1 // © 2016 and later: Unicode, Inc. and others. 2 // License & terms of use: http://www.unicode.org/copyright.html 4 ******************************************************************************* 5 * Copyright (c) 1996-2015, International Business Machines Corporation and others. 7 ******************************************************************************* 15 #if !UCONFIG_NO_COLLATION 23 #if U_SHOW_CPLUSPLUS_API 25 #endif// U_SHOW_CPLUSPLUS_API 104 UCOL_CE_STRENGTH_LIMIT,
131 #ifndef U_HIDE_DEPRECATED_API 137 #endif/* U_HIDE_DEPRECATED_API */ 204 #ifndef U_HIDE_DEPRECATED_API 210 #endif/* U_HIDE_DEPRECATED_API */ 324 #ifndef U_HIDE_DEPRECATED_API 337 #endif/* U_HIDE_DEPRECATED_API */ 358 /* Do not conditionalize the following with #ifndef U_HIDE_DEPRECATED_API, 359 * it is needed for layout of RuleBasedCollator object. */ 360 #ifndef U_FORCE_HIDE_DEPRECATED_API 366 #endif// U_FORCE_HIDE_DEPRECATED_API 452 #ifndef U_HIDE_DEPRECATED_API 491 #endif/* U_HIDE_DEPRECATED_API */ 493 #ifndef U_HIDE_DEPRECATED_API 511 #endif/* U_HIDE_DEPRECATED_API */ 526 USet *contractions,
USet *expansions,
542 #if U_SHOW_CPLUSPLUS_API 579 int32_t sourceLength,
581 int32_t targetLength);
605 int32_t sourceLength,
607 int32_t targetLength,
626 constUChar *source, int32_t sourceLength,
627 constUChar *target, int32_t targetLength);
645 constUChar *source, int32_t sourceLength,
646 constUChar *target, int32_t targetLength);
664 constUChar *source, int32_t sourceLength,
665 constUChar *target, int32_t targetLength);
729 int32_t destCapacity,
773 const int32_t* reorderCodes,
774 int32_t reorderCodesLength,
799 int32_t destCapacity,
818 int32_t resultLength,
830 U_CAPIconstchar* U_EXPORT2
844 #if !UCONFIG_NO_SERVICE 938 constchar* keyword,
constchar* locale,
953 #ifndef U_HIDE_DEPRECATED_API 1006 #endif/* U_HIDE_DEPRECATED_API */ 1034 int32_t sourceLength,
1036 int32_t resultLength);
1063 uint8_t *dest, int32_t count,
1080 #ifndef U_HIDE_DEPRECATED_API 1086 #endif/* U_HIDE_DEPRECATED_API */ 1128 int32_t sourceLength,
1130 uint32_t noOfLevels,
1132 int32_t resultLength,
1205 const uint8_t *src2, int32_t src2Length,
1206 uint8_t *dest, int32_t destCapacity);
1265 #ifndef U_HIDE_DEPRECATED_API 1288 constUChar *varTop, int32_t len,
1290 #endif/* U_HIDE_DEPRECATED_API */ 1305 #ifndef U_HIDE_DEPRECATED_API 1322 #endif/* U_HIDE_DEPRECATED_API */ 1336 #ifndef U_HIDE_DEPRECATED_API 1364 int32_t *pBufferSize,
1371 #define U_COL_SAFECLONE_BUFFERSIZE 1 1373 #endif/* U_HIDE_DEPRECATED_API */ 1393 #ifndef U_HIDE_DEPRECATED_API 1410 #endif/* U_HIDE_DEPRECATED_API */ 1426 U_CAPIconstchar * U_EXPORT2
1442 #ifndef U_HIDE_INTERNAL_API 1481 UBool forceDefaults,
1484 #endif/* U_HIDE_INTERNAL_API */ 1499 uint8_t *buffer, int32_t capacity,
1524 #if U_SHOW_CPLUSPLUS_API || U_SHOW_CPLUSPLUS_HEADER_API 1526 #include <functional> 1527 #include <string_view> 1528 #include <type_traits> 1533 namespaceU_HEADER_ONLY_NAMESPACE {
1549 #if U_SHOW_CPLUSPLUS_API 1552 typename T,
typename U,
1553 typename = std::enable_if_t<ConvertibleToU16StringView<T> && ConvertibleToU16StringView<U>>>
1555 return match(UnicodeString::readOnlyAlias(lhs), UnicodeString::readOnlyAlias(rhs));
1559 booloperator()(std::u16string_view lhs, std::u16string_view rhs)
const{
1560 return match(lhs, rhs);
1563 #if !U_CHAR16_IS_TYPEDEF && (!defined(_LIBCPP_VERSION) || _LIBCPP_VERSION < 180000) 1565 booloperator()(std::basic_string_view<uint16_t> lhs, std::basic_string_view<uint16_t> rhs)
