ICU 72.1 72.1
|
The RuleBasedCollator class provides the implementation of Collator, using data-driven tables. More...
#include <tblcoll.h>
Public Member Functions | |
RuleBasedCollator (const UnicodeString &rules, UErrorCode &status) | |
RuleBasedCollator constructor. More... | |
RuleBasedCollator (const UnicodeString &rules, ECollationStrength collationStrength, UErrorCode &status) | |
RuleBasedCollator constructor. More... | |
RuleBasedCollator (const UnicodeString &rules, UColAttributeValue decompositionMode, UErrorCode &status) | |
RuleBasedCollator constructor. More... | |
RuleBasedCollator (const UnicodeString &rules, ECollationStrength collationStrength, UColAttributeValue decompositionMode, UErrorCode &status) | |
RuleBasedCollator constructor. More... | |
RuleBasedCollator (const UnicodeString &rules, UParseError &parseError, UnicodeString &reason, UErrorCode &errorCode) | |
TODO: document & propose as public API. More... | |
RuleBasedCollator (const RuleBasedCollator &other) | |
Copy constructor. More... | |
RuleBasedCollator (const uint8_t *bin, int32_t length, const RuleBasedCollator *base, UErrorCode &status) | |
Opens a collator from a collator binary image created using cloneBinary. More... | |
virtual | ~RuleBasedCollator () |
Destructor. More... | |
RuleBasedCollator & | operator= (const RuleBasedCollator &other) |
Assignment operator. More... | |
virtual bool | operator== (const Collator &other) const override |
Returns true if argument is the same as this object. More... | |
virtual RuleBasedCollator * | clone () const override |
Makes a copy of this object. More... | |
virtual CollationElementIterator * | createCollationElementIterator (const UnicodeString &source) const |
Creates a collation element iterator for the source string. More... | |
virtual CollationElementIterator * | createCollationElementIterator (const CharacterIterator &source) const |
Creates a collation element iterator for the source. More... | |
virtual UCollationResult | compare (const UnicodeString &source, const UnicodeString &target, UErrorCode &status) const override |
The comparison function compares the character data stored in two different strings. More... | |
virtual UCollationResult | compare (const UnicodeString &source, const UnicodeString &target, int32_t length, UErrorCode &status) const override |
Does the same thing as compare but limits the comparison to a specified length. More... | |
virtual UCollationResult | compare (const char16_t *source, int32_t sourceLength, const char16_t *target, int32_t targetLength, UErrorCode &status) const override |
The comparison function compares the character data stored in two different string arrays. More... | |
virtual UCollationResult | compare (UCharIterator &sIter, UCharIterator &tIter, UErrorCode &status) const override |
Compares two strings using the Collator. More... | |
virtual UCollationResult | compareUTF8 (const StringPiece &source, const StringPiece &target, UErrorCode &status) const override |
Compares two UTF-8 strings using the Collator. More... | |
virtual CollationKey & | getCollationKey (const UnicodeString &source, CollationKey &key, UErrorCode &status) const override |
Transforms the string into a series of characters that can be compared with CollationKey.compare(). More... | |
virtual CollationKey & | getCollationKey (const char16_t *source, int32_t sourceLength, CollationKey &key, UErrorCode &status) const override |
Transforms a specified region of the string into a series of characters that can be compared with CollationKey.compare. More... | |
virtual int32_t | hashCode () const override |
Generates the hash code for the rule-based collation object. More... | |
virtual Locale | getLocale (ULocDataLocaleType type, UErrorCode &status) const override |
Gets the locale of the Collator. More... | |
const UnicodeString & | getRules () const |
Gets the tailoring rules for this collator. More... | |
virtual void | getVersion (UVersionInfo info) const override |
Gets the version information for a Collator. More... | |
int32_t | getMaxExpansion (int32_t order) const |
Returns the maximum length of any expansion sequences that end with the specified comparison order. More... | |
virtual UClassID | getDynamicClassID (void) const override |
Returns a unique class ID POLYMORPHICALLY. More... | |
uint8_t * | cloneRuleData (int32_t &length, UErrorCode &status) const |
Do not use this method: The caller and the ICU library might use different heaps. More... | |
int32_t | cloneBinary (uint8_t *buffer, int32_t capacity, UErrorCode &status) const |
Creates a binary image of a collator. More... | |
void | getRules (UColRuleOption delta, UnicodeString &buffer) const |
Returns current rules. More... | |
virtual void | setAttribute (UColAttribute attr, UColAttributeValue value, UErrorCode &status) override |
Universal attribute setter. More... | |
virtual UColAttributeValue | getAttribute (UColAttribute attr, UErrorCode &status) const override |
Universal attribute getter. More... | |
virtual Collator & | setMaxVariable (UColReorderCode group, UErrorCode &errorCode) override |
Sets the variable top to the top of the specified reordering group. More... | |
virtual UColReorderCode | getMaxVariable () const override |
Returns the maximum reordering group whose characters are affected by UCOL_ALTERNATE_HANDLING. More... | |
