icuSources/test/testdata/rbbitst.txt

   1 # Copyright (c) 2001-2016 International Business Machines
   2 # Corporation and others. All Rights Reserved.
   3 #
   4 # RBBI Test Data
   5 #
   6 #   File:  rbbitst.txt
   7 #
   8 #     The format of this file looks vaguely like some kind of xml-ish markup,
   9 #     but it is NOT.  The syntax is this..
  10 #
  11 #   <word>    any following data is for word break testing
  12 #   <sent>    any following data is for sentence break testing
  13 #   <line>    any following data is for line break testing
  14 #   <char>    any following data is for char break testing
  15 #   <locale local_name>  Switch to the named locale at the next occurence of <word>, <sent>, etc.
  16 #   <data> ... </data>   test data.  May span multiple lines.
  17 #      <>        Break position, status == 0
  18 #      •         Break position, status == 0   (Bullet, \u2022)
  19 #      <nnn>     Break position, status == nnn
  20 #      \         Escape.  Normal ICU unescape applied.
  21 #      \ at end of line  ->  Line Continuation.  Remove both the backslash and the new line
  22 #
  23 # In ICU4C, this test data is run by intltest, rbbi/RBBITest/TestExtended.
  24 # In ICU4J, this test data is run by com.ibm.icu.dev.test.rbbi.RBBITestExtended
  25 #
  26 # There are two copies of this file in the source repository,
  27 #   [ICU4C]   source/test/testdata/rbbitst.txt
  28 #   [ICU4J]   main/tests/core/src/com/ibm/icu/dev/test/rbbi/rbbitst.txt
  29 #
  30 # ICU4C's copy is the master.  If any changes are made to ICU4J's copy, make sure they
  31 #  are merged back into ICU4C's copy of the file, lest they get overwritten later.
  32 # TODO:  figure out how to have a single copy of the file for use by both C and Java.
  33
  34
  35 ## FILTERED BREAK TESTS
  36
  37 # (William Bradford, public domain. http://catalog.hathitrust.org/Record/008651224 ) - edited.
  38 <locale en>
  39 <sent>
  40 <data>\
  41 •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>
  42
  43 <locale en@ss=standard>
  44 <sent>
  45 <data>\
  46 •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>
  47
  48 ## END FILTERED BREAK TESTS
  49
  50 <locale>
  51
  52 #   Temp debugging tests
  53 <sent>
  54 <data>•\u00c0.•</data>
  55
  56 #<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•</data>
  57 ########################################################################################
  58 #
  59 #
  60 #       G r a p h e m e    C l u s t e r     T e s t s
  61 #
  62 #
  63 ##########################################################################################
  64 <char>
  65
  66 <data>•a•b•c• •,•\u0666•</data>                 # Quick Test
  67 <data>•\r•\r•\r\n•\r\n•\n•\r•</data>            # don't break CR/LF
  68
  69 # Always break after controls.  Combining chars don't combine with them.
  70 <data>•\u0003•\N{COMBINING GRAVE ACCENT}•\r•\N{COMBINING GRAVE ACCENT}•</data>
  71 <data>•\u0085•\N{COMBINING MACRON}•A\N{COMBINING MACRON}•</data>
  72
  73 # Surrogates
  74 <data>•\U00011000•\U00010020•\U00010000\N{COMBINING MACRON}•</data>
  75 <data>•\ud800\udc00•\udbff\udfff•a•</data>
  76
  77 # Extend (Combining chars) combine.
  78 <data>•A\N{COMBINING GRAVE ACCENT}•B•</data>
  79 <data>•\N{GREEK SMALL LETTER MU}\N{COMBINING LOW LINE}\N{COMBINING HORN}•</data>
  80 <data>•a\u0301•b\u0302•c\u0303•d\u0304•e\u0305•f\u0306•g\u0307•h\u0308•i\u0309•</data>
  81
  82 <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>
  83
  84 # Don't break Hangul Syllables
  85 #   L : \u1100
  86 #   V : \u1161
  87 #   T : \u11A8
  88 #  LV : \uAC00
  89 # LVT : \uAC01
  90
  91 <data>•\u1100\u1161\u11a8•\u1100\u1161\u11a8•</data>  #LVT
  92 <data>•\u1100\u1161•\u1100\u1161•</data>
  93 <data>•\u1100\u1161\u11a8•\u1161•\u1100•\u11a8•\u1161\u1161\u1161\u11a8•</data>
  94 <data>•\u1100\u1100\uac01•\u1100\uac01•\u1100\uac01\u0301•\uac01•</data>
  95 <data>•\u1100\u0301•\u1161\u11a8\u0301•\u11a8•</data>
  96
  97
  98
  99 # Hindi combining chars.  (An old test)
 100 #   TODO:  Update these tests for Unicode 5.1 Extended Grapheme clusters
 101 #<data>•भ••ा•\u0930•\u0924• •\u0938\u0941\u0902•\u0926•\u0930•
 102 #•\u0939•\u094c•\u0964•</data>
 103 #<data>•\u0916\u0947•\u0938\u0941\u0902•\u0926•\u0930•\u0939•\u094c•\u0964•</data>
 104
 105
 106 #  Bug 1587.  Tamil.  \u0baa\u0bc1 is an Extended Grpaheme Cluster
 107 <data>•\u0baa\u0bc1•\u0baa\u0bc1•</data>
 108
 109 #   Regression test for bug 1889
 110 <data>•\u0f40\u0f7d•\u0000•\u0f7e•</data>
 111
 112
 113 #     0xffff is a legal character, and should not stop the break iterator early.
 114 #     (Requires special casing in implementation, which is why it gets a test.)