const{
1566 return match({uprv_char16PtrFromUint16(lhs.data()), lhs.length()},
1567 {uprv_char16PtrFromUint16(rhs.data()), rhs.length()});
1571 #if U_SIZEOF_WCHAR_T==2 1573 booloperator()(std::wstring_view lhs, std::wstring_view rhs)
const{
1574 return match({uprv_char16PtrFromWchar(lhs.data()), lhs.length()},
1575 {uprv_char16PtrFromWchar(rhs.data()), rhs.length()});
1581 booloperator()(std::string_view lhs, std::string_view rhs)
const{
1582 return match(lhs, rhs);
1585 #if defined(__cpp_char8_t) 1587 booloperator()(std::u8string_view lhs, std::u8string_view rhs)
const{
1588 return match({
reinterpret_cast<constchar*
>(lhs.data()), lhs.length()},
1589 {
reinterpret_cast<constchar*
>(rhs.data()), rhs.length()});
1594 bool match(std::u16string_view lhs, std::u16string_view rhs)
const{
1598 toUCharPtr(lhs.data()),
static_cast<int32_t
>(lhs.length()),
1599 toUCharPtr(rhs.data()),
static_cast<int32_t
>(rhs.length())),
1603 bool match(std::string_view lhs, std::string_view rhs)
const{
1608 lhs.data(),
static_cast<int32_t
>(lhs.length()),
1609 rhs.data(),
static_cast<int32_t
>(rhs.length()),
1615 static constexpr Compare<UCollationResult> compare{};
1618 }
// namespace internal 1662 }
// namespace collator 1664 }
// namespace U_HEADER_ONLY_NAMESPACE 1666 #endif// U_SHOW_CPLUSPLUS_API || U_SHOW_CPLUSPLUS_HEADER_API 1668 #endif/* #if !UCONFIG_NO_COLLATION */ C++ API: char16_t pointer wrappers with implicit conversion from bit-compatible raw pointer types.
"Smart pointer" class, closes a UCollator via ucol_close().
Function object for performing comparisons using a UCollator.
bool operator()(std::string_view lhs, std::string_view rhs) const
bool operator()(const T &lhs, const U &rhs) const
Predicate(const UCollator *ucol)
C++ API: "Smart pointers" for use with and in ICU4C C++ code.
#define U_DEFINE_LOCAL_OPEN_POINTER(LocalPointerClassName, Type, closeFunction)
"Smart pointer" definition macro, deletes objects via the closeFunction.
C API: Parse Error Information.
C API for code unit iteration.
A UParseError struct is used to returned detailed information about parsing errors.
U_CAPI void ucol_setMaxVariable(UCollator *coll, UColReorderCode group, UErrorCode *pErrorCode)
Sets the variable top to the top of the specified reordering group.
U_CAPI UCollator * ucol_openRules(const UChar *rules, int32_t rulesLength, UColAttributeValue normalizationMode, UCollationStrength strength, UParseError *parseError, UErrorCode *status)
Produce a UCollator instance according to the rules supplied.
uint32_t ucol_setVariableTop(UCollator *coll, const UChar *varTop, int32_t len, UErrorCode *status)
Sets the variable top to the primary weight of the specified string.
U_CAPI int32_t ucol_getRulesEx(const UCollator *coll, UColRuleOption delta, UChar *buffer, int32_t bufferLen)
Returns current rules.
U_CAPI void ucol_getVersion(const UCollator *coll, UVersionInfo info)
Gets the version information for a Collator.
U_CAPI void ucol_getContractionsAndExpansions(const UCollator *coll, USet *contractions, USet *expansions, UBool addPrefixes, UErrorCode *status)
Get a set containing the expansions defined by the collator.
U_CAPI USet * ucol_getTailoredSet(const UCollator *coll, UErrorCode *status)
Get a Unicode set that contains all the characters and sequences tailored in this collator.
