Wed, 31 Dec 2014 06:09:35 +0100
Cloned upstream origin tor-browser at tor-browser-31.3.0esr-4.5-1-build1
revision ID fc1c9ff7c1b2defdbc039f12214767608f46423f for hacking purpose.
michael@0 | 1 | /* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*- |
michael@0 | 2 | * vim: set ts=8 sts=4 et sw=4 tw=99: |
michael@0 | 3 | * |
michael@0 | 4 | * Copyright (C) 2012 Apple Inc. All rights reserved. |
michael@0 | 5 | * |
michael@0 | 6 | * Redistribution and use in source and binary forms, with or without |
michael@0 | 7 | * modification, are permitted provided that the following conditions |
michael@0 | 8 | * are met: |
michael@0 | 9 | * 1. Redistributions of source code must retain the above copyright |
michael@0 | 10 | * notice, this list of conditions and the following disclaimer. |
michael@0 | 11 | * 2. Redistributions in binary form must reproduce the above copyright |
michael@0 | 12 | * notice, this list of conditions and the following disclaimer in the |
michael@0 | 13 | * documentation and/or other materials provided with the distribution. |
michael@0 | 14 | * |
michael@0 | 15 | * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY |
michael@0 | 16 | * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE |
michael@0 | 17 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR |
michael@0 | 18 | * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR |
michael@0 | 19 | * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, |
michael@0 | 20 | * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, |
michael@0 | 21 | * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR |
michael@0 | 22 | * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY |
michael@0 | 23 | * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT |
michael@0 | 24 | * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE |
michael@0 | 25 | * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. |
michael@0 | 26 | */ |
michael@0 | 27 | |
michael@0 | 28 | #ifndef yarr_YarrCanonicalizeUCS2_h |
michael@0 | 29 | #define yarr_YarrCanonicalizeUCS2_h |
michael@0 | 30 | |
michael@0 | 31 | #include <stdint.h> |
michael@0 | 32 | |
michael@0 | 33 | #include "yarr/wtfbridge.h" |
michael@0 | 34 | |
michael@0 | 35 | namespace JSC { namespace Yarr { |
michael@0 | 36 | |
michael@0 | 37 | // This set of data (autogenerated using YarrCanonicalizeUCS2.js into YarrCanonicalizeUCS2.cpp) |
michael@0 | 38 | // provides information for each UCS2 code point as to the set of code points that it should |
michael@0 | 39 | // match under the ES5.1 case insensitive RegExp matching rules, specified in 15.10.2.8. |
michael@0 | 40 | enum UCS2CanonicalizationType { |
michael@0 | 41 | CanonicalizeUnique, // No canonically equal values, e.g. 0x0. |
michael@0 | 42 | CanonicalizeSet, // Value indicates a set in characterSetInfo. |
michael@0 | 43 | CanonicalizeRangeLo, // Value is positive delta to pair, E.g. 0x41 has value 0x20, -> 0x61. |
michael@0 | 44 | CanonicalizeRangeHi, // Value is positive delta to pair, E.g. 0x61 has value 0x20, -> 0x41. |
michael@0 | 45 | CanonicalizeAlternatingAligned, // Aligned consequtive pair, e.g. 0x1f4,0x1f5. |
michael@0 | 46 | CanonicalizeAlternatingUnaligned // Unaligned consequtive pair, e.g. 0x241,0x242. |
michael@0 | 47 | }; |
michael@0 | 48 | struct UCS2CanonicalizationRange { uint16_t begin, end, value, type; }; |
michael@0 | 49 | extern const size_t UCS2_CANONICALIZATION_RANGES; |
michael@0 | 50 | extern const uint16_t* const characterSetInfo[]; |
michael@0 | 51 | extern const UCS2CanonicalizationRange rangeInfo[]; |
michael@0 | 52 | |
michael@0 | 53 | // This table is similar to the full rangeInfo table, however this maps from UCS2 codepoints to |
michael@0 | 54 | // the set of Latin1 codepoints that could match. |
michael@0 | 55 | enum LatinCanonicalizationType { |
michael@0 | 56 | CanonicalizeLatinSelf, // This character is in the Latin1 range, but has no canonical equivalent in the range. |
michael@0 | 57 | CanonicalizeLatinMask0x20, // One of a pair of characters, under the mask 0x20. |