 115 <data>•\uffff•\uffff• •a•</data>
 116
 117 #  Treat Japanese Half Width voicing marks as combining
 118 <data>•A\uff9e•B\uff9f\uff9e\uff9f•C•</data>
 119
 120 ########################################################################################
 121 #
 122 #
 123 #       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
 124 #
 125 #
 126 ##########################################################################################
 127 #<xgc>
 128
 129 # Plain Vanilla grapheme clusters
 130 #<data>•a•b•c•</data>
 131 #<data>•a\u0301\u0302• •b\u0303\u0304•</data>
 132
 133 # Assorted Hindi combining marks
 134 #<data>•\u0904\u0903• •\u0937\u093E• •\u0904\u093F• •\u0937\u0940• •\u0937\u0949• •\u0937\u094A• •\u0937\u094B• •\u0937\u094C•</data>
 135
 136 # Thai Clusters
 137 # $Prepend $Extend* $PrependBase $Extend*;
 138 #
 139 #<data>•\u0e40\u0e01•\u0e44\u0301\u0e23\u0302\u0303•\u0e40•\u0e40\u0e02•\u0e02• •</data>
 140
 141
 142 ########################################################################################
 143 #
 144 #
 145 #       W o r d    B o u n d a r y      T e s t s
 146 #
 147 #
 148 ##########################################################################################
 149
 150 <word>
 151 #
 152 # Quick sanity test
 153 #
 154 <data>•hello<200> •there<200> •goodbye<200></data>
 155 <data>•hello<200> •12345<100> •,•</data>
 156
 157
 158 #
 159 # Test data originally in RBBIAPITest::TestFirstNextFollowing() and TestLastPreviousPreceding()
 160 #
 161
 162 <word>
 163 <data>•This<200> •is<200> •a<200> •word<200> •break<200>.• • •Isn't<200> •it<200>?• •2.25<100></data>
 164
 165
 166
 167 #
 168 # Data originally from TestDefaultRuleBasedWordIteration()
 169 #
 170 <data>•Write<200> •wordrules<200>.• •123.456<100> •alpha\u00adbeta\u00adgamma<200> •\u092f\u0939<200> •</data>
 171 <data>• •\u0939\u093f\u0928\u094d\u200d\u0926\u0940<200> •\u0939\u0948<200> •\u0905\u093e\u092a<200> •\u0938\u093f\u0916\u094b\u0917\u0947<200>?•</data>
 172
 173 #Hindi Numbers
 174 <data>• •\u0968\u0966.\u0969\u096f<100> •\u0967\u0966\u0966.\u0966\u0966<100> •\N{RUPEE SIGN}•\u0967,\u0967\u0966\u0966.\u0966\u0966<100> • •\u0905\u092e\u091c<200>\n•</data>
 175
 176 <data>•\u0938\u094d\u200d\u0935\u0924\u0902deadTA\u0930<200>\r•It's<200> •$•30.10<100> •12,34<100>¢•£•¤•¥•alpha\u05f3beta\u05f4gamma<200> •</data>
 177
 178 <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>
 179
 180 #Hangul
 181 <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>
 182
 183 <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>
 184
 185 # Words containing non-BMP letters
 186 <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>
 187
 188 # Unassigned code points
 189 <data>•abc<200>\U0001D800•def<200>\U0001D3FF• •</data>
 190
 191 # Hiragana & Katakana stay together, but separates from each other and Latin.
 192 # *** what to do about theoretical combos of chars? i.e. hiragana + accent
 193 #<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>
 194
 195 # test normalization/dictionary handling of halfwidth katakana: same dictionary phrase in fullwidth and halfwidth
 196 <data>•芽キャベツ<400>芽キャﾍﾞツ<400></data>
 197
 198 # more Japanese tests
 199 # TODO: some script=common characters in the Hiragana and the Katakana block may not be treated correctly
 200 # (was formerly true for U+30FC); need to check and fix if so.
 201 #<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>
 202 <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>
 203
 204 # Testing of word boundary for dictionary word containing both kanji and kana
 205 <data>•中だるみ<400>蔵王の森<400>ウ離島<400></data>
 206
 207 # Testing of Chinese segmentation (taken from a Chinese news article)
 208 <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>
 209
 210 # Words with interior formatting characters
 211 <data>•def\N{COMBINING ACUTE ACCENT}\N{SYRIAC ABBREVIATION MARK}ghi<200> •</data>
 212
 213 # to test for bug #4097779
 214 <data>•aa\N{COMBINING GRAVE ACCENT}a<200> •</data>
 215
 216 # fullwidth numeric, midletter characters etc should be treated like their halfwidth counterparts
 217 # <data>•ＩＳＮ'Ｔ<200> •１９<100>日<400></data>
 218 # why was this added with the dbbi stuff?
 219
 220 #      to test for bug #4098467
 221 #      What follows is a string of Korean characters (I found it in the Yellow Pages
 222 #      ad for the Korean Presbyterian Church of San Francisco, and I hope I transcribed
 223 #      it correctly), first as precomposed syllables, and then as conjoining jamo.
 224 #      Both sequences should be semantically identical and break the same way.
 225 #      precomposed syllables...
 226 <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>
 227
 228 # more Korean tests (Jamo not tested here, not counted as dictionary characters)
 229 # Disable them now because we don't include a Korean dictionary.
 230 #<data>•\ud55c\uad6d<200>\ub300\ud559\uad50<200>\uc790\uc5f0<200>\uacfc\ud559<200>\ub300\ud559<200>\ubb3c\ub9ac\ud559\uacfc<200></data>
 231 #<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>
 232
 233 <data>•abc<200>\u4e01<400>\u4e02<400>\u3005<400>\u4e03\u4e03<400>abc<200> •</data>
 234
 235 <data>•\u06c9<200>\uc799\ufffa•</data>
 236
 237
 238 #
 239 #      Try some words from other scripts.