U_CAPI uint32_t ucol_getVariableTop(const UCollator *coll, UErrorCode *status)
Gets the variable top value of a Collator.
U_CAPI UColAttributeValue ucol_getAttribute(const UCollator *coll, UColAttribute attr, UErrorCode *status)
Universal attribute getter.
U_CAPI void ucol_setStrength(UCollator *coll, UCollationStrength strength)
Set the collation strength used in a UCollator.
U_CAPI int32_t ucol_getUnsafeSet(const UCollator *coll, USet *unsafe, UErrorCode *status)
Calculates the set of unsafe code points, given a collator.
UCollator * ucol_openFromShortString(const char *definition, UBool forceDefaults, UParseError *parseError, UErrorCode *status)
Open a collator defined by a short form string.
struct UCollator UCollator
structure representing a collator object instance
U_CAPI void ucol_setAttribute(UCollator *coll, UColAttribute attr, UColAttributeValue value, UErrorCode *status)
Universal attribute setter.
U_CAPI UColReorderCode ucol_getMaxVariable(const UCollator *coll)
Returns the maximum reordering group whose characters are affected by UCOL_ALTERNATE_HANDLING.
int32_t ucol_normalizeShortDefinitionString(const char *source, char *destination, int32_t capacity, UParseError *parseError, UErrorCode *status)
Verifies and normalizes short definition string.
UColRuleOption
Options for retrieving the rule string.
@ UCOL_FULL_RULES
Retrieves the "UCA rules" concatenated with the tailoring rules.
@ UCOL_TAILORING_ONLY
Retrieves the tailoring rules only.
U_CAPI const UChar * ucol_getRules(const UCollator *coll, int32_t *length)
Get the collation tailoring rules from a UCollator.
UColAttribute
Attributes that collation service understands.
@ UCOL_CASE_FIRST
Controls the ordering of upper and lower case letters.
@ UCOL_NUMERIC_COLLATION
When turned on, this attribute makes substrings of digits sort according to their numeric values.
@ UCOL_STRENGTH
The strength attribute.
@ UCOL_CASE_LEVEL
Controls whether an extra case level (positioned before the third level) is generated or not.
@ UCOL_FRENCH_COLLATION
Attribute for direction of secondary weights - used in Canadian French.
@ UCOL_HIRAGANA_QUATERNARY_MODE
When turned on, this attribute positions Hiragana before all non-ignorables on quaternary level Thi...
@ UCOL_DECOMPOSITION_MODE
An alias for UCOL_NORMALIZATION_MODE attribute.
@ UCOL_ATTRIBUTE_COUNT
One more than the highest normal UColAttribute value.
@ UCOL_ALTERNATE_HANDLING
Attribute for handling variable elements.
@ UCOL_NORMALIZATION_MODE
Controls whether the normalization check and necessary normalizations are performed.
U_CAPI int32_t ucol_getSortKey(const UCollator *coll, const UChar *source, int32_t sourceLength, uint8_t *result, int32_t resultLength)
Get a sort key for a string from a UCollator.
int32_t ucol_getContractions(const UCollator *coll, USet *conts, UErrorCode *status)
Get a set containing the contractions defined by the collator.
U_CAPI UCollationResult ucol_strcoll(const UCollator *coll, const UChar *source, int32_t sourceLength, const UChar *target, int32_t targetLength)
Compare two strings.
UColAttributeValue UCollationStrength
Base letter represents a primary difference.
U_CAPI int32_t ucol_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.
U_CAPI UCollationResult ucol_strcollIter(const UCollator *coll, UCharIterator *sIter, UCharIterator *tIter, UErrorCode *status)
Compare two UTF-8 encoded strings.
int32_t ucol_getShortDefinitionString(const UCollator *coll, const char *locale, char *buffer, int32_t capacity, UErrorCode *status)
Get the short definition string for a collator.
U_CAPI const char * ucol_getAvailable(int32_t localeIndex)
Get a locale for which collation rules are available.
U_CAPI UBool ucol_equal(const UCollator *coll, const UChar *source, int32_t sourceLength, const UChar *target, int32_t targetLength)
Compare two strings for equality.
U_CAPI int32_t ucol_cloneBinary(const UCollator *coll, uint8_t *buffer, int32_t capacity, UErrorCode *status)
Creates a binary image of a collator.
