michael@0: /*
michael@0:  * Copyright (C) 2012 Apple Inc. All rights reserved.
michael@0:  *
michael@0:  * Redistribution and use in source and binary forms, with or without
michael@0:  * modification, are permitted provided that the following conditions
michael@0:  * are met:
michael@0:  * 1. Redistributions of source code must retain the above copyright
michael@0:  *    notice, this list of conditions and the following disclaimer.
michael@0:  * 2. Redistributions in binary form must reproduce the above copyright
michael@0:  *    notice, this list of conditions and the following disclaimer in the
michael@0:  *    documentation and/or other materials provided with the distribution.
michael@0:  *
michael@0:  * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY
michael@0:  * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
michael@0:  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
michael@0:  * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE INC. OR
michael@0:  * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
michael@0:  * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
michael@0:  * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
michael@0:  * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY
michael@0:  * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
michael@0:  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
michael@0:  * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 
michael@0:  */
michael@0: 
michael@0: // See ES 5.1, 15.10.2.8
michael@0: function canonicalize(ch)
michael@0: {
michael@0:     var u = String.fromCharCode(ch).toUpperCase();
michael@0:     if (u.length > 1)
michael@0:         return ch;
michael@0:     var cu = u.charCodeAt(0);
michael@0:     if (ch >= 128 && cu < 128)
michael@0:         return ch;
michael@0:     return cu;
michael@0: }
michael@0: 
michael@0: var MAX_UCS2 = 0xFFFF;
michael@0: var MAX_LATIN = 0xFF;
michael@0: 
michael@0: var groupedCanonically = [];
michael@0: // Pass 1: populate groupedCanonically - this is mapping from canonicalized
michael@0: // values back to the set of character code that canonicalize to them.
michael@0: for (var i = 0; i <= MAX_UCS2; ++i) {
michael@0:     var ch = canonicalize(i);
michael@0:     if (!groupedCanonically[ch])
michael@0:         groupedCanonically[ch] = [];
michael@0:     groupedCanonically[ch].push(i);
michael@0: }
michael@0: 
michael@0: var typeInfo = [];
michael@0: var latinTypeInfo = [];
michael@0: var characterSetInfo = [];
michael@0: // Pass 2: populate typeInfo & characterSetInfo. For every character calculate
michael@0: // a typeInfo value, described by the types above, and a value payload.
michael@0: for (cu in groupedCanonically) {
michael@0:     // The set of characters that canonicalize to cu
michael@0:     var characters = groupedCanonically[cu];
michael@0: 
michael@0:     // If there is only one, it is unique.
michael@0:     if (characters.length == 1) {
michael@0:         typeInfo[characters[0]] = "CanonicalizeUnique:0";
michael@0:         latinTypeInfo[characters[0]] = characters[0] <= MAX_LATIN ? "CanonicalizeLatinSelf:0" : "CanonicalizeLatinInvalid:0";
michael@0:         continue;
michael@0:     }
michael@0: 
michael@0:     // Sort the array.
michael@0:     characters.sort(function(x,y){return x-y;});
michael@0: 
michael@0:     // If there are more than two characters, create an entry in characterSetInfo.
michael@0:     if (characters.length > 2) {
michael@0:         for (i in characters)
michael@0:             typeInfo[characters[i]] = "CanonicalizeSet:" + characterSetInfo.length;
michael@0:         characterSetInfo.push(characters);
michael@0: 
michael@0:         if (characters[1] <= MAX_LATIN)
michael@0:             throw new Error("sets with more than one latin character not supported!");
michael@0:         if (characters[0] <= MAX_LATIN) {
michael@0:             for (i in characters)
michael@0:                 latinTypeInfo[characters[i]] = "CanonicalizeLatinOther:" + characters[0];
michael@0:             latinTypeInfo[characters[0]] = "CanonicalizeLatinSelf:0";
michael@0:         } else {
michael@0:             for (i in characters)
michael@0:                 latinTypeInfo[characters[i]] = "CanonicalizeLatinInvalid:0";
michael@0:         }
michael@0: 
michael@0:         continue;
michael@0:     }
michael@0: 
michael@0:     // We have a pair, mark alternating ranges, otherwise track whether this is the low or high partner.
michael@0:     var lo = characters[0];
michael@0:     var hi = characters[1];
michael@0:     var delta = hi - lo;
michael@0:     if (delta == 1) {
michael@0:         var type = lo & 1 ? "CanonicalizeAlternatingUnaligned:0" : "CanonicalizeAlternatingAligned:0";
michael@0:         typeInfo[lo] = type;
michael@0:         typeInfo[hi] = type;
michael@0:     } else {
michael@0:         typeInfo[lo] = "CanonicalizeRangeLo:" + delta;
michael@0:         typeInfo[hi] = "CanonicalizeRangeHi:" + delta;
michael@0:     }
michael@0: 
michael@0:     if (lo > MAX_LATIN) {
michael@0:         latinTypeInfo[lo] = "CanonicalizeLatinInvalid:0"; 
michael@0:         latinTypeInfo[hi] = "CanonicalizeLatinInvalid:0";
michael@0:     } else if (hi > MAX_LATIN) {
michael@0:         latinTypeInfo[lo] = "CanonicalizeLatinSelf:0"; 
michael@0:         latinTypeInfo[hi] = "CanonicalizeLatinOther:" + lo;
michael@0:     } else {
michael@0:         if (delta != 0x20 || lo & 0x20)
michael@0:             throw new Error("pairs of latin characters that don't mask with 0x20 not supported!");
michael@0:         latinTypeInfo[lo] = "CanonicalizeLatinMask0x20:0";
michael@0:         latinTypeInfo[hi] = "CanonicalizeLatinMask0x20:0";
michael@0:     }
michael@0: }
michael@0: 
michael@0: var rangeInfo = [];
michael@0: // Pass 3: coallesce types into ranges.