 240 #
 241
 242 #       Try some words from other scripts.
 243 #          Greek, Cyrillic, Hebrew, Arabic, Arabic, Georgian, Latin
 244 #
 245 <data>•ΑΒΓ<200> •БВГ<200> •אבג֓<200> •ابت<200> •١٢٣<100> •\u10A0\u10A1\u10A2<200> •ABC<200> •</data>
 246
 247 <data>•\u0301•A<200></data>
 248
 249
 250 #
 251 #  Hindi word break tests, imported from the old RBBI tests.
 252 #    An historical note:  a much earlier version of ICU break iterators had a number
 253 #    of special case rules for Hindi, which were tested by an earlier version of
 254 #    this test data.  The current RBBI rules do not special case Hindi in
 255 #    any way, making this test data much less signfificant.
 256 #
 257 <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>
 258 •\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>
 259
 260 #
 261 #  Failures from monkey tests
 262 #
 263 <data>•\u8527<400>\u02ba<200>\u0027\u0d42•\u00b7•\u09ea<100></data>
 264
 265 #
 266 # Jitterbug 5276 - treat Japanese half width voicing marks as Grapheme Extend
 267 #
 268 <data>•A\uff9e\uff9fBC<200> •1\uff9e\uff9f23<100></data>
 269
 270 # User guide example:
 271 <data>•Parlez<200>-•vous<200> •français<200> •?•</data>
 272
 273 # Test for #11673
 274 <word>
 275 <data>•ジョージア<400> •</data>
 276
 277 ########################################################################################
 278 #
 279 #
 280 #       S e n t e n c e     B o u n d a r y      T e s t s
 281 #
 282 #
 283 ##########################################################################################
 284
 285
 286 #
 287 # Test data originally from RBBI RBBITest::TestDefaultRuleBasedSentenceIteration()
 288 #
 289 <sent>
 290
 291
 292 <sent>
 293 <data>•This\n<100></data>
 294 <data>•Hello! •how are you? •I'am fine. •Thankyou. •How are you \
 295 doing? •This\n<100> costs $20,00,000. •</data>
 296
 297
 298 # Sentence ending in a quote.
 299 <data>•"Sentence ending with a quote." •Bye.•</data>
 300
 301 # Sentence, and test data, ending without a period or other terminator.
 302 <data>•Here is a random sentence, no ending period<100></data>
 303
 304
 305 <data>•  (This is it).  •Testing the sentence iterator. •\
 306 "This isn't it." •Hi! \
 307 •This is a simple sample sentence. •(This is it.) •This is a simple sample sentence. •\
 308 "This isn't it." •\
 309 Hi! •This is a simple sample sentence. •It does not have to make any sense as you can see. •Nel mezzo del cammin di nostra vita, mi ritrovai in una selva oscura. •Che la dritta via aveo smarrita. •He said, that I said, that you said!! •Don't rock the boat.\u2029•Because I am the daddy, that is why.
 310 •Not on my time (el timo.)! •</data>
 311
 312 <data>•Hello.  •So what!!\u2029•"But now," he said, \
 313 "I know!" •\
 314 Harris thumbed down several, including "Away We Go" (which became the huge success Oklahoma!). •One species, B. anthracis, is highly virulent.
 315 •Wolf said about Sounder:\
 316 "Beautifully thought-out and directed." •\
 317 Have you ever said, "This is where\tI shall live"? •He answered, \
 318 "You may not!" •Another popular saying is: "How do you do?". \n•\
 319 Yet another popular saying is: \
 320 'I'm fine thanks.' •\
 321 What is the proper use of the abbreviation pp.? •Yes, I am definatelly 12" tall!!\
 322 •Now\r<100>is\n<100>the\r\n<100>time\n<100>\r<100>for\r<100>\r<100></data>
 323
 324 <data>•No breaks when . is surrounded by UPPER.Case letters.  •</data>
 325 <data>•No breaks when . is followed by Numeric .4 a.4 C.4 3.1 .•</data>
 326 <data>•No breaks when . is followed by a lower, with possible intervening punct .,a .$a .)a. •</data>
 327
 328 #
 329 #  Sentence Breaks: no break at the boundary between CJK and other letters
 330 #
 331 <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>
 332
 333 #
 334 #      Treat fullwidth variants of .!? the same as their
 335 #      normal counterparts
 336 #
 337 <data>•I know I'm right\uff0e •Right\uff1f •Right\uff01 •</data>
 338
 339
 340 #
 341 #       Don't break sentences at boundary between CJK and digits
 342 #
 343 <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>
 344
 345 #
 346 #      Breaks around '(' following a sentence TERM.  (Rule 9)
 347 #
 348 <data>•How do you do?(•Fine). •</data>
 349 <data>•How do you do? •(Fine). •</data>
 350 <data>•How do you do?(•fine). •</data>
 351 <data>•How do you do? •(fine). •</data>
 352
 353 #
 354 <data>•Hello.123<100></data>    # Rule 6
 355 <data>•Hello?•123<100></data>
 356
 357 <data>•HELLO.Bye<100></data>    # Rule 7
 358 <data>•HELLO?•Bye<100></data>
 359
 360 <data>•Hello.goodbye<100></data>  #Rule 8
 361 <data>•Hello. •Goodbye<100></data>
 362 <data>•Hello. goodbye<100></data>
 363
 364
 365
 366 #
 367 #        test for bug #4158381: No breaks when there are no terminators around
 368 #
 369 <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>
 370 <data>•Another test.\u2029•</data>
 371
 372 #        test for bug #4143071: Make sure sentences that end with digits
 373 #        work right
 374 #
 375 <data>•Today is the 27th of May, 1998.  •</data>
 376 <data>•Tomorrow with be 28 May 1998.  •</data>
 377 <data>•The day after will be the 30th.\u2029•</data>
 378
 379 #        test for bug #4152416: Make sure sentences ending with a capital
 380 #        letter are treated correctly
 381 #
 382 <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>
 383
 384 #        test for bug #4152117: Make sure sentence breaking is handling
 385 #        punctuation correctly [COULD NOT REPRODUCE THIS BUG, BUT TEST IS
 386 #        HERE TO MAKE SURE IT DOESN'T CROP UP]
 387 #
 388 <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.
