1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/js/src/yarr/YarrCanonicalizeUCS2.js Wed Dec 31 06:09:35 2014 +0100
1.3 @@ -0,0 +1,218 @@
1.4 +/*
1.5 + * Copyright (C) 2012 Apple Inc. All rights reserved.
1.6 + *
1.7 + * Redistribution and use in source and binary forms, with or without
1.8 + * modification, are permitted provided that the following conditions
1.9 + * are met:
1.10 + * 1. Redistributions of source code must retain the above copyright
1.11 + * notice, this list of conditions and the following disclaimer.
1.12 + * 2. Redistributions in binary form must reproduce the above copyright
1.13 + * notice, this list of conditions and the following disclaimer in the
1.14 + * documentation and/or other materials provided with the distribution.
1.15 + *
1.16 + * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY
1.17 + * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
1.18 + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
1.19 + * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR
1.20 + * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
1.21 + * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
1.22 + * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
1.23 + * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
1.24 + * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
1.25 + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
1.26 + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
1.27 + */
1.28 +
1.29 +// See ES 5.1, 15.10.2.8
1.30 +function canonicalize(ch)
1.31 +{
1.32 + var u = String.fromCharCode(ch).toUpperCase();
1.33 + if (u.length > 1)
1.34 + return ch;
1.35 + var cu = u.charCodeAt(0);
1.36 + if (ch >= 128 && cu < 128)
1.37 + return ch;
1.38 + return cu;
1.39 +}
1.40 +
1.41 +var MAX_UCS2 = 0xFFFF;
1.42 +var MAX_LATIN = 0xFF;
1.43 +
1.44 +var groupedCanonically = [];
1.45 +// Pass 1: populate groupedCanonically - this is mapping from canonicalized
1.46 +// values back to the set of character code that canonicalize to them.
1.47 +for (var i = 0; i <= MAX_UCS2; ++i) {
1.48 + var ch = canonicalize(i);
1.49 + if (!groupedCanonically[ch])
1.50 + groupedCanonically[ch] = [];
1.51 + groupedCanonically[ch].push(i);
1.52 +}
1.53 +
1.54 +var typeInfo = [];
1.55 +var latinTypeInfo = [];
1.56 +var characterSetInfo = [];
1.57 +// Pass 2: populate typeInfo & characterSetInfo. For every character calculate
1.58 +// a typeInfo value, described by the types above, and a value payload.
1.59 +for (cu in groupedCanonically) {
1.60 + // The set of characters that canonicalize to cu
1.61 + var characters = groupedCanonically[cu];
1.62 +
1.63 + // If there is only one, it is unique.
1.64 + if (characters.length == 1) {
1.65 + typeInfo[characters[0]] = "CanonicalizeUnique:0";
1.66 + latinTypeInfo[characters[0]] = characters[0] <= MAX_LATIN ? "CanonicalizeLatinSelf:0" : "CanonicalizeLatinInvalid:0";
1.67 + continue;
1.68 + }
1.69 +
1.70 + // Sort the array.
1.71 + characters.sort(function(x,y){return x-y;});
1.72 +
1.73 + // If there are more than two characters, create an entry in characterSetInfo.
1.74 + if (characters.length > 2) {
1.75 + for (i in characters)
1.76 + typeInfo[characters[i]] = "CanonicalizeSet:" + characterSetInfo.length;
1.77 + characterSetInfo.push(characters);
1.78 +
1.79 + if (characters[1] <= MAX_LATIN)
1.80 + throw new Error("sets with more than one latin character not supported!");
1.81 + if (characters[0] <= MAX_LATIN) {
1.82 + for (i in characters)
1.83 + latinTypeInfo[characters[i]] = "CanonicalizeLatinOther:" + characters[0];
1.84 + latinTypeInfo[characters[0]] = "CanonicalizeLatinSelf:0";
1.85 + } else {
1.86 + for (i in characters)
1.87 + latinTypeInfo[characters[i]] = "CanonicalizeLatinInvalid:0";
1.88 + }
1.89 +
1.90 + continue;
1.91 + }
1.92 +
1.93 + // We have a pair, mark alternating ranges, otherwise track whether this is the low or high partner.
1.94 + var lo = characters[0];
1.95 + var hi = characters[1];
1.96 + var delta = hi - lo;
1.97 + if (delta == 1) {
1.98 + var type = lo & 1 ? "CanonicalizeAlternatingUnaligned:0" : "CanonicalizeAlternatingAligned:0";
1.99 + typeInfo[lo] = type;
1.100 + typeInfo[hi] = type;
1.101 + } else {
1.102 + typeInfo[lo] = "CanonicalizeRangeLo:" + delta;
1.103 + typeInfo[hi] = "CanonicalizeRangeHi:" + delta;
1.104 + }
1.105 +
1.106 + if (lo > MAX_LATIN) {
1.107 + latinTypeInfo[lo] = "CanonicalizeLatinInvalid:0";
1.108 + latinTypeInfo[hi] = "CanonicalizeLatinInvalid:0";
1.109 + } else if (hi > MAX_LATIN) {
1.110 + latinTypeInfo[lo] = "CanonicalizeLatinSelf:0";
1.111 + latinTypeInfo[hi] = "CanonicalizeLatinOther:" + lo;
1.112 + } else {
1.113 + if (delta != 0x20 || lo & 0x20)
1.114 + throw new Error("pairs of latin characters that don't mask with 0x20 not supported!");
1.115 + latinTypeInfo[lo] = "CanonicalizeLatinMask0x20:0";
1.116 + latinTypeInfo[hi] = "CanonicalizeLatinMask0x20:0";
1.117 + }
1.118 +}
1.119 +
1.120 +var rangeInfo = [];
1.121 +// Pass 3: coallesce types into ranges.
