intl/icu/source/common/unisetspan.h@6474c204b198 (annotated)
intl/icu/source/common/unisetspan.h
Wed, 31 Dec 2014 06:09:35 +0100
- author
- Michael Schloh von Bennewitz <michael@schloh.com>
- date
- Wed, 31 Dec 2014 06:09:35 +0100
- changeset 0
- 6474c204b198
- permissions
- -rw-r--r--
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
| michael@0 | 1 | /* |
| michael@0 | 2 | ****************************************************************************** |
| michael@0 | 3 | * |
| michael@0 | 4 | * Copyright (C) 2007, International Business Machines |
| michael@0 | 5 | * Corporation and others. All Rights Reserved. |
| michael@0 | 6 | * |
| michael@0 | 7 | ****************************************************************************** |
| michael@0 | 8 | * file name: unisetspan.h |
| michael@0 | 9 | * encoding: US-ASCII |
| michael@0 | 10 | * tab size: 8 (not used) |
| michael@0 | 11 | * indentation:4 |
| michael@0 | 12 | * |
| michael@0 | 13 | * created on: 2007mar01 |
| michael@0 | 14 | * created by: Markus W. Scherer |
| michael@0 | 15 | */ |
| michael@0 | 16 | |
| michael@0 | 17 | #ifndef __UNISETSPAN_H__ |
| michael@0 | 18 | #define __UNISETSPAN_H__ |
| michael@0 | 19 | |
| michael@0 | 20 | #include "unicode/utypes.h" |
| michael@0 | 21 | #include "unicode/uniset.h" |
| michael@0 | 22 | |
| michael@0 | 23 | U_NAMESPACE_BEGIN |
| michael@0 | 24 | |
| michael@0 | 25 | /* |
| michael@0 | 26 | * Implement span() etc. for a set with strings. |
| michael@0 | 27 | * Avoid recursion because of its exponential complexity. |
| michael@0 | 28 | * Instead, try multiple paths at once and track them with an IndexList. |
| michael@0 | 29 | */ |
| michael@0 | 30 | class UnicodeSetStringSpan : public UMemory { |
| michael@0 | 31 | public: |
| michael@0 | 32 | /* |
| michael@0 | 33 | * Which span() variant will be used? |
| michael@0 | 34 | * The object is either built for one variant and used once, |
| michael@0 | 35 | * or built for all and may be used many times. |
| michael@0 | 36 | */ |
| michael@0 | 37 | enum { |
| michael@0 | 38 | FWD = 0x20, |
| michael@0 | 39 | BACK = 0x10, |
| michael@0 | 40 | UTF16 = 8, |
| michael@0 | 41 | UTF8 = 4, |
| michael@0 | 42 | CONTAINED = 2, |
| michael@0 | 43 | NOT_CONTAINED = 1, |
| michael@0 | 44 | |
| michael@0 | 45 | ALL = 0x3f, |
| michael@0 | 46 | |
| michael@0 | 47 | FWD_UTF16_CONTAINED = FWD | UTF16 | CONTAINED, |
| michael@0 | 48 | FWD_UTF16_NOT_CONTAINED = FWD | UTF16 | NOT_CONTAINED, |
| michael@0 | 49 | FWD_UTF8_CONTAINED = FWD | UTF8 | CONTAINED, |
| michael@0 | 50 | FWD_UTF8_NOT_CONTAINED = FWD | UTF8 | NOT_CONTAINED, |
| michael@0 | 51 | BACK_UTF16_CONTAINED = BACK | UTF16 | CONTAINED, |
| michael@0 | 52 | BACK_UTF16_NOT_CONTAINED= BACK | UTF16 | NOT_CONTAINED, |
| michael@0 | 53 | BACK_UTF8_CONTAINED = BACK | UTF8 | CONTAINED, |
| michael@0 | 54 | BACK_UTF8_NOT_CONTAINED = BACK | UTF8 | NOT_CONTAINED |
| michael@0 | 55 | }; |
| michael@0 | 56 | |
| michael@0 | 57 | UnicodeSetStringSpan(const UnicodeSet &set, const UVector &setStrings, uint32_t which); |
| michael@0 | 58 | |
| michael@0 | 59 | // Copy constructor. Assumes which==ALL for a frozen set. |
| michael@0 | 60 | UnicodeSetStringSpan(const UnicodeSetStringSpan &otherStringSpan, const UVector &newParentSetStrings); |
| michael@0 | 61 | |
| michael@0 | 62 | ~UnicodeSetStringSpan(); |
| michael@0 | 63 | |
| michael@0 | 64 | /* |
| michael@0 | 65 | * Do the strings need to be checked in span() etc.? |
| michael@0 | 66 | * @return TRUE if strings need to be checked (call span() here), |
| michael@0 | 67 | * FALSE if not (use a BMPSet for best performance). |
| michael@0 | 68 | */ |
| michael@0 | 69 | inline UBool needsStringSpanUTF16(); |
| michael@0 | 70 | inline UBool needsStringSpanUTF8(); |
| michael@0 | 71 | |
| michael@0 | 72 | // For fast UnicodeSet::contains(c). |
| michael@0 | 73 | inline UBool contains(UChar32 c) const; |
| michael@0 | 74 | |
| michael@0 | 75 | int32_t span(const UChar *s, int32_t length, USetSpanCondition spanCondition) const; |