U_CAPI UCollationResult ucol_strcollUTF8(const UCollator *coll, const char *source, int32_t sourceLength, const char *target, int32_t targetLength, UErrorCode *status)
Compare two strings in UTF-8.
U_CAPI void ucol_prepareShortStringOpen(const char *definition, UBool forceDefaults, UParseError *parseError, UErrorCode *status)
Touches all resources needed for instantiating a collator from a short string definition,...
U_CAPI UBool ucol_greater(const UCollator *coll, const UChar *source, int32_t sourceLength, const UChar *target, int32_t targetLength)
Determine if one string is greater than another.
UCollationResult
UCOL_LESS is returned if source string is compared to be less than target string in the ucol_strcoll(...
@ UCOL_LESS
string a < string b
@ UCOL_GREATER
string a > string b
@ UCOL_EQUAL
string a == string b
U_CAPI UCollator * ucol_open(const char *loc, UErrorCode *status)
Open a UCollator for comparing strings.
U_CAPI const char * ucol_getLocaleByType(const UCollator *coll, ULocDataLocaleType type, UErrorCode *status)
gets the locale name of the collator.
UCollator * ucol_safeClone(const UCollator *coll, void *stackBuffer, int32_t *pBufferSize, UErrorCode *status)
Thread safe cloning operation.
U_CAPI int32_t ucol_nextSortKeyPart(const UCollator *coll, UCharIterator *iter, uint32_t state[2], uint8_t *dest, int32_t count, UErrorCode *status)
Gets the next count bytes of a sort key.
const char * ucol_getLocale(const UCollator *coll, ULocDataLocaleType type, UErrorCode *status)
gets the locale name of the collator.
UColBoundMode
enum that is taken by ucol_getBound API See below for explanation do not change the values assigned...
@ UCOL_BOUND_LOWER
lower bound
@ UCOL_BOUND_UPPER
upper bound that will match strings of exact size
@ UCOL_BOUND_VALUE_COUNT
One more than the highest normal UColBoundMode value.
@ UCOL_BOUND_UPPER_LONG
upper bound that will match all the strings that have the same initial substring as the given string
U_CAPI int32_t ucol_countAvailable(void)
Determine how many locales have collation rules available.
U_CAPI UBool ucol_greaterOrEqual(const UCollator *coll, const UChar *source, int32_t sourceLength, const UChar *target, int32_t targetLength)
Determine if one string is greater than or equal to another.
U_CAPI int32_t ucol_getEquivalentReorderCodes(int32_t reorderCode, int32_t *dest, int32_t destCapacity, UErrorCode *pErrorCode)
Retrieves the reorder codes that are grouped with the given reorder code.
U_CAPI void ucol_close(UCollator *coll)
Close a UCollator.
U_CAPI int32_t ucol_getFunctionalEquivalent(char *result, int32_t resultCapacity, const char *keyword, const char *locale, UBool *isAvailable, UErrorCode *status)
Return the functionally equivalent locale for the specified input locale, with respect to given keywo...
U_CAPI UEnumeration * ucol_getKeywords(UErrorCode *status)
Create a string enumerator of all possible keywords that are relevant to collation.
U_CAPI UCollator * ucol_clone(const UCollator *coll, UErrorCode *status)
Thread safe cloning operation.
U_CAPI UCollator * ucol_openBinary(const uint8_t *bin, int32_t length, const UCollator *base, UErrorCode *status)
Opens a collator from a collator binary image created using ucol_cloneBinary.
U_CAPI UEnumeration * ucol_openAvailableLocales(UErrorCode *status)
Create a string enumerator of all locales for which a valid collator may be opened.
U_CAPI void ucol_getUCAVersion(const UCollator *coll, UVersionInfo info)
Gets the UCA version information for a Collator.
U_CAPI int32_t ucol_getDisplayName(const char *objLoc, const char *dispLoc, UChar *result, int32_t resultLength, UErrorCode *status)
Get the display name for a UCollator.
UColAttributeValue
Enum containing attribute values for controlling collation behavior.
@ UCOL_UPPER_FIRST
upper case sorts before lower case
@ UCOL_TERTIARY
Tertiary collation strength.
@ UCOL_DEFAULT
accepted by most attributes
@ UCOL_IDENTICAL
Identical collation strength.