virtual uint32_t | setVariableTop (const char16_t *varTop, int32_t len, UErrorCode &status) override |
Sets the variable top to the primary weight of the specified string. More... | |
virtual uint32_t | setVariableTop (const UnicodeString &varTop, UErrorCode &status) override |
Sets the variable top to the primary weight of the specified string. More... | |
virtual void | setVariableTop (uint32_t varTop, UErrorCode &status) override |
Sets the variable top to the specified primary weight. More... | |
virtual uint32_t | getVariableTop (UErrorCode &status) const override |
Gets the variable top value of a Collator. More... | |
virtual UnicodeSet * | getTailoredSet (UErrorCode &status) const override |
Get a UnicodeSet that contains all the characters and sequences tailored in this collator. More... | |
virtual int32_t | getSortKey (const UnicodeString &source, uint8_t *result, int32_t resultLength) const override |
Get the sort key as an array of bytes from a UnicodeString. More... | |
virtual int32_t | getSortKey (const char16_t *source, int32_t sourceLength, uint8_t *result, int32_t resultLength) const override |
Get the sort key as an array of bytes from a char16_t buffer. More... | |
virtual int32_t | getReorderCodes (int32_t *dest, int32_t destCapacity, UErrorCode &status) const override |
Retrieves the reordering codes for this collator. More... | |
virtual void | setReorderCodes (const int32_t *reorderCodes, int32_t reorderCodesLength, UErrorCode &status) override |
Sets the ordering of scripts for this collator. More... | |
virtual UCollationResult | internalCompareUTF8 (const char *left, int32_t leftLength, const char *right, int32_t rightLength, UErrorCode &errorCode) const override |
Implements ucol_strcollUTF8(). More... | |
virtual int32_t | internalGetShortDefinitionString (const char *locale, char *buffer, int32_t capacity, UErrorCode &status) const override |
Get the short definition string for a collator. More... | |
virtual int32_t | internalNextSortKeyPart (UCharIterator *iter, uint32_t state[2], uint8_t *dest, int32_t count, UErrorCode &errorCode) const override |
Implements ucol_nextSortKeyPart(). More... | |
RuleBasedCollator () | |
Only for use in ucol_openRules(). More... | |
const char * | internalGetLocaleID (ULocDataLocaleType type, UErrorCode &errorCode) const |
Implements ucol_getLocaleByType(). More... | |
void | internalGetContractionsAndExpansions (UnicodeSet *contractions, UnicodeSet *expansions, UBool addPrefixes, UErrorCode &errorCode) const |
Implements ucol_getContractionsAndExpansions(). More... | |
void | internalAddContractions (UChar32 c, UnicodeSet &set, UErrorCode &errorCode) const |
Adds the contractions that start with character c to the set. More... | |
void | internalBuildTailoring (const UnicodeString &rules, int32_t strength, UColAttributeValue decompositionMode, UParseError *outParseError, UnicodeString *outReason, UErrorCode &errorCode) |
Implements from-rule constructors, and ucol_openRules(). More... | |
void | internalGetCEs (const UnicodeString &str, UVector64 &ces, UErrorCode &errorCode) const |
Appends the CEs for the string to the vector. More... | |
virtual EComparisonResult | compare (const UnicodeString &source, const UnicodeString &target) const |
The comparison function compares the character data stored in two different strings. More... | |
virtual UCollationResult | compare (const UnicodeString &source, const UnicodeString &target, UErrorCode &status) const=0 |
The comparison function compares the character data stored in two different strings. More... | |
virtual EComparisonResult | compare (const UnicodeString &source, const UnicodeString &target, int32_t length) const |
Does the same thing as compare but limits the comparison to a specified length. More... | |
virtual UCollationResult | compare (const UnicodeString &source, const UnicodeString &target, int32_t length, UErrorCode &status) const=0 |
Does the same thing as compare but limits the comparison to a specified length. More... | |
virtual EComparisonResult | compare (const char16_t *source, int32_t sourceLength, const char16_t *target, int32_t targetLength) const |
The comparison function compares the character data stored in two different string arrays. More... | |
virtual UCollationResult | compare (const char16_t *source, int32_t sourceLength, const char16_t *target, int32_t targetLength, UErrorCode &status) const=0 |
The comparison function compares the character data stored in two different string arrays. More... | |
virtual UCollationResult | compare (UCharIterator &sIter, UCharIterator &tIter, UErrorCode &status) const |
Compares two strings using the Collator. More... | |
Public Member Functions inherited from icu::Collator | |
virtual | ~Collator () |
Destructor. More... | |
virtual bool | operator== (const Collator &other) const |
Returns true if "other" is the same as "this". More... | |
virtual bool | operator!= (const Collator &other) const |
Returns true if "other" is not the same as "this". More... | |
virtual Collator * | clone () const =0 |
Makes a copy of this object. More... | |
virtual EComparisonResult | compare (const UnicodeString &source, const UnicodeString &target) const |
The comparison function compares the character data stored in two different strings. More... | |
