1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/bmpset.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,161 @@
1.4 +/*
1.5 +******************************************************************************
1.6 +*
1.7 +* Copyright (C) 2007, International Business Machines
1.8 +* Corporation and others. All Rights Reserved.
1.9 +*
1.10 +******************************************************************************
1.11 +* file name: bmpset.h
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 2007jan29
1.17 +* created by: Markus W. Scherer
1.18 +*/
1.19 +
1.20 +#ifndef __BMPSET_H__
1.21 +#define __BMPSET_H__
1.22 +
1.23 +#include "unicode/utypes.h"
1.24 +#include "unicode/uniset.h"
1.25 +
1.26 +U_NAMESPACE_BEGIN
1.27 +
1.28 +/*
1.29 + * Helper class for frozen UnicodeSets, implements contains() and span()
1.30 + * optimized for BMP code points. Structured to be UTF-8-friendly.
1.31 + *
1.32 + * ASCII: Look up bytes.
1.33 + * 2-byte characters: Bits organized vertically.
1.34 + * 3-byte characters: Use zero/one/mixed data per 64-block in U+0000..U+FFFF,
1.35 + * with mixed for illegal ranges.
1.36 + * Supplementary characters: Call contains() on the parent set.
1.37 + */
1.38 +class BMPSet : public UMemory {
1.39 +public:
1.40 + BMPSet(const int32_t *parentList, int32_t parentListLength);
1.41 + BMPSet(const BMPSet &otherBMPSet, const int32_t *newParentList, int32_t newParentListLength);
1.42 + virtual ~BMPSet();
1.43 +
1.44 + virtual UBool contains(UChar32 c) const;
1.45 +
1.46 + /*
1.47 + * Span the initial substring for which each character c has spanCondition==contains(c).
1.48 + * It must be s<limit and spanCondition==0 or 1.
1.49 + * @return The string pointer which limits the span.
1.50 + */
1.51 + const UChar *span(const UChar *s, const UChar *limit, USetSpanCondition spanCondition) const;
1.52 +
1.53 + /*
1.54 + * Span the trailing substring for which each character c has spanCondition==contains(c).
1.55 + * It must be s<limit and spanCondition==0 or 1.
1.56 + * @return The string pointer which starts the span.
1.57 + */
1.58 + const UChar *spanBack(const UChar *s, const UChar *limit, USetSpanCondition spanCondition) const;
1.59 +
1.60 + /*
1.61 + * Span the initial substring for which each character c has spanCondition==contains(c).
1.62 + * It must be length>0 and spanCondition==0 or 1.
1.63 + * @return The string pointer which limits the span.
1.64 + */
1.65 + const uint8_t *spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const;
1.66 +
1.67 + /*
1.68 + * Span the trailing substring for which each character c has spanCondition==contains(c).
1.69 + * It must be length>0 and spanCondition==0 or 1.
1.70 + * @return The start of the span.
1.71 + */
1.72 + int32_t spanBackUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const;
1.73 +
1.74 +private:
1.75 + void initBits();
1.76 + void overrideIllegal();
1.77 +
1.78 + /**
1.79 + * Same as UnicodeSet::findCodePoint(UChar32 c) const except that the
1.80 + * binary search is restricted for finding code points in a certain range.
1.81 + *
1.82 + * For restricting the search for finding in the range start..end,
1.83 + * pass in
1.84 + * lo=findCodePoint(start) and
1.85 + * hi=findCodePoint(end)
1.86 + * with 0<=lo<=hi<len.
1.87 + * findCodePoint(c) defaults to lo=0 and hi=len-1.
1.88 + *
1.89 + * @param c a character in a subrange of MIN_VALUE..MAX_VALUE
1.90 + * @param lo The lowest index to be returned.
1.91 + * @param hi The highest index to be returned.
1.92 + * @return the smallest integer i in the range lo..hi,
1.93 + * inclusive, such that c < list[i]
1.94 + */
1.95 + int32_t findCodePoint(UChar32 c, int32_t lo, int32_t hi) const;
1.96 +
1.97 + inline UBool containsSlow(UChar32 c, int32_t lo, int32_t hi) const;
1.98 +
1.99 + /*
1.100 + * One byte per ASCII character, or trail byte in lead position.
1.101 + * 0 or 1 for ASCII characters.
1.102 + * The value for trail bytes is the result of contains(FFFD)
1.103 + * for faster validity checking at runtime.
1.104 + */
1.105 + UBool asciiBytes[0xc0];
1.106 +
1.107 + /*
1.108 + * One bit per code point from U+0000..U+07FF.
1.109 + * The bits are organized vertically; consecutive code points
1.110 + * correspond to the same bit positions in consecutive table words.
1.111 + * With code point parts
1.112 + * lead=c{10..6}
1.113 + * trail=c{5..0}
1.114 + * it is set.contains(c)==(table7FF[trail] bit lead)
1.115 + *
1.116 + * Bits for 0..7F (non-shortest forms) are set to the result of contains(FFFD)
1.117 + * for faster validity checking at runtime.
1.118 + */
1.119 + uint32_t table7FF[64];
1.120 +
1.121 + /*
1.122 + * One bit per 64 BMP code points.
1.123 + * The bits are organized vertically; consecutive 64-code point blocks
1.124 + * correspond to the same bit position in consecutive table words.
1.125 + * With code point parts
1.126 + * lead=c{15..12}
1.127 + * t1=c{11..6}
1.128 + * test bits (lead+16) and lead in bmpBlockBits[t1].
1.129 + * If the upper bit is 0, then the lower bit indicates if contains(c)
1.130 + * for all code points in the 64-block.
1.131 + * If the upper bit is 1, then the block is mixed and set.contains(c)
1.132 + * must be called.
1.133 + *
1.134 + * Bits for 0..7FF (non-shortest forms) and D800..DFFF are set to
1.135 + * the result of contains(FFFD) for faster validity checking at runtime.
1.136 + */
1.137 + uint32_t bmpBlockBits[64];
1.138 +
1.139 + /*
1.140 + * Inversion list indexes for restricted binary searches in
1.141 + * findCodePoint(), from
1.142 + * findCodePoint(U+0800, U+1000, U+2000, .., U+F000, U+10000).
1.143 + * U+0800 is the first 3-byte-UTF-8 code point. Code points below U+0800 are
1.144 + * always looked up in the bit tables.
1.145 + * The last pair of indexes is for finding supplementary code points.
1.146 + */
1.147 + int32_t list4kStarts[18];
1.148 +
1.149 + /*
1.150 + * The inversion list of the parent set, for the slower contains() implementation
1.151 + * for mixed BMP blocks and for supplementary code points.
1.152 + * The list is terminated with list[listLength-1]=0x110000.
1.153 + */
1.154 + const int32_t *list;
1.155 + int32_t listLength;
1.156 +};
1.157 +
1.158 +inline UBool BMPSet::containsSlow(UChar32 c, int32_t lo, int32_t hi) const {
1.159 + return (UBool)(findCodePoint(c, lo, hi) & 1);
1.160 +}
1.161 +
1.162 +U_NAMESPACE_END
1.163 +
1.164 +#endif
