golden hour
/opt/alt/libicu65/usr/include/unicode
⬆️ Go Up
Upload
File/Folder
Size
Actions
alphaindex.h
26.48 KB
Del
OK
appendable.h
8.49 KB
Del
OK
basictz.h
9.15 KB
Del
OK
brkiter.h
27.8 KB
Del
OK
bytestream.h
9.6 KB
Del
OK
bytestrie.h
20.77 KB
Del
OK
bytestriebuilder.h
7.08 KB
Del
OK
calendar.h
105.74 KB
Del
OK
caniter.h
7.43 KB
Del
OK
casemap.h
25.33 KB
Del
OK
char16ptr.h
7.22 KB
Del
OK
chariter.h
24.05 KB
Del
OK
choicfmt.h
23.91 KB
Del
OK
coleitr.h
13.76 KB
Del
OK
coll.h
56.23 KB
Del
OK
compactdecimalformat.h
6.88 KB
Del
OK
curramt.h
3.76 KB
Del
OK
currpinf.h
7.3 KB
Del
OK
currunit.h
4.05 KB
Del
OK
datefmt.h
40.67 KB
Del
OK
dbbi.h
1.19 KB
Del
OK
dcfmtsym.h
20.13 KB
Del
OK
decimfmt.h
87.38 KB
Del
OK
docmain.h
6.97 KB
Del
OK
dtfmtsym.h
37.7 KB
Del
OK
dtintrv.h
3.84 KB
Del
OK
dtitvfmt.h
46.63 KB
Del
OK
dtitvinf.h
18.51 KB
Del
OK
dtptngen.h
25.08 KB
Del
OK
dtrule.h
8.68 KB
Del
OK
edits.h
20.74 KB
Del
OK
enumset.h
2.08 KB
Del
OK
errorcode.h
4.84 KB
Del
OK
fieldpos.h
8.69 KB
Del
OK
filteredbrk.h
5.37 KB
Del
OK
fmtable.h
24.42 KB
Del
OK
format.h
12.5 KB
Del
OK
formattedvalue.h
10.27 KB
Del
OK
fpositer.h
3.04 KB
Del
OK
gender.h
3.33 KB
Del
OK
gregocal.h
31.71 KB
Del
OK
icudataver.h
1.03 KB
Del
OK
icuplug.h
11.88 KB
Del
OK
idna.h
12.7 KB
Del
OK
listformatter.h
9.47 KB
Del
OK
localebuilder.h
11.27 KB
Del
OK
localematcher.h
22.5 KB
Del
OK
localpointer.h
19.69 KB
Del
OK
locdspnm.h
7.12 KB
Del
OK
locid.h
47.4 KB
Del
OK
measfmt.h
11.33 KB
Del
OK
measunit.h
93.31 KB
Del
OK
measure.h
4.32 KB
Del
OK
messagepattern.h
33.71 KB
Del
OK
msgfmt.h
44.11 KB
Del
OK
normalizer2.h
34.03 KB
Del
OK
normlzr.h
30.94 KB
Del
OK
nounit.h
2.69 KB
Del
OK
numberformatter.h
86.31 KB
Del
OK
numberrangeformatter.h
30.14 KB
Del
OK
numfmt.h
49.81 KB
Del
OK
numsys.h
7.19 KB
Del
OK
parseerr.h
3.08 KB
Del
OK
parsepos.h
5.56 KB
Del
OK
platform.h
28.08 KB
Del
OK
plurfmt.h
25.2 KB
Del
OK
plurrule.h
18.39 KB
Del
OK
ptypes.h
3.49 KB
Del
OK
putil.h
6.33 KB
Del
OK
rbbi.h
26.58 KB
Del
OK
rbnf.h
48.73 KB
Del
OK
rbtz.h
15.6 KB
Del
OK
regex.h
84.36 KB
Del
OK
region.h
9.18 KB
Del
OK
reldatefmt.h
22.62 KB
Del
OK
rep.h
9.37 KB
Del
OK
resbund.h
18.07 KB
Del
OK
schriter.h
6.32 KB
Del
OK
scientificnumberformatter.h
6.4 KB
Del
OK
search.h
22.22 KB
Del
OK
selfmt.h
14.3 KB
Del
OK
simpleformatter.h
12.59 KB
Del
OK
simpletz.h
45.44 KB
Del
OK
smpdtfmt.h
70.97 KB
Del
OK
sortkey.h
11.18 KB
Del
OK
std_string.h
1.05 KB
Del
OK
strenum.h
9.92 KB
Del
OK
stringoptions.h
5.79 KB
Del
OK
stringpiece.h
7.38 KB
Del
OK
stringtriebuilder.h
15.33 KB
Del
OK
stsearch.h
21.3 KB
Del
OK
symtable.h
4.27 KB
Del
OK
tblcoll.h
36.61 KB
Del
OK
timezone.h
41.02 KB
Del
OK
tmunit.h
3.38 KB
Del
OK
tmutamt.h
4.9 KB
Del
OK
tmutfmt.h
7.85 KB
Del
OK
translit.h
65.82 KB
Del
OK
tzfmt.h
42.89 KB
Del
OK
tznames.h
16.85 KB
Del
OK
tzrule.h
35.37 KB
Del
OK
tztrans.h
6.12 KB
Del
OK
ubidi.h
89.56 KB
Del
OK
ubiditransform.h
12.65 KB
Del
OK
ubrk.h
23.97 KB
Del
OK
ucal.h
56.9 KB
Del
OK
ucasemap.h
15.18 KB
Del
OK
ucat.h
5.36 KB
Del
OK
uchar.h
140.56 KB
Del
OK
ucharstrie.h
22.58 KB
Del
OK
ucharstriebuilder.h
7.21 KB
Del
OK
uchriter.h
13.2 KB
Del
OK
uclean.h
11.21 KB
Del
OK
ucnv.h
83.09 KB
Del
OK
ucnv_cb.h
6.59 KB
Del
OK
ucnv_err.h
20.99 KB
Del
OK
ucnvsel.h
6.14 KB
Del
OK
ucol.h
