Go to the documentation of this file.
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
521 virtual int32_t hashCode(
void)
const;
550 inline USet *toUSet();
560 inline const USet * toUSet()
const;
575 inline UBool isFrozen()
const;
640 #ifndef U_HIDE_INTERNAL_API
714 UBool escapeUnprintable =
false)
const override;
787 virtual int32_t size(
void)
const;
795 virtual UBool isEmpty(
void)
const;
797 #ifndef U_HIDE_DRAFT_API
802 UBool hasStrings()
const;
803 #endif // U_HIDE_DRAFT_API
925 int32_t span(
const char16_t *s, int32_t length,
USetSpanCondition spanCondition)
const;
958 int32_t spanBack(
const char16_t *s, int32_t length,
USetSpanCondition spanCondition)
const;
993 int32_t spanUTF8(
const char *s, int32_t length,
USetSpanCondition spanCondition)
const;
1012 int32_t spanBackUTF8(
const char *s, int32_t length,
USetSpanCondition spanCondition)
const;
1021 UBool incremental)
override;
1047 int32_t start, int32_t limit,
1059 int32_t findCodePoint(
UChar32 c)
const;
1080 int32_t indexOf(
UChar32 c)
const;
1097 UChar32 charAt(int32_t index)
const;
1232 #ifndef U_HIDE_DRAFT_API
1244 #endif // U_HIDE_DRAFT_API
1437 virtual int32_t getRangeCount(
void)
const;
1446 virtual UChar32 getRangeStart(int32_t index)
const;
1455 virtual UChar32 getRangeEnd(int32_t index)
const;
1505 int32_t serialize(uint16_t *dest, int32_t destCapacity,
UErrorCode& ec)
const;
1542 friend class USetAccess;
1560 friend class RBBIRuleScanner;
1578 void applyPattern(RuleCharacterIterator& chars,
1590 static int32_t nextCapacity(int32_t minCapacity);
1592 bool ensureCapacity(int32_t newLen);
1594 bool ensureBufferCapacity(int32_t newLen);
1596 void swapBuffers(
void);
1599 int32_t stringsSize()
const;
1603 UBool escapeUnprintable)
const;
1606 UBool escapeUnprintable)
const;
1613 UBool escapeUnprintable);
1619 void exclusiveOr(
const UChar32* other, int32_t otherLen, int8_t polarity);
1621 void add(
const UChar32* other, int32_t otherLen, int8_t polarity);
1623 void retain(
const UChar32* other, int32_t otherLen, int8_t polarity);
1633 static UBool resemblesPropertyPattern(RuleCharacterIterator& chars,
1679 void applyPropertyPattern(RuleCharacterIterator& chars,
1689 typedef UBool (*Filter)(
UChar32 codePoint,
void* context);
1700 void applyFilter(Filter filter,
1706 void applyIntPropertyValue(
const UCPMap *map,
1716 void setPattern(
const char16_t *newPat, int32_t newPatLen);
1720 void releasePattern();
1722 friend class UnicodeSetIterator;
1731 inline UBool UnicodeSet::isFrozen()
const {
1736 return !containsNone(start, end);
1740 return !containsNone(s);
1744 return !containsNone(s);
1747 inline UBool UnicodeSet::isBogus()
const {
1748 return (
UBool)(fFlags & kIsBogus);
1756 return reinterpret_cast<const UnicodeSet *
>(uset);
1759 inline USet *UnicodeSet::toUSet() {
1760 return reinterpret_cast<USet *
>(
this);
1763 inline const USet *UnicodeSet::toUSet()
const {
1764 return reinterpret_cast<const USet *
>(
this);
1768 int32_t sLength=s.
length();
1771 }
else if(start>sLength) {
1774 return start+span(s.
getBuffer()+start, sLength-start, spanCondition);
1778 int32_t sLength=s.
length();
1781 }
else if(limit>sLength) {
1784 return spanBack(s.
getBuffer(), limit, spanCondition);
char16_t * getBuffer(int32_t minCapacity)
Get a read/write pointer to the internal buffer.
Basic definitions for ICU, for both C and C++ APIs.
A mutable set of Unicode characters and multicharacter strings.
int8_t UBool
The ICU boolean type, a signed-byte integer.
Replaceable is an abstract base class representing a string of characters that supports the replaceme...
U_EXPORT UBool operator==(const StringPiece &x, const StringPiece &y)
Global operator == for StringPiece.
virtual UBool matchesIndexValue(uint8_t v) const =0
Returns true if this matcher will match a character c, where c & 0xFF == v, at offset,...
UProperty
Selection constants for Unicode properties.
virtual UMatchDegree matches(const Replaceable &text, int32_t &offset, int32_t limit, UBool incremental) override
Implement UnicodeMatcher API.
An interface that defines both lookup protocol and parsing of symbolic names.
USetSpanCondition
Argument values for whether span() and similar functions continue while the current character is cont...
int32_t length(void) const
Return the length of the UnicodeString object.
static UClassID getStaticClassID()
ICU "poor man's RTTI", returns a UClassID for this class.
virtual UnicodeFilter * clone() const override=0
Clones this object polymorphically.
virtual UClassID getDynamicClassID(void) const override=0
Returns a unique class ID polymorphically.
UnicodeString is a string class that stores Unicode characters directly and provides similar function...
int32_t UChar32
Define UChar32 as a type for single Unicode code points.
void * UClassID
UClassID is used to identify classes without using the compiler's RTTI.
UErrorCode
Standard ICU4C error code type, a substitute for exceptions.
uint32_t UCPMapValueFilter(const void *context, uint32_t value)
Callback function type: Modifies a map value.
struct UCPMap UCPMap
Abstract map from Unicode code points (U+0000..U+10FFFF) to integer values.
UMatchDegree
Constants returned by UnicodeMatcher::matches() indicating the degree of match.
virtual UBool contains(UChar32 c) const =0
Returns true for characters that are in the selected subset.
bool operator!=(const StringPiece &x, const StringPiece &y)
Global operator != for StringPiece.
UnicodeFilter defines a protocol for selecting a subset of the full range (U+0000 to U+10FFFF) of Uni...
ParsePosition is a simple class used by Format and its subclasses to keep track of the current positi...
virtual UnicodeString & toPattern(UnicodeString &result, UBool escapeUnprintable=false) const =0
Returns a string representation of this matcher.
struct USet USet
USet is the C API type corresponding to C++ class UnicodeSet.
virtual void addMatchSetTo(UnicodeSet &toUnionTo) const =0
Union the set of all characters that may be matched by this object into the given set.
#define INITIAL_CAPACITY
The initial size of an array if it is unspecified.