@ UCOL_OFF
Turn the feature off - works for UCOL_FRENCH_COLLATION, UCOL_CASE_LEVEL, UCOL_HIRAGANA_QUATERNARY_MOD...
@ UCOL_QUATERNARY
Quaternary collation strength.
@ UCOL_SHIFTED
Valid for UCOL_ALTERNATE_HANDLING.
@ UCOL_PRIMARY
Primary collation strength.
@ UCOL_LOWER_FIRST
Valid for UCOL_CASE_FIRST - lower case sorts before upper case.
@ UCOL_ATTRIBUTE_VALUE_COUNT
One more than the highest normal UColAttributeValue value.
@ UCOL_SECONDARY
Secondary collation strength.
@ UCOL_NON_IGNORABLE
Valid for UCOL_ALTERNATE_HANDLING.
@ UCOL_ON
Turn the feature on - works for UCOL_FRENCH_COLLATION, UCOL_CASE_LEVEL, UCOL_HIRAGANA_QUATERNARY_MODE...
@ UCOL_DEFAULT_STRENGTH
Default collation strength.
void ucol_restoreVariableTop(UCollator *coll, const uint32_t varTop, UErrorCode *status)
Sets the variable top to the specified primary weight.
U_CAPI UCollationStrength ucol_getStrength(const UCollator *coll)
Get the collation strength used in a UCollator.
U_CAPI int32_t ucol_getReorderCodes(const UCollator *coll, int32_t *dest, int32_t destCapacity, UErrorCode *pErrorCode)
Retrieves the reordering codes for this collator.
U_CAPI void ucol_setReorderCodes(UCollator *coll, const int32_t *reorderCodes, int32_t reorderCodesLength, UErrorCode *pErrorCode)
Sets the reordering codes for this collator.
UColReorderCode
Enum containing the codes for reordering segments of the collation table that are not script codes.
@ UCOL_REORDER_CODE_OTHERS
A special reordering code that is used to specify all other codes used for reordering except for the ...
@ UCOL_REORDER_CODE_NONE
A special reordering code that is used to specify no reordering codes.
@ UCOL_REORDER_CODE_DEFAULT
A special reordering code that is used to specify the default reordering codes for a locale.
@ UCOL_REORDER_CODE_LIMIT
One more than the highest normal UColReorderCode value.
@ UCOL_REORDER_CODE_SYMBOL
Characters with the symbol property.
@ UCOL_REORDER_CODE_DIGIT
Characters with the digit property.
@ UCOL_REORDER_CODE_PUNCTUATION
Characters with the punctuation property.
@ UCOL_REORDER_CODE_SPACE
Characters with the space property.
@ UCOL_REORDER_CODE_FIRST
The first entry in the enumeration of reordering groups.
@ UCOL_REORDER_CODE_CURRENCY
Characters with the currency property.
U_CAPI UEnumeration * ucol_getKeywordValuesForLocale(const char *key, const char *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 di...
U_CAPI UEnumeration * ucol_getKeywordValues(const char *keyword, UErrorCode *status)
Given a keyword, create a string enumeration of all values for that keyword that are currently in use...
U_CAPI int32_t ucol_mergeSortkeys(const uint8_t *src1, int32_t src1Length, const uint8_t *src2, int32_t src2Length, uint8_t *dest, int32_t destCapacity)
Merges two sort keys.
struct UEnumeration UEnumeration
structure representing an enumeration object instance
C API: Locale ID functionality similar to C++ class Locale.
ULocDataLocaleType
Constants for *_getLocale() Allow user to select whether she wants information on requested,...
#define U_DEPRECATED
This is used to declare a function as a deprecated public ICU C API
int8_t UBool
The ICU boolean type, a signed-byte integer.
#define U_CAPI
This is used to declare a function as a public ICU C API.
char16_t UChar
The base type for UTF-16 code units and pointers.
C API: Unicode Normalization.
C API: Unicode Script Information.
struct USet USet
USet is the C API type corresponding to C++ class UnicodeSet.
Basic definitions for ICU, for both C and C++ APIs.
UErrorCode
Standard ICU4C error code type, a substitute for exceptions.
@ U_ZERO_ERROR
No error, no warning.
uint8_t UVersionInfo[U_MAX_VERSION_LENGTH]
The binary form of a version on ICU APIs is an array of 4 uint8_t.