61.46 KB
Del
OK
ucoleitr.h
9.46 KB
Del
OK
uconfig.h
12.07 KB
Del
OK
ucpmap.h
5.53 KB
Del
OK
ucptrie.h
22.46 KB
Del
OK
ucsdet.h
14.67 KB
Del
OK
ucurr.h
16.12 KB
Del
OK
udat.h
60.88 KB
Del
OK
udata.h
15.56 KB
Del
OK
udateintervalformat.h
10.03 KB
Del
OK
udatpg.h
26.01 KB
Del
OK
udisplaycontext.h
5.89 KB
Del
OK
uenum.h
7.78 KB
Del
OK
ufieldpositer.h
4.36 KB
Del
OK
uformattable.h
10.94 KB
Del
OK
uformattedvalue.h
12.14 KB
Del
OK
ugender.h
2 KB
Del
OK
uidna.h
33.37 KB
Del
OK
uiter.h
22.77 KB
Del
OK
uldnames.h
10.45 KB
Del
OK
ulistformatter.h
8.83 KB
Del
OK
uloc.h
52.54 KB
Del
OK
ulocdata.h
11.26 KB
Del
OK
umachine.h
14.53 KB
Del
OK
umisc.h
1.33 KB
Del
OK
umsg.h
24.23 KB
Del
OK
umutablecptrie.h
8.24 KB
Del
OK
unifilt.h
3.96 KB
Del
OK
unifunct.h
4.04 KB
Del
OK
unimatch.h
6.1 KB
Del
OK
unirepl.h
3.38 KB
Del
OK
uniset.h
64.9 KB
Del
OK
unistr.h
170.43 KB
Del
OK
unorm.h
20.52 KB
Del
OK
unorm2.h
24.66 KB
Del
OK
unum.h
53.62 KB
Del
OK
unumberformatter.h
25.36 KB
Del
OK
unumsys.h
7.21 KB
Del
OK
uobject.h
10.68 KB
Del
OK
upluralrules.h
7.88 KB
Del
OK
uregex.h
72.05 KB
Del
OK
uregion.h
9.84 KB
Del
OK
ureldatefmt.h
17.26 KB
Del
OK
urename.h
130.97 KB
Del
OK
urep.h
5.38 KB
Del
OK
ures.h
36.54 KB
Del
OK
uscript.h
26.87 KB
Del
OK
usearch.h
38.12 KB
Del
OK
uset.h
40 KB
Del
OK
usetiter.h
9.55 KB
Del
OK
ushape.h
18 KB
Del
OK
uspoof.h
65.9 KB
Del
OK
usprep.h
8.14 KB
Del
OK
ustdio.h
38.54 KB
Del
OK
ustream.h
1.89 KB
Del
OK
ustring.h
72.47 KB
Del
OK
ustringtrie.h
3.15 KB
Del
OK
utext.h
58.13 KB
Del
OK
utf.h
7.86 KB
Del
OK
utf16.h
23.32 KB
Del
OK
utf32.h
763 B
Del
OK
utf8.h
30.96 KB
Del
OK
utf_old.h
45.83 KB
Del
OK
utmscale.h
13.78 KB
Del
OK
utrace.h
15.73 KB
Del
OK
utrans.h
25.52 KB
Del
OK
utypes.h
30.74 KB
Del
OK
uvernum.h
6.67 KB
Del
OK
uversion.h
6 KB
Del
OK
vtzone.h
20.3 KB
Del
OK
Edit: bytestrie.h
// © 2016 and later: Unicode, Inc. and others. // License & terms of use: http://www.unicode.org/copyright.html /* ******************************************************************************* * Copyright (C) 2010-2012, International Business Machines * Corporation and others. All Rights Reserved. ******************************************************************************* * file name: bytestrie.h * encoding: UTF-8 * tab size: 8 (not used) * indentation:4 * * created on: 2010sep25 * created by: Markus W. Scherer */ #ifndef __BYTESTRIE_H__ #define __BYTESTRIE_H__ /** * \file * \brief C++ API: Trie for mapping byte sequences to integer values. */ #include "unicode/utypes.h" #if U_SHOW_CPLUSPLUS_API #include "unicode/stringpiece.h" #include "unicode/uobject.h" #include "unicode/ustringtrie.h" U_NAMESPACE_BEGIN class ByteSink; class BytesTrieBuilder; class CharString; class UVector32; /** * Light-weight, non-const reader class for a BytesTrie. * Traverses a byte-serialized data structure with minimal state, * for mapping byte sequences to non-negative integer values. * * This class owns the serialized trie data only if it was constructed by * the builder's build() method. * The public constructor and the copy constructor only alias the data (only copy the pointer). * There is no assignment operator. * * This class is not intended for public subclassing. * @stable ICU 4.8 */ class U_COMMON_API