virtual UCollationResult | compare (const UnicodeString &source, const UnicodeString &target, UErrorCode &status) const =0 |
The comparison function compares the character data stored in two different strings. More... | |
virtual EComparisonResult | compare (const UnicodeString &source, const UnicodeString &target, int32_t length) const |
Does the same thing as compare but limits the comparison to a specified length. More... | |
virtual UCollationResult | compare (const UnicodeString &source, const UnicodeString &target, int32_t length, UErrorCode &status) const =0 |
Does the same thing as compare but limits the comparison to a specified length. More... | |
virtual EComparisonResult | compare (const char16_t *source, int32_t sourceLength, const char16_t *target, int32_t targetLength) const |
The comparison function compares the character data stored in two different string arrays. More... | |
virtual UCollationResult | compare (const char16_t *source, int32_t sourceLength, const char16_t *target, int32_t targetLength, UErrorCode &status) const =0 |
The comparison function compares the character data stored in two different string arrays. More... | |
virtual UCollationResult | compare (UCharIterator &sIter, UCharIterator &tIter, UErrorCode &status) const |
Compares two strings using the Collator. More... | |
virtual UCollationResult | compareUTF8 (const StringPiece &source, const StringPiece &target, UErrorCode &status) const |
Compares two UTF-8 strings using the Collator. More... | |
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. More... | |
virtual CollationKey & | getCollationKey (const char16_t *source, int32_t sourceLength, CollationKey &key, UErrorCode &status) const =0 |
Transforms the string into a series of characters that can be compared with CollationKey::compareTo. More... | |
virtual int32_t | hashCode (void) const =0 |
Generates the hash code for the collation object. More... | |
virtual Locale | getLocale (ULocDataLocaleType type, UErrorCode &status) const =0 |
Gets the locale of the Collator. More... | |
UBool | greater (const UnicodeString &source, const UnicodeString &target) const |
Convenience method for comparing two strings based on the collation rules. More... | |
UBool | greaterOrEqual (const UnicodeString &source, const UnicodeString &target) const |
Convenience method for comparing two strings based on the collation rules. More... | |
UBool | equals (const UnicodeString &source, const UnicodeString &target) const |
Convenience method for comparing two strings based on the collation rules. More... | |
virtual ECollationStrength | getStrength (void) const |
Determines the minimum strength that will be used in comparison or transformation. More... | |
virtual void | setStrength (ECollationStrength newStrength) |
Sets the minimum strength to be used in comparison or transformation. More... | |
virtual int32_t | getReorderCodes (int32_t *dest, int32_t destCapacity, UErrorCode &status) const |
Retrieves the reordering codes for this collator. More... | |
virtual void | setReorderCodes (const int32_t *reorderCodes, int32_t reorderCodesLength, UErrorCode &status) |
Sets the ordering of scripts for this collator. More... | |
virtual void | getVersion (UVersionInfo info) const =0 |
Gets the version information for a Collator. More... | |
virtual UClassID | getDynamicClassID (void) const override=0 |
Returns a unique class ID POLYMORPHICALLY. More... | |
virtual void | setAttribute (UColAttribute attr, UColAttributeValue value, UErrorCode &status)=0 |
Universal attribute setter. More... | |
virtual UColAttributeValue | getAttribute (UColAttribute attr, UErrorCode &status) const =0 |
Universal attribute getter. More... | |
virtual Collator & | setMaxVariable (UColReorderCode group, UErrorCode &errorCode) |
Sets the variable top to the top of the specified reordering group. More... | |
virtual UColReorderCode | getMaxVariable () const |
Returns the maximum reordering group whose characters are affected by UCOL_ALTERNATE_HANDLING. More... | |
virtual uint32_t | setVariableTop (const char16_t *varTop, int32_t len, UErrorCode &status)=0 |
Sets the variable top to the primary weight of the specified string. More... | |
virtual uint32_t | setVariableTop (const UnicodeString &varTop, UErrorCode &status)=0 |
Sets the variable top to the primary weight of the specified string. More... | |
virtual void | setVariableTop (uint32_t varTop, UErrorCode &status)=0 |
Sets the variable top to the specified primary weight. More... | |
virtual uint32_t | getVariableTop (UErrorCode &status) const =0 |
Gets the variable top value of a Collator. More... | |
virtual UnicodeSet * | getTailoredSet (UErrorCode &status) const |
Get a UnicodeSet that contains all the characters and sequences tailored in this collator. More... | |
virtual Collator * | safeClone () const |
Same as clone(). More... | |
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. More... | |
virtual int32_t | getSortKey (const char16_t *source, int32_t sourceLength, uint8_t *result, int32_t resultLength) const =0 |
Get the sort key as an array of bytes from a char16_t buffer. More... | |
virtual void | setLocales (const Locale &requestedLocale, const Locale &validLocale, const Locale &actualLocale) |
Used internally by registration to define the requested and valid locales. More... | |
