1 /*
2 ******************************************************************************
3 * Copyright (C) 1996-2012, International Business Machines Corporation and
4 * others. All Rights Reserved.
5 ******************************************************************************
6 */
7
8 /**
9 * \file
10 * \brief C++ API: RuleBasedCollator class provides the simple implementation of Collator.
11 */
12
13 /**
14 * File tblcoll.h
15 *
16 * Created by: Helena Shih
17 *
18 * Modification History:
19 *
20 * Date Name Description
21 * 2/5/97 aliu Added streamIn and streamOut methods. Added
22 * constructor which reads RuleBasedCollator object from
23 * a binary file. Added writeToFile method which streams
24 * RuleBasedCollator out to a binary file. The streamIn
25 * and streamOut methods use istream and ostream objects
26 * in binary mode.
27 * 2/12/97 aliu Modified to use TableCollationData sub-object to
28 * hold invariant data.
29 * 2/13/97 aliu Moved several methods into this class from Collation.
30 * Added a private RuleBasedCollator(Locale&) constructor,
31 * to be used by Collator::createDefault(). General
32 * clean up.
33 * 2/20/97 helena Added clone, operator==, operator!=, operator=, and copy
34 * constructor and getDynamicClassID.
35 * 3/5/97 aliu Modified constructFromFile() to add parameter
36 * specifying whether or not binary loading is to be
37 * attempted. This is required for dynamic rule loading.
38 * 05/07/97 helena Added memory allocation error detection.
39 * 6/17/97 helena Added IDENTICAL strength for compare, changed getRules to
40 * use MergeCollation::getPattern.
41 * 6/20/97 helena Java class name change.
42 * 8/18/97 helena Added internal API documentation.
43 * 09/03/97 helena Added createCollationKeyValues().
44 * 02/10/98 damiba Added compare with "length" parameter
45 * 08/05/98 erm Synched with 1.2 version of RuleBasedCollator.java
46 * 04/23/99 stephen Removed EDecompositionMode, merged with
47 * Normalizer::EMode
48 * 06/14/99 stephen Removed kResourceBundleSuffix
49 * 11/02/99 helena Collator performance enhancements. Eliminates the
50 * UnicodeString construction and special case for NO_OP.
51 * 11/23/99 srl More performance enhancements. Updates to NormalizerIterator
52 * internal state management.
53 * 12/15/99 aliu Update to support Thai collation. Move NormalizerIterator
54 * to implementation file.
55 * 01/29/01 synwee Modified into a C++ wrapper which calls C API
56 * (ucol.h)
57 */
58
59 #ifndef TBLCOLL_H
60 #define TBLCOLL_H
61
62 #include "unicode/utypes.h"
63
64
65 #if !UCONFIG_NO_COLLATION
66
67 #include "unicode/coll.h"
68 #include "unicode/ucol.h"
69 #include "unicode/sortkey.h"
70 #include "unicode/normlzr.h"
71
72 U_NAMESPACE_BEGIN
73
74 /**
75 * @stable ICU 2.0
76 */
77 class StringSearch;
78 /**
79 * @stable ICU 2.0
80 */
81 class CollationElementIterator;
82
83 /**
84 * The RuleBasedCollator class provides the simple implementation of
85 * Collator, using data-driven tables. The user can create a customized
86 * table-based collation.
87 * <P>
88 * <em>Important: </em>The ICU collation service has been reimplemented
89 * in order to achieve better performance and UCA compliance.
90 * For details, see the
91 * <a href="http://source.icu-project.org/repos/icu/icuhtml/trunk/design/collation/ICU_collation_design.htm">
92 * collation design document</a>.
93 * <p>
94 * RuleBasedCollator is a thin C++ wrapper over the C implementation.
95 * <p>
96 * For more information about the collation service see
97 * <a href="http://icu-project.org/userguide/Collate_Intro.html">the users guide</a>.
98 * <p>
99 * Collation service provides correct sorting orders for most locales supported in ICU.
100 * If specific data for a locale is not available, the orders eventually falls back
101 * to the <a href="http://www.unicode.org/unicode/reports/tr10/">UCA sort order</a>.