BytesTrie : public UMemory { public: /** * Constructs a BytesTrie reader instance. * * The trieBytes must contain a copy of a byte sequence from the BytesTrieBuilder, * starting with the first byte of that sequence. * The BytesTrie object will not read more bytes than * the BytesTrieBuilder generated in the corresponding build() call. * * The array is not copied/cloned and must not be modified while * the BytesTrie object is in use. * * @param trieBytes The byte array that contains the serialized trie. * @stable ICU 4.8 */ BytesTrie(const void *trieBytes) : ownedArray_(NULL), bytes_(static_cast<const uint8_t *>(trieBytes)), pos_(bytes_), remainingMatchLength_(-1) {} /** * Destructor. * @stable ICU 4.8 */ ~BytesTrie(); /** * Copy constructor, copies the other trie reader object and its state, * but not the byte array which will be shared. (Shallow copy.) * @param other Another BytesTrie object. * @stable ICU 4.8 */ BytesTrie(const BytesTrie &other) : ownedArray_(NULL), bytes_(other.bytes_), pos_(other.pos_), remainingMatchLength_(other.remainingMatchLength_) {} /** * Resets this trie to its initial state. * @return *this * @stable ICU 4.8 */ BytesTrie &reset() { pos_=bytes_; remainingMatchLength_=-1; return *this; } #ifndef U_HIDE_DRAFT_API /** * Returns the state of this trie as a 64-bit integer. * The state value is never 0. * * @return opaque state value * @see resetToState64 * @draft ICU 65 */ uint64_t getState64() const { return (static_cast<uint64_t>(remainingMatchLength_ + 2) << kState64RemainingShift) | (uint64_t)(pos_ - bytes_); } /** * Resets this trie to the saved state. * Unlike resetToState(State), the 64-bit state value * must be from getState64() from the same trie object or * from one initialized the exact same way. * Because of no validation, this method is faster. * * @param state The opaque trie state value from getState64(). * @return *this * @see getState64 * @see resetToState * @see reset * @draft ICU 65 */ BytesTrie &resetToState64(uint64_t state) { remainingMatchLength_ = static_cast<int32_t>(state >> kState64RemainingShift) - 2; pos_ = bytes_ + (state & kState64PosMask); return *this; } #endif /* U_HIDE_DRAFT_API */ /** * BytesTrie state object, for saving a trie's current state * and resetting the trie back to this state later. * @stable ICU 4.8 */ class State : public UMemory { public: /** * Constructs an empty State. * @stable ICU 4.8 */ State() { bytes=NULL; } private: friend class BytesTrie; const uint8_t *bytes; const uint8_t *pos; int32_t remainingMatchLength; }; /** * Saves the state of this trie. * @param state The State object to hold the trie's state. * @return *this * @see resetToState * @stable ICU 4.8 */ const BytesTrie &saveState(State &state) const { state.bytes=bytes_; state.pos=pos_; state.remainingMatchLength=remainingMatchLength_; return *this; } /** * Resets this trie to the saved state. * If the state object contains no state, or the state of a different trie, * then this trie remains unchanged. * @param state The State object which holds a saved trie state. * @return *this * @see saveState * @see reset * @stable ICU 4.8 */ BytesTrie &resetToState(const State &state) { if(bytes_==state.bytes && bytes_!