virtual int32_t | internalGetShortDefinitionString (const char *locale, char *buffer, int32_t capacity, UErrorCode &status) const |
Get the short definition string for a collator. More... | |
virtual UCollationResult | internalCompareUTF8 (const char *left, int32_t leftLength, const char *right, int32_t rightLength, UErrorCode &errorCode) const |
Implements ucol_strcollUTF8(). More... | |
virtual int32_t | internalNextSortKeyPart (UCharIterator *iter, uint32_t state[2], uint8_t *dest, int32_t count, UErrorCode &errorCode) const |
Implements ucol_nextSortKeyPart(). More... | |
UCollator * | toUCollator () |
const UCollator * | toUCollator () const |
Public Member Functions inherited from icu::UObject | |
virtual | ~UObject () |
Destructor. More... | |
virtual UClassID | getDynamicClassID () const |
ICU4C "poor man's RTTI", returns a UClassID for the actual ICU class. More... | |
Static Public Member Functions | |
static UClassID | getStaticClassID (void) |
Returns the class ID for this class. More... | |
static RuleBasedCollator * | rbcFromUCollator (UCollator *uc) |
static const RuleBasedCollator * | rbcFromUCollator (const UCollator *uc) |
Static Public Member Functions inherited from icu::Collator | |
static Collator * | createInstance (UErrorCode &err) |
Creates the Collator object for the current default locale. More... | |
static Collator * | createInstance (const Locale &loc, UErrorCode &err) |
Gets the collation object for the desired locale. More... | |
static int32_t | getEquivalentReorderCodes (int32_t reorderCode, int32_t *dest, int32_t destCapacity, UErrorCode &status) |
Retrieves the reorder codes that are grouped with the given reorder code. More... | |
static UnicodeString & | getDisplayName (const Locale &objectLocale, const Locale &displayLocale, UnicodeString &name) |
Get name of the object for the desired Locale, in the desired language. More... | |
static UnicodeString & | getDisplayName (const Locale &objectLocale, UnicodeString &name) |
Get name of the object for the desired Locale, in the language of the default locale. More... | |
static const Locale * | getAvailableLocales (int32_t &count) |
Get the set of Locales for which Collations are installed. More... | |
static StringEnumeration * | getAvailableLocales (void) |
Return a StringEnumeration over the locales available at the time of the call, including registered locales. More... | |
static StringEnumeration * | getKeywords (UErrorCode &status) |
Create a string enumerator of all possible keywords that are relevant to collation. More... | |
static StringEnumeration * | getKeywordValues (const char *keyword, UErrorCode &status) |
Given a keyword, create a string enumeration of all values for that keyword that are currently in use. More... | |
static StringEnumeration * | getKeywordValuesForLocale (const char *keyword, const Locale &locale, UBool commonlyUsed, UErrorCode &status) |
Given a key and a locale, returns an array of string values in a preferred order that would make a difference. More... | |
static Locale | getFunctionalEquivalent (const char *keyword, const Locale &locale, UBool &isAvailable, UErrorCode &status) |
Return the functionally equivalent locale for the given requested locale, with respect to given keyword, for the collation service. More... | |
static URegistryKey | registerInstance (Collator *toAdopt, const Locale &locale, UErrorCode &status) |
Register a new Collator. More... | |
static URegistryKey | registerFactory (CollatorFactory *toAdopt, UErrorCode &status) |
Register a new CollatorFactory. More... | |
static UBool | unregister (URegistryKey key, UErrorCode &status) |
Unregister a previously-registered Collator or CollatorFactory using the key returned from the register call. More... | |
static int32_t | getBound (const uint8_t *source, int32_t sourceLength, UColBoundMode boundType, uint32_t noOfLevels, uint8_t *result, int32_t resultLength, UErrorCode &status) |
Produce a bound for a given sortkey and a number of levels. More... | |
static Collator * | fromUCollator (UCollator *uc) |
static const Collator * | fromUCollator (const UCollator *uc) |
Protected Member Functions | |
virtual void | setLocales (const Locale &requestedLocale, const Locale &validLocale, const Locale &actualLocale) override |
Used internally by registration to define the requested and valid locales. More... | |
Protected Member Functions inherited from icu::Collator | |
Collator () | |
Default constructor. More... | |
Collator (UCollationStrength collationStrength, UNormalizationMode decompositionMode) | |
Constructor. More... | |
Collator (const Collator &other) | |
Copy constructor. More... | |
Friends | |
class | CollationElementIterator |
class | Collator |
Additional Inherited Members | |
Public Types inherited from icu::Collator | |
enum | ECollationStrength { PRIMARY = UCOL_PRIMARY , SECONDARY = UCOL_SECONDARY , TERTIARY = UCOL_TERTIARY , QUATERNARY = UCOL_QUATERNARY , IDENTICAL = UCOL_IDENTICAL } |
Base letter represents a primary difference. More... | |
enum | EComparisonResult { LESS = UCOL_LESS , EQUAL = UCOL_EQUAL , GREATER = UCOL_GREATER } |
LESS is returned if source string is compared to be less than target string in the compare() method. More... | |
The RuleBasedCollator class provides the implementation of Collator, using data-driven tables.
The user can create a customized table-based collation.
For more information about the collation service see the User Guide.
Collation service provides correct sorting orders for most locales supported in ICU. If specific data for a locale is not available, the orders eventually falls back to the CLDR root sort order.
Sort ordering may be customized by providing your own set of rules. For more on this subject see the Collation Customization section of the User Guide.
Note, RuleBasedCollator is not to be subclassed.
icu::RuleBasedCollator::RuleBasedCollator | ( | const UnicodeString & | rules, |
UErrorCode & | status | ||
) |
RuleBasedCollator constructor.
This takes the table rules and builds a collation table out of them. Please see RuleBasedCollator class description for more details on the collation rule syntax.
rules | the collation rules to build the collation table from. |
status | reporting a success or an error. |
icu::RuleBasedCollator::RuleBasedCollator | ( | const UnicodeString & | rules, |
ECollationStrength | collationStrength, | ||
UErrorCode & | status | ||
) |
RuleBasedCollator constructor.
This takes the table rules and builds a collation table out of them. Please see RuleBasedCollator class description for more details on the collation rule syntax.
rules | the collation rules to build the collation table from. |
collationStrength | strength for comparison |
status | reporting a success or an error. |
icu::RuleBasedCollator::RuleBasedCollator | ( | const UnicodeString & | rules, |
UColAttributeValue | decompositionMode, | ||
UErrorCode & | status | ||
) |
RuleBasedCollator constructor.
This takes the table rules and builds a collation table out of them. Please see RuleBasedCollator class description for more details on the collation rule syntax.
rules | the collation rules to build the collation table from. |
decompositionMode | the normalisation mode |
status | reporting a success or an error. |
icu::RuleBasedCollator::RuleBasedCollator | ( | const UnicodeString & | rules, |
ECollationStrength | collationStrength, | ||
UColAttributeValue | decompositionMode, | ||
UErrorCode & | status | ||
) |
RuleBasedCollator constructor.
This takes the table rules and builds a collation table out of them. Please see RuleBasedCollator class description for more details on the collation rule syntax.
rules | the collation rules to build the collation table from. |
collationStrength | strength for comparison |
decompositionMode | the normalisation mode |
status | reporting a success or an error. |
icu::RuleBasedCollator::RuleBasedCollator | ( | const UnicodeString & | rules, |
UParseError & | parseError, | ||
UnicodeString & | reason, | ||
UErrorCode & | errorCode | ||
) |
TODO: document & propose as public API.
icu::RuleBasedCollator::RuleBasedCollator | ( | const RuleBasedCollator & | other | ) |
icu::RuleBasedCollator::RuleBasedCollator | ( | const uint8_t * | bin, |
int32_t | length, | ||
const RuleBasedCollator * | base, | ||
UErrorCode & | status | ||
) |
Opens a collator from a collator binary image created using cloneBinary.
Binary image used in instantiation of the collator remains owned by the user and should stay around for the lifetime of the collator. The API also takes a base collator which must be the root collator.
bin | binary image owned by the user and required through the lifetime of the collator |
length | size of the image. If negative, the API will try to figure out the length of the image |
base | Base collator, for lookup of untailored characters. Must be the root collator, must not be NULL. The base is required to be present through the lifetime of the collator. |
status | for catching errors |
|
virtual |
Destructor.
icu::RuleBasedCollator::RuleBasedCollator | ( | ) |
Only for use in ucol_openRules().
|
overridevirtual |
Makes a copy of this object.
Implements icu::Collator.
int32_t icu::RuleBasedCollator::cloneBinary | ( | uint8_t * | buffer, |
int32_t | capacity, | ||
UErrorCode & | status | ||
) | const |
Creates a binary image of a collator.
This binary image can be stored and later used to instantiate a collator using ucol_openBinary. This API supports preflighting.
buffer | a fill-in buffer to receive the binary image |
capacity | capacity of the destination buffer |
status | for catching errors |
uint8_t * icu::RuleBasedCollator::cloneRuleData | ( | int32_t & | length, |
UErrorCode & | status | ||
) | const |
Do not use this method: The caller and the ICU library might use different heaps.
Use cloneBinary() instead which writes to caller-provided memory.
Returns a binary format of this collator.
length | Returns the length of the data, in bytes |
status | the error code status. |
|
virtual |
The comparison function compares the character data stored in two different string arrays.
Returns information about whether a string array is less than, greater than or equal to another string array.
Example of use:
. char16_t ABC[] = {0x41, 0x42, 0x43, 0}; // = "ABC" . char16_t abc[] = {0x61, 0x62, 0x63, 0}; // = "abc" . UErrorCode status = U_ZERO_ERROR; . Collator *myCollation = . Collator::createInstance(Locale::getUS(), status); . if (U_FAILURE(status)) return; . myCollation->setStrength(Collator::PRIMARY); . // result would be Collator::EQUAL ("abc" == "ABC") . // (no primary difference between "abc" and "ABC") . Collator::EComparisonResult result = . myCollation->compare(abc, 3, ABC, 3); . myCollation->setStrength(Collator::TERTIARY); . // result would be Collator::LESS ("abc" <<< "ABC") . // (with tertiary difference between "abc" and "ABC") . result = myCollation->compare(abc, 3, ABC, 3);
source | the source string array to be compared with. |
sourceLength | the length of the source string array. If this value is equal to -1, the string array is null-terminated. |
target | the string that is to be compared with the source string. |
targetLength | the length of the target string array. If this value is equal to -1, the string array is null-terminated. |
Reimplemented from icu::Collator.
|
overridevirtual |
The comparison function compares the character data stored in two different string arrays.
