Utility::appendToRule(rule, (UChar) 0x003B /*;*/, TRUE, escapeUnprintable, quoteBuf);
513 
514     return rule;
515 }
516 
setData(const TransliterationRuleData * d)517 void TransliterationRule::setData(const TransliterationRuleData* d) {
518     data = d;
519     if (anteContext != NULL) anteContext->setData(d);
520     if (postContext != NULL) postContext->setData(d);
521     if (key != NULL) key->setData(d);
522     // assert(output != NULL);
523     output->setData(d);
524     // Don't have to do segments since they are in the context or key
525 }
526 
527 /**
528  * Union the set of all characters that may be modified by this rule
529  * into the given set.
530  */
addSourceSetTo(UnicodeSet & toUnionTo) const531 void TransliterationRule::addSourceSetTo(UnicodeSet& toUnionTo) const {
532     int32_t limit = anteContextLength + keyLength;
533     for (int32_t i=anteContextLength; i<limit; ) {
534         UChar32 ch = pattern.char32At(i);
535         i += U16_LENGTH(ch);
536         const UnicodeMatcher* matcher = data->lookupMatcher(ch);
537         if (matcher == NULL) {
538             toUnionTo.add(ch);
539         } else {
540             matcher->addMatchSetTo(toUnionTo);
541         }
542     }
543 }
544 
545 /**
546  * Union the set of all characters that may be emitted by this rule
547  * into the given set.
548  */
addTargetSetTo(UnicodeSet & toUnionTo) const549 void TransliterationRule::addTargetSetTo(UnicodeSet& toUnionTo) const {
550     output->toReplacer()->addReplacementSetTo(toUnionTo);
551 }
552 
553 U_NAMESPACE_END
554 
555 #endif /* #if !UCONFIG_NO_TRANSLITERATION */
556 
557 //eof
558