=NULL) { pos_=state.pos; remainingMatchLength_=state.remainingMatchLength; } return *this; } /** * Determines whether the byte sequence so far matches, whether it has a value, * and whether another input byte can continue a matching byte sequence. * @return The match/value Result. * @stable ICU 4.8 */ UStringTrieResult current() const; /** * Traverses the trie from the initial state for this input byte. * Equivalent to reset().next(inByte). * @param inByte Input byte value. Values -0x100..-1 are treated like 0..0xff. * Values below -0x100 and above 0xff will never match. * @return The match/value Result. * @stable ICU 4.8 */ inline UStringTrieResult first(int32_t inByte) { remainingMatchLength_=-1; if(inByte<0) { inByte+=0x100; } return nextImpl(bytes_, inByte); } /** * Traverses the trie from the current state for this input byte. * @param inByte Input byte value. Values -0x100..-1 are treated like 0..0xff. * Values below -0x100 and above 0xff will never match. * @return The match/value Result. * @stable ICU 4.8 */ UStringTrieResult next(int32_t inByte); /** * Traverses the trie from the current state for this byte sequence. * Equivalent to * \code * Result result=current(); * for(each c in s) * if(!USTRINGTRIE_HAS_NEXT(result)) return USTRINGTRIE_NO_MATCH; * result=next(c); * return result; * \endcode * @param s A string or byte sequence. Can be NULL if length is 0. * @param length The length of the byte sequence. Can be -1 if NUL-terminated. * @return The match/value Result. * @stable ICU 4.8 */ UStringTrieResult next(const char *s, int32_t length); /** * Returns a matching byte sequence's value if called immediately after * current()/first()/next() returned USTRINGTRIE_INTERMEDIATE_VALUE or USTRINGTRIE_FINAL_VALUE. * getValue() can be called multiple times. * * Do not call getValue() after USTRINGTRIE_NO_MATCH or USTRINGTRIE_NO_VALUE! * @return The value for the byte sequence so far. * @stable ICU 4.8 */ inline int32_t getValue() const { const uint8_t *pos=pos_; int32_t leadByte=*pos++; // U_ASSERT(leadByte>=kMinValueLead); return readValue(pos, leadByte>>1); } /** * Determines whether all byte sequences reachable from the current state * map to the same value. * @param uniqueValue Receives the unique value, if this function returns TRUE. * (output-only) * @return TRUE if all byte sequences reachable from the current state * map to the same value. * @stable ICU 4.8 */ inline UBool hasUniqueValue(int32_t &uniqueValue) const { const uint8_t *pos=pos_; // Skip the rest of a pending linear-match node. return pos!=NULL && findUniqueValue(pos+remainingMatchLength_+1, FALSE, uniqueValue); } /** * Finds each byte which continues the byte sequence from the current state. * That is, each byte b for which it would be next(b)!