Returns information about whether a string array is less than, greater than or equal to another string array.
source | the source string array to be compared with. |
sourceLength | the length of the source string array. If this value is equal to -1, the string array is null-terminated. |
target | the string that is to be compared with the source string. |
targetLength | the length of the target string array. If this value is equal to -1, the string array is null-terminated. |
status | possible error code |
Implements icu::Collator.
|
virtual |
The comparison function compares the character data stored in two different string arrays.
Returns information about whether a string array is less than, greater than or equal to another string array.
source | the source string array to be compared with. |
sourceLength | the length of the source string array. If this value is equal to -1, the string array is null-terminated. |
target | the string that is to be compared with the source string. |
targetLength | the length of the target string array. If this value is equal to -1, the string array is null-terminated. |
status | possible error code |
Implements icu::Collator.
|
virtual |
The comparison function compares the character data stored in two different strings.
Returns information about whether a string is less than, greater than or equal to another string.
source | the source string to be compared with. |
target | the string that is to be compared with the source string. |
Reimplemented from icu::Collator.
|
virtual |
Does the same thing as compare but limits the comparison to a specified length.
source | the source string to be compared with. |
target | the string that is to be compared with the source string. |
length | the length the comparison is limited to |
Reimplemented from icu::Collator.
|
overridevirtual |
Does the same thing as compare but limits the comparison to a specified length.
source | the source string to be compared with. |
target | the string that is to be compared with the source string. |
length | the length the comparison is limited to |
status | possible error code |
Implements icu::Collator.
|
virtual |
Does the same thing as compare but limits the comparison to a specified length.
source | the source string to be compared with. |
target | the string that is to be compared with the source string. |
length | the length the comparison is limited to |
status | possible error code |
Implements icu::Collator.
|
overridevirtual |
The comparison function compares the character data stored in two different strings.
Returns information about whether a string is less than, greater than or equal to another string.
source | the source string to be compared with. |
target | the string that is to be compared with the source string. |
status | possible error code |
Implements icu::Collator.
|
virtual |
The comparison function compares the character data stored in two different strings.
Returns information about whether a string is less than, greater than or equal to another string.
source | the source string to be compared with. |
target | the string that is to be compared with the source string. |
status | possible error code |
Implements icu::Collator.
|
virtual |
Compares two strings using the Collator.
Returns whether the first one compares less than/equal to/greater than the second one. This version takes UCharIterator input.
sIter | the first ("source") string iterator |
tIter | the second ("target") string iterator |
status | ICU status |
Reimplemented from icu::Collator.
|
overridevirtual |
Compares two strings using the Collator.
Returns whether the first one compares less than/equal to/greater than the second one. This version takes UCharIterator input.
sIter | the first ("source") string iterator |
tIter | the second ("target") string iterator |
status | ICU status |
Reimplemented from icu::Collator.
|
overridevirtual |
Compares two UTF-8 strings using the Collator.
Returns whether the first one compares less than/equal to/greater than the second one. This version takes UTF-8 input. Note that a StringPiece can be implicitly constructed from a std::string or a NUL-terminated const char * string.
source | the first UTF-8 string |
target | the second UTF-8 string |
status | ICU status |
Reimplemented from icu::Collator.
|
virtual |
Creates a collation element iterator for the source.
The caller of this method is responsible for the memory management of the returned pointer.
source | the CharacterIterator which produces the characters over which the CollationElementItgerator will iterate. |
|
virtual |
Creates a collation element iterator for the source string.
The caller of this method is responsible for the memory management of the return pointer.
source | the string over which the CollationElementIterator will iterate. |
|
overridevirtual |
Universal attribute getter.
attr | attribute type |
status | to indicate whether the operation went on smoothly or there were errors |
Implements icu::Collator.
|
overridevirtual |
Transforms a specified region of the string into a series of characters that can be compared with CollationKey.compare.
Note that sort keys are often less efficient than simply doing comparison. For more details, see the ICU User Guide.
source | the source string. |
sourceLength | the length of the source string. |
key | the transformed key of the source string. |
status | the error code status. |
Implements icu::Collator.
|
overridevirtual |
Transforms the string into a series of characters that can be compared with CollationKey.compare().
Note that sort keys are often less efficient than simply doing comparison. For more details, see the ICU User Guide.
source | the source string. |
key | the transformed key of the source string. |
status | the error code status. |
Implements icu::Collator.
|
overridevirtual |
Returns a unique class ID POLYMORPHICALLY.
