1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/intl/icu/source/common/unisetspan.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,155 @@
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: unisetspan.h
1.12 +* encoding: US-ASCII
1.13 +* tab size: 8 (not used)
1.14 +* indentation:4
1.15 +*
1.16 +* created on: 2007mar01
1.17 +* created by: Markus W. Scherer
1.18 +*/
1.19 +
1.20 +#ifndef __UNISETSPAN_H__
1.21 +#define __UNISETSPAN_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 + * Implement span() etc. for a set with strings.
1.30 + * Avoid recursion because of its exponential complexity.
1.31 + * Instead, try multiple paths at once and track them with an IndexList.
1.32 + */
1.33 +class UnicodeSetStringSpan : public UMemory {
1.34 +public:
1.35 + /*
1.36 + * Which span() variant will be used?
1.37 + * The object is either built for one variant and used once,
1.38 + * or built for all and may be used many times.
1.39 + */
1.40 + enum {
1.41 + FWD = 0x20,
1.42 + BACK = 0x10,
1.43 + UTF16 = 8,
1.44 + UTF8 = 4,
1.45 + CONTAINED = 2,
1.46 + NOT_CONTAINED = 1,
1.47 +
1.48 + ALL = 0x3f,
1.49 +
1.50 + FWD_UTF16_CONTAINED = FWD | UTF16 | CONTAINED,
1.51 + FWD_UTF16_NOT_CONTAINED = FWD | UTF16 | NOT_CONTAINED,
1.52 + FWD_UTF8_CONTAINED = FWD | UTF8 | CONTAINED,
1.53 + FWD_UTF8_NOT_CONTAINED = FWD | UTF8 | NOT_CONTAINED,
1.54 + BACK_UTF16_CONTAINED = BACK | UTF16 | CONTAINED,
1.55 + BACK_UTF16_NOT_CONTAINED= BACK | UTF16 | NOT_CONTAINED,
1.56 + BACK_UTF8_CONTAINED = BACK | UTF8 | CONTAINED,
1.57 + BACK_UTF8_NOT_CONTAINED = BACK | UTF8 | NOT_CONTAINED
1.58 + };
1.59 +
1.60 + UnicodeSetStringSpan(const UnicodeSet &set, const UVector &setStrings, uint32_t which);
1.61 +
1.62 + // Copy constructor. Assumes which==ALL for a frozen set.
1.63 + UnicodeSetStringSpan(const UnicodeSetStringSpan &otherStringSpan, const UVector &newParentSetStrings);
1.64 +
1.65 + ~UnicodeSetStringSpan();
1.66 +
1.67 + /*
1.68 + * Do the strings need to be checked in span() etc.?
1.69 + * @return TRUE if strings need to be checked (call span() here),
1.70 + * FALSE if not (use a BMPSet for best performance).
1.71 + */
1.72 + inline UBool needsStringSpanUTF16();
1.73 + inline UBool needsStringSpanUTF8();
1.74 +
1.75 + // For fast UnicodeSet::contains(c).
1.76 + inline UBool contains(UChar32 c) const;
1.77 +
1.78 + int32_t span(const UChar *s, int32_t length, USetSpanCondition spanCondition) const;
1.79 +
1.80 + int32_t spanBack(const UChar *s, int32_t length, USetSpanCondition spanCondition) const;
1.81 +
1.82 + int32_t spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const;
1.83 +
1.84 + int32_t spanBackUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const;
1.85 +
1.86 +private:
1.87 + // Special spanLength byte values.
1.88 + enum {
1.89 + // The spanLength is >=0xfe.
1.90 + LONG_SPAN=0xfe,
1.91 + // All code points in the string are contained in the parent set.
1.92 + ALL_CP_CONTAINED=0xff
1.93 + };
1.94 +
1.95 + // Add a starting or ending string character to the spanNotSet
1.96 + // so that a character span ends before any string.
1.97 + void addToSpanNotSet(UChar32 c);
1.98 +
1.99 + int32_t spanNot(const UChar *s, int32_t length) const;
1.100 + int32_t spanNotBack(const UChar *s, int32_t length) const;
1.101 + int32_t spanNotUTF8(const uint8_t *s, int32_t length) const;
1.102 + int32_t spanNotBackUTF8(const uint8_t *s, int32_t length) const;
1.103 +
1.104 + // Set for span(). Same as parent but without strings.
1.105 + UnicodeSet spanSet;
1.106 +
1.107 + // Set for span(not contained).
1.108 + // Same as spanSet, plus characters that start or end strings.
1.109 + UnicodeSet *pSpanNotSet;
1.110 +
1.111 + // The strings of the parent set.
1.112 + const UVector &strings;
1.113 +
1.114 + // Pointer to the UTF-8 string lengths.
1.115 + // Also pointer to further allocated storage for meta data and
1.116 + // UTF-8 string contents as necessary.
1.117 + int32_t *utf8Lengths;
1.118 +
1.119 + // Pointer to the part of the (utf8Lengths) memory block that stores
1.120 + // the lengths of span(), spanBack() etc. for each string.
1.121 + uint8_t *spanLengths;
1.122 +
1.123 + // Pointer to the part of the (utf8Lengths) memory block that stores
1.124 + // the UTF-8 versions of the parent set's strings.
1.125 + uint8_t *utf8;
1.126 +
1.127 + // Number of bytes for all UTF-8 versions of strings together.
1.128 + int32_t utf8Length;
1.129 +
1.130 + // Maximum lengths of relevant strings.
1.131 + int32_t maxLength16;
1.132 + int32_t maxLength8;
1.133 +
1.134 + // Set up for all variants of span()?
1.135 + UBool all;
1.136 +
1.137 + // Memory for small numbers and lengths of strings.
1.138 + // For example, for 8 strings:
1.139 + // 8 UTF-8 lengths, 8*4 bytes span lengths, 8*2 3-byte UTF-8 characters
1.140 + // = 112 bytes = int32_t[28].
1.141 + int32_t staticLengths[32];
1.142 +};
1.143 +
1.144 +UBool UnicodeSetStringSpan::needsStringSpanUTF16() {
1.145 + return (UBool)(maxLength16!=0);
1.146 +}
1.147 +
1.148 +UBool UnicodeSetStringSpan::needsStringSpanUTF8() {
1.149 + return (UBool)(maxLength8!=0);
1.150 +}
1.151 +
1.152 +UBool UnicodeSetStringSpan::contains(UChar32 c) const {
1.153 + return spanSet.contains(c);
1.154 +}
1.155 +
1.156 +U_NAMESPACE_END
1.157 +
1.158 +#endif