1.122 +for (var end = 0; end <= MAX_UCS2; ++end) {
1.123 + var begin = end;
1.124 + var type = typeInfo[end];
1.125 + while (end < MAX_UCS2 && typeInfo[end + 1] == type)
1.126 + ++end;
1.127 + rangeInfo.push({begin:begin, end:end, type:type});
1.128 +}
1.129 +
1.130 +var latinRangeInfo = [];
1.131 +// Pass 4: coallesce latin-1 types into ranges.
1.132 +for (var end = 0; end <= MAX_UCS2; ++end) {
1.133 + var begin = end;
1.134 + var type = latinTypeInfo[end];
1.135 + while (end < MAX_UCS2 && latinTypeInfo[end + 1] == type)
1.136 + ++end;
1.137 + latinRangeInfo.push({begin:begin, end:end, type:type});
1.138 +}
1.139 +
1.140 +
1.141 +// Helper function to convert a number to a fixed width hex representation of a C uint16_t.
1.142 +function hex(x)
1.143 +{
1.144 + var s = Number(x).toString(16);
1.145 + while (s.length < 4)
1.146 + s = 0 + s;
1.147 + return "0x" + s + "u";
1.148 +}
1.149 +
1.150 +var copyright = (
1.151 + "/*" + "\n" +
1.152 + " * Copyright (C) 2012 Apple Inc. All rights reserved." + "\n" +
1.153 + " *" + "\n" +
1.154 + " * Redistribution and use in source and binary forms, with or without" + "\n" +
1.155 + " * modification, are permitted provided that the following conditions" + "\n" +
1.156 + " * are met:" + "\n" +
1.157 + " * 1. Redistributions of source code must retain the above copyright" + "\n" +
1.158 + " * notice, this list of conditions and the following disclaimer." + "\n" +
1.159 + " * 2. Redistributions in binary form must reproduce the above copyright" + "\n" +
1.160 + " * notice, this list of conditions and the following disclaimer in the" + "\n" +
1.161 + " * documentation and/or other materials provided with the distribution." + "\n" +
1.162 + " *" + "\n" +
1.163 + " * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY" + "\n" +
1.164 + " * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE" + "\n" +
1.165 + " * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR" + "\n" +
1.166 + " * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL APPLE INC. OR" + "\n" +
1.167 + " * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL," + "\n" +
1.168 + " * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO," + "\n" +
1.169 + " * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR" + "\n" +
1.170 + " * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY" + "\n" +
1.171 + " * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT" + "\n" +
1.172 + " * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE" + "\n" +
1.173 + " * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. " + "\n" +
1.174 + " */");
1.175 +
1.176 +print(copyright);
1.177 +print();
1.178 +print("// DO NOT EDIT! - this file autogenerated by YarrCanonicalizeUCS2.js");
1.179 +print();
1.180 +print('#include "yarr/YarrCanonicalizeUCS2.h"');
1.181 +print();
1.182 +print('#include <stddef.h>');
1.183 +print();
1.184 +print("namespace JSC { namespace Yarr {");
1.185 +print();
1.186 +
1.187 +for (i in characterSetInfo) {
1.188 + var characters = ""
1.189 + var set = characterSetInfo[i];
1.190 + for (var j in set)
1.191 + characters += hex(set[j]) + ", ";
1.192 + print("uint16_t ucs2CharacterSet" + i + "[] = { " + characters + "0 };");
1.193 +}
1.194 +print();
1.195 +print("static const size_t UCS2_CANONICALIZATION_SETS = " + characterSetInfo.length + ";");
1.196 +print("uint16_t* characterSetInfo[UCS2_CANONICALIZATION_SETS] = {");
1.197 +for (i in characterSetInfo)
1.198 +print(" ucs2CharacterSet" + i + ",");
1.199 +print("};");
1.200 +print();
1.201 +print("const size_t UCS2_CANONICALIZATION_RANGES = " + rangeInfo.length + ";");
1.202 +print("UCS2CanonicalizationRange rangeInfo[UCS2_CANONICALIZATION_RANGES] = {");
1.203 +for (i in rangeInfo) {
1.204 + var info = rangeInfo[i];
1.205 + var typeAndValue = info.type.split(':');
1.206 + print(" { " + hex(info.begin) + ", " + hex(info.end) + ", " + hex(typeAndValue[1]) + ", " + typeAndValue[0] + " },");
1.207 +}
1.208 +print("};");
1.209 +print();
1.210 +print("const size_t LATIN_CANONICALIZATION_RANGES = " + latinRangeInfo.length + ";");
1.211 +print("LatinCanonicalizationRange latinRangeInfo[LATIN_CANONICALIZATION_RANGES] = {");
1.212 +for (i in latinRangeInfo) {
1.213 + var info = latinRangeInfo[i];
1.214 + var typeAndValue = info.type.split(':');
1.215 + print(" { " + hex(info.begin) + ", " + hex(info.end) + ", " + hex(typeAndValue[1]) + ", " + typeAndValue[0] + " },");
1.216 +}
1.217 +print("};");
1.218 +print();
1.219 +print("} } // JSC::Yarr");
1.220 +print();
1.221 +