| michael@0 | 76 | |
| michael@0 | 77 | int32_t spanBack(const UChar *s, int32_t length, USetSpanCondition spanCondition) const; |
| michael@0 | 78 | |
| michael@0 | 79 | int32_t spanUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const; |
| michael@0 | 80 | |
| michael@0 | 81 | int32_t spanBackUTF8(const uint8_t *s, int32_t length, USetSpanCondition spanCondition) const; |
| michael@0 | 82 | |
| michael@0 | 83 | private: |
| michael@0 | 84 | // Special spanLength byte values. |
| michael@0 | 85 | enum { |
| michael@0 | 86 | // The spanLength is >=0xfe. |
| michael@0 | 87 | LONG_SPAN=0xfe, |
| michael@0 | 88 | // All code points in the string are contained in the parent set. |
| michael@0 | 89 | ALL_CP_CONTAINED=0xff |
| michael@0 | 90 | }; |
| michael@0 | 91 | |
| michael@0 | 92 | // Add a starting or ending string character to the spanNotSet |
| michael@0 | 93 | // so that a character span ends before any string. |
| michael@0 | 94 | void addToSpanNotSet(UChar32 c); |
| michael@0 | 95 | |
| michael@0 | 96 | int32_t spanNot(const UChar *s, int32_t length) const; |
| michael@0 | 97 | int32_t spanNotBack(const UChar *s, int32_t length) const; |
| michael@0 | 98 | int32_t spanNotUTF8(const uint8_t *s, int32_t length) const; |
| michael@0 | 99 | int32_t spanNotBackUTF8(const uint8_t *s, int32_t length) const; |
| michael@0 | 100 | |
| michael@0 | 101 | // Set for span(). Same as parent but without strings. |
| michael@0 | 102 | UnicodeSet spanSet; |
| michael@0 | 103 | |
| michael@0 | 104 | // Set for span(not contained). |
| michael@0 | 105 | // Same as spanSet, plus characters that start or end strings. |
| michael@0 | 106 | UnicodeSet *pSpanNotSet; |
| michael@0 | 107 | |
| michael@0 | 108 | // The strings of the parent set. |
| michael@0 | 109 | const UVector &strings; |
| michael@0 | 110 | |
| michael@0 | 111 | // Pointer to the UTF-8 string lengths. |
| michael@0 | 112 | // Also pointer to further allocated storage for meta data and |
| michael@0 | 113 | // UTF-8 string contents as necessary. |
| michael@0 | 114 | int32_t *utf8Lengths; |
| michael@0 | 115 | |
| michael@0 | 116 | // Pointer to the part of the (utf8Lengths) memory block that stores |
| michael@0 | 117 | // the lengths of span(), spanBack() etc. for each string. |
| michael@0 | 118 | uint8_t *spanLengths; |
| michael@0 | 119 | |
| michael@0 | 120 | // Pointer to the part of the (utf8Lengths) memory block that stores |
| michael@0 | 121 | // the UTF-8 versions of the parent set's strings. |
| michael@0 | 122 | uint8_t *utf8; |
| michael@0 | 123 | |
| michael@0 | 124 | // Number of bytes for all UTF-8 versions of strings together. |
| michael@0 | 125 | int32_t utf8Length; |
| michael@0 | 126 | |
| michael@0 | 127 | // Maximum lengths of relevant strings. |
| michael@0 | 128 | int32_t maxLength16; |
| michael@0 | 129 | int32_t maxLength8; |
| michael@0 | 130 | |
| michael@0 | 131 | // Set up for all variants of span()? |
| michael@0 | 132 | UBool all; |
| michael@0 | 133 | |
| michael@0 | 134 | // Memory for small numbers and lengths of strings. |
| michael@0 | 135 | // For example, for 8 strings: |
| michael@0 | 136 | // 8 UTF-8 lengths, 8*4 bytes span lengths, 8*2 3-byte UTF-8 characters |
| michael@0 | 137 | // = 112 bytes = int32_t[28]. |
| michael@0 | 138 | int32_t staticLengths[32]; |
| michael@0 | 139 | }; |
| michael@0 | 140 | |
| michael@0 | 141 | UBool UnicodeSetStringSpan::needsStringSpanUTF16() { |
| michael@0 | 142 | return (UBool)(maxLength16!=0); |
| michael@0 | 143 | } |
| michael@0 | 144 | |
| michael@0 | 145 | UBool UnicodeSetStringSpan::needsStringSpanUTF8() { |
| michael@0 | 146 | return (UBool)(maxLength8!=0); |
| michael@0 | 147 | } |
| michael@0 | 148 | |
| michael@0 | 149 | UBool UnicodeSetStringSpan::contains(UChar32 c) const { |
| michael@0 | 150 | return spanSet.contains(c); |
| michael@0 | 151 | } |
| michael@0 | 152 | |
| michael@0 | 153 | U_NAMESPACE_END |
| michael@0 | 154 | |
| michael@0 | 155 | #endif |
