18 #if U_SHOW_CPLUSPLUS_API
35 class RBBIRuleScanner;
37 class UnicodeSetStringSpan;
39 class RuleCharacterIterator;
293 static constexpr uint8_t kIsBogus = 1;
300 BMPSet *bmpSet =
nullptr;
302 int32_t bufferCapacity = 0;
313 char16_t *pat =
nullptr;
316 UVector* strings =
nullptr;
317 UnicodeSetStringSpan *stringSpan =
nullptr;
336 inline UBool isBogus(
void)
const;
394 #ifndef U_HIDE_INTERNAL_API
412 UnicodeSet(
const uint16_t buffer[], int32_t bufferLen,
413 ESerialization serialization,
UErrorCode &status);
427 #ifndef U_HIDE_INTERNAL_API
525 virtual int32_t hashCode(
void)
const;
554 inline USet *toUSet();
564 inline const USet * toUSet()
const;
579 inline UBool isFrozen()
const;
644 #ifndef U_HIDE_INTERNAL_API
722 UBool escapeUnprintable =
false)
const override;
795 virtual int32_t size(
void)
const;
803 virtual UBool isEmpty(
void)
const;
809 UBool hasStrings()
const;
931 int32_t span(
const char16_t *s, int32_t length,
USetSpanCondition spanCondition)
const;
964 int32_t spanBack(
const char16_t *s, int32_t length,
USetSpanCondition spanCondition)
const;
999 int32_t spanUTF8(
const char *s, int32_t length,
USetSpanCondition spanCondition)
const;
1018 int32_t spanBackUTF8(
const char *s, int32_t length,
USetSpanCondition spanCondition)
const;
1027 UBool incremental)
override;
1053 int32_t start, int32_t limit,
1065 int32_t findCodePoint(
UChar32 c)
const;
1086 int32_t indexOf(
UChar32 c)
const;
1103 UChar32 charAt(int32_t index)
const;
1443 virtual int32_t getRangeCount(
void)
const;
1452 virtual UChar32 getRangeStart(int32_t index)
const;
1461 virtual UChar32 getRangeEnd(int32_t index)
const;
1511 int32_t serialize(uint16_t *dest, int32_t destCapacity,
UErrorCode& ec)
const;
1548 friend class USetAccess;
1566 friend class RBBIRuleScanner;
1584 void applyPattern(RuleCharacterIterator& chars,
1592 void closeOverCaseInsensitive(
bool simple);
1593 void closeOverAddCaseMappings();
1599 static int32_t nextCapacity(int32_t minCapacity);
1601 bool ensureCapacity(int32_t newLen);
1603 bool ensureBufferCapacity(int32_t newLen);
1605 void swapBuffers(
void);
1608 int32_t stringsSize()
const;
1612 UBool escapeUnprintable)
const;
1615 UBool escapeUnprintable)
const;
1622 UBool escapeUnprintable);
1628 void exclusiveOr(
const UChar32* other, int32_t otherLen, int8_t polarity);
1630 void add(
const UChar32* other, int32_t otherLen, int8_t polarity);
1632 void retain(
const UChar32* other, int32_t otherLen, int8_t polarity);
1642 static UBool resemblesPropertyPattern(RuleCharacterIterator& chars,
1688 void applyPropertyPattern(RuleCharacterIterator& chars,
1696 typedef UBool (*Filter)(
UChar32 codePoint,
void* context);
1707 void applyFilter(Filter filter,
1718 void setPattern(
const char16_t *newPat, int32_t newPatLen);
1722 void releasePattern();
1724 friend class UnicodeSetIterator;
1733 inline UBool UnicodeSet::isFrozen()
const {
1734 return (
UBool)(bmpSet!=
nullptr || stringSpan!=
nullptr);
1738 return !containsNone(start, end);
1742 return !containsNone(s);
1746 return !containsNone(s);
1749 inline UBool UnicodeSet::isBogus()
const {
1750 return (
UBool)(fFlags & kIsBogus);
1758 return reinterpret_cast<const UnicodeSet *
>(uset);
1761 inline USet *UnicodeSet::toUSet() {
1762 return reinterpret_cast<USet *
>(
this);
1765 inline const USet *UnicodeSet::toUSet()
const {
1766 return reinterpret_cast<const USet *
>(
this);
1770 int32_t sLength=s.
length();
1773 }
else if(start>sLength) {
1776 return start+span(s.
getBuffer()+start, sLength-start, spanCondition);
1780 int32_t sLength=s.
length();
1783 }
else if(limit>sLength) {
1786 return spanBack(s.
getBuffer(), limit, spanCondition);
#define INITIAL_CAPACITY
The initial size of an array if it is unspecified.
bool operator!=(const StringPiece &x, const StringPiece &y)
Global operator != for StringPiece.
virtual UMatchDegree matches(const Replaceable &text, int32_t &offset, int32_t limit, UBool incremental) override
Implement UnicodeMatcher API.
static UClassID getStaticClassID()
ICU "poor man's RTTI", returns a UClassID for this class.
UMatchDegree
Constants returned by UnicodeMatcher::matches() indicating the degree of match.
U_EXPORT UBool operator==(const StringPiece &x, const StringPiece &y)
Global operator == for StringPiece.
void * UClassID
UClassID is used to identify classes without using the compiler's RTTI.
C API: This file defines an abstract map from Unicode code points to integer values.
virtual UBool matchesIndexValue(uint8_t v) const =0
Returns true if this matcher will match a character c, where c & 0xFF == v, at offset, in the forward direction (with limit > offset).
An interface that defines both lookup protocol and parsing of symbolic names.
virtual UnicodeString & toPattern(UnicodeString &result, UBool escapeUnprintable=false) const =0
Returns a string representation of this matcher.
Replaceable is an abstract base class representing a string of characters that supports the replaceme...
UnicodeFilter defines a protocol for selecting a subset of the full range (U+0000 to U+10FFFF) of Uni...
virtual void addMatchSetTo(UnicodeSet &toUnionTo) const =0
Union the set of all characters that may be matched by this object into the given set...
int32_t UChar32
Define UChar32 as a type for single Unicode code points.
A mutable set of Unicode characters and multicharacter strings.
USetSpanCondition
Argument values for whether span() and similar functions continue while the current character is cont...
UProperty
Selection constants for Unicode properties.
UErrorCode
Standard ICU4C error code type, a substitute for exceptions.
struct USet USet
USet is the C API type corresponding to C++ class UnicodeSet.
ParsePosition is a simple class used by Format and its subclasses to keep track of the current positi...
char16_t * getBuffer(int32_t minCapacity)
Get a read/write pointer to the internal buffer.
virtual UnicodeFilter * clone() const override=0
Clones this object polymorphically.
Basic definitions for ICU, for both C and C++ APIs.
virtual UBool contains(UChar32 c) const =0
Returns true for characters that are in the selected subset.
#define U_COMMON_API
Set to export library symbols from inside the common library, and to import them from outside...
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
int32_t length(void) const
Return the length of the UnicodeString object.
int8_t UBool
The ICU boolean type, a signed-byte integer.
virtual UClassID getDynamicClassID(void) const override=0
Returns a unique class ID polymorphically.