1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/js/src/yarr/YarrCanonicalizeUCS2.h Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,141 @@
1.4 +/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 4 -*-
1.5 + * vim: set ts=8 sts=4 et sw=4 tw=99:
1.6 + *
1.7 + * Copyright (C) 2012 Apple Inc. All rights reserved.
1.8 + *
1.9 + * Redistribution and use in source and binary forms, with or without
1.10 + * modification, are permitted provided that the following conditions
1.11 + * are met:
1.12 + * 1. Redistributions of source code must retain the above copyright
1.13 + * notice, this list of conditions and the following disclaimer.
1.14 + * 2. Redistributions in binary form must reproduce the above copyright
1.15 + * notice, this list of conditions and the following disclaimer in the
1.16 + * documentation and/or other materials provided with the distribution.
1.17 + *
1.18 + * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY
1.19 + * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
1.20 + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
1.21 + * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR
1.22 + * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
1.23 + * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
1.24 + * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
1.25 + * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
1.26 + * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
1.27 + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
1.28 + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
1.29 + */
1.30 +
1.31 +#ifndef yarr_YarrCanonicalizeUCS2_h
1.32 +#define yarr_YarrCanonicalizeUCS2_h
1.33 +
1.34 +#include <stdint.h>
1.35 +
1.36 +#include "yarr/wtfbridge.h"
1.37 +
1.38 +namespace JSC { namespace Yarr {
1.39 +
1.40 +// This set of data (autogenerated using YarrCanonicalizeUCS2.js into YarrCanonicalizeUCS2.cpp)
1.41 +// provides information for each UCS2 code point as to the set of code points that it should
1.42 +// match under the ES5.1 case insensitive RegExp matching rules, specified in 15.10.2.8.
1.43 +enum UCS2CanonicalizationType {
1.44 + CanonicalizeUnique, // No canonically equal values, e.g. 0x0.
1.45 + CanonicalizeSet, // Value indicates a set in characterSetInfo.
1.46 + CanonicalizeRangeLo, // Value is positive delta to pair, E.g. 0x41 has value 0x20, -> 0x61.
1.47 + CanonicalizeRangeHi, // Value is positive delta to pair, E.g. 0x61 has value 0x20, -> 0x41.
1.48 + CanonicalizeAlternatingAligned, // Aligned consequtive pair, e.g. 0x1f4,0x1f5.
1.49 + CanonicalizeAlternatingUnaligned // Unaligned consequtive pair, e.g. 0x241,0x242.
1.50 +};
1.51 +struct UCS2CanonicalizationRange { uint16_t begin, end, value, type; };
1.52 +extern const size_t UCS2_CANONICALIZATION_RANGES;
1.53 +extern const uint16_t* const characterSetInfo[];
1.54 +extern const UCS2CanonicalizationRange rangeInfo[];
1.55 +
1.56 +// This table is similar to the full rangeInfo table, however this maps from UCS2 codepoints to
1.57 +// the set of Latin1 codepoints that could match.
1.58 +enum LatinCanonicalizationType {
1.59 + CanonicalizeLatinSelf, // This character is in the Latin1 range, but has no canonical equivalent in the range.
1.60 + CanonicalizeLatinMask0x20, // One of a pair of characters, under the mask 0x20.
1.61 + CanonicalizeLatinOther, // This character is not in the Latin1 range, but canonicalizes to another that is.
1.62 + CanonicalizeLatinInvalid // Cannot match against Latin1 input.
1.63 +};
1.64 +struct LatinCanonicalizationRange { uint16_t begin, end, value, type; };
1.65 +extern const size_t LATIN_CANONICALIZATION_RANGES;
1.66 +extern const LatinCanonicalizationRange latinRangeInfo[];
1.67 +
1.68 +// This searches in log2 time over ~364 entries, so should typically result in 8 compares.
1.69 +inline const UCS2CanonicalizationRange* rangeInfoFor(UChar ch)
1.70 +{
1.71 + const UCS2CanonicalizationRange* info = rangeInfo;
1.72 + size_t entries = UCS2_CANONICALIZATION_RANGES;
1.73 +
1.74 + while (true) {
1.75 + size_t candidate = entries >> 1;
1.76 + const UCS2CanonicalizationRange* candidateInfo = info + candidate;
1.77 + if (ch < candidateInfo->begin)
1.78 + entries = candidate;
1.79 + else if (ch <= candidateInfo->end)
1.80 + return candidateInfo;
1.81 + else {
1.82 + info = candidateInfo + 1;
1.83 + entries -= (candidate + 1);
1.84 + }
1.85 + }
1.86 +}
1.87 +
1.88 +// Should only be called for characters that have one canonically matching value.
1.89 +inline UChar getCanonicalPair(const UCS2CanonicalizationRange* info, UChar ch)
1.90 +{
1.91 + ASSERT(ch >= info->begin && ch <= info->end);
1.92 + switch (info->type) {
1.93 + case CanonicalizeRangeLo:
1.94 + return ch + info->value;
1.95 + case CanonicalizeRangeHi:
1.96 + return ch - info->value;
1.97 + case CanonicalizeAlternatingAligned:
1.98 + return ch ^ 1;
1.99 + case CanonicalizeAlternatingUnaligned:
1.100 + return ((ch - 1) ^ 1) + 1;
1.101 + default:
1.102 + ASSERT_NOT_REACHED();
1.103 + }
1.104 + ASSERT_NOT_REACHED();
1.105 + return 0;
1.106 +}
1.107 +
1.108 +// Returns true if no other UCS2 codepoint can match this value.
1.109 +inline bool isCanonicallyUnique(UChar ch)
1.110 +{
1.111 + return rangeInfoFor(ch)->type == CanonicalizeUnique;
1.112 +}
1.113 +
1.114 +// Returns true if values are equal, under the canonicalization rules.
1.115 +inline bool areCanonicallyEquivalent(UChar a, UChar b)
1.116 +{
1.117 + const UCS2CanonicalizationRange* info = rangeInfoFor(a);
1.118 + switch (info->type) {
1.119 + case CanonicalizeUnique:
1.120 + return a == b;
1.121 + case CanonicalizeSet: {
1.122 + for (const uint16_t* set = characterSetInfo[info->value]; (a = *set); ++set) {
1.123 + if (a == b)
1.124 + return true;
1.125 + }
1.126 + return false;
1.127 + }
1.128 + case CanonicalizeRangeLo:
1.129 + return (a == b) || (a + info->value == b);
1.130 + case CanonicalizeRangeHi:
1.131 + return (a == b) || (a - info->value == b);
1.132 + case CanonicalizeAlternatingAligned:
1.133 + return (a | 1) == (b | 1);
1.134 + case CanonicalizeAlternatingUnaligned:
1.135 + return ((a - 1) | 1) == ((b - 1) | 1);
1.136 + }
1.137 +
1.138 + ASSERT_NOT_REACHED();
1.139 + return false;
1.140 +}
1.141 +
1.142 +} } // JSC::Yarr
1.143 +
1.144 +#endif /* yarr_YarrCanonicalizeUCS2_h */