Pure virtual override. This method is to implement a simple version of RTTI, since not all C++ compilers support genuine RTTI. Polymorphic operator==() and clone() methods call this method.
Implements icu::Collator.
|
overridevirtual |
Gets the locale of the Collator.
type | can be either requested, valid or actual locale. For more information see the definition of ULocDataLocaleType in uloc.h |
status | the error code status. |
Implements icu::Collator.
int32_t icu::RuleBasedCollator::getMaxExpansion | ( | int32_t | order | ) | const |
Returns the maximum length of any expansion sequences that end with the specified comparison order.
This is specific to the kind of collation element values and sequences returned by the CollationElementIterator. Call CollationElementIterator::getMaxExpansion() instead.
order | a collation order returned by CollationElementIterator::previous or CollationElementIterator::next. |
|
overridevirtual |
Returns the maximum reordering group whose characters are affected by UCOL_ALTERNATE_HANDLING.
Reimplemented from icu::Collator.
|
overridevirtual |
Retrieves the reordering codes for this collator.
dest | The array to fill with the script ordering. |
destCapacity | The length of dest. If it is 0, then dest may be NULL and the function will only return the length of the result without writing any codes (pre-flighting). |
status | A reference to an error code value, which must not indicate a failure before the function call. |
Reimplemented from icu::Collator.
const UnicodeString & icu::RuleBasedCollator::getRules | ( | ) | const |
Gets the tailoring rules for this collator.
void icu::RuleBasedCollator::getRules | ( | UColRuleOption | delta, |
UnicodeString & | buffer | ||
) | const |
Returns current rules.
Delta defines whether full rules are returned or just the tailoring.
getRules(void) should normally be used instead. See https://unicode-org.github.io/icu/userguide/collation/customization#building-on-existing-locales
delta | one of UCOL_TAILORING_ONLY, UCOL_FULL_RULES. |
buffer | UnicodeString to store the result rules |
|
overridevirtual |
Get the sort key as an array of bytes from a char16_t buffer.
Note that sort keys are often less efficient than simply doing comparison. For more details, see the ICU User Guide.
source | string to be processed. |
sourceLength | length of string to be processed. If -1, the string is 0 terminated and length will be decided by the function. |
result | buffer to store result in. If NULL, number of bytes needed will be returned. |
resultLength | length of the result buffer. If if not enough the buffer will be filled to capacity. |
Implements icu::Collator.
|
overridevirtual |
Get the sort key as an array of bytes from a UnicodeString.
Note that sort keys are often less efficient than simply doing comparison. For more details, see the ICU User Guide.
source | string to be processed. |
result | buffer to store result in. If NULL, number of bytes needed will be returned. |
resultLength | length of the result buffer. If if not enough the buffer will be filled to capacity. |
Implements icu::Collator.
|
static |
Returns the class ID for this class.
This is useful only for comparing to a return value from getDynamicClassID(). For example:
Base* polymorphic_pointer = createPolymorphicObject(); if (polymorphic_pointer->getDynamicClassID() == Derived::getStaticClassID()) ...
|
overridevirtual |
Get a UnicodeSet that contains all the characters and sequences tailored in this collator.
status | error code of the operation |
Reimplemented from icu::Collator.
|
overridevirtual |
Gets the variable top value of a Collator.
status | error code (not changed by function). If error code is set, the return value is undefined. |
Implements icu::Collator.
|
overridevirtual |
Gets the version information for a Collator.
info | the version # information, the result will be filled in |
Implements icu::Collator.
|
overridevirtual |
Generates the hash code for the rule-based collation object.
Implements icu::Collator.
void icu::RuleBasedCollator::internalAddContractions | ( | UChar32 | c, |
UnicodeSet & | set, | ||
UErrorCode & | errorCode | ||
) | const |
Adds the contractions that start with character c to the set.
Ignores prefixes. Used by AlphabeticIndex.
void icu::RuleBasedCollator::internalBuildTailoring | ( | const UnicodeString & | rules, |
int32_t | strength, | ||
UColAttributeValue | decompositionMode, | ||
UParseError * | outParseError, | ||
UnicodeString * | outReason, | ||
UErrorCode & | errorCode | ||
) |
Implements from-rule constructors, and ucol_openRules().
|
overridevirtual |
Implements ucol_strcollUTF8().
Reimplemented from icu::Collator.
void icu::RuleBasedCollator::internalGetCEs | ( | const UnicodeString & | str, |
UVector64 & | ces, | ||
UErrorCode & | errorCode | ||
) | const |
Appends the CEs for the string to the vector.
void icu::RuleBasedCollator::internalGetContractionsAndExpansions | ( | UnicodeSet * | contractions, |
UnicodeSet * | expansions, | ||
UBool | addPrefixes, | ||
UErrorCode & | errorCode | ||
) | const |
Implements ucol_getContractionsAndExpansions().
