ICU 54.1  54.1
tblcoll.h
Go to the documentation of this file.
1 /*
2 ******************************************************************************
3 * Copyright (C) 1996-2014, International Business Machines Corporation and
4 * others. All Rights Reserved.
5 ******************************************************************************
6 */
7 
60 #ifndef TBLCOLL_H
61 #define TBLCOLL_H
62 
63 #include "unicode/utypes.h"
64 
65 #if !UCONFIG_NO_COLLATION
66 
67 #include "unicode/coll.h"
68 #include "unicode/locid.h"
69 #include "unicode/uiter.h"
70 #include "unicode/ucol.h"
71 
73 
74 struct CollationCacheEntry;
75 struct CollationData;
76 struct CollationSettings;
77 struct CollationTailoring;
81 class StringSearch;
85 class CollationElementIterator;
86 class CollationKey;
87 class SortKeyByteSink;
88 class UnicodeSet;
89 class UnicodeString;
90 class UVector64;
91 
112 public:
121  RuleBasedCollator(const UnicodeString& rules, UErrorCode& status);
122 
132  RuleBasedCollator(const UnicodeString& rules,
133  ECollationStrength collationStrength,
134  UErrorCode& status);
135 
145  RuleBasedCollator(const UnicodeString& rules,
146  UColAttributeValue decompositionMode,
147  UErrorCode& status);
148 
159  RuleBasedCollator(const UnicodeString& rules,
160  ECollationStrength collationStrength,
161  UColAttributeValue decompositionMode,
162  UErrorCode& status);
163 
164 #ifndef U_HIDE_INTERNAL_API
165 
169  RuleBasedCollator(const UnicodeString &rules,
170  UParseError &parseError, UnicodeString &reason,
171  UErrorCode &errorCode);
172 #endif /* U_HIDE_INTERNAL_API */
173 
179  RuleBasedCollator(const RuleBasedCollator& other);
180 
181 
199  RuleBasedCollator(const uint8_t *bin, int32_t length,
200  const RuleBasedCollator *base,
201  UErrorCode &status);
202 
207  virtual ~RuleBasedCollator();
208 
214  RuleBasedCollator& operator=(const RuleBasedCollator& other);
215 
222  virtual UBool operator==(const Collator& other) const;
223 
229  virtual Collator* clone(void) const;
230 
241  virtual CollationElementIterator* createCollationElementIterator(
242  const UnicodeString& source) const;
243 
253  virtual CollationElementIterator* createCollationElementIterator(
254  const CharacterIterator& source) const;
255 
256  // Make deprecated versions of Collator::compare() visible.
257  using Collator::compare;
258 
271  virtual UCollationResult compare(const UnicodeString& source,
272  const UnicodeString& target,
273  UErrorCode &status) const;
274 
288  virtual UCollationResult compare(const UnicodeString& source,
289  const UnicodeString& target,
290  int32_t length,
291  UErrorCode &status) const;
292 
309  virtual UCollationResult compare(const UChar* source, int32_t sourceLength,
310  const UChar* target, int32_t targetLength,
311  UErrorCode &status) const;
312 
324  virtual UCollationResult compare(UCharIterator &sIter,
325  UCharIterator &tIter,
326  UErrorCode &status) const;
327 
341  virtual UCollationResult compareUTF8(const StringPiece &source,
342  const StringPiece &target,
343  UErrorCode &status) const;
344 
359  virtual CollationKey& getCollationKey(const UnicodeString& source,
360  CollationKey& key,
361  UErrorCode& status) const;
362 
378  virtual CollationKey& getCollationKey(const UChar *source,
379  int32_t sourceLength,
380  CollationKey& key,
381  UErrorCode& status) const;
382 
388  virtual int32_t hashCode() const;
389 
400  virtual Locale getLocale(ULocDataLocaleType type, UErrorCode& status) const;
401 
407  const UnicodeString& getRules() const;
408 
414  virtual void getVersion(UVersionInfo info) const;
415 
416 #ifndef U_HIDE_DEPRECATED_API
417 
433  int32_t getMaxExpansion(int32_t order) const;