 389 •</data>
 390
 391 #        sentence breaks for hindi which used Devanagari script
 392 #        make sure there is sentence break after ?,danda(hindi phrase separator),
 393 #        fullstop followed by space.  (VERY old test)
 394 #
 395 <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>\
 396 \u0915\u0948\u0938\u0947 \u0939\u0948?•\u0935\u0939 \u0915\u094d\u200d\u092f\u093e\n\
 397 <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>
 398
 399 #         Regression test for bug #1984, Sentence break in Arabic text.
 400
 401 <data>\
 402 •\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>
 403
 404 #         Try a few more of the less common sentence endings.
 405 <data>•Hello, world\u3002 •Hello, world\u1803 •Hello, world\u2048 •Hello, world\u203c •Let's end here. •</data>
 406
 407
 408
 409
 410 ################################################################
 411 #
 412 #
 413 #        L I N E     B R E A K
 414 #
 415 #
 416 ################################################################
 417
 418 <line>
 419 #
 420 # Test Character for each of the line break classes.
 421 #
 422 # 00A1;AI # INVERTED EXCLAMATION MARK  ¡
 423 # 0041;AL # LATIN CAPITAL LETTER A
 424 # 0009;BA # <control>
 425 # 00B4;BB # ACUTE ACCENT
 426 # 000C;BK # <control>
 427 # 2014;B2 # EM DASH
 428 # FFFC;CB # OBJECT REPLACEMENT CHARACTER
 429 # 0029;CL # RIGHT PARENTHESIS
 430 # 0301;CM # COMBINING ACUTE ACCENT
 431 # 0021;EX # EXCLAMATION MARK
 432 # 00A0;GL # NO-BREAK SPACE
 433 # 002D;HY # HYPHEN-MINUS
 434 # 4E00;ID # <CJK Ideograph, First>
 435 # 2024;IN # ONE DOT LEADER
 436 # 002C;IS # COMMA
 437 # 000A;LF # <control>
 438 # 0E5A;NS # THAI CHARACTER ANGKHANKHU
 439 # 0032;NU # DIGIT TWO
 440 # 0028;OP # LEFT PARENTHESIS
 441 # 0025;PO # PERCENT SIGN
 442 # 0024;PR # DOLLAR SIGN
 443 # 0022;QU # QUOTATION MARK
 444 # 0E01;SA # THAI CHARACTER KO KAI
 445 # DB7F;SG # Surrogate
 446 # 0020;SP # SPACE
 447 # 002F;SY # SOLIDUS /
 448 # F8FF;XX # Private Use
 449 # 200B;ZW # ZERO WIDTH SPACE
 450
 451
 452 # 2b  Always break at end of text
 453
 454 <data>•  •\u00A1•</data>
 455 <data>•  •\u0041•</data>
 456 <data>•  •\u0009•</data>
 457 <data>•  •\u00B4•</data>
 458 <data>•    \u000C<100></data>    # LB3C  × BK
 459 <data>•  •\u2014•</data>
 460 <data>•  •\uFFFC•</data>
 461 <data>•   \u0029•</data>    # LB 8   × CL
 462 # <data>• • \u0301•</data>    # LB 7a   Treat SP CM* as if it were ID  #TODO:  SP CM
 463 <data>•   \u0021•</data>    # LB 8   × EX
 464 #<data>•   \u00A0•</data>    # LB 11b   × GL   TODO:  fix.
 465 <data>•  •\u002D•</data>
 466 <data>•  •\u4E00•</data>
 467 <data>•  •\u2024•</data>
 468 <data>•   \u002C•</data>    # LB 8   × IS
 469 <data>•   \u000A<100></data>    # LB3C   × ( BK | CR | LF | NL )
 470 <data>•  •\u0E5A•</data>
 471 <data>•  •\u0032•</data>
 472 <data>•  •\u0028•</data>
 473 <data>•  •\u0025•</data>
 474 <data>•  •\u0024•</data>
 475 <data>•  •\u0022•</data>
 476 <data>•  •\u0E01•</data>
 477 <data>•  •\uDB7F•</data>
 478 <data>•   \u0020•</data>     # LB4 - don't break before space.
 479 <data>•   \u002F•</data>     # LB 8  × SY
 480 <data>•  •\uF8FF•</data>
 481 <data>•   \u200B•</data>     # LB4 - don't break before ZA
 482
 483
 484 # 3a  Always break after hard line breaks.
 485 # 3c  Never break before hard line breaks.
 486
 487 <data>•  •\u00A1\u2028<100>\u00A1•</data>
 488 <data>•  •\u0041\u2028<100>\u0041•</data>
 489 <data>•  •\u0009\u2028<100>\u0009•</data>
 490 <data>•  •\u00B4\u2028<100>\u00B4•</data>
 491 <data>•   \u000C<100>\u2028<100>\u000C<100></data>
 492 <data>•  •\u2014\u2028<100>\u2014•</data>
 493 <data>•  •\uFFFC\u2028<100>\uFFFC•</data>
 494 <data>•   \u0029\u2028<100>\u0029•</data>
 495 #<data>•   \u0301\u2028<100>\u0301•</data>    # TODO:  fix.