Gets this collator's sets of contraction strings and/or characters and strings that map to multiple collation elements (expansions). If addPrefixes is true, then contractions that are expressed as prefix/pre-context rules are included.
contractions | if not NULL, the set to hold the contractions |
expansions | if not NULL, the set to hold the expansions |
addPrefixes | include prefix contextual mappings |
errorCode | in/out ICU error code |
const char * icu::RuleBasedCollator::internalGetLocaleID | ( | ULocDataLocaleType | type, |
UErrorCode & | errorCode | ||
) | const |
Implements ucol_getLocaleByType().
Needed because the lifetime of the locale ID string must match that of the collator. getLocale() returns a copy of a Locale, with minimal lifetime in a C wrapper.
|
overridevirtual |
Get the short definition string for a collator.
This internal API harvests the collator's locale and the attribute set and produces a string that can be used for opening a collator with the same attributes using the ucol_openFromShortString API. This string will be normalized. The structure and the syntax of the string is defined in the "Naming collators" section of the users guide: https://unicode-org.github.io/icu/userguide/collation/concepts#collator-naming-scheme This function supports preflighting.
This is internal, and intended to be used with delegate converters.
locale | a locale that will appear as a collators locale in the resulting short string definition. If NULL, the locale will be harvested from the collator. |
buffer | space to hold the resulting string |
capacity | capacity of the buffer |
status | for returning errors. All the preflighting errors are featured |
Reimplemented from icu::Collator.
|
overridevirtual |
Implements ucol_nextSortKeyPart().
Reimplemented from icu::Collator.
RuleBasedCollator & icu::RuleBasedCollator::operator= | ( | const RuleBasedCollator & | other | ) |
|
overridevirtual |
Returns true if argument is the same as this object.
other | Collator object to be compared. |
Reimplemented from icu::Collator.
|
inlinestatic |
Definition at line 790 of file tblcoll.h.
References icu::Collator::fromUCollator().
|
inlinestatic |
Definition at line 786 of file tblcoll.h.
References icu::Collator::fromUCollator().
|
overridevirtual |
Universal attribute setter.
attr | attribute type |
value | attribute value |
status | to indicate whether the operation went on smoothly or there were errors |
Implements icu::Collator.
|
overrideprotectedvirtual |
Used internally by registration to define the requested and valid locales.
requestedLocale | the requested locale |
validLocale | the valid locale |
actualLocale | the actual locale |
Reimplemented from icu::Collator.
|
overridevirtual |
Sets the variable top to the top of the specified reordering group.
The variable top determines the highest-sorting character which is affected by UCOL_ALTERNATE_HANDLING. If that attribute is set to UCOL_NON_IGNORABLE, then the variable top has no effect.
group | one of UCOL_REORDER_CODE_SPACE, UCOL_REORDER_CODE_PUNCTUATION, UCOL_REORDER_CODE_SYMBOL, UCOL_REORDER_CODE_CURRENCY; or UCOL_REORDER_CODE_DEFAULT to restore the default max variable group |
errorCode | Standard ICU error code. Its input value must pass the U_SUCCESS() test, or else the function returns immediately. Check for U_FAILURE() on output or use with function chaining. (See User Guide for details.) |
Reimplemented from icu::Collator.
|
overridevirtual |
Sets the ordering of scripts for this collator.
reorderCodes | An array of script codes in the new order. This can be NULL if the length is also set to 0. An empty array will clear any reordering codes on the collator. |
reorderCodesLength | The length of reorderCodes. |
status | error code |
Reimplemented from icu::Collator.
|
overridevirtual |
Sets the variable top to the primary weight of the specified string.
Beginning with ICU 53, the variable top is pinned to the top of one of the supported reordering groups, and it must not be beyond the last of those groups. See setMaxVariable().
varTop | one or more (if contraction) char16_ts to which the variable top should be set |
len | length of variable top string. If -1 it is considered to be zero terminated. |
status | error code. If error code is set, the return value is undefined. Errors set by this function are: U_CE_NOT_FOUND_ERROR if more than one character was passed and there is no such contraction U_ILLEGAL_ARGUMENT_ERROR if the variable top is beyond the last reordering group supported by setMaxVariable() |
Implements icu::Collator.
|
overridevirtual |
Sets the variable top to the primary weight of the specified string.
Beginning with ICU 53, the variable top is pinned to the top of one of the supported reordering groups, and it must not be beyond the last of those groups. See setMaxVariable().
varTop | a UnicodeString size 1 or more (if contraction) of char16_ts to which the variable top should be set |
status | error code. If error code is set, the return value is undefined. Errors set by this function are: U_CE_NOT_FOUND_ERROR if more than one character was passed and there is no such contraction U_ILLEGAL_ARGUMENT_ERROR if the variable top is beyond the last reordering group supported by setMaxVariable() |
Implements icu::Collator.
|
overridevirtual |
Sets the variable top to the specified primary weight.
Beginning with ICU 53, the variable top is pinned to the top of one of the supported reordering groups, and it must not be beyond the last of those groups. See setMaxVariable().
varTop | primary weight, as returned by setVariableTop or ucol_getVariableTop |
status | error code |
Implements icu::Collator.
|
friend |