434 #endif /* U_HIDE_DEPRECATED_API */
435 
446  virtual UClassID getDynamicClassID(void) const;
447 
459  static UClassID U_EXPORT2 getStaticClassID(void);
460 
461 #ifndef U_HIDE_DEPRECATED_API
462 
472  uint8_t *cloneRuleData(int32_t &length, UErrorCode &status) const;
473 #endif /* U_HIDE_DEPRECATED_API */
474 
485  int32_t cloneBinary(uint8_t *buffer, int32_t capacity, UErrorCode &status) const;
486 
498  void getRules(UColRuleOption delta, UnicodeString &buffer) const;
499 
507  virtual void setAttribute(UColAttribute attr, UColAttributeValue value,
508  UErrorCode &status);
509 
518  UErrorCode &status) const;
519 
536  virtual Collator &setMaxVariable(UColReorderCode group, UErrorCode &errorCode);
537 
544  virtual UColReorderCode getMaxVariable() const;
545 
562  virtual uint32_t setVariableTop(const UChar *varTop, int32_t len, UErrorCode &status);
563 
579  virtual uint32_t setVariableTop(const UnicodeString &varTop, UErrorCode &status);
580 
592  virtual void setVariableTop(uint32_t varTop, UErrorCode &status);
593 
601  virtual uint32_t getVariableTop(UErrorCode &status) const;
602 
612  virtual UnicodeSet *getTailoredSet(UErrorCode &status) const;
613 
628  virtual int32_t getSortKey(const UnicodeString& source, uint8_t *result,
629  int32_t resultLength) const;
630 
647  virtual int32_t getSortKey(const UChar *source, int32_t sourceLength,
648  uint8_t *result, int32_t resultLength) const;
649 
663  virtual int32_t getReorderCodes(int32_t *dest,
664  int32_t destCapacity,
665  UErrorCode& status) const;
666 
677  virtual void setReorderCodes(const int32_t* reorderCodes,
678  int32_t reorderCodesLength,
679  UErrorCode& status) ;
680 
686  const char *left, int32_t leftLength,
687  const char *right, int32_t rightLength,
688  UErrorCode &errorCode) const;
689 
713  virtual int32_t internalGetShortDefinitionString(const char *locale,
714  char *buffer,
715  int32_t capacity,
716  UErrorCode &status) const;
717 
722  virtual int32_t internalNextSortKeyPart(
723  UCharIterator *iter, uint32_t state[2],
724  uint8_t *dest, int32_t count, UErrorCode &errorCode) const;
725 
731 
732 #ifndef U_HIDE_INTERNAL_API
733 
739  const char *internalGetLocaleID(ULocDataLocaleType type, UErrorCode &errorCode) const;
740 
753  void internalGetContractionsAndExpansions(
754  UnicodeSet *contractions, UnicodeSet *expansions,
755  UBool addPrefixes, UErrorCode &errorCode) const;
756 
762  void internalAddContractions(UChar32 c, UnicodeSet &set, UErrorCode &errorCode) const;
763 
768  void internalBuildTailoring(
769  const UnicodeString &rules,
770  int32_t strength,
771  UColAttributeValue decompositionMode,
772  UParseError *outParseError, UnicodeString *outReason,
773  UErrorCode &errorCode);
774 
777  return dynamic_cast<RuleBasedCollator *>(fromUCollator(uc));
778  }
780  static inline const RuleBasedCollator *rbcFromUCollator(const UCollator *uc) {
781  return dynamic_cast<const RuleBasedCollator *>(fromUCollator(uc));
782  }
783 
788  void internalGetCEs(const UnicodeString &str, UVector64 &ces, UErrorCode &errorCode) const;
789 #endif // U_HIDE_INTERNAL_API
790 
791 protected:
799  virtual void setLocales(const Locale& requestedLocale, const Locale& validLocale, const Locale& actualLocale);
800 
801 private:
802  friend class CollationElementIterator;
803  friend class Collator;
804 
805  RuleBasedCollator(const CollationCacheEntry *entry);
806 
812  enum Attributes {
813  ATTR_VARIABLE_TOP = UCOL_ATTRIBUTE_COUNT,
814  ATTR_LIMIT
815  };
816 
817  void adoptTailoring(CollationTailoring *t, UErrorCode &errorCode);
818 
819  // Both lengths must be <0 or else both must be >=0.