 496 <data>•   \u0021\u2028<100>\u0021•</data>
 497 #<data>•   \u00A0\u2028<100>\u00A0•</data>    # TODO:  fix
 498 <data>•  •\u002D\u2028<100>\u002D•</data>
 499 <data>•  •\u4E00\u2028<100>\u4E00•</data>
 500 <data>•  •\u2024\u2028<100>\u2024•</data>
 501 <data>•   \u002C\u2028<100>\u002C•</data>
 502 <data>•   \u000A<100>\u2028<100>\u000A<100></data>
 503 <data>•  •\u0E5A\u2028<100>\u0E5A•</data>
 504 <data>•  •\u0032\u2028<100>\u0032•</data>
 505 <data>•  •\u0028\u2028<100>\u0028•</data>
 506 <data>•  •\u0025\u2028<100>\u0025•</data>
 507 <data>•  •\u0024\u2028<100>\u0024•</data>
 508 <data>•  •\u0022\u2028<100>\u0022•</data>
 509 <data>•  •\u0E01\u2028<100>\u0E01•</data>
 510 <data>•  •\uDB7F\u2028<100>\uDB7F•</data>
 511 <data>•   \u0020\u2028<100>\u0020•</data>
 512 <data>•   \u002F\u2028<100>\u002F•</data>
 513 <data>•  •\uF8FF\u2028<100>\uF8FF•</data>
 514 <data>•   \u200B\u2028<100>\u200B•</data>
 515
 516 # Regional Indicator sequences. They group in pairs. The reverse rules are tricky.
 517 #    Sequences are long enough that the non-exaustive monkey test won't reliably pick up problems.
 518
 519 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•</data>
 520 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6•</data>
 521
 522 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•</data>
 523 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6•</data>
 524 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•</data>
 525 <data>•\U0001F1E6\U0001F1E6•\U0001F1E6\u00a0\U0001F1E6\U0001F1E6•\U0001F1E6\U0001F1E6•\U0001F1E6•</data>
 526
 527
 528 # User Guide example
 529
 530 <data>•Parlez-•vous •français ?•</data>
 531
 532 #
 533 #  Old Line Break Test data.  Orginally located in RBBITest::TestDefaultRuleBasedLineIteration()
 534 #
 535
 536 <line>
 537
 538 <data>•Multi-•Level •example •of •a •semi-•idiotic •non-•sensical •(non-•important) •sentence.
 539 <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>
 540
 541 <line>
 542 <data>•Hello! •how\r\n<100> •(are)\r<100> •you? •I'am •fine- •Thankyou. •foo\u00a0bar
 543 <100>How, •are, •you? •This, •costs •$20,00,000.•</data>
 544
 545 #     test for bug #4068133
 546 #
 547 <data>•\u96f6•\u4e00\u3002•\u4e8c\u3001•\u4e09\u3002\u3001•\u56db\u3001\u3002\u3001•\u4e94,•\u516d.•\u4e03.\u3001,\u3002•\u516b•</data>
 548
 549 #     to test for bug #4086052
 550 <data>•foo\u00a0bar•</data>
 551
 552 #     to test for bug #4097920
 553 <data>•dog,cat,mouse •(one)•(two)\n<100></data>
 554
 555 #     to test for bug #4035266
 556 <data>•The •balance •is •$-23,456.78, •not •-•$32,456.78!\n<100></data>
 557
 558
 559 #      to test for bug #4098467
 560 #      What follows is a string of Korean characters (I found it in the Yellow Pages
 561 #      ad for the Korean Presbyterian Church of San Francisco, and I hope I transcribed
 562 #      it correctly), first as precomposed syllables, and then as conjoining jamo.
 563 #      Both sequences should be semantically identical and break the same way.
 564 #      precomposed syllables...  (I == Rich Gillam?)
 565 #
 566 <data>•\uc0c1•\ud56d •\ud55c•\uc778 •\uc5f0•\ud569 •\uc7a5•\ub85c•\uad50•\ud68c•</data>
 567
 568 #      conjoining jamo...
 569 <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>
 570
 571 #      to test for bug #4117554: Fullwidth .!? should be treated as postJwrd
 572 <data>•\u4e01\uff0e•\u4e02\uff01•\u4e03\uff1f•</data>
 573
 574 #      Surrogate line break tests.
 575 #
 576 <data>•\u4e01•\ud840\udc01•\u4e02•abc •\ue000 •\udb80\udc01•</data>  #This line and the following are equivalent.
 577 <data>•\u4e01•\U00020001•\u4e02•abc •\ue000 •\U000f0001•</data>
 578
 579 #      Regression for bug 836
 580 #        Note:  Unicode 5.1 changed this behavior
 581 #               Unicode 5.2 changed it again, there is no break following the '('
 582 <data>•AAA(AAA •</data>
 583
 584 #       Try some words from other scripts.
 585 #          Greek, Cyrillic, Hebrew, Arabic, Arabic, Georgian, Latin
 586 #
 587 <data>•ΑΒΓ •БВГ •אבג֓ •ابت •١٢٣ •\u10A0\u10A1\u10A2 •ABC •</data>
 588
 589 #
 590 #       ticket #4853:  unpaired surrogates should behave like AL
 591 #
 592 <data>•abc\ud801xyz•</data>
 593
 594 #
 595 #     Regression tests for failures that originally came from the monkey test.
 596 #     Monkey test failure lines can, with slight reformatting, be copied into this section
 597 #     as test cases.  The error display from here is more informative.