michael@0 | 58 | CanonicalizeLatinOther, // This character is not in the Latin1 range, but canonicalizes to another that is. |
michael@0 | 59 | CanonicalizeLatinInvalid // Cannot match against Latin1 input. |
michael@0 | 60 | }; |
michael@0 | 61 | struct LatinCanonicalizationRange { uint16_t begin, end, value, type; }; |
michael@0 | 62 | extern const size_t LATIN_CANONICALIZATION_RANGES; |
michael@0 | 63 | extern const LatinCanonicalizationRange latinRangeInfo[]; |
michael@0 | 64 | |
michael@0 | 65 | // This searches in log2 time over ~364 entries, so should typically result in 8 compares. |
michael@0 | 66 | inline const UCS2CanonicalizationRange* rangeInfoFor(UChar ch) |
michael@0 | 67 | { |
michael@0 | 68 | const UCS2CanonicalizationRange* info = rangeInfo; |
michael@0 | 69 | size_t entries = UCS2_CANONICALIZATION_RANGES; |
michael@0 | 70 | |
michael@0 | 71 | while (true) { |
michael@0 | 72 | size_t candidate = entries >> 1; |
michael@0 | 73 | const UCS2CanonicalizationRange* candidateInfo = info + candidate; |
michael@0 | 74 | if (ch < candidateInfo->begin) |
michael@0 | 75 | entries = candidate; |
michael@0 | 76 | else if (ch <= candidateInfo->end) |
michael@0 | 77 | return candidateInfo; |
michael@0 | 78 | else { |
michael@0 | 79 | info = candidateInfo + 1; |
michael@0 | 80 | entries -= (candidate + 1); |
michael@0 | 81 | } |
michael@0 | 82 | } |
michael@0 | 83 | } |
michael@0 | 84 | |
michael@0 | 85 | // Should only be called for characters that have one canonically matching value. |
michael@0 | 86 | inline UChar getCanonicalPair(const UCS2CanonicalizationRange* info, UChar ch) |
michael@0 | 87 | { |
michael@0 | 88 | ASSERT(ch >= info->begin && ch <= info->end); |
michael@0 | 89 | switch (info->type) { |
michael@0 | 90 | case CanonicalizeRangeLo: |
michael@0 | 91 | return ch + info->value; |
michael@0 | 92 | case CanonicalizeRangeHi: |
michael@0 | 93 | return ch - info->value; |
michael@0 | 94 | case CanonicalizeAlternatingAligned: |
michael@0 | 95 | return ch ^ 1; |
michael@0 | 96 | case CanonicalizeAlternatingUnaligned: |
michael@0 | 97 | return ((ch - 1) ^ 1) + 1; |
michael@0 | 98 | default: |
michael@0 | 99 | ASSERT_NOT_REACHED(); |
michael@0 | 100 | } |
michael@0 | 101 | ASSERT_NOT_REACHED(); |
michael@0 | 102 | return 0; |
michael@0 | 103 | } |
michael@0 | 104 | |
michael@0 | 105 | // Returns true if no other UCS2 codepoint can match this value. |
michael@0 | 106 | inline bool isCanonicallyUnique(UChar ch) |
michael@0 | 107 | { |
michael@0 | 108 | return rangeInfoFor(ch)->type == CanonicalizeUnique; |
michael@0 | 109 | } |
michael@0 | 110 | |
michael@0 | 111 | // Returns true if values are equal, under the canonicalization rules. |
michael@0 | 112 | inline bool areCanonicallyEquivalent(UChar a, UChar b) |
michael@0 | 113 | { |
michael@0 | 114 | const UCS2CanonicalizationRange* info = rangeInfoFor(a); |
michael@0 | 115 | switch (info->type) { |
michael@0 | 116 | case CanonicalizeUnique: |
michael@0 | 117 | return a == b; |
michael@0 | 118 | case CanonicalizeSet: { |
michael@0 | 119 | for (const uint16_t* set = characterSetInfo[info->value]; (a = *set); ++set) { |
michael@0 | 120 | if (a == b) |
michael@0 | 121 | return true; |
michael@0 | 122 | } |
michael@0 | 123 | return false; |
michael@0 | 124 | } |
michael@0 | 125 | case CanonicalizeRangeLo: |
michael@0 | 126 | return (a == b) || (a + info->value == b); |
michael@0 | 127 | case CanonicalizeRangeHi: |
michael@0 | 128 | return (a == b) || (a - info->value == b); |
michael@0 | 129 | case CanonicalizeAlternatingAligned: |
michael@0 | 130 | return (a | 1) == (b | 1); |
michael@0 | 131 | case CanonicalizeAlternatingUnaligned: |
michael@0 | 132 | return ((a - 1) | 1) == ((b - 1) | 1); |
michael@0 | 133 | } |
michael@0 | 134 | |
michael@0 | 135 | ASSERT_NOT_REACHED(); |
michael@0 | 136 | return false; |
michael@0 | 137 | } |
michael@0 | 138 | |
michael@0 | 139 | } } // JSC::Yarr |
michael@0 | 140 | |
michael@0 | 141 | #endif /* yarr_YarrCanonicalizeUCS2_h */ |