820  UCollationResult doCompare(const UChar *left, int32_t leftLength,
821  const UChar *right, int32_t rightLength,
822  UErrorCode &errorCode) const;
823  UCollationResult doCompare(const uint8_t *left, int32_t leftLength,
824  const uint8_t *right, int32_t rightLength,
825  UErrorCode &errorCode) const;
826 
827  void writeSortKey(const UChar *s, int32_t length,
828  SortKeyByteSink &sink, UErrorCode &errorCode) const;
829 
830  void writeIdenticalLevel(const UChar *s, const UChar *limit,
831  SortKeyByteSink &sink, UErrorCode &errorCode) const;
832 
833  const CollationSettings &getDefaultSettings() const;
834 
835  void setAttributeDefault(int32_t attribute) {
836  explicitlySetAttributes &= ~((uint32_t)1 << attribute);
837  }
838  void setAttributeExplicitly(int32_t attribute) {
839  explicitlySetAttributes |= (uint32_t)1 << attribute;
840  }
841  UBool attributeHasBeenSetExplicitly(int32_t attribute) const {
842  // assert(0 <= attribute < ATTR_LIMIT);
843  return (UBool)((explicitlySetAttributes & ((uint32_t)1 << attribute)) != 0);
844  }
845 
853  UBool isUnsafe(UChar32 c) const;
854 
855  static void computeMaxExpansions(const CollationTailoring *t, UErrorCode &errorCode);
856  UBool initMaxExpansions(UErrorCode &errorCode) const;
857 
858  void setFastLatinOptions(CollationSettings &ownedSettings) const;
859 
860  const CollationData *data;
861  const CollationSettings *settings; // reference-counted
862  const CollationTailoring *tailoring; // alias of cacheEntry->tailoring
863  const CollationCacheEntry *cacheEntry; // reference-counted
864  Locale validLocale;
865  uint32_t explicitlySetAttributes;
866 
867  UBool actualLocaleIsSameAsValid;
868 };
869 
871 
872 #endif // !UCONFIG_NO_COLLATION
873 #endif // TBLCOLL_H
uint8_t UVersionInfo[U_MAX_VERSION_LENGTH]
The binary form of a version on ICU APIs is an array of 4 uint8_t.
Definition: uversion.h:57
virtual UBool operator==(const Collator &other) const
Returns TRUE if "other" is the same as "this".
virtual void setReorderCodes(const int32_t *reorderCodes, int32_t reorderCodesLength, UErrorCode &status)
Sets the ordering of scripts for this collator.
virtual int32_t hashCode(void) const =0
Generates the hash code for the collation object.
The Collator class performs locale-sensitive string comparison.
Definition: coll.h:163
virtual UCollationResult internalCompareUTF8(const char *left, int32_t leftLength, const char *right, int32_t rightLength, UErrorCode &errorCode) const
Implements ucol_strcollUTF8().
virtual Collator & setMaxVariable(UColReorderCode group, UErrorCode &errorCode)
Sets the variable top to the top of the specified reordering group.
UCollationResult
UCOL_LESS is returned if source string is compared to be less than target string in the ucol_strcoll(...
Definition: ucol.h:71
virtual uint32_t getVariableTop(UErrorCode &status) const =0
Gets the variable top value of a Collator.
void * UClassID
UClassID is used to identify classes without using the compiler's RTTI.
Definition: uobject.h:91
C API for code unit iteration.
Definition: uiter.h:339
virtual int32_t internalGetShortDefinitionString(const char *locale, char *buffer, int32_t capacity, UErrorCode &status) const
Get the short definition string for a collator.
ECollationStrength
Base letter represents a primary difference.
Definition: coll.h:193
static const RuleBasedCollator * rbcFromUCollator(const UCollator *uc)
Definition: tblcoll.h:780
UColReorderCode
Enum containing the codes for reordering segments of the collation table that are not script codes...
Definition: ucol.h:139
UColAttribute
Attributes that collation service understands.
Definition: ucol.h:234
#define U_I18N_API
Set to export library symbols from inside the i18n library, and to import them from outside...
Definition: utypes.h:358
virtual uint32_t setVariableTop(const UChar *varTop, int32_t len, UErrorCode &status)=0
Sets the variable top to the primary weight of the specified string.
virtual UColAttributeValue getAttribute(UColAttribute attr, UErrorCode &status) const =0
Universal attribute getter.