 598 #
 599 <data>•\ufffc•\u30e3\u000c<100>\u1b39\u300a\u002f\u203a\u200b•\ufffc•\uaf64•\udcfb•</data>
 600 <data>•\u114d\u31f3•\ube44\u002d•\u0362\u24e2\u276e\u2014\u205f\ufe16•\uc877•\u0fd0\u000a<100>\u20a3•</data>
 601 <data>•\u080a\u215b\U0001d7d3\u002c•\u2025\U000e012e•\u02df\u118d\u0029\ua8d6\u0085<100>\u6cc4\u2024\u202f\ufffc•</data>
 602
 603 # Test for #10176 (in root)
 604 <line>
 605 <data>•abc/•s •def•</data>
 606 <data>•abc/\u05D9 •def•</data>
 607 <data>•\u05E7\u05D7/\u05D9 •\u05DE\u05E2\u05D9\u05DC•</data>
 608 <data>•\u05D3\u05E8\u05D5\u05E9\u05D9\u05DD •\u05E9\u05D7\u05E7\u05E0\u05D9\u05DD/\u05D9\u05D5\u05EA•</data>
 609
 610
 611
 612 ########################################################################################
 613 #
 614 #
 615 #       T i t l e   B o u n d a r y      T e s t s
 616 #
 617 #
 618 ##########################################################################################
 619 <title>
 620 <data>•Here •is •a •short •sample •sentence.  •And •another.•</data>
 621 <data>•HERE •IS •A •SHORT •SAMPLE •SENTENCE.  •AND •ANOTHER.•</data>
 622 <data>•  •Start •and •end •with •spaces   •</data>
 623 <data>•Include 123 456 ^&  •some 54332 •numbers 4445•abc123•abc •ending 1223 •</data>
 624
 625 <data>•Combining\u0301 \u0301•ma\u0306rks  •bye  •</data>
 626 <data>•123  •Start •with •a •number.•</data>
 627
 628 <data>•'•start •with •a •case-•ignorable •cha'r'a'cter•</data>
 629 <data>•'  '' •start •with •case-•ignorable & •case-•insensitive •cha'r'a'cter•</data>
 630 <data>•  ''•aaa' •bbb '•ccc' '•ddd''' '''•eee   '''•fff'''   •ggg  ''•</data>
 631 # Note: apostrophe is case-ignorable. space is not cased.
 632
 633 ##########################################################################################
 634 #
 635 #   Thai Tests
 636 #
 637 ##########################################################################################
 638 <locale th>
 639 <word>
 640 #
 641 # Test data originally from the test code source file
 642 #      // @suwit -- Thai sample data from GVT Guideline
 643 #
 644 <data>•\u0E2B\u0E19\u0E36\u0E48\u0E07<200>\u0E04\u0E33<200>\u0E44\u0E17\u0E22<200>\
 645 \u0E2A\u0E32\u0E21\u0E32\u0E23\u0E16<200>\u0E1B\u0E23\u0E30\u0E01\u0E2D\u0E1A<200>\
 646 \u0E14\u0E49\u0E27\u0E22<200>\u0e2b\u0e25\u0e32\u0e22<200>\
 647 \u0e1e\u0e22\u0e32\u0e07\u0e04\u0e4c<200></data>
 648
 649 # Test data originally from http://bugs.icu-project.org/trac/search?q=r30327
 650 <data>•กู<200> •กิน<200>กุ้ง<200> •ปิ้่<200>งอ<200>ยู่<200>ใน<200>ถ้ำ<200></data>
 651
 652 <data>•\u0E01\u0E39<200>\u0020•\u0E01\u0E34\u0E19<200>\u0E01\u0E38\u0E49\u0E07<200>\
 653 \u0020•\u0E1B\u0E34\u0E49\u0E48<200>\u0E07\u0E2D<200>\u0E22\u0E39\u0E48<200>\
 654 \u0E43\u0E19<200>\u0E16\u0E49\u0E33<200></data>
 655
 656 <line>
 657 <data>•0E01\u0E39\u0020•\u0E01\u0E34\u0E19•\u0E01\u0E38\u0E49\u0E07\
 658 \u0020•\u0E1B\u0E34\u0E49\u0E48•\u0E07\u0E2D•\u0E22\u0E39\u0E48•\
 659 \u0E43\u0E19•\u0E16\u0E49\u0E33•</data>
 660
 661 # Data originally from intltest RBBITest::TestThaiLineBreak()
 662 #
 663 #  \u0e2f-- the Thai paiyannoi character-- isn't a letter.  It's a symbol that
 664 #  represents elided letters at the end of a long word.  It should be bound to
 665 #  the end of the word and not treated as an independent punctuation mark.
 666 #
 667 #  the one time where the paiyannoi occurs somewhere other than at the end
 668 #  of a word is in the Thai abbrevation for "etc.", which both begins and
 669 #  ends with a paiyannoi
 670 #
 671 <line>
 672 <data>•\u0e2a\u0e16\u0e32\u0e19\u0e35\u0e2f•\
 673 \u0e08\u0e30•\
 674 \u0e23\u0e30\u0e14\u0e21•\
 675 \u0e40\u0e08\u0e49\u0e32•\
 676 \u0e2b\u0e19\u0e49\u0e32\u0e17\u0e35\u0e48•\
 677 \u0e2d\u0e2d\u0e01•\
 678 \u0e21\u0e32•\
 679 \u0e40\u0e23\u0e48\u0e07•\
 680 \u0e23\u0e30\u0e1a\u0e32\u0e22•\
 681 \u0e2d\u0e22\u0e48\u0e32\u0e07•\
 682 \u0e40\u0e15\u0e47\u0e21•\
 683 \u0e2f\u0e25\u0e2f•\
 684 \u0e17\u0e35\u0e48•\
 685 \u0e19\u0e31\u0e49\u0e19•</data>
 686
 687 # Data originally from RBBITest::TestMixedThaiLineBreak()
 688 #   @suwit -- Test Arabic numerals, Thai numerals, Punctuation and English characters start
 689 #
 690 <line>
 691 <data>•\u0E1B\u0E35•\
 692 \u0E1E\u0E38\u0E17\u0E18\u0E28\u0E31\u0E01\u0E23\u0E32\u0E0A •\
 693 2545 •\
 694 \u0E40\u0E1B\u0E47\u0E19•\
 695 \u0E1B\u0E35•\
 696 \u0E09\u0E25\u0E2D\u0E07•\
 697 \u0E04\u0E23\u0E1A•\
 698 \u0E23\u0E2D\u0E1A •\
 699 \"\u0E52\u0E52\u0E50 •\
 700 \u0E1b\u0E35\" •\
 701 \u0E02\u0E2d\u0E07•\
 702 \u0E01\u0E23\u0E38\u0E07•\
 703 \u0E23\u0E31\u0E15\u0E19\u0E42\u0E01\u0E2A\u0E34\u0E19\u0E17\u0E23\u0E4C •\
 704 (\u0E01\u0E23\u0E38\u0E07\u0E40\u0E17\u0E1e\u0E2F•\
 705 \u0E2B\u0E23\u0E37\u0E2D •\
 706 Bangkok)•</data>
 707
 708 # Data originally from RBBITest::TestMaiyamok()
 709 #   The Thai maiyamok character is a shorthand symbol that means "repeat the previous
 710 #   word".  Instead of appearing as a word unto itself, however, it's kept together
 711 #   with the word before it.