=USTRINGTRIE_NO_MATCH now. * @param out Each next byte is appended to this object. * (Only uses the out.Append(s, length) method.) * @return the number of bytes which continue the byte sequence from here * @stable ICU 4.8 */ int32_t getNextBytes(ByteSink &out) const; /** * Iterator for all of the (byte sequence, value) pairs in a BytesTrie. * @stable ICU 4.8 */ class U_COMMON_API Iterator : public UMemory { public: /** * Iterates from the root of a byte-serialized BytesTrie. * @param trieBytes The trie bytes. * @param maxStringLength If 0, the iterator returns full strings/byte sequences. * Otherwise, the iterator returns strings with this maximum length. * @param 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.) * @stable ICU 4.8 */ Iterator(const void *trieBytes, int32_t maxStringLength, UErrorCode &errorCode); /** * Iterates from the current state of the specified BytesTrie. * @param trie The trie whose state will be copied for iteration. * @param maxStringLength If 0, the iterator returns full strings/byte sequences. * Otherwise, the iterator returns strings with this maximum length. * @param 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.) * @stable ICU 4.8 */ Iterator(const BytesTrie &trie, int32_t maxStringLength, UErrorCode &errorCode); /** * Destructor. * @stable ICU 4.8 */ ~Iterator(); /** * Resets this iterator to its initial state. * @return *this * @stable ICU 4.8 */ Iterator &reset(); /** * @return TRUE if there are more elements. * @stable ICU 4.8 */ UBool hasNext() const; /** * Finds the next (byte sequence, value) pair if there is one. * * If the byte sequence is truncated to the maximum length and does not * have a real value, then the value is set to -1. * In this case, this "not a real value" is indistinguishable from * a real value of -1. * @param 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.) * @return TRUE if there is another element. * @stable ICU 4.8 */ UBool next(UErrorCode &errorCode); /** * @return The NUL-terminated byte sequence for the last successful next(). * @stable ICU 4.8 */ StringPiece getString() const; /** * @return The value for the last successful next(). * @stable ICU 4.8 */ int32_t getValue() const { return value_; } private: UBool truncateAndStop(); const uint8_t *branchNext(const uint8_t *pos, int32_t length, UErrorCode &errorCode); const uint8_t *bytes_; const uint8_t *pos_; const uint8_t *initialPos_; int32_t remainingMatchLength_; int32_t initialRemainingMatchLength_; CharString *str_; int32_t maxLength_; int32_t value_; // The stack stores pairs of integers for backtracking to another // outbound edge of a branch node. // The first integer is an offset from bytes_. // The second integer has the str_->length() from before the node in bits 15..0, // and the remaining branch length in bits 24..16. (Bits 31..25 are unused.) // (We could store the remaining branch length minus 1 in bits 23..16 and not use bits 31..24, // but the code looks more confusing that way.) UVector32 *stack_; }; private: friend class BytesTrieBuilder; /** * Constructs a BytesTrie reader instance. * Unlike the public constructor which just aliases an array, * this constructor adopts the builder's array. * This constructor is only called by the builder. */ BytesTrie(void *adoptBytes, const void *trieBytes) : ownedArray_(static_cast<uint8_t *>(adoptBytes)), bytes_(static_cast<const uint8_t *>(trieBytes)), pos_(bytes_), remainingMatchLength_(-1) {} // No assignment operator. BytesTrie &operator=(const BytesTrie &other); inline void stop() { pos_=NULL; } // Reads a compact 32-bit integer. // pos is already after the leadByte, and the lead byte is already shifted right by 1. static int32_t readValue(const