#define U_NAMESPACE_BEGIN
This is used to begin a declaration of a public ICU C++ API.
Definition: uversion.h:129
The RuleBasedCollator class provides the implementation of Collator, using data-driven tables...
Definition: tblcoll.h:111
C++ API: Collation Service.
Abstract class that defines an API for iteration on text objects.
Definition: chariter.h:356
virtual Locale getLocale(ULocDataLocaleType type, UErrorCode &status) const =0
Gets the locale of the Collator.
virtual Collator * clone(void) const =0
Makes a copy of this object.
virtual UClassID getDynamicClassID(void) const =0
Returns a unique class ID POLYMORPHICALLY.
The CollationElementIterator class is used as an iterator to walk through each character of an intern...
Definition: coleitr.h:116
int32_t UChar32
Define UChar32 as a type for single Unicode code points.
Definition: umachine.h:332
virtual void getVersion(UVersionInfo info) const =0
Gets the version information for a Collator.
virtual void setLocales(const Locale &requestedLocale, const Locale &validLocale, const Locale &actualLocale)
Used internally by registration to define the requested and valid locales.
C API: Collator.
Collation keys are generated by the Collator class.
Definition: sortkey.h:97
virtual UColReorderCode getMaxVariable() const
Returns the maximum reordering group whose characters are affected by UCOL_ALTERNATE_HANDLING.
virtual int32_t getSortKey(const UnicodeString &source, uint8_t *result, int32_t resultLength) const =0
Get the sort key as an array of bytes from a UnicodeString.
virtual CollationKey & getCollationKey(const UnicodeString &source, CollationKey &key, UErrorCode &status) const =0
Transforms the string into a series of characters that can be compared with CollationKey::compareTo.
virtual EComparisonResult compare(const UnicodeString &source, const UnicodeString &target) const
The comparison function compares the character data stored in two different strings.
A mutable set of Unicode characters and multicharacter strings.
Definition: uniset.h:276
C API: Unicode Character Iteration.
virtual int32_t getReorderCodes(int32_t *dest, int32_t destCapacity, UErrorCode &status) const
Retrieves the reordering codes for this collator.
uint16_t UChar
Define UChar to be UCHAR_TYPE, if that is #defined (for example, to char16_t), or wchar_t if that is ...
Definition: umachine.h:312
#define U_NAMESPACE_END
This is used to end a declaration of a public ICU C++ API.
Definition: uversion.h:130
virtual UnicodeSet * getTailoredSet(UErrorCode &status) const
Get a UnicodeSet that contains all the characters and sequences tailored in this collator.
UErrorCode
Error code to replace exception handling, so that the code is compatible with all C++ compilers...
Definition: utypes.h:476
ULocDataLocaleType
Constants for *_getLocale() Allow user to select whether she wants information on requested...
Definition: uloc.h:336
static RuleBasedCollator * rbcFromUCollator(UCollator *uc)
Definition: tblcoll.h:776
C++ API: Locale ID object.
struct UCollator UCollator
structure representing a collator object instance
Definition: ucol.h:56
virtual void setAttribute(UColAttribute attr, UColAttributeValue value, UErrorCode &status)=0
Universal attribute setter.
A UParseError struct is used to returned detailed information about parsing errors.
Definition: parseerr.h:56
Basic definitions for ICU, for both C and C++ APIs.
static Collator * fromUCollator(UCollator *uc)
Definition: coll.h:1159
virtual int32_t internalNextSortKeyPart(UCharIterator *iter, uint32_t state[2], uint8_t *dest, int32_t count, UErrorCode &errorCode) const
Implements ucol_nextSortKeyPart().
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
Definition: unistr.h:245
UColAttributeValue
Enum containing attribute values for controling collation behavior.
Definition: ucol.h:87
A string-like object that points to a sized piece of memory.
Definition: stringpiece.h:52
UColRuleOption
Options for retrieving the rule string.
Definition: ucol.h:346
The number of UColAttribute constants.
Definition: ucol.h:340
int8_t UBool
The ICU boolean type.
Definition: umachine.h:234
virtual UCollationResult compareUTF8(const StringPiece &source, const StringPiece &target, UErrorCode &status) const
Compares two UTF-8 strings using the Collator.
A Locale object represents a specific geographical, political, or cultural region.
Definition: locid.h:185