michael@0: for (var end = 0; end <= MAX_UCS2; ++end) {
michael@0:     var begin = end;
michael@0:     var type = typeInfo[end];
michael@0:     while (end < MAX_UCS2 && typeInfo[end + 1] == type)
michael@0:         ++end;
michael@0:     rangeInfo.push({begin:begin, end:end, type:type});
michael@0: }
michael@0: 
michael@0: var latinRangeInfo = [];
michael@0: // Pass 4: coallesce latin-1 types into ranges.
michael@0: for (var end = 0; end <= MAX_UCS2; ++end) {
michael@0:     var begin = end;
michael@0:     var type = latinTypeInfo[end];
michael@0:     while (end < MAX_UCS2 && latinTypeInfo[end + 1] == type)
michael@0:         ++end;
michael@0:     latinRangeInfo.push({begin:begin, end:end, type:type});
michael@0: }
michael@0: 
michael@0: 
michael@0: // Helper function to convert a number to a fixed width hex representation of a C uint16_t.
michael@0: function hex(x)
michael@0: {
michael@0:     var s = Number(x).toString(16);
michael@0:     while (s.length < 4)
michael@0:         s = 0 + s;
michael@0:     return "0x" + s + "u";
michael@0: }
michael@0: 
michael@0: var copyright = (
michael@0:     "/*"                                                                            + "\n" +
michael@0:     " * Copyright (C) 2012 Apple Inc. All rights reserved."                         + "\n" +
michael@0:     " *"                                                                            + "\n" +
michael@0:     " * Redistribution and use in source and binary forms, with or without"         + "\n" +
michael@0:     " * modification, are permitted provided that the following conditions"         + "\n" +
michael@0:     " * are met:"                                                                   + "\n" +
michael@0:     " * 1. Redistributions of source code must retain the above copyright"          + "\n" +
michael@0:     " *    notice, this list of conditions and the following disclaimer."           + "\n" +
michael@0:     " * 2. Redistributions in binary form must reproduce the above copyright"       + "\n" +
michael@0:     " *    notice, this list of conditions and the following disclaimer in the"     + "\n" +
michael@0:     " *    documentation and/or other materials provided with the distribution."    + "\n" +
michael@0:     " *"                                                                            + "\n" +
michael@0:     " * THIS SOFTWARE IS PROVIDED BY APPLE INC. ``AS IS'' AND ANY"                  + "\n" +
michael@0:     " * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE"          + "\n" +
michael@0:     " * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR"         + "\n" +
michael@0:     " * PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL APPLE INC. OR"                   + "\n" +
michael@0:     " * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,"      + "\n" +
michael@0:     " * EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,"        + "\n" +
michael@0:     " * PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR"         + "\n" +
michael@0:     " * PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY"        + "\n" +
michael@0:     " * OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT"               + "\n" +
michael@0:     " * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE"      + "\n" +
michael@0:     " * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. "      + "\n" +
michael@0:     " */");
michael@0: 
michael@0: print(copyright);
michael@0: print();
michael@0: print("// DO NOT EDIT! - this file autogenerated by YarrCanonicalizeUCS2.js");
michael@0: print();
michael@0: print('#include "yarr/YarrCanonicalizeUCS2.h"');
michael@0: print();
michael@0: print('#include <stddef.h>');
michael@0: print();
michael@0: print("namespace JSC { namespace Yarr {");
michael@0: print();
michael@0: 
michael@0: for (i in characterSetInfo) {
michael@0:     var characters = ""
michael@0:     var set = characterSetInfo[i];
michael@0:     for (var j in set)
michael@0:         characters += hex(set[j]) + ", ";
michael@0:     print("uint16_t ucs2CharacterSet" + i + "[] = { " + characters + "0 };");
michael@0: }
michael@0: print();
michael@0: print("static const size_t UCS2_CANONICALIZATION_SETS = " + characterSetInfo.length + ";");
michael@0: print("uint16_t* characterSetInfo[UCS2_CANONICALIZATION_SETS] = {");
michael@0: for (i in characterSetInfo)
michael@0: print("    ucs2CharacterSet" + i + ",");
michael@0: print("};");
michael@0: print();
michael@0: print("const size_t UCS2_CANONICALIZATION_RANGES = " + rangeInfo.length + ";");
michael@0: print("UCS2CanonicalizationRange rangeInfo[UCS2_CANONICALIZATION_RANGES] = {");
michael@0: for (i in rangeInfo) {
michael@0:     var info = rangeInfo[i];
michael@0:     var typeAndValue = info.type.split(':');
michael@0:     print("    { " + hex(info.begin) + ", " + hex(info.end) + ", " + hex(typeAndValue[1]) + ", " + typeAndValue[0] + " },");
michael@0: }
michael@0: print("};");
michael@0: print();
michael@0: print("const size_t LATIN_CANONICALIZATION_RANGES = " + latinRangeInfo.length + ";");
michael@0: print("LatinCanonicalizationRange latinRangeInfo[LATIN_CANONICALIZATION_RANGES] = {");
michael@0: for (i in latinRangeInfo) {
michael@0:     var info = latinRangeInfo[i];
michael@0:     var typeAndValue = info.type.split(':');
michael@0:     print("    { " + hex(info.begin) + ", " + hex(info.end) + ", " + hex(typeAndValue[1]) + ", " + typeAndValue[0] + " },");
michael@0: }
michael@0: print("};");
michael@0: print();
michael@0: print("} } // JSC::Yarr");
michael@0: print();
michael@0: