icuSources/test/testdata/rbbitst.txt

   1 # Copyright (C) 2016 and later: Unicode, Inc. and others.
   2 # License & terms of use: http://www.unicode.org/copyright.html
   3 # Copyright (c) 2001-2016 International Business Machines
   4 # Corporation and others. All Rights Reserved.
   5 #
   6 # RBBI Test Data
   7 #
   8 #   File:  rbbitst.txt
   9 #
  10 #     The format of this file looks vaguely like some kind of xml-ish markup,
  11 #     but it is NOT.  The syntax is this..
  12 #
  13 #   <word>    any following data is for word break testing
  14 #   <sent>    any following data is for sentence break testing
  15 #   <line>    any following data is for line break testing
  16 #   <char>    any following data is for char break testing
  17 #   <locale local_name>  Switch to the named locale at the next occurence of <word>, <sent>, etc.
  18 #   <data> ... </data>   test data.  May span multiple lines.
  19 #      <>        Break position, status == 0
  20 #      •         Break position, status == 0   (Bullet, \u2022)
  21 #      <nnn>     Break position, status == nnn
  22 #      \         Escape.  Normal ICU unescape applied.
  23 #      \ at end of line  ->  Line Continuation.  Remove both the backslash and the new line
  24 #
  25 # In ICU4C, this test data is run by intltest, rbbi/RBBITest/TestExtended.
  26 # In ICU4J, this test data is run by com.ibm.icu.dev.test.rbbi.RBBITestExtended
  27 #
  28 # There are two copies of this file in the source repository,
  29 #   [ICU4C]   source/test/testdata/rbbitst.txt
  30 #   [ICU4J]   main/tests/core/src/com/ibm/icu/dev/test/rbbi/rbbitst.txt
  31 #
  32 # ICU4C's copy is the master.  If any changes are made to ICU4J's copy, make sure they
  33 #  are merged back into ICU4C's copy of the file, lest they get overwritten later.
  34 # TODO:  figure out how to have a single copy of the file for use by both C and Java.
  35
  36
  37 #   Temp debugging tests
  38 <locale en>
  39 <word>
  40 <data><0>１•２•３•４•</data>
  41 # <data><0>ク<400>ライアン<400>トサーバー<400></data>
  42
  43 ## FILTERED BREAK TESTS
  44
  45 # (William Bradford, public domain. http://catalog.hathitrust.org/Record/008651224 ) - edited.
  46 <locale en>
  47 <sent>
  48 <data>\
  49 •In the meantime Mr. •Weston arrived with his small ship, which he had now recovered. •Capt. •Gorges, who informed the Sgt. here that one purpose of his going east was to meet with Mr. •Weston, took this opportunity to call him to account for some abuses he had to lay to his charge.•</data>
  50
  51 <locale en@ss=standard>
  52 <sent>
  53 <data>\
  54 •In the meantime Mr. Weston arrived with his small ship, which he had now recovered. •Capt. Gorges, who informed the Sgt. here that one purpose of his going east was to meet with Mr. Weston, took this opportunity to call him to account for some abuses he had to lay to his charge.•</data>
  55
  56 # The standard ICU sample as follows uses D. but that is an abbreviation in Apple exceptions:
  57 #  This hits the case where "D." would match the end of "Ph.D.".
  58 #  •Doctor with a D. •As in, Ph.D., you know.•</data>
  59 # So we use a different example, with "C.".
  60 <locale en@ss=standard>
  61 <sent>
  62 <data>\
  63 •Sample with a C. •As in, D.C., you know.•</data>
  64
  65 # same as root (unless some exceptions are added!)
  66 <locale tfg@ss=standard>
  67 <sent>
  68 <data>\
  69 •In the meantime Mr. •Weston arrived with his small ship, which he had now recovered. •Capt. •Gorges, who informed the Sgt. here that one purpose of his going east was to meet with Mr. •Weston, took this opportunity to call him to account for some abuses he had to lay to his charge.•</data>
  70
  71 # same as root (unless some exceptions are added!)
  72 <locale ja@ss=standard>
  73 <sent>
  74 <data>\
  75 •In the meantime Mr. •Weston arrived with his small ship, which he had now recovered. •Capt. •Gorges, who informed the Sgt. here that one purpose of his going east was to meet with Mr. •Weston, took this opportunity to call him to account for some abuses he had to lay to his charge.•</data>
  76
  77 ## END FILTERED BREAK TESTS
  78
  79
  80 ########################################################################################
  81 #
  82 #
  83 #       G r a p h e m e    C l u s t e r     T e s t s
  84 #
  85 #
  86 ##########################################################################################
  87 <char>
  88
  89 <data>•a•b•c• •,•\u0666•</data>                 # Quick Test
  90 <data>•\r•\r•\r\n•\r\n•\n•\r•</data>            # don't break CR/LF
  91
  92 # Always break after controls.  Combining chars don't combine with them.
  93 <data>•\u0003•\N{COMBINING GRAVE ACCENT}•\r•\N{COMBINING GRAVE ACCENT}•</data>
  94 <data>•\u0085•\N{COMBINING MACRON}•A\N{COMBINING MACRON}•</data>
  95
  96 # Surrogates
  97 <data>•\U00011000•\U00010020•\U00010000\N{COMBINING MACRON}•</data>
  98 <data>•\ud800\udc00•\udbff\udfff•a•</data>
  99
 100 # Extend (Combining chars) combine.
 101 <data>•A\N{COMBINING GRAVE ACCENT}•B•</data>
 102 <data>•\N{GREEK SMALL LETTER MU}\N{COMBINING LOW LINE}\N{COMBINING HORN}•</data>
 103 <data>•a\u0301•b\u0302•c\u0303•d\u0304•e\u0305•f\u0306•g\u0307•h\u0308•i\u0309•</data>
 104
 105 <data>•a\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304\u0301\u0302\u0303\u0304•</data>
 106
 107 # Don't break Hangul Syllables
 108 #   L : \u1100
 109 #   V : \u1161
 110 #   T : \u11A8
 111 #  LV : \uAC00
 112 # LVT : \uAC01
 113
 114 <data>•\u1100\u1161\u11a8•\u1100\u1161\u11a8•</data>  #LVT
 115 <data>•\u1100\u1161•\u1100\u1161•</data>
 116 <data>•\u1100\u1161\u11a8•\u1161•\u1100•\u11a8•\u1161\u1161\u1161\u11a8•</data>
 117 <data>•\u1100\u1100\uac01•\u1100\uac01•\u1100\uac01\u0301•\uac01•</data>
 118 <data>•\u1100\u0301•\u1161\u11a8\u0301•\u11a8•</data>
 119
 120
 121
 122 # Hindi combining chars.  (An old test)
 123 #   TODO:  Update these tests for Unicode 5.1 Extended Grapheme clusters
 124 #<data>•भ••ा•\u0930•\u0924• •\u0938\u0941\u0902•\u0926•\u0930•
 125 #•\u0939•\u094c•\u0964•</data>
 126 #<data>•\u0916\u0947•\u0938\u0941\u0902•\u0926•\u0930•\u0939•\u094c•\u0964•</data>
 127
 128
 129 #  Bug 1587.  Tamil.  \u0baa\u0bc1 is an Extended Grpaheme Cluster
 130 <data>•\u0baa\u0bc1•\u0baa\u0bc1•</data>
 131
 132 #   Regression test for bug 1889
 133 <data>•\u0f40\u0f7d•\u0000•\u0f7e•</data>
 134
 135
 136 #     0xffff is a legal character, and should not stop the break iterator early.
 137 #     (Requires special casing in implementation, which is why it gets a test.)
 138 <data>•\uffff•\uffff• •a•</data>
 139
 140 #  Treat Japanese Half Width voicing marks as combining
 141 <data>•A\uff9e•B\uff9f\uff9e\uff9f•C•</data>
 142
 143 ########################################################################################
 144 #
 145 #
 146 #       E x t e n d e d    G r a p h e m e    C l u s t e r     T e s t s
 147 #
 148 #
 149 ##########################################################################################
 150 #<xgc>
 151
 152 # Plain Vanilla grapheme clusters
 153 #<data>•a•b•c•</data>
 154 #<data>•a\u0301\u0302• •b\u0303\u0304•</data>
 155
 156 # Assorted Hindi combining marks
 157 #<data>•\u0904\u0903• •\u0937\u093E• •\u0904\u093F• •\u0937\u0940• •\u0937\u0949• •\u0937\u094A• •\u0937\u094B• •\u0937\u094C•</data>
 158
 159 # Thai Clusters
 160 # $Prepend $Extend* $PrependBase $Extend*;
 161 #
 162 #<data>•\u0e40\u0e01•\u0e44\u0301\u0e23\u0302\u0303•\u0e40•\u0e40\u0e02•\u0e02• •</data>
 163
 164
 165 ########################################################################################
 166 #
 167 #
 168 #       W o r d    B o u n d a r y      T e s t s
 169 #
 170 #
 171 ##########################################################################################
 172
 173 <word>
 174 #
 175 # Quick sanity test
 176 #
 177 <data>•hello<200> •there<200> •goodbye<200></data>
 178 <data>•hello<200> •12345<100> •,•</data>
 179
 180
 181 #
 182 # Test data originally in RBBIAPITest::TestFirstNextFollowing() and TestLastPreviousPreceding()
 183 #
 184
 185 <word>
 186 <data>•This<200> •is<200> •a<200> •word<200> •break<200>.• • •Isn't<200> •it<200>?• •2.25<100></data>
 187
 188
 189
 190 #
 191 # Data originally from TestDefaultRuleBasedWordIteration()
 192 #
 193 <data>•Write<200> •wordrules<200>.• •123.456<100> •alpha\u00adbeta\u00adgamma<200> •\u092f\u0939<200> •</data>
 194 <data>• •\u0939\u093f\u0928\u094d\u200d\u0926\u0940<200> •\u0939\u0948<200> •\u0905\u093e\u092a<200> •\u0938\u093f\u0916\u094b\u0917\u0947<200>?•</data>
 195
 196 #Hindi Numbers
 197 <data>• •\u0968\u0966.\u0969\u096f<100> •\u0967\u0966\u0966.\u0966\u0966<100> •\N{RUPEE SIGN}•\u0967,\u0967\u0966\u0966.\u0966\u0966<100> • •\u0905\u092e\u091c<200>\n•</data>
 198
 199 <data>•\u0938\u094d\u200d\u0935\u0924\u0902deadTA\u0930<200>\r•It's<200> •$•30.10<100> •12,34<100>¢•£•¤•¥•alpha\u05f3beta\u05f4gamma<200> •</data>
 200
 201 <data>•Badges<200>?• •BADGES<200>!•?•!• •We<200> •don't<200> •need<200> •no<200> •STINKING<200> •BADGES<200>!•!•1000,233,456.000<100> •1,23.322<100>%•123.1222<100>$•123,000.20<100> •179.01<100>%•X<200> •Now<200>\r•is<200>\n•the<200>\r\n•time<200> •</data>
 202
 203 #Hangul
 204 <data>•\uc5f0\ud569<200> •\uc7a5\ub85c\uad50\ud68c<200> •\u1109\u1161\u11bc\u1112\u1161\u11bc<200> •\u1112\u1161\u11ab\u110b\u1175\u11ab<200> •Hello<200>,• •how<200> •are<200> •you<200> •</data>
 205
 206 <data>•Hello<200>,• •how<200> •are<200> •you<200> •\uc5f0\ud569<200> •\uc7a5\ub85c\uad50\ud68c<200> •\u1109\u1161\u11bc\u1112\u1161\u11bc<200> •\u1112\u1161\u11ab\u110b\u1175\u11ab<200> •</data>
 207
 208 # Words containing non-BMP letters
 209 <data>•abc\U00010300<200> •abc\N{DESERET SMALL LETTER ENG}<200> •abc\N{MATHEMATICAL BOLD SMALL Z}<200> •abc\N{MATHEMATICAL SANS-SERIF BOLD ITALIC PI SYMBOL}<200> •</data>
 210
 211 # Unassigned code points
 212 <data>•abc<200>\U0001D800•def<200>\U0001D3FF• •</data>
 213
 214 # Hiragana & Katakana stay together, but separates from each other and Latin.
 215 # *** what to do about theoretical combos of chars? i.e. hiragana + accent
 216 #<data>•abc<200>\N{HIRAGANA LETTER SMALL A}<400>\N{HIRAGANA LETTER VU}\N{COMBINING ACUTE ACCENT}<400>\N{HIRAGANA ITERATION MARK}<400>\N{KATAKANA LETTER SMALL A}\N{KATAKANA ITERATION MARK}\N{HALFWIDTH KATAKANA LETTER WO}\N{HALFWIDTH KATAKANA LETTER N}<400>def<200>#•</data>
 217
 218 # test normalization/dictionary handling of halfwidth katakana: same dictionary phrase in fullwidth and halfwidth
 219 <data>•芽キャベツ<400>芽キャﾍﾞツ<400></data>
 220
 221 # more Japanese tests
 222 # TODO: some script=common characters in the Hiragana and the Katakana block may not be treated correctly
 223 # (was formerly true for U+30FC); need to check and fix if so.
 224 #<data>•どー<400>せ<400>日本語<400>を<400>勉強<400>する<400>理由<400>について<400>　•て<400>こと<400>は<400>我<400>でも<400>知<400>ら<400>も<400>い<400>こと<400>なん<400>だ<400>。•</data>
 225 <data>•日本語<400>を<400>勉強<400>する<400>理由<400>について<400>　•て<400>こと<400>は<400>我<400>でも<400>知<400>ら<400>も<400>い<400>こと<400>なん<400>だ<400>。•</data>
 226
 227 # Testing of word boundary for dictionary word containing both kanji and kana
 228 <data>•中だるみ<400>蔵王の森<400>ウ離島<400></data>
 229
 230 # Testing of Chinese segmentation (taken from a Chinese news article)
 231 <data>•400<100>余<400>名<400>中央<400>委员<400>和<400>中央<400>候补<400>委员<400>都<400>领<400>到了<400>“•推荐<400>票<400>”•，•有<400>资格<400>在<400>200<100>多<400>名<400>符合<400>条件<400>的<400>63<100>岁<400>以下<400>中共<400>正<400>部<400>级<400>干部<400>中<400>，•选出<400>他们<400>属意<400>的<400>中央<400>政治局<400>委员<400>以<400>向<400>政治局<400>常委<400>会<400>举荐<400>。•</data>
 232
 233 # Words with interior formatting characters
 234 <data>•def\N{COMBINING ACUTE ACCENT}\N{SYRIAC ABBREVIATION MARK}ghi<200> •</data>
 235
 236 # to test for bug #4097779
 237 <data>•aa\N{COMBINING GRAVE ACCENT}a<200> •</data>
 238
 239 # fullwidth numeric, midletter characters etc should be treated like their halfwidth counterparts
 240 # <data>•ＩＳＮ'Ｔ<200> •１９<100>日<400></data>
 241 # why was this added with the dbbi stuff?
 242
 243 #      to test for bug #4098467
 244 #      What follows is a string of Korean characters (I found it in the Yellow Pages
 245 #      ad for the Korean Presbyterian Church of San Francisco, and I hope I transcribed
 246 #      it correctly), first as precomposed syllables, and then as conjoining jamo.
 247 #      Both sequences should be semantically identical and break the same way.
 248 #      precomposed syllables...
 249 <data>•\uc0c1\ud56d<200> •\ud55c\uc778<200> •\uc5f0\ud569<200> •\uc7a5\ub85c\uad50\ud68c<200> •\u1109\u1161\u11bc\u1112\u1161\u11bc<200> •\u1112\u1161\u11ab\u110b\u1175\u11ab<200> •\u110b\u1167\u11ab\u1112\u1161\u11b8<200> •\u110c\u1161\u11bc\u1105\u1169\u1100\u116d\u1112\u116c<200> •</data>
 250
 251 # more Korean tests (Jamo not tested here, not counted as dictionary characters)
 252 # Disable them now because we don't include a Korean dictionary.
 253 #<data>•\ud55c\uad6d<200>\ub300\ud559\uad50<200>\uc790\uc5f0<200>\uacfc\ud559<200>\ub300\ud559<200>\ubb3c\ub9ac\ud559\uacfc<200></data>
 254 #<data>•\ud604\uc7ac<200>\ub294<200> •\uac80\ucc30<200>\uc774<200> •\ubd84\uc2dd<200>\ud68c\uacc4<200>\ubb38\uc81c<200>\ub97c<200> •\uc870\uc0ac<200>\ud560<200> •\uac00\ub2a5\uc131<200>\uc740<200> •\uc5c6\ub2e4<200>\u002e•</data>
 255
 256 <data>•abc<200>\u4e01<400>\u4e02<400>\u3005<400>\u4e03\u4e03<400>abc<200> •</data>
 257
 258 <data>•\u06c9<200>\uc799\ufffa•</data>
 259
 260
 261 #
 262 #      Try some words from other scripts.
 263 #
 264
 265 #       Try some words from other scripts.
 266 #          Greek, Cyrillic, Hebrew, Arabic, Arabic, Georgian, Latin
 267 #
 268 <data>•ΑΒΓ<200> •БВГ<200> •אבג֓<200> •ابت<200> •١٢٣<100> •\u10A0\u10A1\u10A2<200> •ABC<200> •</data>
 269
 270 <data>•\u0301•A<200></data>
 271
 272
 273 #
 274 #  Hindi word break tests, imported from the old RBBI tests.
 275 #    An historical note:  a much earlier version of ICU break iterators had a number
 276 #    of special case rules for Hindi, which were tested by an earlier version of
 277 #    this test data.  The current RBBI rules do not special case Hindi in
 278 #    any way, making this test data much less signfificant.
 279 #
 280 <data>•\u0917\u092a\u00ad\u0936\u092a<200>!•\u092f\u0939<200> •\u0939\u093f\u0928\u094d\u200d\u0926\u0940<200> •\u0939\u0948<200> •\u0905\u093e\u092a<200> •\u0938\u093f\u0916\u094b\u0917\u0947<200>?•\n•:•\u092a\u094d\u0930\u093e\u092f\u0903<200>
 281 •\u0935\u0930\u094d\u0937\u093e<200>\r\n•\u092a\u094d\u0930\u0915\u093e\u0936<200>,•\u0924\u0941\u092e\u093e\u0930\u094b<200> •\u092e\u093f\u0924\u094d\u0930<200> •\u0915\u093e<200> •\u092a\u0924\u094d\u0930<200> •\u092a\u095d\u094b<200> •\u0938\u094d\u0924\u094d\u0930\u093f<200>.• •\u0968\u0966.\u0969\u096f<100> •\u0967\u0966\u0966.\u0966\u0966<100>\u20a8•\u0967,\u0967\u0966\u0966.\u0966\u0966<100> •\u0905\u092e\u091c<200>\n•\u0938\u094d\u200d\u0935\u0924\u0902\u0924\u094d\u0930<200>\r•</data>
 282
 283 #
 284 #  Failures from monkey tests
 285 #
 286 <data>•\u8527<400>\u02ba<200>\u0027\u0d42•\u00b7•\u09ea<100></data>
 287
 288 #
 289 # Jitterbug 5276 - treat Japanese half width voicing marks as Grapheme Extend
 290 #
 291 <data>•A\uff9e\uff9fBC<200> •1\uff9e\uff9f23<100></data>
 292
 293 # User guide example:
 294 <data>•Parlez<200>-•vous<200> •français<200> •?•</data>
 295
 296 # Test for #11673
 297 <word>
 298 <data>•ジョージア<400> •</data>
 299
 300 # Test for #11723
 301 <word>
 302 <data>•アレルギー性<400>結膜炎<400></data>
 303 <data>•アテ<400>ローム<400>性<400>動脈硬化<400></data>
 304
 305 # Ticket #11996
 306 <locale en>
 307 <word>
 308 <data>•栃木<400>県<400>足利<400>市<400>で<400>の<400>撮影<400>が<400>公開<400></data>
 309 <data>•栃木<400>県<400>足利<400>市<400>で<400>の<400>撮影<400>が<400>公開<400>さ<400>れ<400>た<400></data>
 310
 311 # Ticket #11999
 312 #   Unhandled Break Engine was consuming all characters, not just unhandled.
 313 #   \U00011700 is AHOM LETTER KA. There is no dictionary for AHOM, triggering the unhandled engine,
 314 #       which then incorrectly also consumed the following Japanese text. (ICU4J only)
 315 <word>
 316 <locale en>
 317 <data>•ロ<400>から<400>売却<400>完了<400>時<400>の<400>時価<400>が<400>提示<400>さ<400>れ<400>て<400>いる<400></data>
 318 <data>•\U00011700<200>ロ<400>から<400>売却<400>完了<400>時<400>の<400>時価<400>が<400>提示<400>さ<400>れ<400>て<400>いる<400></data>
 319
 320 #
 321 # What Is Unicode in Japanese
 322 # From http://unicode.org/standard/translations/japanese.html
 323
 324 <locale en>
 325 <word>
 326 <data><0>ユニ<400>コード<400>と<400>は<400>何<400>か<400>？<0></data>
 327 <data><0>ユニ<400>コード<400>は<400>、<0>すべて<400>の<400>文字<400>に<400>固有<400>の<400>番号<400>を<400>付与<400>し<400>ます<400></data>
 328 <data><0>プラットフォーム<400>に<400>は<400>依存<400>しま<400>せん<400></data>
 329 <data><0>プログラム<400>に<400>も<400>依存<400>しま<400>せん<400></data>
 330 <data><0>言語<400>に<400>も<400>依存<400>しま<400>せん<400></data>
 331
 332 <data><0>コンピューター<400>は<400>、<0>本質<400>的<400>に<400>は<400>数字<400>しか<400>扱う<400>こと<400>が<400>でき<400>ま<400>せん<400>。<0>\
 333 コンピューター<400>は<400>、<0>文字<400>や<400>記号<400>など<400>の<400>それぞれに<400>番号<400>を<400>割り振る<400>こと<400>によって<400>扱える<400>\
 334 よう<400>にし<400>ます<400>。<0>ユニ<400>コード<400>が<400>出来る<400>まで<400>は<400>、<0>これらの<400>番号<400>を<400>割り振る<400>仕組み<400>が<400>\
 335 何<400>百<400>種類<400>も<400>存在<400>しま<400>した<400>。<0>どの<400>一つ<400>を<400>とっても<400>、<0>十分<400>な<400>文字<400>を<400>含<400>\
 336 んで<400>は<400>いま<400>せん<400>で<400>した<400>。<0>例えば<400>、<0>欧州<400>連合<400>一つ<400>を<400>見<400>て<400>も<400>、<0>その<400>\
 337 すべて<400>の<400>言語<400>を<400>カバー<400>する<400>ため<400>に<400>は<400>、<0>いくつか<400>の<400>異なる<400>符号<400>化<400>の<400>仕組み<400>\
 338 が<400>必要<400>で<400>した<400>。<0>英語<400>の<400>よう<400>な<400>一つ<400>の<400>言語<400>に<400>限<400>って<400>も<400>、<0>一つ<400>だけ<400>\
 339 の<400>符号<400>化<400>の<400>仕組み<400>では<400>、<0>一般<400>的<400>に<400>使<400>われる<400>すべて<400>の<400>文字<400>、<0>句読点<400>、<0>\
 340 技術<400>的<400>な<400>記号<400>など<400>を<400>扱う<400>に<400>は<400>不十分<400>で<400>した<400>。<0></data>
 341
 342 <data><0>これらの<400>符号<400>化<400>の<400>仕組み<400>は<400>、<0>相互<400>に<400>矛盾<400>する<400>もの<400>でも<400>ありま<400>した<400>。<0>\
 343 二つ<400>の<400>異なる<400>符号<400>化<400>の<400>仕組み<400>が<400>、<0>二つ<400>の<400>異なる<400>文字<400>に<400>同一<400>の<400>番号<400>\
 344 を<400>付ける<400>こと<400>も<400>できる<400>し<400>、<0>同じ<400>文字<400>に<400>異なる<400>番号<400>を<400>付ける<400>こと<400>も<400>できる<400>\
 345 の<400>です<400>。<0>どの<400>よう<400>な<400>コンピューター<400>も<400>（<0>特に<400>サーバー<400>は<400>）<0>多く<400>の<400>異<400>な<400>っ<400>\
 346 た<400>符号<400>化<400>の<400>仕組み<400>を<400>サポート<400>する<400>必要<400>が<400>あり<400>ます<400>。<0>たとえ<400>データ<400>が<400>異なる<400>\
 347 符号<400>化<400>の<400>仕組み<400>や<400>プラットフォーム<400>を<400>通過<400>し<400>て<400>も<400>、<0>いつ<400>どこ<400>で<400>データ<400>が<400>\
 348 乱れる<400>か<400>分<400>から<400>ない<400>危険<400>を<400>冒す<400>こと<400>の<400>なる<400>の<400>です<400>。<0></data>
 349
 350 <data><0>ユニ<400>コード<400>は<400>すべて<400>を<400>変<400>え<400>ます<400></data>
 351
 352 <data><0>ユニ<400>コード<400>は<400>、<0>プラットフォーム<400>に<400>係<400>わら<400>ず<400>、<0>プログラム<400>に<400>係<400>わら<400>ず<400>、<0>\
 353 言語<400>に<400>係<400>わら<400>ず<400>、<0>すべて<400>の<400>文字<400>に<400>独立<400>した<400>番号<400>を<400>与<400>え<400>ます<400>。<0>\
 354 ユニ<400>コード<400>標準<400>は<400>、<0>アップル<400>、<0>ヒュー<400>レット<400>パッ<400>カード<400>、<0>IBM<200>、<0>ジャスト<400>システム<400>\
 355 、<0>マイクロ<400>ソフト<400>、<0>オラクル<400>、<0>SAＰ<200>、<0>サン<400>、<0>サイ<400>ベース<400>など<400>の<400>産業<400>界<400>の<400>\
 356 主導<400>的<400>企業<400>と<400>他の<400>多く<400>の<400>企業<400>に<400>採用<400>さ<400>れ<400>てい<400>ます<400>。<0>ユニ<400>コード<400>\
 357 は<400>、<0>XML<200>、<0>Java<200>、<0>ECMAScript<200>(<0>JavaScript<200>)<0>、<0>LDAP<200>、<0>CORBA<200> <0>3.0<100>など<400>\
 358 の<400>最先端<400>の<400>標準<400>の<400>前提<400>と<400>な<400>って<400>おり<400>、<0>ユニ<400>コード<400>を<400>実装<400>す<400>れ<400>ば<400>\
 359 、<0>ISO<200>/<0>IEC<200></data>
 360 <data><0> <0>10646<100>に<400>適合<400>する<400>ことに<400>なり<400>ます<400>。<0>ユニ<400>コード<400>は<400>、<0>多く<400>の<400>\
 361 オペレーティングシステム<400>と<400>すべて<400>の<400>最新<400>の<400>ブラウザー<400>と<400>他の<400>多く<400>の<400>製品<400>で<400>サポート<400>\
 362 さ<400>れ<400>てい<400>ます<400>。<0>ユニ<400>コード<400>標準<400>の<400>出現<400>と<400>ユニ<400>コード<400>を<400>サポート<400>する<400>\
 363 ツール<400>類<400>は<400>、<0>昨今<400>顕著<400>に<400>な<400>って<400>いる<400>ソフトウエア<400>技術<400>の<400>グローバル<400>化<400>の<400>\
 364 流れ<400>に対して<400>、<0>特に<400>役<400>に<400>立<400>って<400>い<400>ます<400>。<0></data>
 365
 366 <data><0>ユニ<400>コード<400>を<400>ク<400>ライアン<400>ト<400>サーバー<400>型<400>の<400>アプリケーション<400>や<400>、<0>多層<400>構造<400>\
 367 を<400>持つ<400>アプリケーション<400>、<0>ウェブサイト<400>など<400>に<400>に<400>組み込む<400>こと<400>で<400>、<0>従来<400>の<400>文字<400>\
 368 コードセット<400>を<400>用いる<400>より<400>も<400>明らか<400>な<400>コスト<400>削減<400>が<400>可能<400>です<400>。<0>ユニ<400>コード<400>は<400>\
 369 、<0>単一<400>の<400>ソフトウエア<400>製品<400>、<0>単一<400>の<400>ウェブサイト<400>に<400>、<0>何ら<400>手<400>を<400>加える<400>こと<400>なく<400>\
 370 、<0>複数<400>の<400>プラットフォーム<400>、<0>複数<400>の<400>言語<400>、<0>複数<400>の<400>国<400>を<400>カバー<400>する<400>こと<400>が<400>\
 371 出来る<400>の<400>です<400>。<0>ユニ<400>コード<400>は<400>、<0>データ<400>が<400>多く<400>の<400>異なる<400>システム<400>の<400>間<400>を<400>、<0>\
 372 何<400>の<400>乱れ<400>も<400>なし<400>に<400>転送<400>する<400>こと<400>を<400>可能<400>と<400>する<400>の<400>です<400>。<0></data>
 373
 374 <data><0>ユニ<400>コード<400>コンソーシアム<400>について<400></data>
 375
 376 <data><0>ユニ<400>コード<400>コンソーシアム<400>は<400>、<0>最新<400>の<400>ソフトウエア<400>製品<400>と<400>標準<400>において<400>テキスト<400>\
 377 を<400>表現<400>する<400>こと<400>を<400>意味<400>する<400>“<0>ユニ<400>コード<400>標準<400>”<0>の<400>構築<400>、<0>発展<400>、<0>普及<400>、<0>\
 378 利用<400>促進<400>を<400>目的<400>として<400>設立<400>さ<400>れ<400>た<400>非<400>営利<400>組織<400>です<400>。<0>同<400>コンソーシアム<400>\
 379 の<400>会員<400>は<400>、<0>コンピューター<400>と<400>情報処理<400>に<400>係わる<400>広汎<400>な<400>企業<400>や<400>組織<400>から<400>構成<400>\
 380 さ<400>れ<400>てい<400>ます<400>。<0>同<400>コンソーシアム<400>は<400>、<0>財政<400>的<400>に<400>は<400>、<0>純粋<400>に<400>会費<400>のみ<400>\
 381 によって<400>運営<400>さ<400>れ<400>てい<400>ます<400>。<0>ユニ<400>コード<400>標準<400>を<400>支持<400>し<400>、<0>その<400>拡張<400>と<400>\
 382 実装<400>を<400>支援<400>する<400>世界中<400>の<400>組織<400>や<400>個人<400>は<400>、<0>だれ<400>も<400>が<400>ユニ<400>コード<400>\
 383 コンソーシアム<400>の<400>会員<400>なる<400>こと<400>が<400>でき<400>ます<400>。<0></data>
 384
 385 <data><0>より<400>詳しい<400>こと<400>を<400>お<400>知<400>り<400>に<400>なり<400>たい<400>方<400>は<400>、<0>Glossary<200>,<0> <0>\
 386 Technical<200> <0>Introduction<200> <0>および<400> <0>Useful<200> <0>Resources<200>を<400>ご<400>参照<400>くだ<400>さい<400>。<0></data>
 387
 388
 389 ########################################################################################
 390 #
 391 #
 392 #       S e n t e n c e     B o u n d a r y      T e s t s
 393 #
 394 #
 395 ##########################################################################################
 396
 397
 398 #
 399 # Test data originally from RBBI RBBITest::TestDefaultRuleBasedSentenceIteration()
 400 #
 401 <sent>
 402
 403
 404 <sent>
 405 <data>•This\n<100></data>
 406 <data>•Hello! •how are you? •I'am fine. •Thankyou. •How are you \
 407 doing? •This\n<100> costs $20,00,000. •</data>
 408
 409
 410 # Sentence ending in a quote.
 411 <data>•"Sentence ending with a quote." •Bye.•</data>
 412
 413 # Sentence, and test data, ending without a period or other terminator.
 414 <data>•Here is a random sentence, no ending period<100></data>
 415
 416
 417 <data>•  (This is it).  •Testing the sentence iterator. •\
 418 "This isn't it." •Hi! \
 419 •This is a simple sample sentence. •(This is it.) •This is a simple sample sentence. •\
 420 "This isn't it." •\
 421 Hi! •This is a simple sample sentence. •It does not have to make any sense as you can see. •Nel mezzo del cammin di nostra vita, mi ritrovai in una selva oscura. •Che la dritta via aveo smarrita. •He said, that I said, that you said!! •Don't rock the boat.\u2029•Because I am the daddy, that is why.
 422 •Not on my time (el timo.)! •</data>
 423
 424 <data>•Hello.  •So what!!\u2029•"But now," he said, \
 425 "I know!" •\
 426 Harris thumbed down several, including "Away We Go" (which became the huge success Oklahoma!). •One species, B. anthracis, is highly virulent.
 427 •Wolf said about Sounder:\
 428 "Beautifully thought-out and directed." •\
 429 Have you ever said, "This is where\tI shall live"? •He answered, \
 430 "You may not!" •Another popular saying is: "How do you do?". \n•\
 431 Yet another popular saying is: \
 432 'I'm fine thanks.' •\
 433 What is the proper use of the abbreviation pp.? •Yes, I am definatelly 12" tall!!\
 434 •Now\r<100>is\n<100>the\r\n<100>time\n<100>\r<100>for\r<100>\r<100></data>
 435
 436 <data>•No breaks when . is surrounded by UPPER.Case letters.  •</data>
 437 <data>•No breaks when . is followed by Numeric .4 a.4 C.4 3.1 .•</data>
 438 <data>•No breaks when . is followed by a lower, with possible intervening punct .,a .$a .)a. •</data>
 439
 440 #
 441 #  Sentence Breaks: no break at the boundary between CJK and other letters
 442 #
 443 <data>•\u5487\u67ff\ue591\u5017\u61b3\u60a1\u9510\u8165:"JAVA\u821c\u8165\u7fc8\u51ce\u306d,\u2494\u56d8\u4ec0\u60b1\u8560\u51ba\u611d\u57b6\u2510\u5d46".\u2029•\u5487\u67ff\ue591\u5017\u61b3\u60a1\u9510\u8165\u9de8\u97e4JAVA\u821c\u8165\u7fc8\u51ce\u306d\ue30b\u2494\u56d8\u4ec0\u60b1\u8560\u51ba\u611d\u57b6\u2510\u5d46\u97e5\u7751\u3002•\u5487\u67ff\ue591\u5017\u61b3\u60a1\u9510\u8165\u9de8\u97e4\u6470\u8790JAVA\u821c\u8165\u7fc8\u51ce\u306d\ue30b\u2494\u56d8\u4ec0\u60b1\u8560\u51ba\u611d\u57b6\u2510\u5d46\u97e5\u7751\u2048•He said, "I can go there."\u2029•Bye, now.•</data>
 444
 445 #
 446 #      Treat fullwidth variants of .!? the same as their
 447 #      normal counterparts
 448 #
 449 <data>•I know I'm right\uff0e •Right\uff1f •Right\uff01 •</data>
 450
 451
 452 #
 453 #       Don't break sentences at boundary between CJK and digits
 454 #
 455 <data>•\u5487\u67ff\ue591\u5017\u61b3\u60a1\u9510\u8165\u9de8\u97e48888\u821c\u8165\u7fc8\u51ce\u306d\ue30b\u2494\u56d8\u4ec0\u60b1\u8560\u51ba\u611d\u57b6\u2510\u5d46\u97e5\u7751\u3002•Bye, now<100></data>
 456
 457 #
 458 #      Breaks around '(' following a sentence TERM.  (Rule 9)
 459 #
 460 <data>•How do you do?(•Fine). •</data>
 461 <data>•How do you do? •(Fine). •</data>
 462 <data>•How do you do?(•fine). •</data>
 463 <data>•How do you do? •(fine). •</data>
 464
 465 #
 466 <data>•Hello.123<100></data>    # Rule 6
 467 <data>•Hello?•123<100></data>
 468
 469 <data>•HELLO.Bye<100></data>    # Rule 7
 470 <data>•HELLO?•Bye<100></data>
 471
 472 <data>•Hello.goodbye<100></data>  #Rule 8
 473 <data>•Hello. •Goodbye<100></data>
 474 <data>•Hello. goodbye<100></data>
 475
 476
 477
 478 #
 479 #        test for bug #4158381: No breaks when there are no terminators around
 480 #
 481 <data>•\<P>Provides a set of &quot;lightweight&quot; (all-java\<FONT SIZE="-2">\<SUP>TM\</SUP>\</FONT> language) components that, to the maximum degree possible, work the same on all platforms.  •</data>
 482 <data>•Another test.\u2029•</data>
 483
 484 #        test for bug #4143071: Make sure sentences that end with digits
 485 #        work right
 486 #
 487 <data>•Today is the 27th of May, 1998.  •</data>
 488 <data>•Tomorrow with be 28 May 1998.  •</data>
 489 <data>•The day after will be the 30th.\u2029•</data>
 490
 491 #        test for bug #4152416: Make sure sentences ending with a capital
 492 #        letter are treated correctly
 493 #
 494 <data>•The type of all primitive \<code>boolean\</code> values accessed in the target VM.  •Calls to xxx will return an implementor of this interface.  \u2029•</data>
 495
 496 #        test for bug #4152117: Make sure sentence breaking is handling
 497 #        punctuation correctly [COULD NOT REPRODUCE THIS BUG, BUT TEST IS
 498 #        HERE TO MAKE SURE IT DOESN'T CROP UP]
 499 #
 500 <data>•Constructs a randomly generated BigInteger, uniformly distributed over the range \<tt>0\</tt> to \<tt>(2\<sup>numBits\</sup> - 1\)\</tt>, inclusive.  •The uniformity of the distribution assumes that a fair source of random bits is provided in \<tt>rnd\</tt>.  •Note that this constructor always constructs a non-negative biginteger.  \n•Ahh abc.
 501 •</data>
 502
 503 #        sentence breaks for hindi which used Devanagari script
 504 #        make sure there is sentence break after ?,danda(hindi phrase separator),
 505 #        fullstop followed by space.  (VERY old test)
 506 #
 507 <data>•\u0928\u092e\u0938\u094d\u200d\u0924\u0947 \u0930\u092e\u0947\u0936\u0905\u093e\u092a\u0915\u0948\u0938\u0947 \u0939\u0948?•\u092e\u0948 \u0905\u091a\u094d\u200d \u091b\u093e \u0939\u0942\u0901\u0964 •\u0905\u093e\u092a\r\n<100>\
 508 \u0915\u0948\u0938\u0947 \u0939\u0948?•\u0935\u0939 \u0915\u094d\u200d\u092f\u093e\n\
 509 <100>\u0939\u0948?•\u092f\u0939 \u0905\u093e\u092e \u0939\u0948. •\u092f\u0939 means "this". •"\u092a\u095d\u093e\u0908" meaning "education" or "studies". •\u0905\u093e\u091c(\u0938\u094d\u200d\u0935\u0924\u0902\u0924\u094d\u0930 \u0926\u093f\u0935\u093e\u0938) \u0939\u0948\u0964 •Let's end here. •</data>
 510
 511 #         Regression test for bug #1984, Sentence break in Arabic text.
 512
 513 <data>\
 514 •\u0623\u0633\u0627\u0633\u064b\u0627\u060c\u0020\u062a\u062a\u0639\u0627"\u0645\u0644\u0020\u0627\u0644\u062d\u0648\u0627\u0633\u064a\u0628\u0020"\u0641\u0642\u0637\u0020\u0645\u0639\u0020\u0627\u0644\u0623\u0631\u0642\u0627\u0645\u060c\u0648\u062a\u0642\u0648\u0645\u0020\u0628\u062a\u062e\u0632\u064a\u0646\u0020\u0627\u0644\u0623\u062d\u0631\u0641\u0020\u0648\u0627\u0644\u0645\u062d\u0627\u0631\u0641\u0020\u0627\u0644\u0623\u062e\u0631\u0649\u0020\u0628\u0639\u062f\u0020\u0623\u0646\u062a\u064f\u0639\u0637\u064a\u0020\u0631\u0642\u0645\u0627\u0020\u0645\u0639\u064a\u0646\u0627\u0020\u0644\u0643\u0644\u0020\u0648\u0627\u062d\u062f\u0020\u0645\u0646\u0647\u0627\u002e\u0020•\u0648\u0642\u0628\u0644\u0020\u0627\u062e\u062a\u0631\u0627\u0639\u0022\u064a\u0648\u0646\u0650\u0643\u0648\u062f\u0022\u060c\u0020\u0643\u0627\u0646\u0020\u0647\u0646\u0627\u0643\u0020\u0645\u0626\u0627\u062a\u0020\u0627\u0644\u0623\u0646\u0638\u0645\u0629\u0020\u0644\u0644\u062a\u0634\u0641\u064a\u0631\u0648\u062a\u062e\u0635\u064a\u0635\u0020\u0647\u0630\u0647\u0020\u0627\u0644\u0623\u0631\u0642\u0627\u0645\u0020\u0644\u0644\u0645\u062d\u0627\u0631\u0641\u060c\u0020\u0648\u0644\u0645\u0020\u064a\u0648\u062c\u062f\u0020\u0646\u0638\u0627\u0645\u062a\u0634\u0641\u064a\u0020\u0639\u0644\u0649\u0020\u062c\u0645\u064a\u0639\u0020\u0627\u0644\u0645\u062d\u0627\u0631\u0641\u0020\u0627\u0644\u0636\u0631\u0648\u0631\u064a\u0629.  •</data>
 515
 516 #         Try a few more of the less common sentence endings.
 517 <data>•Hello, world\u3002 •Hello, world\u1803 •Hello, world\u2048 •Hello, world\u203c •Let's end here. •</data>
 518
 519
 520
 521
 522 ################################################################
 523 #
 524 #
 525 #        L I N E     B R E A K
 526 #
 527 #
 528 ################################################################
 529
 530 <line>
 531 #
 532 # Test Character for each of the line break classes.
 533 #
 534 # 00A1;AI # INVERTED EXCLAMATION MARK  ¡
 535 # 0041;AL # LATIN CAPITAL LETTER A
 536 # 0009;BA # <control>
 537 # 00B4;BB # ACUTE ACCENT
 538 # 000C;BK # <control>
 539 # 2014;B2 # EM DASH
 540 # FFFC;CB # OBJECT REPLACEMENT CHARACTER
 541 # 0029;CL # RIGHT PARENTHESIS
 542 # 0301;CM # COMBINING ACUTE ACCENT
 543 # 0021;EX # EXCLAMATION MARK
 544 # 00A0;GL # NO-BREAK SPACE
 545 # 002D;HY # HYPHEN-MINUS
 546 # 4E00;ID # <CJK Ideograph, First>
 547 # 2024;IN # ONE DOT LEADER
 548 # 002C;IS # COMMA
 549 # 000A;LF # <control>
 550 # 0E5A;NS # THAI CHARACTER ANGKHANKHU
 551 # 0032;NU # DIGIT TWO
 552 # 0028;OP # LEFT PARENTHESIS
 553 # 0025;PO # PERCENT SIGN
 554 # 0024;PR # DOLLAR SIGN
 555 # 0022;QU # QUOTATION MARK
 556 # 0E01;SA # THAI CHARACTER KO KAI
 557 # DB7F;SG # Surrogate
 558 # 0020;SP # SPACE
 559 # 002F;SY # SOLIDUS /
 560 # F8FF;XX # Private Use
 561 # 200B;ZW # ZERO WIDTH SPACE
 562
 563
 564 # 2b  Always break at end of text
 565
 566 <data>•  •\u00A1•</data>
 567 <data>•  •\u0041•</data>
 568 <data>•  •\u0009•</data>
 569 <data>•  •\u00B4•</data>
 570 <data>•    \u000C<100></data>    # LB3C  × BK
 571 <data>•  •\u2014•</data>
 572 <data>•  •\uFFFC•</data>
 573 <data>•   \u0029•</data>    # LB 8   × CL
 574 # <data>• • \u0301•</data>    # LB 7a   Treat SP CM* as if it were ID  #TODO:  SP CM
 575 <data>•   \u0021•</data>    # LB 8   × EX
 576 #<data>•   \u00A0•</data>    # LB 11b   × GL   TODO:  fix.
 577 <data>•  •\u002D•</data>
 578 <data>•  •\u4E00•</data>
 579 <data>•  •\u2024•</data>
 580 <data>•   \u002C•</data>    # LB 8   × IS
 581 <data>•   \u000A<100></data>    # LB3C   × ( BK | CR | LF | NL )
 582 <data>•  •\u0E5A•</data>
 583 <data>•  •\u0032•</data>
 584 <data>•  •\u0028•</data>
 585 <data>•  •\u0025•</data>
 586 <data>•  •\u0024•</data>
 587 <data>•  •\u0022•</data>
 588 <data>•  •\u0E01•</data>
 589 <data>•  •\uDB7F•</data>
 590 <data>•   \u0020•</data>     # LB4 - don't break before space.
 591 <data>•   \u002F•</data>     # LB 8  × SY
 592 <data>•  •\uF8FF•</data>
 593 <data>•   \u200B•</data>     # LB4 - don't break before ZA
 594
 595
 596 # 3a  Always break after hard line breaks.
 597 # 3c  Never break before hard line breaks.
 598
 599 <data>•  •\u00A1\u2028<100>\u00A1•</data>
 600 <data>•  •\u0041\u2028<100>\u0041•</data>
 601 <data>•  •\u0009\u2028<100>\u0009•</data>
 602 <data>•  •\u00B4\u2028<100>\u00B4•</data>
 603 <data>•   \u000C<100>\u2028<100>\u000C<100></data>
 604 <data>•  •\u2014\u2028<100>\u2014•</data>
 605 <data>•  •\uFFFC\u2028<100>\uFFFC•</data>
 606 <data>•   \u0029\u2028<100>\u0029•</data>
 607 #<data>•   \u0301\u2028<100>\u0301•</data>    # TODO:  fix.
 608 <data>•   \u0021\u2028<100>\u0021•</data>
 609 #<data>•   \u00A0\u2028<100>\u00A0•</data>    # TODO:  fix
 610 <data>•  •\u002D\u2028<100>\u002D•</data>
 611 <data>•  •\u4E00\u2028<100>\u4E00•</data>
 612 <data>•  •\u2024\u2028<100>\u2024•</data>
 613 <data>•   \u002C\u2028<100>\u002C•</data>
 614 <data>•   \u000A<100>\u2028<100>\u000A<100></data>
 615 <data>•  •\u0E5A\u2028<100>\u0E5A•</data>
 616 <data>•  •\u0032\u2028<100>\u0032•</data>
 617 <data>•  •\u0028\u2028<100>\u0028•</data>
 618 <data>•  •\u0025\u2028<100>\u0025•</data>
 619 <data>•  •\u0024\u2028<100>\u0024•</data>
 620 <data>•  •\u0022\u2028<100>\u0022•</data>
 621 <data>•  •\u0E01\u2028<100>\u0E01•</data>
 622 <data>•  •\uDB7F\u2028<100>\uDB7F•</data>
 623 <data>•   \u0020\u2028<100>\u0020•</data>
 624 <data>•   \u002F\u2028<100>\u002F•</data>
 625 <data>•  •\uF8FF\u2028<100>\uF8FF•</data>
 626 <data>•   \u200B\u2028<100>\u200B•</data>
 627
 628 # Regional Indicator sequences. They group in pairs. The reverse rules are tricky.
 629 #    Sequences are long enough that the non-exaustive monkey test won't reliably pick up problems.
 630
 631 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•</data>
 632 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6•</data>
 633
 634 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•</data>
 635 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6•</data>
 636 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•</data>
 637 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6•</data>
 638
 639
 640 # User Guide example
 641
 642 <data>•Parlez-•vous •français ?•</data>
 643
 644 #
 645 #  Old Line Break Test data.  Orginally located in RBBITest::TestDefaultRuleBasedLineIteration()
 646 #
 647
 648 <line>
 649
 650 <data>•Multi-•Level •example •of •a •semi-•idiotic •non-•sensical •(non-•important) •sentence.
 651 <100>Hi  •Hello •How\n<100>are\r<100>you\u2028<100>fine.\t•good.  •Now\r<100>is\n<100>the\r\n<100>time\n<100>\r<100>for\r<100>\r<100>all•</data>
 652
 653 <line>
 654 <data>•Hello! •how\r\n<100> •(are)\r<100> •you? •I'am •fine- •Thankyou. •foo\u00a0bar
 655 <100>How, •are, •you? •This, •costs •$20,00,000.•</data>
 656
 657 #     test for bug #4068133
 658 #
 659 <data>•\u96f6•\u4e00\u3002•\u4e8c\u3001•\u4e09\u3002\u3001•\u56db\u3001\u3002\u3001•\u4e94,•\u516d.•\u4e03.\u3001,\u3002•\u516b•</data>
 660
 661 #     to test for bug #4086052
 662 <data>•foo\u00a0bar•</data>
 663
 664 #     to test for bug #4097920
 665 <data>•dog,cat,mouse •(one)•(two)\n<100></data>
 666
 667 #     to test for bug #4035266
 668 <data>•The •balance •is •$-23,456.78, •not •-•$32,456.78!\n<100></data>
 669
 670
 671 #      to test for bug #4098467
 672 #      What follows is a string of Korean characters (I found it in the Yellow Pages
 673 #      ad for the Korean Presbyterian Church of San Francisco, and I hope I transcribed
 674 #      it correctly), first as precomposed syllables, and then as conjoining jamo.
 675 #      Both sequences should be semantically identical and break the same way.
 676 #      precomposed syllables...  (I == Rich Gillam?)
 677 #
 678 <data>•\uc0c1•\ud56d •\ud55c•\uc778 •\uc5f0•\ud569 •\uc7a5•\ub85c•\uad50•\ud68c•</data>
 679
 680 #      conjoining jamo...
 681 <data>•\u1109\u1161\u11bc•\u1112\u1161\u11bc •\u1112\u1161\u11ab•\u110b\u1175\u11ab •\u110b\u1167\u11ab•\u1112\u1161\u11b8 •\u110c\u1161\u11bc•\u1105\u1169•\u1100\u116d•\u1112\u116c•</data>
 682
 683 #      to test for bug #4117554: Fullwidth .!? should be treated as postJwrd
 684 <data>•\u4e01\uff0e•\u4e02\uff01•\u4e03\uff1f•</data>
 685
 686 #      Surrogate line break tests.
 687 #
 688 <data>•\u4e01•\ud840\udc01•\u4e02•abc •\ue000 •\udb80\udc01•</data>  #This line and the following are equivalent.
 689 <data>•\u4e01•\U00020001•\u4e02•abc •\ue000 •\U000f0001•</data>
 690
 691 #      Regression for bug 836
 692 #        Note:  Unicode 5.1 changed this behavior
 693 #               Unicode 5.2 changed it again, there is no break following the '('
 694 <data>•AAA(AAA •</data>
 695
 696 #       Try some words from other scripts.
 697 #          Greek, Cyrillic, Hebrew, Arabic, Arabic, Georgian, Latin
 698 #
 699 <data>•ΑΒΓ •БВГ •אבג֓ •ابت •١٢٣ •\u10A0\u10A1\u10A2 •ABC •</data>
 700
 701 #
 702 #       ticket #4853:  unpaired surrogates should behave like AL
 703 #
 704 <data>•abc\ud801xyz•</data>
 705
 706 #
 707 #     Regression tests for failures that originally came from the monkey test.
 708 #     Monkey test failure lines can, with slight reformatting, be copied into this section
 709 #     as test cases.  The error display from here is more informative.
 710 #
 711 <data>•\ufffc•\u30e3\u000c<100>\u1b39\u300a\u002f\u203a\u200b•\ufffc•\uaf64•\udcfb•</data>
 712 <data>•\u114d\u31f3•\ube44\u002d•\u0362\u24e2\u276e\u2014\u205f\ufe16•\uc877•\u0fd0\u000a<100>\u20a3•</data>
 713 <data>•\u080a\u215b\U0001d7d3\u002c•\u2025\U000e012e•\u02df\u118d\u0029\ua8d6\u0085<100>\u6cc4\u2024\u202f\ufffc•</data>
 714
 715 # Test for #10176 (in root)
 716 <line>
 717 <data>•abc/•s •def•</data>
 718 <data>•abc/\u05D9 •def•</data>
 719 <data>•\u05E7\u05D7/\u05D9 •\u05DE\u05E2\u05D9\u05DC•</data>
 720 <data>•\u05D3\u05E8\u05D5\u05E9\u05D9\u05DD •\u05E9\u05D7\u05E7\u05E0\u05D9\u05DD/\u05D9\u05D5\u05EA•</data>
 721
 722 # Ticket #11556 don't break "R$" or "JP¥"
 723 <locale en>
 724 <line>
 725 <data>•R$ •JP¥ •a9 •3a •H% •CA$ •Travi$ •Scott •Ke$ha •Curren$y •A$AP •Rocky•</data>
 726
 727
 728
 729 ########################################################################################
 730 #
 731 #
 732 #       T i t l e   B o u n d a r y      T e s t s
 733 #
 734 #
 735 ##########################################################################################
 736 <title>
 737 <data>•Here •is •a •short •sample •sentence.  •And •another.•</data>
 738 <data>•HERE •IS •A •SHORT •SAMPLE •SENTENCE.  •AND •ANOTHER.•</data>
 739 <data>•  •Start •and •end •with •spaces   •</data>
 740 <data>•Include 123 456 ^&  •some 54332 •numbers 4445•abc123•abc •ending 1223 •</data>
 741
 742 <data>•Combining\u0301 \u0301•ma\u0306rks  •bye  •</data>
 743 <data>•123  •Start •with •a •number.•</data>
 744
 745 <data>•'•start •with •a •case-•ignorable •cha'r'a'cter•</data>
 746 <data>•'  '' •start •with •case-•ignorable & •case-•insensitive •cha'r'a'cter•</data>
 747 <data>•  ''•aaa' •bbb '•ccc' '•ddd''' '''•eee   '''•fff'''   •ggg  ''•</data>
 748 # Note: apostrophe is case-ignorable. space is not cased.
 749
 750 ##########################################################################################
 751 #
 752 #   Thai Tests
 753 #
 754 ##########################################################################################
 755 <locale th>
 756 <word>
 757 #
 758 # Test data originally from the test code source file
 759 #      // @suwit -- Thai sample data from GVT Guideline
 760 #
 761 <data>•\u0E2B\u0E19\u0E36\u0E48\u0E07<200>\u0E04\u0E33<200>\u0E44\u0E17\u0E22<200>\
 762 \u0E2A\u0E32\u0E21\u0E32\u0E23\u0E16<200>\u0E1B\u0E23\u0E30\u0E01\u0E2D\u0E1A<200>\
 763 \u0E14\u0E49\u0E27\u0E22<200>\u0e2b\u0e25\u0e32\u0e22<200>\
 764 \u0e1e\u0e22\u0e32\u0e07\u0e04\u0e4c<200></data>
 765
 766 # Test data originally from http://bugs.icu-project.org/trac/search?q=r30327
 767 <data>•กู<200> •กิน<200>กุ้ง<200> •ปิ้่<200>งอ<200>ยู่<200>ใน<200>ถ้ำ<200></data>
 768
 769 <data>•\u0E01\u0E39<200>\u0020•\u0E01\u0E34\u0E19<200>\u0E01\u0E38\u0E49\u0E07<200>\
 770 \u0020•\u0E1B\u0E34\u0E49\u0E48<200>\u0E07\u0E2D<200>\u0E22\u0E39\u0E48<200>\
 771 \u0E43\u0E19<200>\u0E16\u0E49\u0E33<200></data>
 772
 773 <line>
 774 <data>•0E01\u0E39\u0020•\u0E01\u0E34\u0E19•\u0E01\u0E38\u0E49\u0E07\
 775 \u0020•\u0E1B\u0E34\u0E49\u0E48•\u0E07\u0E2D•\u0E22\u0E39\u0E48•\
 776 \u0E43\u0E19•\u0E16\u0E49\u0E33•</data>
 777
 778 # Data originally from intltest RBBITest::TestThaiLineBreak()
 779 #
 780 #  \u0e2f-- the Thai paiyannoi character-- isn't a letter.  It's a symbol that
 781 #  represents elided letters at the end of a long word.  It should be bound to
 782 #  the end of the word and not treated as an independent punctuation mark.
 783 #
 784 #  the one time where the paiyannoi occurs somewhere other than at the end
 785 #  of a word is in the Thai abbrevation for "etc.", which both begins and
 786 #  ends with a paiyannoi
 787 #
 788 <line>
 789 <data>•\u0e2a\u0e16\u0e32\u0e19\u0e35\u0e2f•\
 790 \u0e08\u0e30•\
 791 \u0e23\u0e30\u0e14\u0e21•\
 792 \u0e40\u0e08\u0e49\u0e32•\
 793 \u0e2b\u0e19\u0e49\u0e32\u0e17\u0e35\u0e48•\
 794 \u0e2d\u0e2d\u0e01•\
 795 \u0e21\u0e32•\
 796 \u0e40\u0e23\u0e48\u0e07•\
 797 \u0e23\u0e30\u0e1a\u0e32\u0e22•\
 798 \u0e2d\u0e22\u0e48\u0e32\u0e07•\
 799 \u0e40\u0e15\u0e47\u0e21•\
 800 \u0e2f\u0e25\u0e2f•\
 801 \u0e17\u0e35\u0e48•\
 802 \u0e19\u0e31\u0e49\u0e19•</data>
 803
 804 # Data originally from RBBITest::TestMixedThaiLineBreak()
 805 #   @suwit -- Test Arabic numerals, Thai numerals, Punctuation and English characters start
 806 #
 807 <line>
 808 <data>•\u0E1B\u0E35•\
 809 \u0E1E\u0E38\u0E17\u0E18\u0E28\u0E31\u0E01\u0E23\u0E32\u0E0A •\
 810 2545 •\
 811 \u0E40\u0E1B\u0E47\u0E19•\
 812 \u0E1B\u0E35•\
 813 \u0E09\u0E25\u0E2D\u0E07•\
 814 \u0E04\u0E23\u0E1A•\
 815 \u0E23\u0E2D\u0E1A •\
 816 \"\u0E52\u0E52\u0E50 •\
 817 \u0E1b\u0E35\" •\
 818 \u0E02\u0E2d\u0E07•\
 819 \u0E01\u0E23\u0E38\u0E07•\
 820 \u0E23\u0E31\u0E15\u0E19\u0E42\u0E01\u0E2A\u0E34\u0E19\u0E17\u0E23\u0E4C •\
 821 (\u0E01\u0E23\u0E38\u0E07\u0E40\u0E17\u0E1e\u0E2F•\
 822 \u0E2B\u0E23\u0E37\u0E2D •\
 823 Bangkok)•</data>
 824
 825 # Data originally from RBBITest::TestMaiyamok()
 826 #   The Thai maiyamok character is a shorthand symbol that means "repeat the previous
 827 #   word".  Instead of appearing as a word unto itself, however, it's kept together
 828 #   with the word before it.
 829 #
 830 <line>
 831 <data>•\u0e44\u0e1b\u0e46•\
 832 \u0e21\u0e32\u0e46•\
 833 \u0e23\u0e30\u0e2b\u0e27\u0e48\u0e32\u0e07•\
 834 \u0e01\u0e23\u0e38\u0e07•\
 835 \u0e40\u0e17\u0e1e•\
 836 \u0e41\u0e25\u0e30•\
 837 \u0e40\u0e03\u0e35•\
 838 \u0e22\u0e07•\
 839 \u0e43\u0e2b\u0e21\u0e48•</data>
 840
 841 # Test for #10296
 842 <line>
 843 <data>•ใช•มั้ย•</data>
 844 <data>•มั๊ยล่ะ•ที่รัก•</data>
 845
 846 # Test for #10593
 847 <line>
 848 <data>•เล่น•ผ่าน•ทาง•บลูทูธ•บน•อุปกรณ์•</data>
 849
 850 # Test for city names #10691
 851 <line>
 852 <data>•ไป•ที่•ซานฟรานซิสโก•</data>
 853
 854 # Test for #10630, #10631
 855 <line>
 856 <data>•แท็ก•แอปพลิเคชัน•เป็น•พิเศษ•</data>
 857
 858 # Test for #11019
 859 <line>
 860 <data>•เบ•เบราว์เซอร์•โพ•โพสต์•โพสท์•</data>
 861
 862 # Test for #11688
 863 <line>
 864 <data>•อัปเดต•อีเวนต์•</data>
 865
 866 ##########################################################################################
 867 #
 868 #   Lao Tests
 869 #
 870 ##########################################################################################
 871 <locale en>
 872 # Basic check for #7647
 873 <line>
 874 <data>•ສະບາຍດີ•</data>
 875 <data>•ດີ•ຂອບໃຈ•</data>
 876 <data>•ເຈົ້າ•ເວົ້າ•ພາສາ•ອັງກິດ•ໄດ້•ບໍ່•</data>
 877 <data>•ກະລຸນາ•ເວົ້າ•ຊ້າ•ໆ•</data>
 878
 879 ##########################################################################################
 880 #
 881 #   Burmese/Myanmar Tests
 882 #
 883 ##########################################################################################
 884 <locale en>
 885 # Basic sanity check for #10326 (some text from http://www.unicode.org/udhr/d/udhr_mya.txt)
 886 <line>
 887 <data>•လူ•တိုင်း•သည် •တူညီ •လွတ်လပ်•သော •ဂုဏ်•သိ•က္•ခါ•ဖြ•င့် •လည်းကောင်း၊ •</data>
 888 <data>•တူညီ•လွတ်လပ်•သော •အ•ခွ•င့်•အရေး•များ•ဖြ•င့် •လည်းကောင်း၊ •မွေး•ဖွား•လာ•သူများ •ဖြစ်သည်။•</data>
 889 <data>•ထို•သူ•တို့၌ •ပိုင်းခြား •ဝေဖန်•တတ်•သော •ဉာဏ်•နှ•င့် •ကျ•င့်•ဝတ် •သိတတ်•သော •စိတ်•တို့•ရှိ•ကြ၍ •</data>
 890 <data>•ထို•သူ•တို့သည် •အချင်းချင်း •မေတ္တာ•ထား၍ •ဆက်ဆံ•ကျ•င့်•သုံး•</data>
 891
 892 ##########################################################################################
 893 #
 894 #   Khmer Tests
 895 #
 896 ##########################################################################################
 897
 898 # Test data originally from http://bugs.icu-project.org/trac/search?q=r30327
 899 #  from the file testdata/wordsegments.txt
 900 <locale en>
 901 <word>
 902
 903 <data>•តើ<200>លោក<200>មក<200>ពី<200>ប្រទេស<200>ណា<200></data>
 904 <data>•សណ្ដូក<200>ក<200>បណ្ដែត<200>ខ្លួន<200></data>
 905 <data>•ពណ៌ស<200>ម្ដេច<200>ថា<200>ខ្មៅ<200></data>
 906 #ប្រយោគ|ពី|របៀប|រួបរួម|និង|ភាព|ផ្សេងគ្នា|ដែល|អាច|ចូល<200></data>
 907 <data>•ប្រយោគ<200>ពី<200>របៀប<200>ដែល<200>និង<200>ភាព<200>ផ្សេងគ្នា<200>ដែល<200>អាច<200>ចូល<200></data>
 908 #ប្រយោគ|ពី|របៀប|ជា|មួយ|និង|ភាព|ផ្សេងគ្នា|ដែល|អាច|ចូល<200></data>
 909 <data>•សូម<200>ចំណាយពេល<200>បន្តិច<200>ដើម្បី<200>អធិស្ឋាន<200>អរព្រះគុណ<200>ដល់<200>ព្រះអង្គ<200></data>
 910 <data>•ការ<200>ថោកទាប<200>បរិប្បូណ៌<200>ដោយ<200></data>
 911 <data>•ប្រើប្រាស់<200>ស្អាត<200>ទាំង<200>ចិត្ត<200>សិស្ស<200>នោះ<200></data>
 912 <data>•បើ<200>អ្នក<200>ប្រព្រឺត្ត<200>អំពើអាក្រក់<200>មុខ<200>ជា<200>មាន<200></data>
 913 <data>•ប្រដាប់<200>ប្រដា<200>រ<200>រៀនសូត្រ<200>បន្ទប់<200>រៀន<200></data>
 914 <data>•ដើរតួ<200>មនុស្សគ<200>ឥត<200>បញ្ចេញ<200>យោបល់<200>សោះ<200>ឡើយ<200></data>
 915 <data>•មិន<200>អាច<200>ឲ្យ<200>យើង<200>ធ្វើ<200>កសិកម្ម<200>បាន<200>ឡើយ<200></data>
 916 <data>•បន្ត<200>សេចក្ត<200>ទៅទៀត<200></data>
 917 <data>•ក្រុម<200>ប៉ូលិស<200>បណ្តាក់<200>គ្នា<200></data>
 918 <data>•គ្មាន<200>សុខ<200>សំរាន្ត<200>ដង<200>ណា<200></data>
 919 <data>•បាន<200>សុខភាព<200>បរិប្បូណ៌<200></data>
 920 <data>•ជា<200>មេចោរ<200>ខ្ញុំ<200>នឹង<200>ស្លាប់<200>ទៅវិញ<200>ជា<200>មេចោរ<200></data>
 921 <data>•ឯ<200>ការ<200>វាយ<200>ផ្ចាល<200>ដែល<200>នាំ<200></data>
 922 <data>•គេ<200>ដឹក<200>ទៅ<200>សំឡាប់<200></data>
 923 #អ្នក|ដែល|ជា|មន្ត្រី|ធំ|លើ|គាត់|ទេ<200></data>
 924 <data>•យក<200>ទៅ<200>សម្លាប់ចោល<200>ស្ងាត់<200></data>
 925 <data>•ត្រូវ<200>បាន<200>គេ<200>សម្លាប់<200></data>
 926 <data>•នៅក្នុង<200>ស្រុក<200>ខ្ល<200>ងហ្ស៊ុន<200></data>
 927
 928
 929 #
 930 #  Jitterbug 3671 Test Case
 931 #
 932 <data>•สวัสดี<200>ครับ<200>สบาย<200>ดี<200>ไหม<200> •ครับ<200></data>
 933
 934 #
 935 #  Trac ticket 5595 Test Case
 936 <data>•บท<200>ที่๑พายุ<200>ไซโคลน<200>โด<200>โรธี<200>อาศัย<200>อยู่<200>ท่ามกลาง<200>\
 937 ทุ่งใหญ่<200>ใน<200>แคนซัส<200>กับ<200>ลุง<200>เฮ<200>นรี<200>ชาวไร่<200>และ<200>ป้า<200>เอ็ม<200>\
 938 ภรรยา<200>ชาวไร่<200>บ้าน<200>ของ<200>พวก<200>เขา<200>หลัง<200>เล็ก<200>เพราะ<200>ไม้<200>\
 939 สร้าง<200>บ้าน<200>ต้อง<200>ขน<200>มา<200>ด้วย<200>เกวียน<200>เป็น<200>ระยะ<200>ทาง<200>หลาย<200>\
 940 ไมล์<200></data>
 941
 942 ####################################################################################
 943 #
 944 #  Tailored (locale specific) breaking.
 945 #
 946 ####################################################################################
 947
 948 # Japanese line break tailoring test
 949
 950 <locale ja>
 951 <line>
 952 <data>•\u3041•\u3043•\u3045•\u31f1•</data>
 953 <locale en>
 954 <line>
 955 <data>•\u3041\u3043\u3045\u31f1•</data>
 956
 957 # The following data was originally in RBBITest::TestJapaneseWordBreak()
 958 <locale ja>
 959 <word>
 960 <data>•\u4ECA\u65E5<400>\u306F<400>\u3044\u3044<400>\u5929\u6C17<400>\u3067\u3059<400>\u306D<400>\u3002•\u000D\u000A•</data>
 961
 962 # UBreakIteratorType UBRK_WORD, Locale "ja"
 963 # Don't break in runs of hiragana or runs of ideograph, where the latter includes \u3005 \u3007 \u303B (cldrbug #2009).
 964 # \u79C1\u9054\u306B\u4E00\u3007\u3007\u3007\u306E\u30B3\u30F3\u30D4\u30E5\u30FC\u30BF\u304C\u3042\u308B\u3002\u5948\u3005\u306F\u30EF\u30FC\u30C9\u3067\u3042\u308B\u3002
 965 # modified to work with dbbi code - should verify
 966
 967 <locale ja>
 968 <word>
 969 <data>•私<400>達<400>に<400>一<400>〇<400>〇〇<400>の<400>コンピュータ<400>が<400>ある<400>。<0>奈々<400>は<400>ワード<400>で<400>ある<400>。•</data>
 970
 971 # Test for #10176 (in ja)
 972 <line>
 973 <data>•abc/•s •def•</data>
 974 <data>•abc/\u05D9 •def•</data>
 975 <data>•\u05E7\u05D7/\u05D9 •\u05DE\u05E2\u05D9\u05DC•</data>
 976 <data>•\u05D3\u05E8\u05D5\u05E9\u05D9\u05DD •\u05E9\u05D7\u05E7\u05E0\u05D9\u05DD/\u05D9\u05D5\u05EA•</data>
 977
 978
 979 <locale root>
 980 <word>
 981 <data>•私<400>達<400>に<400>一<400>〇<400>〇〇<400>の<400>コンピュータ<400>が<400>ある<400>。<0>奈々<400>は<400>ワード<400>で<400>ある<400>。•</data>
 982 # The following test is for #10300
 983 <data>•例えば<400>オーストラリア<400>。•</data>
 984 # The following test is for #10571
 985 <data>•一部<400>の<400>地域<400>では<400>、<0>ブラジル<400>、<0>インドネシア<400>、<0>オーストリア<400>、<0>ニュージーランド<400>で<400>ある<400>。•</data>
 986
 987 # UBreakIteratorType UBRK_SENTENCE, Locale "el"
 988 # Add break after Greek question mark (cldrbug #2069).
 989 # "\u0391\u03B2, \u03B3\u03B4; \u0395 \u03B6\u03B7\u037E \u0398 \u03B9\u03BA. "
 990 # "\u039B\u03BC \u03BD\u03BE! \u039F\u03C0, \u03A1\u03C2? \u03A3"
 991 # which is "Αβ, γδ; Ε ζη; Θ ικ. Λμ νξ! Οπ, Ρς? Σ"
 992
 993 <locale root>
 994 <sent>
 995 <data>•Αβ, γδ; Ε ζη; Θ ικ. •Λμ νξ! •Οπ, Ρς? •Σ<100></data>
 996
 997 <locale el>
 998 <sent>
 999 <data>•Αβ, γδ; •Ε ζη; •Θ ικ. •Λμ νξ! •Οπ, Ρς? •Σ<100></data>
1000
1001 #  UBreakIteratorType UBRK_WORD, Locale "en_US_POSIX"
1002 #  Words don't include colon or period (cldrbug #1969).
1003
1004 <locale sv>
1005 <word>
1006 <data>•Can't<200> •have<200> •breaks<200> •in<200> •xx:yy<200> •or<200> •struct.field<200> \
1007 •for<200> •CS<200>-•types<200>.•</data>
1008 <data>•\uFF92\uFF76\uFF9E<400> •</data>
1009
1010 <locale en_US>
1011 <word>
1012 <data>•Can't<200> •have<200> •breaks<200> •in<200> •xx<200>:•yy<200> •or<200> •struct.field<200> \
1013 •for<200> •CS<200>-•types<200>.•</data>
1014 <data>•\uFF92\uFF76\uFF9E<400> •</data>
1015
1016 <locale en_US_POSIX>
1017 <word>
1018 <data>•Can't<200> •have<200> •breaks<200> •in<200> •xx<200>:•yy<200> •or<200> •struct<200>.•field<200> \
1019 •for<200> •CS<200>-•types<200>.•</data>
1020 <data>•\u06c9<200>\uc799\ufffa•</data>
1021 <data>•\uFF92\uFF76\uFF9E<400> •</data>
1022
1023
1024 # UBreakIteratorType UBRK_CHARACTER, Locale "th"
1025 # Clusters should not include spacing Thai/Lao vowels (prefix or postfix), except for [SARA] AM (cldrbug #2161).
1026 # Update:  As of Unicode 6.1 root has same behavior as th for this.
1027 #
1028 # "\u0E01\u0E23\u0E30\u0E17\u0E48\u0E2D\u0E21\u0E23\u0E08\u0E19\u0E32 "
1029 # "(\u0E2A\u0E38\u0E0A\u0E32\u0E15\u0E34-\u0E08\u0E38\u0E11\u0E32\u0E21\u0E32\u0E28) "
1030 # "\u0E40\u0E14\u0E47\u0E01\u0E21\u0E35\u0E1B\u0E31\u0E0D\u0E2B\u0E32 "
1031 # which is "กระท่อมรจนา (สุชาติ-จุฑามาศ) เด็กมีปัญหา "
1032
1033 <locale th>
1034 <char>
1035 <data>•\u0E01•\u0E23•\u0E30•\u0E17\u0E48•\u0E2D•\u0E21•\u0E23•\u0E08•\u0E19•\u0E32• •\
1036 (•\u0E2A\u0E38•\u0E0A•\u0E32•\u0E15\u0E34•-•\u0E08\u0E38•\u0E11•\u0E32•\u0E21•\u0E32•\u0E28•)• •\
1037 \u0E40•\u0E14\u0E47•\u0E01•\u0E21\u0E35•\u0E1B\u0E31•\u0E0D•\u0E2B•\u0E32• •</data>
1038
1039 # Finnish line breaking
1040 #
1041 # These rules deal with hyphens when there is a space on the leading side.
1042 # There should be a break opportunity between the space and the hyphen, and not after the hyphen.
1043 # See CLDR ticket 3029.
1044 # See ICU ticket 8151
1045
1046 <locale root>
1047 <line>
1048 <data>•abc •- •def    •abc •-•def    •abc- •def   •abc-•def•</data>   # With ASCII hyphen
1049 <data>•abc •‐ •def    •abc •‐•def    •abc‐ •def   •abc‐•def•</data>   # With Unicode u2010 hyphen
1050
1051 <locale fi>
1052 <line>
1053 # TODO: problems with Finnish line break rules cause these two lines to fail.
1054 #<data>•abc •- •def    •abc •-def    •abc- •def   •abc-•def•</data>   # With ASCII hyphen
1055 #<data>•abc •‐ •def    •abc •‐def    •abc‐ •def   •abc‐•def•</data>   # With Unicode u2010 hyphen
1056
1057 <data>•abc •- •def    •abc •-def    •abc- •def   •</data>   # With ASCII hyphen
1058 <data>•abc •‐ •def    •abc •‐def    •abc‐ •def   •</data>   # With Unicode u2010 hyphen
1059
1060 # Test for #10176 (in fi)
1061 <line>
1062 <data>•abc/•s •def•</data>
1063 <data>•abc/\u05D9 •def•</data>
1064 <data>•\u05E7\u05D7/\u05D9 •\u05DE\u05E2\u05D9\u05DC•</data>
1065 <data>•\u05D3\u05E8\u05D5\u05E9\u05D9\u05DD •\u05E9\u05D7\u05E7\u05E0\u05D9\u05DD/\u05D9\u05D5\u05EA•</data>
1066
1067 ####################################################################################
1068 #
1069 #  Test CSS line break variants: strict, normal, loose
1070 #
1071 ####################################################################################
1072
1073 <locale ja@lb=strict>
1074 <line>
1075 #     •no brk before 3063              •no brk before 301C•no brk btw 2026   •no brk before FF01•
1076 <data>•\u3084\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
1077
1078 <locale ja@lb=normal>
1079 <line>
1080 #     •brk OK before 3063               •brk OK before 301C •no brk btw 2026   •no brk before FF01•
1081 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031•\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
1082
1083 <locale ja@lb=loose>
1084 <line>
1085 #     •brk OK before 3063               •brk OK before 301C •brk OK btw 2026    •brk OK before FF01•
1086 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031•\u301C\u0020•\u2026•\u2026\u0020•u30A2•\uFF01\u0020•</data>
1087
1088 <locale en@lb=strict>
1089 <line>
1090 #     •no brk before 3063              •no brk before 301C•no brk btw 2026   •no brk before FF01•
1091 <data>•\u3084\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
1092
1093 <locale en@lb=normal>
1094 <line>
1095 #     •brk OK before 3063               •no brk before 301C •no brk btw 2026  •no brk before FF01•
1096 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
1097
1098 <locale en@lb=loose>
1099 <line>
1100 #     •brk OK before 3063               •no brk before 301C •brk OK btw 2026   •no brk before FF01•
1101 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026•\u2026\u0020•u30A2\uFF01\u0020•</data>
1102
1103 ####################################################################################
1104 #
1105 #  Test Apple early change of lb class for 22EF
1106 #
1107 ####################################################################################
1108
1109 <locale en>
1110 <line>
1111 <data>•\u4E00\u2026\u2026•\u4E00\u22EF\u22EF•\u4E00\u0020•</data>
1112
1113 ####################################################################################
1114 #
1115 #  Test Apple early change of cjdict
1116 #
1117 ####################################################################################
1118
1119 <locale en>
1120 <word>
1121 <data>•ジョージア<400> •</data>
1122 <data>•主场<400>客场<400>干练<400>条码<400>杯具<400>温婉<400>猕猴桃<400>肌肤<400>黑头<400>话唠<400>话痨<400> •</data>
1123
1124 ####################################################################################
1125 #
1126 #  Test Apple early change of thaidict
1127 #
1128 ####################################################################################
1129
1130 <locale th>
1131 <line>
1132 <data>•อัปเดต•อีเวนต์•</data>
1133
1134 ####################################################################################
1135 #
1136 #  Apple tests for currency break per <rdar://problem/24359173>
1137 #
1138 ####################################################################################
1139
1140 <locale en>
1141 <line>
1142 <data>•CN¥522.5•</data>
1143 <locale zh>
1144 <line>
1145 <data>•￥522.5•</data>
1146
1147 ####################################################################################
1148 #
1149 #  Apple tests for @lw=keepall per <rdar://problem/11679343&16501490>
1150 #  Default changes per <rdar://problem/23224717>
1151 #
1152 ####################################################################################
1153
1154 <locale ko>
1155 <line>
1156 <data>•\uBAA8\uB4E0 •\uC778\uB958 •\uAD6C\uC131\uC6D0\uC758 •\uCC9C\uBD80\uC758 •\uC874\uC5C4\uC131\uACFC •\uB3D9\uB4F1\uD558\uACE0 •\uC591\uB3C4\uD560•</data>
1157
1158 <locale ko@lw=keepall>
1159 <line>
1160 <data>•\uBAA8\uB4E0 •\uC778\uB958 •\uAD6C\uC131\uC6D0\uC758 •\uCC9C\uBD80\uC758 •\uC874\uC5C4\uC131\uACFC •\uB3D9\uB4F1\uD558\uACE0 •\uC591\uB3C4\uD560•</data>
1161
1162 <locale ko@lw=normal>
1163 <line>
1164 <data>•\uBAA8•\uB4E0 •\uC778•\uB958 •\uAD6C•\uC131•\uC6D0•\uC758 •\uCC9C•\uBD80•\uC758 •\uC874•\uC5C4•\uC131•\uACFC •\uB3D9•\uB4F1•\uD558•\uACE0 •\uC591•\uB3C4•\uD560•</data>
1165
1166 ####################################################################################
1167 #
1168 #  Test breaks for emoji clusters (same for all locales, mostly the same for all break types)
1169 #
1170 ####################################################################################
1171
1172 <locale root>
1173
1174 <char>
1175 # woman zwj woman zwj girl zwj girl, woman/fitz-1-2 zwj woman/fitz-4 zwj boy/fitz-6
1176 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•\U0001F469\U0001F3FB\u200D\U0001F469\U0001F3FD\u200D\U0001F466\U0001F3FF•</data>
1177 # woman zwj baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1178 <data>•\U0001F469\u200D\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1179 # man zwj hvy_blk_heart zwj man, woman, man zwj hvy_blk_heart/evar zwj man, woman
1180 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1181 # woman zwj hvy_blk_heart/evar zwj kiss_mark zwj woman, sleuth/fitz-4, horse_racing/fitz-5
1182 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1183 # victory_hand/evar, victory_hand/evar/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5, space,
1184 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE•\u0020•</data>
1185 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1186 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8•\u0020•</data>
1187 # family (man, boy, boy), family (woman, girl), space
1188 <data>•\U0001F468\u200D\U0001F466\u200D\U0001F466•\U0001F469\u200D\U0001F467•\u0020•</data>
1189 # rainbow flag using evar, rainbow flag no evar, space
1190 <data>•\U0001F3F3\uFE0F\u200D\U0001F308•\U0001F3F3\u200D\U0001F308•\u0020•</data>
1191 # woman with ball using evar, woman with ball no evar, woman with ball fitz 4 using evar, woman with ball fitz 4 no evar, space
1192 <data>•\u26F9\uFE0F\u200D\u2640\uFE0F•\u26F9\u200D\u2640•\u26F9\U0001F3FD\u200D\u2640\uFE0F•\u26F9\U0001F3FD\u200D\u2640•\u0020•</data>
1193 # woman runner using evar, woman runner no evar, woman runner fitz 4 using evar, woman runner fitz 4 no evar, space
1194 <data>•\U0001F3C3\u200D\u2640\uFE0F•\U0001F3C3\u200D\u2640•\U0001F3C3\U0001F3FD\u200D\u2640\uFE0F•\U0001F3C3\U0001F3FD\u200D\u2640•\u0020•</data>
1195 # 9.0 + professions
1196 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1197 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD•\u0020•</data>
1198 # facepalm, facepalm / fitz4, facepalm + woman +var, facepalm + woman -var, facepalm/fitz4 + woman +var, facepalm/fitz4 + woman -var
1199 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640•\u0020•</data>
1200 # handball, handball / fitz4, handball + man +var, handball + man -var, handball/fitz4 + man +var, handball/fitz4 + man -var
1201 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642•\u0020•</data>
1202 # man health worker +var, man health worker -var, man health worker / fitz4 +var, man health worker / fitz4 -var
1203 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695•\u0020•</data>
1204 # woman astronaut, woman astronaut / fitz4
1205 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680•\u0020•</data>
1206 # modifier bases added in emoji 4.0
1207 # SNOWBOARDER/fitz-3, HORSE RACING/fitz-4, GOLFER/fitz-5, MAN IN BUSINESS SUIT LEVITATING/fitz-6, SLEEPING ACCOMMODATION/fitz-12, space
1208 <data>•\U0001F3C2\U0001F3FC•\U0001F3C7\U0001F3FD•\U0001F3CC\U0001F3FE•\U0001F574\U0001F3FF•\U0001F6CC\U0001F3FB•\u0020•</data>
1209 # test kiss: man, man with variation selector <rdar://problem/30614982>
1210 <data>•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F468•\u0020•</data>
1211 # modifier bases we do not support, but other platforms do
1212 # family/fitz-3, TWO WOMEN HOLDING HANDS/fitz-4, WOMAN WITH BUNNY EARS/fitz-5, HANDSHAKE/fitz-6, WRESTLERS/fitz-12, space
1213 <data>•\U0001F46A\U0001F3FC•\U0001F46D\U0001F3FD•\U0001F46F\U0001F3FE•\U0001F91D\U0001F3FF•\U0001F93C\U0001F3FB•\u0020•</data>
1214
1215 # flags1    AE AF AL AM AO AR AT
1216 <data>•\U0001F1E6\U0001F1EA•\U0001F1E6\U0001F1EB•\U0001F1E6\U0001F1F1•\U0001F1E6\U0001F1F2•\U0001F1E6\U0001F1F4•\U0001F1E6\U0001F1F7•\U0001F1E6\U0001F1F9•</data>
1217 # flags2    AU AZ BA BD BE BF BG
1218 <data>•\U0001F1E6\U0001F1FA•\U0001F1E6\U0001F1FF•\U0001F1E7\U0001F1E6•\U0001F1E7\U0001F1E9•\U0001F1E7\U0001F1EA•\U0001F1E7\U0001F1EB•\U0001F1E7\U0001F1EC•</data>
1219 # flags3    BH BJ BN BO BR BS BT
1220 <data>•\U0001F1E7\U0001F1ED•\U0001F1E7\U0001F1EF•\U0001F1E7\U0001F1F3•\U0001F1E7\U0001F1F4•\U0001F1E7\U0001F1F7•\U0001F1E7\U0001F1F8•\U0001F1E7\U0001F1F9•</data>
1221 # flags4    BW BY BZ CA CD CF CG
1222 <data>•\U0001F1E7\U0001F1FC•\U0001F1E7\U0001F1FE•\U0001F1E7\U0001F1FF•\U0001F1E8\U0001F1E6•\U0001F1E8\U0001F1E9•\U0001F1E8\U0001F1EB•\U0001F1E8\U0001F1EC•</data>
1223 # flags5    CH CI CL CM CN CO CR
1224 <data>•\U0001F1E8\U0001F1ED•\U0001F1E8\U0001F1EE•\U0001F1E8\U0001F1F1•\U0001F1E8\U0001F1F2•\U0001F1E8\U0001F1F3•\U0001F1E8\U0001F1F4•\U0001F1E8\U0001F1F7•</data>
1225 # flags6    CU CV CY CZ DE DJ DK
1226 <data>•\U0001F1E8\U0001F1FA•\U0001F1E8\U0001F1FB•\U0001F1E8\U0001F1FE•\U0001F1E8\U0001F1FF•\U0001F1E9\U0001F1EA•\U0001F1E9\U0001F1EF•\U0001F1E9\U0001F1F0•</data>
1227 # flags7    DM DO DZ EC EE EG ER
1228 <data>•\U0001F1E9\U0001F1F2•\U0001F1E9\U0001F1F4•\U0001F1E9\U0001F1FF•\U0001F1EA\U0001F1E8•\U0001F1EA\U0001F1EA•\U0001F1EA\U0001F1EC•\U0001F1EA\U0001F1F7•</data>
1229 # flags8    ES ET FI FJ FR GA GB
1230 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F9•\U0001F1EB\U0001F1EE•\U0001F1EB\U0001F1EF•\U0001F1EB\U0001F1F7•\U0001F1EC\U0001F1E6•\U0001F1EC\U0001F1E7•</data>
1231 # flags9    GE GH GM GN GR GT GW
1232 <data>•\U0001F1EC\U0001F1EA•\U0001F1EC\U0001F1ED•\U0001F1EC\U0001F1F2•\U0001F1EC\U0001F1F3•\U0001F1EC\U0001F1F7•\U0001F1EC\U0001F1F9•\U0001F1EC\U0001F1FC•</data>
1233 # flags10   GY HK HN HR HT HU ID
1234 <data>•\U0001F1EC\U0001F1FE•\U0001F1ED\U0001F1F0•\U0001F1ED\U0001F1F3•\U0001F1ED\U0001F1F7•\U0001F1ED\U0001F1F9•\U0001F1ED\U0001F1FA•\U0001F1EE\U0001F1E9•</data>
1235 # flags11   IE IL IN IQ IR IS IT
1236 <data>•\U0001F1EE\U0001F1EA•\U0001F1EE\U0001F1F1•\U0001F1EE\U0001F1F3•\U0001F1EE\U0001F1F6•\U0001F1EE\U0001F1F7•\U0001F1EE\U0001F1F8•\U0001F1EE\U0001F1F9•</data>
1237 # flags12   JM JO JP KE KG KH KR
1238 <data>•\U0001F1EF\U0001F1F2•\U0001F1EF\U0001F1F4•\U0001F1EF\U0001F1F5•\U0001F1F0\U0001F1EA•\U0001F1F0\U0001F1EC•\U0001F1F0\U0001F1ED•\U0001F1F0\U0001F1F7•</data>
1239 # flags13   MX MY NL NO PL PT
1240 <data>•\U0001F1F2\U0001F1FD•\U0001F1F2\U0001F1FE•\U0001F1F3\U0001F1F1•\U0001F1F3\U0001F1F4•\U0001F1F5\U0001F1F1•\U0001F1F5\U0001F1F9•</data>
1241 # flags14   RO RU SA SE SK TH TR
1242 <data>•\U0001F1F7\U0001F1F4•\U0001F1F7\U0001F1FA•\U0001F1F8\U0001F1E6•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1F0•\U0001F1F9\U0001F1ED•\U0001F1F9\U0001F1F7•</data>
1243 # flags15   UA US VN XK ZW
1244 <data>•\U0001F1FA\U0001F1E6•\U0001F1FA\U0001F1F8•\U0001F1FB\U0001F1F3•\U0001F1FD\U0001F1F0•\U0001F1FF\U0001F1FC•</data>
1245 # flagsX1   ES ES ES SE SE SE
1246 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•</data>
1247 # flagsX2   GB GB GB BG BG BG
1248 <data>•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•</data>
1249 # flagsXtnd AE AF AL AM AO AR
1250 <data>•\U0001F1E6\U0001F1EA\u200C•\U0001F1E6\U0001F1EB\u200C•\U0001F1E6\U0001F1F1\u200C•\U0001F1E6\U0001F1F2\u0300•\U0001F1E6\U0001F1F4\u20DE•\U0001F1E6\U0001F1F7\u200C•</data>
1251
1252 <word>
1253 # woman zwj woman zwj girl zwj girl, woman/fitz-1-2 zwj woman/fitz-4 zwj boy/fitz-6
1254 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•\U0001F469\U0001F3FB\u200D\U0001F469\U0001F3FD\u200D\U0001F466\U0001F3FF•</data>
1255 # woman zwj baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1256 <data>•\U0001F469\u200D\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1257 # man zwj hvy_blk_heart zwj man, woman, man zwj hvy_blk_heart/evar zwj man, woman
1258 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1259 # woman zwj hvy_blk_heart/evar zwj kiss_mark zwj woman, sleuth/fitz-4, horse_racing/fitz-5
1260 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1261 # victory_hand/evar, victory_hand/evar/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5, space,
1262 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE•\u0020•</data>
1263 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1264 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8•\u0020•</data>
1265 # family (man, boy, boy), family (woman, girl), space
1266 <data>•\U0001F468\u200D\U0001F466\u200D\U0001F466•\U0001F469\u200D\U0001F467•\u0020•</data>
1267 # rainbow flag using evar, rainbow flag no evar, space
1268 <data>•\U0001F3F3\uFE0F\u200D\U0001F308•\U0001F3F3\u200D\U0001F308•\u0020•</data>
1269 # woman with ball using evar, woman with ball no evar, woman with ball fitz 4 using evar, woman with ball fitz 4 no evar, space
1270 <data>•\u26F9\uFE0F\u200D\u2640\uFE0F•\u26F9\u200D\u2640•\u26F9\U0001F3FD\u200D\u2640\uFE0F•\u26F9\U0001F3FD\u200D\u2640•\u0020•</data>
1271 # woman runner using evar, woman runner no evar, woman runner fitz 4 using evar, woman runner fitz 4 no evar, space
1272 <data>•\U0001F3C3\u200D\u2640\uFE0F•\U0001F3C3\u200D\u2640•\U0001F3C3\U0001F3FD\u200D\u2640\uFE0F•\U0001F3C3\U0001F3FD\u200D\u2640•\u0020•</data>
1273 # 9.0 + professions
1274 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1275 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD•\u0020•</data>
1276 # facepalm, facepalm / fitz4, facepalm + woman +evar, facepalm + woman -evar, facepalm/fitz4 + woman +evar, facepalm/fitz4 + woman -evar
1277 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640•\u0020•</data>
1278 # handball, handball / fitz4, handball + man +evar, handball + man -evar, handball/fitz4 + man +evar, handball/fitz4 + man -evar
1279 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642•\u0020•</data>
1280 # man health worker +evar, man health worker -evar, man health worker / fitz4 +evar, man health worker / fitz4 -evar
1281 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695•\u0020•</data>
1282 # woman astronaut, woman astronaut / fitz4
1283 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680•\u0020•</data>
1284 # modifier bases added in emoji 4.0
1285 # SNOWBOARDER/fitz-3, HORSE RACING/fitz-4, GOLFER/fitz-5, MAN IN BUSINESS SUIT LEVITATING/fitz-6, SLEEPING ACCOMMODATION/fitz-12, space
1286 <data>•\U0001F3C2\U0001F3FC•\U0001F3C7\U0001F3FD•\U0001F3CC\U0001F3FE•\U0001F574\U0001F3FF•\U0001F6CC\U0001F3FB•\u0020•</data>
1287 # modifier bases we do not support, but other platforms do
1288 # family/fitz-3, TWO WOMEN HOLDING HANDS/fitz-4, WOMAN WITH BUNNY EARS/fitz-5, HANDSHAKE/fitz-6, WRESTLERS/fitz-12, space
1289 <data>•\U0001F46A\U0001F3FC•\U0001F46D\U0001F3FD•\U0001F46F\U0001F3FE•\U0001F91D\U0001F3FF•\U0001F93C\U0001F3FB•\u0020•</data>
1290
1291 # flags1    AE AF AL AM AO AR AT
1292 <data>•\U0001F1E6\U0001F1EA•\U0001F1E6\U0001F1EB•\U0001F1E6\U0001F1F1•\U0001F1E6\U0001F1F2•\U0001F1E6\U0001F1F4•\U0001F1E6\U0001F1F7•\U0001F1E6\U0001F1F9•</data>
1293 # flags2    AU AZ BA BD BE BF BG
1294 <data>•\U0001F1E6\U0001F1FA•\U0001F1E6\U0001F1FF•\U0001F1E7\U0001F1E6•\U0001F1E7\U0001F1E9•\U0001F1E7\U0001F1EA•\U0001F1E7\U0001F1EB•\U0001F1E7\U0001F1EC•</data>
1295 # flags3    BH BJ BN BO BR BS BT
1296 <data>•\U0001F1E7\U0001F1ED•\U0001F1E7\U0001F1EF•\U0001F1E7\U0001F1F3•\U0001F1E7\U0001F1F4•\U0001F1E7\U0001F1F7•\U0001F1E7\U0001F1F8•\U0001F1E7\U0001F1F9•</data>
1297 # flags4    BW BY BZ CA CD CF CG
1298 <data>•\U0001F1E7\U0001F1FC•\U0001F1E7\U0001F1FE•\U0001F1E7\U0001F1FF•\U0001F1E8\U0001F1E6•\U0001F1E8\U0001F1E9•\U0001F1E8\U0001F1EB•\U0001F1E8\U0001F1EC•</data>
1299 # flags5    CH CI CL CM CN CO CR
1300 <data>•\U0001F1E8\U0001F1ED•\U0001F1E8\U0001F1EE•\U0001F1E8\U0001F1F1•\U0001F1E8\U0001F1F2•\U0001F1E8\U0001F1F3•\U0001F1E8\U0001F1F4•\U0001F1E8\U0001F1F7•</data>
1301 # flags6    CU CV CY CZ DE DJ DK
1302 <data>•\U0001F1E8\U0001F1FA•\U0001F1E8\U0001F1FB•\U0001F1E8\U0001F1FE•\U0001F1E8\U0001F1FF•\U0001F1E9\U0001F1EA•\U0001F1E9\U0001F1EF•\U0001F1E9\U0001F1F0•</data>
1303 # flags7    DM DO DZ EC EE EG ER
1304 <data>•\U0001F1E9\U0001F1F2•\U0001F1E9\U0001F1F4•\U0001F1E9\U0001F1FF•\U0001F1EA\U0001F1E8•\U0001F1EA\U0001F1EA•\U0001F1EA\U0001F1EC•\U0001F1EA\U0001F1F7•</data>
1305 # flags8    ES ET FI FJ FR GA GB
1306 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F9•\U0001F1EB\U0001F1EE•\U0001F1EB\U0001F1EF•\U0001F1EB\U0001F1F7•\U0001F1EC\U0001F1E6•\U0001F1EC\U0001F1E7•</data>
1307 # flags9    GE GH GM GN GR GT GW
1308 <data>•\U0001F1EC\U0001F1EA•\U0001F1EC\U0001F1ED•\U0001F1EC\U0001F1F2•\U0001F1EC\U0001F1F3•\U0001F1EC\U0001F1F7•\U0001F1EC\U0001F1F9•\U0001F1EC\U0001F1FC•</data>
1309 # flags10   GY HK HN HR HT HU ID
1310 <data>•\U0001F1EC\U0001F1FE•\U0001F1ED\U0001F1F0•\U0001F1ED\U0001F1F3•\U0001F1ED\U0001F1F7•\U0001F1ED\U0001F1F9•\U0001F1ED\U0001F1FA•\U0001F1EE\U0001F1E9•</data>
1311 # flags11   IE IL IN IQ IR IS IT
1312 <data>•\U0001F1EE\U0001F1EA•\U0001F1EE\U0001F1F1•\U0001F1EE\U0001F1F3•\U0001F1EE\U0001F1F6•\U0001F1EE\U0001F1F7•\U0001F1EE\U0001F1F8•\U0001F1EE\U0001F1F9•</data>
1313 # flags12   JM JO JP KE KG KH KR
1314 <data>•\U0001F1EF\U0001F1F2•\U0001F1EF\U0001F1F4•\U0001F1EF\U0001F1F5•\U0001F1F0\U0001F1EA•\U0001F1F0\U0001F1EC•\U0001F1F0\U0001F1ED•\U0001F1F0\U0001F1F7•</data>
1315 # flags13   MX MY NL NO PL PT
1316 <data>•\U0001F1F2\U0001F1FD•\U0001F1F2\U0001F1FE•\U0001F1F3\U0001F1F1•\U0001F1F3\U0001F1F4•\U0001F1F5\U0001F1F1•\U0001F1F5\U0001F1F9•</data>
1317 # flags14   RO RU SA SE SK TH TR
1318 <data>•\U0001F1F7\U0001F1F4•\U0001F1F7\U0001F1FA•\U0001F1F8\U0001F1E6•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1F0•\U0001F1F9\U0001F1ED•\U0001F1F9\U0001F1F7•</data>
1319 # flags15   UA US VN XK ZW
1320 <data>•\U0001F1FA\U0001F1E6•\U0001F1FA\U0001F1F8•\U0001F1FB\U0001F1F3•\U0001F1FD\U0001F1F0•\U0001F1FF\U0001F1FC•</data>
1321 # flagsX1   ES ES ES SE SE SE
1322 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•</data>
1323 # flagsX2   GB GB GB BG BG BG
1324 <data>•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•</data>
1325 # flagsXtnd AE AF AL AM AO AR
1326 <data>•\U0001F1E6\U0001F1EA\u200C•\U0001F1E6\U0001F1EB\u200C•\U0001F1E6\U0001F1F1\u200C•\U0001F1E6\U0001F1F2\u0300•\U0001F1E6\U0001F1F4\u20DE•\U0001F1E6\U0001F1F7\u200C•</data>
1327
1328 <line>
1329 # woman zwj woman zwj girl zwj girl
1330 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•</data>
1331 # woman zwj baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1332 <data>•\U0001F469\u200D\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1333 # man zwj hvy_blk_heart zwj man, woman, man zwj hvy_blk_heart/evar zwj man, woman
1334 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1335 # woman zwj hvy_blk_heart/evar zwj kiss_mark zwj woman, sleuth/fitz-4, horse_racing/fitz-5
1336 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1337 # victory_hand/evar, victory_hand/evar/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5 space,
1338 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE\u0020•</data>
1339 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1340 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8\u0020•</data>
1341 # family (man, boy, boy), family (woman, girl), space
1342 <data>•\U0001F468\u200D\U0001F466\u200D\U0001F466•\U0001F469\u200D\U0001F467\u0020•</data>
1343 # rainbow flag using evar, rainbow flag no evar, space
1344 <data>•\U0001F3F3\uFE0F\u200D\U0001F308•\U0001F3F3\u200D\U0001F308\u0020•</data>
1345 # woman with ball using evar, woman with ball no evar, woman with ball fitz 4 using evar, woman with ball fitz 4 no evar, space
1346 <data>•\u26F9\uFE0F\u200D\u2640\uFE0F•\u26F9\u200D\u2640•\u26F9\U0001F3FD\u200D\u2640\uFE0F•\u26F9\U0001F3FD\u200D\u2640\u0020•</data>
1347 # woman runner using evar, woman runner no evar, woman runner fitz 4 using evar, woman runner fitz 4 no evar, space
1348 <data>•\U0001F3C3\u200D\u2640\uFE0F•\U0001F3C3\u200D\u2640•\U0001F3C3\U0001F3FD\u200D\u2640\uFE0F•\U0001F3C3\U0001F3FD\u200D\u2640\u0020•</data>
1349 # 9.0 + professions
1350 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1351 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD\u0020•</data>
1352 # facepalm, facepalm / fitz4, facepalm + woman +evar, facepalm + woman -evar, facepalm/fitz4 + woman +evar, facepalm/fitz4 + woman -evar
1353 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640\u0020•</data>
1354 # handball, handball / fitz4, handball + man +evar, handball + man -evar, handball/fitz4 + man +evar, handball/fitz4 + man -evar
1355 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642\u0020•</data>
1356 # man health worker +evar, man health worker -evar, man health worker / fitz4 +evar, man health worker / fitz4 -evar
1357 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695\u0020•</data>
1358 # woman astronaut, woman astronaut / fitz4
1359 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680\u0020•</data>
1360 # modifier bases added in emoji 4.0
1361 # SNOWBOARDER/fitz-3, HORSE RACING/fitz-4, GOLFER/fitz-5, MAN IN BUSINESS SUIT LEVITATING/fitz-6, SLEEPING ACCOMMODATION/fitz-12 space
1362 <data>•\U0001F3C2\U0001F3FC•\U0001F3C7\U0001F3FD•\U0001F3CC\U0001F3FE•\U0001F574\U0001F3FF•\U0001F6CC\U0001F3FB\u0020•</data>
1363 # modifier bases we do not support, but other platforms do
1364 # family/fitz-3, TWO WOMEN HOLDING HANDS/fitz-4, WOMAN WITH BUNNY EARS/fitz-5, HANDSHAKE/fitz-6, WRESTLERS/fitz-12 space
1365 <data>•\U0001F46A\U0001F3FC•\U0001F46D\U0001F3FD•\U0001F46F\U0001F3FE•\U0001F91D\U0001F3FF•\U0001F93C\U0001F3FB\u0020•</data>
1366
1367 # no special flags handling for line
1368
1369 <locale ja@lb=loose>
1370 <line>
1371 # woman zwj woman zwj girl zwj girl
1372 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•</data>
1373 # woman zwj baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1374 <data>•\U0001F469\u200D\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1375 # man zwj hvy_blk_heart zwj man, woman, man zwj hvy_blk_heart/evar zwj man, woman
1376 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1377 # woman zwj hvy_blk_heart/evar zwj kiss_mark zwj woman, sleuth/fitz-4, horseRacing/fitz-5
1378 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1379 # victory_hand/evar, victory_hand/evar/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5 space,
1380 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE\u0020•</data>
1381 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1382 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8\u0020•</data>
1383 # 9.0 + professions
1384 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1385 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD\u0020•</data>
1386 # facepalm, facepalm / fitz4, facepalm + woman +evar, facepalm + woman -evar, facepalm/fitz4 + woman +evar, facepalm/fitz4 + woman -evar
1387 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640\u0020•</data>
1388 # handball, handball / fitz4, handball + man +evar, handball + man -evar, handball/fitz4 + man +evar, handball/fitz4 + man -evar
1389 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642\u0020•</data>
1390 # man health worker +evar, man health worker -evar, man health worker / fitz4 +evar, man health worker / fitz4 -evar
1391 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695\u0020•</data>
1392 # woman astronaut, woman astronaut / fitz4
1393 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680\u0020•</data>
1394 # modifiers we do not support, but other platforms do
1395 # family/fitz-3, wrestlers/fitz-4, space
1396 <data>•\U0001F46A\U0001F3FC•\U0001F93C\U0001F3FD\u0020•</data>
1397
1398 # no special flags handling for line