uint8_t *pos, int32_t leadByte); static inline const uint8_t *skipValue(const uint8_t *pos, int32_t leadByte) { // U_ASSERT(leadByte>=kMinValueLead); if(leadByte>=(kMinTwoByteValueLead<<1)) { if(leadByte<(kMinThreeByteValueLead<<1)) { ++pos; } else if(leadByte<(kFourByteValueLead<<1)) { pos+=2; } else { pos+=3+((leadByte>>1)&1); } } return pos; } static inline const uint8_t *skipValue(const uint8_t *pos) { int32_t leadByte=*pos++; return skipValue(pos, leadByte); } // Reads a jump delta and jumps. static const uint8_t *jumpByDelta(const uint8_t *pos); static inline const uint8_t *skipDelta(const uint8_t *pos) { int32_t delta=*pos++; if(delta>=kMinTwoByteDeltaLead) { if(delta<kMinThreeByteDeltaLead) { ++pos; } else if(delta<kFourByteDeltaLead) { pos+=2; } else { pos+=3+(delta&1); } } return pos; } static inline UStringTrieResult valueResult(int32_t node) { return (UStringTrieResult)(USTRINGTRIE_INTERMEDIATE_VALUE-(node&kValueIsFinal)); } // Handles a branch node for both next(byte) and next(string). UStringTrieResult branchNext(const uint8_t *pos, int32_t length, int32_t inByte); // Requires remainingLength_<0. UStringTrieResult nextImpl(const uint8_t *pos, int32_t inByte); // Helper functions for hasUniqueValue(). // Recursively finds a unique value (or whether there is not a unique one) // from a branch. static const uint8_t *findUniqueValueFromBranch(const uint8_t *pos, int32_t length, UBool haveUniqueValue, int32_t &uniqueValue); // Recursively finds a unique value (or whether there is not a unique one) // starting from a position on a node lead byte. static UBool findUniqueValue(const uint8_t *pos, UBool haveUniqueValue, int32_t &uniqueValue); // Helper functions for getNextBytes(). // getNextBytes() when pos is on a branch node. static void getNextBranchBytes(const uint8_t *pos, int32_t length, ByteSink &out); static void append(ByteSink &out, int c); // BytesTrie data structure // // The trie consists of a series of byte-serialized nodes for incremental // string/byte sequence matching. The root node is at the beginning of the trie data. // // Types of nodes are distinguished by their node lead byte ranges. // After each node, except a final-value node, another node follows to // encode match values or continue matching further bytes. // // Node types: // - Value node: Stores a 32-bit integer in a compact, variable-length format. // The value is for the string/byte sequence so far. // One node bit indicates whether the value is final or whether // matching continues with the next node. // - Linear-match node: Matches a number of bytes. // - Branch node: Branches to other nodes according to the current input byte. // The node byte is the length of the branch (number of bytes to select from) // minus 1. It is followed by a sub-node: // - If the length is at most kMaxBranchLinearSubNodeLength, then // there are length-1 (key, value) pairs and then one more comparison byte. // If one of the key bytes matches, then the value is either a final value for // the string/byte sequence so far, or a "jump" delta to the next node. // If the last byte matches, then matching continues with the next node. // (Values have the same encoding as value nodes.) // - If the length is greater than kMaxBranchLinearSubNodeLength, then // there is one byte and one "jump" delta. // If the input byte is less than the sub-node byte, then "jump" by delta to // the next sub-node which will have a length of length/2. // (The delta has its own compact encoding.) // Otherwise, skip the "jump" delta to the next sub-node // which will have a length of length-length/2. // Node lead byte values. // 00..0f: Branch node. If node!=0 then the length is node+1, otherwise // the length is one more than the next byte. // For a branch sub-node with at most this many entries, we drop down // to a linear search. static const int32_t kMaxBranchLinearSubNodeLength=5; // 10..1f: Linear-match node, match 1..16 bytes and continue reading the next node. static const int32_t kMinLinearMatch=0x10; static const int32_t kMaxLinearMatchLength=0x10; // 20..ff: Variable-length value node. // If odd, the value is final. (Otherwise, intermediate value or jump delta.) // Then shift-right by 1 bit. // The remaining lead byte value indicates the number of following bytes (0..4) // and contains the value's top bits. static const int32_t kMinValueLead=kMinLinearMatch+kMaxLinearMatchLength; // 0x20 // It is a final value if bit 0 is set. static const int32_t kValueIsFinal=1; // Compact value: After testing bit 0, shift right by 1 and then use the following thresholds. static const int32_t kMinOneByteValueLead=kMinValueLead/2; // 0x10 static const int32_t kMaxOneByteValue=0x40; // At least 6 bits in the first byte. static const int32_t kMinTwoByteValueLead=kMinOneByteValueLead+kMaxOneByteValue+1; // 0x51 static const int32_t kMaxTwoByteValue=0x1aff; static const int32_t kMinThreeByteValueLead=kMinTwoByteValueLead+(kMaxTwoByteValue>>8)+1; // 0x6c static const int32_t kFourByteValueLead=0x7e; // A little more than Unicode code points. (0x11ffff) static const int32_t kMaxThreeByteValue=((kFourByteValueLead-kMinThreeByteValueLead)<<16)-1; static const int32_t kFiveByteValueLead=0x7f; // Compact delta integers. static const int32_t kMaxOneByteDelta=0xbf; static const int32_t kMinTwoByteDeltaLead=kMaxOneByteDelta+1; // 0xc0 static const int32_t kMinThreeByteDeltaLead=0xf0; static const int32_t kFourByteDeltaLead=0xfe; static const int32_t kFiveByteDeltaLead=0xff; static const int32_t kMaxTwoByteDelta=((kMinThreeByteDeltaLead-kMinTwoByteDeltaLead)<<8)-1; // 0x2fff static const int32_t kMaxThreeByteDelta=((kFourByteDeltaLead-kMinThreeByteDeltaLead)<<16)-1; // 0xdffff // For getState64(): // The remainingMatchLength_ is -1..14=(kMaxLinearMatchLength=0x10)-2 // so we need at least 5 bits for that. // We add 2 to store it as a positive value 1..16=kMaxLinearMatchLength. static constexpr int32_t kState64RemainingShift = 59; static constexpr uint64_t kState64PosMask = (UINT64_C(1) << kState64RemainingShift) - 1; uint8_t *ownedArray_; // Fixed value referencing the BytesTrie bytes. const uint8_t *bytes_; // Iterator variables. // Pointer to next trie byte to read. NULL if no more matches. const uint8_t *pos_; // Remaining length of a linear-match node, minus 1. Negative if not in such a node. int32_t remainingMatchLength_; }; U_NAMESPACE_END #endif /* U_SHOW_CPLUSPLUS_API */ #endif // __BYTESTRIE_H__
Save