102 * <p>
103 * Sort ordering may be customized by providing your own set of rules. For more on
104 * this subject see the <a href="http://icu-project.org/userguide/Collate_Customization.html">
105 * Collation customization</a> section of the users guide.
106 * <p>
107 * Note, RuleBasedCollator is not to be subclassed.
108 * @see Collator
109 * @version 2.0 11/15/2001
110 */
111 class U_I18N_API RuleBasedCollator : public Collator
112 {
113 public:
114
115 // constructor -------------------------------------------------------------
116
117 /**
118 * RuleBasedCollator constructor. This takes the table rules and builds a
119 * collation table out of them. Please see RuleBasedCollator class
120 * description for more details on the collation rule syntax.
121 * @param rules the collation rules to build the collation table from.
122 * @param status reporting a success or an error.
123 * @see Locale
124 * @stable ICU 2.0
125 */
126 RuleBasedCollator(const UnicodeString& rules, UErrorCode& status);
127
128 /**
129 * RuleBasedCollator constructor. This takes the table rules and builds a
130 * collation table out of them. Please see RuleBasedCollator class
131 * description for more details on the collation rule syntax.
132 * @param rules the collation rules to build the collation table from.
133 * @param collationStrength default strength for comparison
134 * @param status reporting a success or an error.
135 * @see Locale
136 * @stable ICU 2.0
137 */
138 RuleBasedCollator(const UnicodeString& rules,
139 ECollationStrength collationStrength,
140 UErrorCode& status);
141
142 /**
143 * RuleBasedCollator constructor. This takes the table rules and builds a
144 * collation table out of them. Please see RuleBasedCollator class
145 * description for more details on the collation rule syntax.
146 * @param rules the collation rules to build the collation table from.
147 * @param decompositionMode the normalisation mode
148 * @param status reporting a success or an error.
149 * @see Locale
150 * @stable ICU 2.0
151 */
152 RuleBasedCollator(const UnicodeString& rules,
153 UColAttributeValue decompositionMode,
154 UErrorCode& status);
155
156 /**
157 * RuleBasedCollator constructor. This takes the table rules and builds a
158 * collation table out of them. Please see RuleBasedCollator class
159 * description for more details on the collation rule syntax.
160 * @param rules the collation rules to build the collation table from.
161 * @param collationStrength default strength for comparison
162 * @param decompositionMode the normalisation mode
163 * @param status reporting a success or an error.
164 * @see Locale
165 * @stable ICU 2.0
166 */
167 RuleBasedCollator(const UnicodeString& rules,
168 ECollationStrength collationStrength,
169 UColAttributeValue decompositionMode,
170 UErrorCode& status);
171
172 /**
173 * Copy constructor.
174 * @param other the RuleBasedCollator object to be copied
175 * @see Locale
176 * @stable ICU 2.0
177 */
178 RuleBasedCollator(const RuleBasedCollator& other);
179
180
181 /** Opens a collator from a collator binary image created using
182 * cloneBinary. Binary image used in instantiation of the
183 * collator remains owned by the user and should stay around for
184 * the lifetime of the collator. The API also takes a base collator
185 * which usualy should be UCA.
186 * @param bin binary image owned by the user and required through the
187 * lifetime of the collator
188 * @param length size of the image. If negative, the API will try to
189 * figure out the length of the image
190 * @param base fallback collator, usually UCA. Base is required to be
191 * present through the lifetime of the collator. Currently
192 * it cannot be NULL.
193 * @param status for catching errors
194 * @return newly created collator
195 * @see cloneBinary
196 * @stable ICU 3.4
197 */
198 RuleBasedCollator(const uint8_t *bin, int32_t length,
199 const RuleBasedCollator *base,
200 UErrorCode &status);
201 // destructor --------------------------------------------------------------
202
203 /**
204 * Destructor.
205 * @stable ICU 2.0
206 */
207 virtual ~RuleBasedCollator();
208
209 // public methods ----------------------------------------------------------
210
211 /**
212 * Assignment operator.