 712 #
 713 <line>
 714 <data>•\u0e44\u0e1b\u0e46•\
 715 \u0e21\u0e32\u0e46•\
 716 \u0e23\u0e30\u0e2b\u0e27\u0e48\u0e32\u0e07•\
 717 \u0e01\u0e23\u0e38\u0e07•\
 718 \u0e40\u0e17\u0e1e•\
 719 \u0e41\u0e25\u0e30•\
 720 \u0e40\u0e03\u0e35•\
 721 \u0e22\u0e07•\
 722 \u0e43\u0e2b\u0e21\u0e48•</data>
 723
 724 # Test for #10296
 725 <line>
 726 <data>•ใช•มั้ย•</data>
 727 <data>•มั๊ยล่ะ•ที่รัก•</data>
 728
 729 # Test for #10593
 730 <line>
 731 <data>•เล่น•ผ่าน•ทาง•บลูทูธ•บน•อุปกรณ์•</data>
 732
 733 # Test for city names #10691
 734 <line>
 735 <data>•ไป•ที่•ซานฟรานซิสโก•</data>
 736
 737 # Test for #10630, #10631
 738 <line>
 739 <data>•แท็ก•แอปพลิเคชัน•เป็น•พิเศษ•</data>
 740
 741 # Test for #11019
 742 <line>
 743 <data>•เบ•เบราว์เซอร์•โพ•โพสต์•โพสท์•</data>
 744
 745 # Test for #11688
 746 <line>
 747 <data>•อัปเดต•อีเวนต์•</data>
 748
 749 ##########################################################################################
 750 #
 751 #   Lao Tests
 752 #
 753 ##########################################################################################
 754 <locale en>
 755 # Basic check for #7647
 756 <line>
 757 <data>•ສະບາຍດີ•</data>
 758 <data>•ດີ•ຂອບໃຈ•</data>
 759 <data>•ເຈົ້າ•ເວົ້າ•ພາສາ•ອັງກິດ•ໄດ້•ບໍ່•</data>
 760 <data>•ກະລຸນາ•ເວົ້າ•ຊ້າ•ໆ•</data>
 761
 762 ##########################################################################################
 763 #
 764 #   Burmese/Myanmar Tests
 765 #
 766 ##########################################################################################
 767 <locale en>
 768 # Basic sanity check for #10326 (some text from http://www.unicode.org/udhr/d/udhr_mya.txt)
 769 <line>
 770 <data>•လူ•တိုင်း•သည် •တူညီ •လွတ်လပ်•သော •ဂုဏ်•သိ•က္•ခါ•ဖြ•င့် •လည်းကောင်း၊ •</data>
 771 <data>•တူညီ•လွတ်လပ်•သော •အ•ခွ•င့်•အရေး•များ•ဖြ•င့် •လည်းကောင်း၊ •မွေး•ဖွား•လာ•သူများ •ဖြစ်သည်။•</data>
 772 <data>•ထို•သူ•တို့၌ •ပိုင်းခြား •ဝေဖန်•တတ်•သော •ဉာဏ်•နှ•င့် •ကျ•င့်•ဝတ် •သိတတ်•သော •စိတ်•တို့•ရှိ•ကြ၍ •</data>
 773 <data>•ထို•သူ•တို့သည် •အချင်းချင်း •မေတ္တာ•ထား၍ •ဆက်ဆံ•ကျ•င့်•သုံး•</data>
 774
 775 ##########################################################################################
 776 #
 777 #   Khmer Tests
 778 #
 779 ##########################################################################################
 780
 781 # Test data originally from http://bugs.icu-project.org/trac/search?q=r30327
 782 #  from the file testdata/wordsegments.txt
 783 <locale en>
 784 <word>
 785
 786 <data>•តើ<200>លោក<200>មក<200>ពី<200>ប្រទេស<200>ណា<200></data>
 787 <data>•សណ្ដូក<200>ក<200>បណ្ដែត<200>ខ្លួន<200></data>
 788 <data>•ពណ៌ស<200>ម្ដេច<200>ថា<200>ខ្មៅ<200></data>
 789 #ប្រយោគ|ពី|របៀប|រួបរួម|និង|ភាព|ផ្សេងគ្នា|ដែល|អាច|ចូល<200></data>
 790 <data>•ប្រយោគ<200>ពី<200>របៀប<200>ដែល<200>និង<200>ភាព<200>ផ្សេងគ្នា<200>ដែល<200>អាច<200>ចូល<200></data>
 791 #ប្រយោគ|ពី|របៀប|ជា|មួយ|និង|ភាព|ផ្សេងគ្នា|ដែល|អាច|ចូល<200></data>
 792 <data>•សូម<200>ចំណាយពេល<200>បន្តិច<200>ដើម្បី<200>អធិស្ឋាន<200>អរព្រះគុណ<200>ដល់<200>ព្រះអង្គ<200></data>
 793 <data>•ការ<200>ថោកទាប<200>បរិប្បូណ៌<200>ដោយ<200></data>
 794 <data>•ប្រើប្រាស់<200>ស្អាត<200>ទាំង<200>ចិត្ត<200>សិស្ស<200>នោះ<200></data>
