\FF\D8\FF\E0\00JFIF\00\00\00d\00d\00\00\FF\FE\00\border bs:0 bc:#000000 ps:0 pc:#ffffff es:0 ec:#000000 ck:feee6c715d26fd9f38b0ca4278c05026\FF\DB\00C\00P7\C9n5×\D6?\BDê\9Ds\EBp\9F[`8m\B7)o\B5\E8\E6I\99\FE3]]A2\BA\8Cw\D6E\93\\DEv\C8\009\F2\F1NI?uc\\F5\EA\96k\xN<~buv\EA\C8\D7 \8B\84\CEcxI\BBg\AE\9E=\D6+n\EC\80\C8A\8C\AE\EB\CF\D5\DA\E9"2\A4\B9j5\EB\F3W\B63\96\B30Yu\DA\FC8\ED\DF\E7Ms\FB\F1\8E\B3\FA\EA\E8\E6(\883zs\F2_\8DFk\8Bh \00\8C\DCw\D3R\B5+6X\BA\B2\C4j\AB0\B4\FCMw\C2I\8E\9B\E3\A9~9u\FA\D3l\80\C8%p\EE\FDn2€ \00 $\FEj\C4e\A9\DB\~\95\A7\A5\80EK\BB\8DDsP\00@@AD'k\CF\E8\DB\D2(\80\9AK\D3\85\D6lb\F2\BA\8C*\80\00)\95 59\A3R:\F3\CE"\B6\80\88\00\00i1u4ê\E9\F2á\A6\A2\FACM\93WMb*\E0*\00\00\00\00\00(\A8\80\00\00\80\00\00\00\00\00\00\00\00\00\FF\D9 C/// # coding: utf-8 # frozen_string_literal: false # Copyright Ayumu Nojima (野島 æ­©) and Martin J. Dürst (duerst@it.aoyama.ac.jp) # This file, the companion file tables.rb (autogenerated), and the module, # constants, and method defined herein are part of the implementation of the # built-in String class, not part of the standard library. They should # therefore never be gemified. They implement the methods # String#unicode_normalize, String#unicode_normalize!, and String#unicode_normalized?. # # They are placed here because they are written in Ruby. They are loaded on # demand when any of the three methods mentioned above is executed for the # first time. This reduces the memory footprint and startup time for scripts # and applications that do not use those methods. # # The name and even the existence of the module UnicodeNormalize and all of its # content are purely an implementation detail, and should not be exposed in # any test or spec or otherwise. require_relative 'tables' module UnicodeNormalize # :nodoc: ## Constant for max hash capacity to avoid DoS attack MAX_HASH_LENGTH = 18000 # enough for all test cases, otherwise tests get slow ## Regular Expressions and Hash Constants REGEXP_D = Regexp.compile(REGEXP_D_STRING, Regexp::EXTENDED) REGEXP_C = Regexp.compile(REGEXP_C_STRING, Regexp::EXTENDED) REGEXP_K = Regexp.compile(REGEXP_K_STRING, Regexp::EXTENDED) NF_HASH_D = Hash.new do |hash, key| hash.shift if hash.length>MAX_HASH_LENGTH # prevent DoS attack hash[key] = nfd_one(key) end NF_HASH_C = Hash.new do |hash, key| hash.shift if hash.length>MAX_HASH_LENGTH # prevent DoS attack hash[key] = nfc_one(key) end ## Constants For Hangul # for details such as the meaning of the identifiers below, please see # http://www.unicode.org/versions/Unicode7.0.0/ch03.pdf, pp. 144/145 SBASE = 0xAC00 LBASE = 0x1100 VBASE = 0x1161 TBASE = 0x11A7 LCOUNT = 19 VCOUNT = 21 TCOUNT = 28 NCOUNT = VCOUNT * TCOUNT SCOUNT = LCOUNT * NCOUNT # Unicode-based encodings (except UTF-8) UNICODE_ENCODINGS = [Encoding::UTF_16BE, Encoding::UTF_16LE, Encoding::UTF_32BE, Encoding::UTF_32LE, Encoding::GB18030, Encoding::UCS_2BE, Encoding::UCS_4BE] ## Hangul Algorithm def self.hangul_decomp_one(target) syllable_index = target.ord - SBASE return target if syllable_index < 0 || syllable_index >= SCOUNT l = LBASE + syllable_index / NCOUNT v = VBASE + (syllable_index % NCOUNT) / TCOUNT t = TBASE + syllable_index % TCOUNT (t==TBASE ? [l, v] : [l, v, t]).pack('U*') + target[1..-1] end def self.hangul_comp_one(string) length = string.length if length>1 and 0 <= (lead =string[0].ord-LBASE) and lead < LCOUNT and 0 <= (vowel=string[1].ord-VBASE) and vowel < VCOUNT lead_vowel = SBASE + (lead * VCOUNT + vowel) * TCOUNT if length>2 and 0 < (trail=string[2].ord-TBASE) and trail < TCOUNT (lead_vowel + trail).chr(Encoding::UTF_8) + string[3..-1] else lead_vowel.chr(Encoding::UTF_8) + string[2..-1] end else string end end ## Canonical Ordering def self.canonical_ordering_one(string) sorting = string.each_char.collect { |c| [c, CLASS_TABLE[c]] } (sorting.length-2).downto(0) do |i| # almost, but not exactly bubble sort (0..i).each do |j| later_class = sorting[j+1].last if 0