213 * @param other other RuleBasedCollator object to compare with.
214 * @stable ICU 2.0
215 */
216 RuleBasedCollator& operator=(const RuleBasedCollator& other);
217
218 /**
219 * Returns true if argument is the same as this object.
220 * @param other Collator object to be compared.
221 * @return true if arguments is the same as this object.
222 * @stable ICU 2.0
223 */
224 virtual UBool operator==(const Collator& other) const;
225
226 /**
227 * Makes a copy of this object.
228 * @return a copy of this object, owned by the caller
229 * @stable ICU 2.0
230 */
231 virtual Collator* clone(void) const;
232
233 /**
234 * Creates a collation element iterator for the source string. The caller of
235 * this method is responsible for the memory management of the return
236 * pointer.
237 * @param source the string over which the CollationElementIterator will
238 * iterate.
239 * @return the collation element iterator of the source string using this as
240 * the based Collator.
241 * @stable ICU 2.2
242 */
243 virtual CollationElementIterator* createCollationElementIterator(
244 const UnicodeString& source) const;
245
246 /**
247 * Creates a collation element iterator for the source. The caller of this
248 * method is responsible for the memory management of the returned pointer.
249 * @param source the CharacterIterator which produces the characters over
250 * which the CollationElementItgerator will iterate.
251 * @return the collation element iterator of the source using this as the
252 * based Collator.
253 * @stable ICU 2.2
254 */
255 virtual CollationElementIterator* createCollationElementIterator(
256 const CharacterIterator& source) const;
257
258 // Make deprecated versions of Collator::compare() visible.
259 using Collator::compare;
260
261 /**
262 * The comparison function compares the character data stored in two
263 * different strings. Returns information about whether a string is less
264 * than, greater than or equal to another string.
265 * @param source the source string to be compared with.
266 * @param target the string that is to be compared with the source string.
267 * @param status possible error code
268 * @return Returns an enum value. UCOL_GREATER if source is greater
269 * than target; UCOL_EQUAL if source is equal to target; UCOL_LESS if source is less
270 * than target
271 * @stable ICU 2.6
272 **/
273 virtual UCollationResult compare(const UnicodeString& source,
274 const UnicodeString& target,
275 UErrorCode &status) const;
276
277 /**
278 * Does the same thing as compare but limits the comparison to a specified
279 * length
280 * @param source the source string to be compared with.
281 * @param target the string that is to be compared with the source string.
282 * @param length the length the comparison is limited to
283 * @param status possible error code
284 * @return Returns an enum value. UCOL_GREATER if source (up to the specified
285 * length) is greater than target; UCOL_EQUAL if source (up to specified
286 * length) is equal to target; UCOL_LESS if source (up to the specified
287 * length) is less than target.
288 * @stable ICU 2.6
289 */
290 virtual UCollationResult compare(const UnicodeString& source,
291 const UnicodeString& target,
292 int32_t length,
293 UErrorCode &status) const;
294
295 /**
296 * The comparison function compares the character data stored in two
297 * different string arrays. Returns information about whether a string array
298 * is less than, greater than or equal to another string array.
299 * @param source the source string array to be compared with.
300 * @param sourceLength the length of the source string array. If this value
301 * is equal to -1, the string array is null-terminated.
302 * @param target the string that is to be compared with the source string.
303 * @param targetLength the length of the target string array. If this value
304 * is equal to -1, the string array is null-terminated.
305 * @param status possible error code
306 * @return Returns an enum value. UCOL_GREATER if source is greater
307 * than target; UCOL_EQUAL if source is equal to target; UCOL_LESS if source is less
308 * than target
309 * @stable ICU 2.6
310 */
311 virtual UCollationResult compare(const UChar* source, int32_t sourceLength,
312 const UChar* target, int32_t targetLength,
313 UErrorCode &status) const;
314
315 /**
316 * Compares two strings using the Collator.
317 * Returns whether the first one compares less than/equal to/greater than
318 * the second one.
319 * This version takes UCharIterator input.
320 * @param sIter the first ("source") string iterator
321 * @param tIter the second ("target") string iterator
322 * @param status ICU status
323 * @return UCOL_LESS, UCOL_EQUAL or UCOL_GREATER
324 * @stable ICU 4.2
325 */
326 virtual UCollationResult compare(UCharIterator &sIter,
327 UCharIterator &tIter,
328 UErrorCode &status) const;
329
330 /**
331 * Transforms a specified region of the string into a series of characters
332 * that can be compared with CollationKey.compare. Use a CollationKey when
333 * you need to do repeated comparisions on the same string. For a single
334 * comparison the compare method will be faster.
335 * @param source the source string.
336 * @param key the transformed key of the source string.
337 * @param status the error code status.
338 * @return the transformed key.
339 * @see CollationKey
340 * @stable ICU 2.0
341 */
342 virtual CollationKey& getCollationKey(const UnicodeString& source,
343 CollationKey& key,
344 UErrorCode& status) const;
345
346 /**
347 * Transforms a specified region of the string into a series of characters
348 * that can be compared with CollationKey.compare. Use a CollationKey when
349 * you need to do repeated comparisions on the same string. For a single
350 * comparison the compare method will be faster.
351 * @param source the source string.
352 * @param sourceLength the length of the source string.
353 * @param key the transformed key of the source string.
354 * @param status the error code status.
355 * @return the transformed key.
356 * @see CollationKey
357 * @stable ICU 2.0
358 */
359 virtual CollationKey& getCollationKey(const UChar *source,
360 int32_t sourceLength,
361 CollationKey& key,
362 UErrorCode& status) const;
363
364 /**
365 * Generates the hash code for the rule-based collation object.
366 * @return the hash code.
367 * @stable ICU 2.0
368 */
369 virtual int32_t hashCode(void) const;
370
371 /**
372 * Gets the locale of the Collator
373 * @param type can be either requested, valid or actual locale. For more
374 * information see the definition of ULocDataLocaleType in
375 * uloc.h
376 * @param status the error code status.
377 * @return locale where the collation data lives. If the collator
378 * was instantiated from rules, locale is empty.
379 * @deprecated ICU 2.8 likely to change in ICU 3.0, based on feedback
380 */
381 virtual Locale getLocale(ULocDataLocaleType type, UErrorCode& status) const;
382
383 /**
384 * Gets the tailoring rules for this collator.
385 * @return the collation tailoring from which this collator was created
386 * @stable ICU 2.0
387 */
388 const UnicodeString& getRules(void) const;
389
390 /**
391 * Gets the version information for a Collator.
392 * @param info the version # information, the result will be filled in
393 * @stable ICU 2.0
394 */
395 virtual void getVersion(UVersionInfo info) const;
396
397 /**
398 * Return the maximum length of any expansion sequences that end with the
399 * specified comparison order.
400 * @param order a collation order returned by previous or next.
401 * @return maximum size of the expansion sequences ending with the collation
402 * element or 1 if collation element does not occur at the end of
403 * any expansion sequence
404 * @see CollationElementIterator#getMaxExpansion
405 * @stable ICU 2.0
406 */
407 int32_t getMaxExpansion(int32_t order) const;
408
409 /**
410 * Returns a unique class ID POLYMORPHICALLY. Pure virtual override. This
411 * method is to implement a simple version of RTTI, since not all C++
412 * compilers support genuine RTTI. Polymorphic operator==() and clone()
413 * methods call this method.
414 * @return The class ID for this object. All objects of a given class have
415 * the same class ID. Objects of other classes have different class
416 * IDs.
417 * @stable ICU 2.0
418 */
419 virtual UClassID getDynamicClassID(void) const;
420
421 /**
422 * Returns the class ID for this class. This is useful only for comparing to
423 * a return value from getDynamicClassID(). For example:
424 * <pre>
425 * Base* polymorphic_pointer = createPolymorphicObject();
426 * if (polymorphic_pointer->getDynamicClassID() ==
427 * Derived::getStaticClassID()) ...
428 * </pre>
429 * @return The class ID for all objects of this class.
430 * @stable ICU 2.0
431 */
432 static UClassID U_EXPORT2 getStaticClassID(void);
433
434 /**
435 * Returns the binary format of the class's rules. The format is that of
436 * .col files.
437 * @param length Returns the length of the data, in bytes
438 * @param status the error code status.
439 * @return memory, owned by the caller, of size 'length' bytes.
440 * @stable ICU 2.2
441 */
442 uint8_t *cloneRuleData(int32_t &length, UErrorCode &status);
443
444
445 /** Creates a binary image of a collator. This binary image can be stored and
446 * later used to instantiate a collator using ucol_openBinary.
447 * This API supports preflighting.
448 * @param buffer a fill-in buffer to receive the binary image
449 * @param capacity capacity of the destination buffer
450 * @param status for catching errors
451 * @return size of the image
452 * @see ucol_openBinary
453 * @stable ICU 3.4
454 */
455 int32_t cloneBinary(uint8_t *buffer, int32_t capacity, UErrorCode &status);
456
457 /**
458 * Returns current rules. Delta defines whether full rules are returned or
459 * just the tailoring.
460 *
461 * getRules(void) should normally be used instead.
462 * See http://userguide.icu-project.org/collation/customization#TOC-Building-on-Existing-Locales
463 * @param delta one of UCOL_TAILORING_ONLY, UCOL_FULL_RULES.
464 * @param buffer UnicodeString to store the result rules
465 * @stable ICU 2.2
466 * @see UCOL_FULL_RULES
467 */
468 void getRules(UColRuleOption delta, UnicodeString &buffer);
469
470 /**
471 * Universal attribute setter
472 * @param attr attribute type
473 * @param value attribute value
474 * @param status to indicate whether the operation went on smoothly or there were errors
475 * @stable ICU 2.2
476 */
477 virtual void setAttribute(UColAttribute attr, UColAttributeValue value,
478 UErrorCode &status);
479
480 /**
481 * Universal attribute getter.
482 * @param attr attribute type
483 * @param status to indicate whether the operation went on smoothly or there were errors
484 * @return attribute value
485 * @stable ICU 2.2
486 */
487 virtual UColAttributeValue getAttribute(UColAttribute attr,
488 UErrorCode &status) const;
489
490 /**
491 * Sets the variable top to a collation element value of a string supplied.
492 * @param varTop one or more (if contraction) UChars to which the variable top should be set
493 * @param len length of variable top string. If -1 it is considered to be zero terminated.
494 * @param status error code. If error code is set, the return value is undefined. Errors set by this function are: <br>
495 * U_CE_NOT_FOUND_ERROR if more than one character was passed and there is no such a contraction<br>
496 * U_PRIMARY_TOO_LONG_ERROR if the primary for the variable top has more than two bytes
497 * @return a 32 bit value containing the value of the variable top in upper 16 bits. Lower 16 bits are undefined
498 * @stable ICU 2.0
499 */
500 virtual uint32_t setVariableTop(const UChar *varTop, int32_t len, UErrorCode &status);
501
502 /**
503 * Sets the variable top to a collation element value of a string supplied.
504 * @param varTop an UnicodeString size 1 or more (if contraction) of UChars to which the variable top should be set
505 * @param status error code. If error code is set, the return value is undefined. Errors set by this function are: <br>
506 * U_CE_NOT_FOUND_ERROR if more than one character was passed and there is no such a contraction<br>
507 * U_PRIMARY_TOO_LONG_ERROR if the primary for the variable top has more than two bytes
508 * @return a 32 bit value containing the value of the variable top in upper 16 bits. Lower 16 bits are undefined
509 * @stable ICU 2.0
510 */
511 virtual uint32_t setVariableTop(const UnicodeString &varTop, UErrorCode &status);
512
513 /**
514 * Sets the variable top to a collation element value supplied. Variable top is set to the upper 16 bits.
515 * Lower 16 bits are ignored.
516 * @param varTop CE value, as returned by setVariableTop or ucol)getVariableTop
517 * @param status error code (not changed by function)
518 * @stable ICU 2.0
519 */
520 virtual void setVariableTop(uint32_t varTop, UErrorCode &status);
521
522 /**
523 * Gets the variable top value of a Collator.
524 * Lower 16 bits are undefined and should be ignored.
525 * @param status error code (not changed by function). If error code is set, the return value is undefined.
526 * @stable ICU 2.0
527 */
528 virtual uint32_t getVariableTop(UErrorCode &status) const;
529
530 /**
531 * Get an UnicodeSet that contains all the characters and sequences tailored in
532 * this collator.
533 * @param status error code of the operation
534 * @return a pointer to a UnicodeSet object containing all the
535 * code points and sequences that may sort differently than
536 * in the UCA. The object must be disposed of by using delete
537 * @stable ICU 2.4
538 */
539 virtual UnicodeSet *getTailoredSet(UErrorCode &status) const;
540
541 /**
542 * Get the sort key as an array of bytes from an UnicodeString.
543 * @param source string to be processed.
544 * @param result buffer to store result in. If NULL, number of bytes needed
545 * will be returned.
546 * @param resultLength length of the result buffer. If if not enough the
547 * buffer will be filled to capacity.
548 * @return Number of bytes needed for storing the sort key
549 * @stable ICU 2.0
550 */
551 virtual int32_t getSortKey(const UnicodeString& source, uint8_t *result,
552 int32_t resultLength) const;
553
554 /**
555 * Get the sort key as an array of bytes from an UChar buffer.
556 * @param source string to be processed.
557 * @param sourceLength length of string to be processed. If -1, the string
558 * is 0 terminated and length will be decided by the function.
559 * @param result buffer to store result in. If NULL, number of bytes needed
560 * will be returned.
561 * @param resultLength length of the result buffer. If if not enough the
562 * buffer will be filled to capacity.
563 * @return Number of bytes needed for storing the sort key
564 * @stable ICU 2.2
565 */
566 virtual int32_t getSortKey(const UChar *source, int32_t sourceLength,
567 uint8_t *result, int32_t resultLength) const;
568
569 /**
570 * Retrieves the reordering codes for this collator.
571 * @param dest The array to fill with the script ordering.
572 * @param destCapacity The length of dest. If it is 0, then dest may be NULL and the function
573 * will only return the length of the result without writing any of the result string (pre-flighting).
574 * @param status A reference to an error code value, which must not indicate
575 * a failure before the function call.
576 * @return The length of the script ordering array.
577 * @see ucol_setReorderCodes
578 * @see Collator#getEquivalentReorderCodes
579 * @see Collator#setReorderCodes
580 * @stable ICU 4.8
581 */
582 virtual int32_t getReorderCodes(int32_t *dest,
583 int32_t destCapacity,
584 UErrorCode& status) const;
585
586 /**
587 * Sets the ordering of scripts for this collator.
588 * @param reorderCodes An array of script codes in the new order. This can be NULL if the
589 * length is also set to 0. An empty array will clear any reordering codes on the collator.
590 * @param reorderCodesLength The length of reorderCodes.
591 * @param status error code
592 * @see Collator#getReorderCodes
593 * @see Collator#getEquivalentReorderCodes
594 * @stable ICU 4.8
595 */
596 virtual void setReorderCodes(const int32_t* reorderCodes,
597 int32_t reorderCodesLength,
598 UErrorCode& status) ;
599
600 /**
601 * Retrieves the reorder codes that are grouped with the given reorder code. Some reorder
602 * codes will be grouped and must reorder together.
603 * @param reorderCode The reorder code to determine equivalence for.
604 * @param dest The array to fill with the script equivalene reordering codes.
605 * @param destCapacity The length of dest. If it is 0, then dest may be NULL and the
606 * function will only return the length of the result without writing any of the result
607 * string (pre-flighting).
608 * @param status A reference to an error code value, which must not indicate
609 * a failure before the function call.
610 * @return The length of the of the reordering code equivalence array.
611 * @see ucol_setReorderCodes
612 * @see Collator#getReorderCodes
613 * @see Collator#setReorderCodes
614 * @stable ICU 4.8
615 */
616 static int32_t U_EXPORT2 getEquivalentReorderCodes(int32_t reorderCode,
617 int32_t* dest,
618 int32_t destCapacity,
619 UErrorCode& status);
620
621 private:
622
623 // private static constants -----------------------------------------------
624
625 enum {
626 /* need look up in .commit() */
627 CHARINDEX = 0x70000000,
628 /* Expand index follows */
629 EXPANDCHARINDEX = 0x7E000000,
630 /* contract indexes follows */
631 CONTRACTCHARINDEX = 0x7F000000,
632 /* unmapped character values */
633 UNMAPPED = 0xFFFFFFFF,
634 /* primary strength increment */
635 PRIMARYORDERINCREMENT = 0x00010000,
636 /* secondary strength increment */
637 SECONDARYORDERINCREMENT = 0x00000100,
638 /* tertiary strength increment */
639 TERTIARYORDERINCREMENT = 0x00000001,
640 /* mask off anything but primary order */
641 PRIMARYORDERMASK = 0xffff0000,
642 /* mask off anything but secondary order */
643 SECONDARYORDERMASK = 0x0000ff00,
644 /* mask off anything but tertiary order */
645 TERTIARYORDERMASK = 0x000000ff,
646 /* mask off ignorable char order */
647 IGNORABLEMASK = 0x0000ffff,
648 /* use only the primary difference */
649 PRIMARYDIFFERENCEONLY = 0xffff0000,
650 /* use only the primary and secondary difference */
651 SECONDARYDIFFERENCEONLY = 0xffffff00,
652 /* primary order shift */
653 PRIMARYORDERSHIFT = 16,
654 /* secondary order shift */
655 SECONDARYORDERSHIFT = 8,
656 /* starting value for collation elements */
657 COLELEMENTSTART = 0x02020202,
658 /* testing mask for primary low element */
659 PRIMARYLOWZEROMASK = 0x00FF0000,
660 /* reseting value for secondaries and tertiaries */
661 RESETSECONDARYTERTIARY = 0x00000202,
662 /* reseting value for tertiaries */
663 RESETTERTIARY = 0x00000002,
664
665 PRIMIGNORABLE = 0x0202
666 };
667
668 // private data members ---------------------------------------------------
669
670 UBool dataIsOwned;
671
672 UBool isWriteThroughAlias;
673
674 /**
675 * c struct for collation. All initialisation for it has to be done through
676 * setUCollator().
677 */
678 UCollator *ucollator;
679
680 /**
681 * Rule UnicodeString
682 */
683 UnicodeString urulestring;
684
685 // friend classes --------------------------------------------------------
686
687 /**
688 * Used to iterate over collation elements in a character source.
689 */
690 friend class CollationElementIterator;
691
692 /**
693 * Collator ONLY needs access to RuleBasedCollator(const Locale&,
694 * UErrorCode&)
695 */
696 friend class Collator;
697
698 /**
699 * Searching over collation elements in a character source
700 */
701 friend class StringSearch;
702
703 // private constructors --------------------------------------------------
704
705 /**
706 * Default constructor
707 */
708 RuleBasedCollator();
709
710 /**
711 * RuleBasedCollator constructor. This constructor takes a locale. The
712 * only caller of this class should be Collator::createInstance(). If
713 * createInstance() happens to know that the requested locale's collation is
714 * implemented as a RuleBasedCollator, it can then call this constructor.
715 * OTHERWISE IT SHOULDN'T, since this constructor ALWAYS RETURNS A VALID
716 * COLLATION TABLE. It does this by falling back to defaults.
717 * @param desiredLocale locale used
718 * @param status error code status
719 */
720 RuleBasedCollator(const Locale& desiredLocale, UErrorCode& status);
721
722 /**
723 * common constructor implementation
724 *
725 * @param rules the collation rules to build the collation table from.
726 * @param collationStrength default strength for comparison
727 * @param decompositionMode the normalisation mode
728 * @param status reporting a success or an error.
729 */
730 void
731 construct(const UnicodeString& rules,
732 UColAttributeValue collationStrength,
733 UColAttributeValue decompositionMode,
734 UErrorCode& status);
735
736 // private methods -------------------------------------------------------
737
738 /**
739 * Creates the c struct for ucollator
740 * @param locale desired locale
741 * @param status error status
742 */
743 void setUCollator(const Locale& locale, UErrorCode& status);
744
745 /**
746 * Creates the c struct for ucollator
747 * @param locale desired locale name
748 * @param status error status
749 */
750 void setUCollator(const char* locale, UErrorCode& status);
751
752 /**
753 * Creates the c struct for ucollator. This used internally by StringSearch.
754 * Hence the responsibility of cleaning up the ucollator is not done by
755 * this RuleBasedCollator. The isDataOwned flag is set to FALSE.
756 * @param collator new ucollator data
757 */
758 void setUCollator(UCollator *collator);
759
760 public:
761 #ifndef U_HIDE_INTERNAL_API
762 /**
763 * Get UCollator data struct. Used only by StringSearch & intltest.
764 * @return UCollator data struct
765 * @internal
766 */
767 const UCollator * getUCollator();
768 #endif /* U_HIDE_INTERNAL_API */
769
770 protected:
771 /**
772 * Used internally by registraton to define the requested and valid locales.
773 * @param requestedLocale the requsted locale
774 * @param validLocale the valid locale
775 * @param actualLocale the actual locale
776 * @internal
777 */
778 virtual void setLocales(const Locale& requestedLocale, const Locale& validLocale, const Locale& actualLocale);
779
780 private:
781 // if not owned and not a write through alias, copy the ucollator
782 void checkOwned(void);
783
784 // utility to init rule string used by checkOwned and construct
785 void setRuleStringFromCollator();
786
787 public:
788 /** Get the short definition string for a collator. This internal API harvests the collator's
789 * locale and the attribute set and produces a string that can be used for opening
790 * a collator with the same properties using the ucol_openFromShortString API.
791 * This string will be normalized.
792 * The structure and the syntax of the string is defined in the "Naming collators"
793 * section of the users guide:
794 * http://icu-project.org/userguide/Collate_Concepts.html#Naming_Collators
795 * This function supports preflighting.
796 *
797 * This is internal, and intended to be used with delegate converters.
798 *
799 * @param locale a locale that will appear as a collators locale in the resulting
800 * short string definition. If NULL, the locale will be harvested
801 * from the collator.
802 * @param buffer space to hold the resulting string
803 * @param capacity capacity of the buffer
804 * @param status for returning errors. All the preflighting errors are featured
805 * @return length of the resulting string
806 * @see ucol_openFromShortString
807 * @see ucol_normalizeShortDefinitionString
808 * @see ucol_getShortDefinitionString
809 * @internal
810 */
811 virtual int32_t internalGetShortDefinitionString(const char *locale,
812 char *buffer,
813 int32_t capacity,
814 UErrorCode &status) const;
815 };
816
817 // inline method implementation ---------------------------------------------
818
setUCollator(const Locale & locale,UErrorCode & status)819 inline void RuleBasedCollator::setUCollator(const Locale &locale,
820 UErrorCode &status)
821 {
822 setUCollator(locale.getName(), status);
823 }
824
825
setUCollator(UCollator * collator)826 inline void RuleBasedCollator::setUCollator(UCollator *collator)
827 {
828
829 if (ucollator && dataIsOwned) {
830 ucol_close(ucollator);
831 }
832 ucollator = collator;
833 dataIsOwned = FALSE;
834 isWriteThroughAlias = TRUE;
835 setRuleStringFromCollator();
836 }
837
838 #ifndef U_HIDE_INTERNAL_API
getUCollator()839 inline const UCollator * RuleBasedCollator::getUCollator()
840 {
841 return ucollator;
842 }
843 #endif
844
845 U_NAMESPACE_END
846
847 #endif /* #if !UCONFIG_NO_COLLATION */
848
849 #endif
850