 795 <data>•បើ<200>អ្នក<200>ប្រព្រឺត្ត<200>អំពើអាក្រក់<200>មុខ<200>ជា<200>មាន<200></data>
 796 <data>•ប្រដាប់<200>ប្រដា<200>រ<200>រៀនសូត្រ<200>បន្ទប់<200>រៀន<200></data>
 797 <data>•ដើរតួ<200>មនុស្សគ<200>ឥត<200>បញ្ចេញ<200>យោបល់<200>សោះ<200>ឡើយ<200></data>
 798 <data>•មិន<200>អាច<200>ឲ្យ<200>យើង<200>ធ្វើ<200>កសិកម្ម<200>បាន<200>ឡើយ<200></data>
 799 <data>•បន្ត<200>សេចក្ត<200>ទៅទៀត<200></data>
 800 <data>•ក្រុម<200>ប៉ូលិស<200>បណ្តាក់<200>គ្នា<200></data>
 801 <data>•គ្មាន<200>សុខ<200>សំរាន្ត<200>ដង<200>ណា<200></data>
 802 <data>•បាន<200>សុខភាព<200>បរិប្បូណ៌<200></data>
 803 <data>•ជា<200>មេចោរ<200>ខ្ញុំ<200>នឹង<200>ស្លាប់<200>ទៅវិញ<200>ជា<200>មេចោរ<200></data>
 804 <data>•ឯ<200>ការ<200>វាយ<200>ផ្ចាល<200>ដែល<200>នាំ<200></data>
 805 <data>•គេ<200>ដឹក<200>ទៅ<200>សំឡាប់<200></data>
 806 #អ្នក|ដែល|ជា|មន្ត្រី|ធំ|លើ|គាត់|ទេ<200></data>
 807 <data>•យក<200>ទៅ<200>សម្លាប់ចោល<200>ស្ងាត់<200></data>
 808 <data>•ត្រូវ<200>បាន<200>គេ<200>សម្លាប់<200></data>
 809 <data>•នៅក្នុង<200>ស្រុក<200>ខ្ល<200>ងហ្ស៊ុន<200></data>
 810
 811
 812 #
 813 #  Jitterbug 3671 Test Case
 814 #
 815 <data>•สวัสดี<200>ครับ<200>สบาย<200>ดี<200>ไหม<200> •ครับ<200></data>
 816
 817 #
 818 #  Trac ticket 5595 Test Case
 819 <data>•บท<200>ที่๑พายุ<200>ไซโคลน<200>โด<200>โรธี<200>อาศัย<200>อยู่<200>ท่ามกลาง<200>\
 820 ทุ่งใหญ่<200>ใน<200>แคนซัส<200>กับ<200>ลุง<200>เฮ<200>นรี<200>ชาวไร่<200>และ<200>ป้า<200>เอ็ม<200>\
 821 ภรรยา<200>ชาวไร่<200>บ้าน<200>ของ<200>พวก<200>เขา<200>หลัง<200>เล็ก<200>เพราะ<200>ไม้<200>\
 822 สร้าง<200>บ้าน<200>ต้อง<200>ขน<200>มา<200>ด้วย<200>เกวียน<200>เป็น<200>ระยะ<200>ทาง<200>หลาย<200>\
 823 ไมล์<200></data>
 824
 825 ####################################################################################
 826 #
 827 #  Tailored (locale specific) breaking.
 828 #
 829 ####################################################################################
 830
 831 # Japanese line break tailoring test
 832
 833 <locale ja>
 834 <line>
 835 <data>•\u3041•\u3043•\u3045•\u31f1•</data>
 836 <locale en>
 837 <line>
 838 <data>•\u3041\u3043\u3045\u31f1•</data>
 839
 840 # The following data was originally in RBBITest::TestJapaneseWordBreak()
 841 <locale ja>
 842 <word>
 843 <data>•\u4ECA\u65E5<400>\u306F<400>\u3044\u3044<400>\u5929\u6C17<400>\u3067\u3059<400>\u306D<400>\u3002•\u000D\u000A•</data>
 844
 845 # UBreakIteratorType UBRK_WORD, Locale "ja"
 846 # Don't break in runs of hiragana or runs of ideograph, where the latter includes \u3005 \u3007 \u303B (cldrbug #2009).
 847 # \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
 848 # modified to work with dbbi code - should verify
 849
 850 <locale ja>
 851 <word>
 852 <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>
 853
 854 # Test for #10176 (in ja)
 855 <line>
 856 <data>•abc/•s •def•</data>
 857 <data>•abc/\u05D9 •def•</data>
 858 <data>•\u05E7\u05D7/\u05D9 •\u05DE\u05E2\u05D9\u05DC•</data>
 859 <data>•\u05D3\u05E8\u05D5\u05E9\u05D9\u05DD •\u05E9\u05D7\u05E7\u05E0\u05D9\u05DD/\u05D9\u05D5\u05EA•</data>
 860
 861
 862 <locale root>
 863 <word>
 864 <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>
 865 # The following test is for #10300
 866 <data>•例えば<400>オーストラリア<400>。•</data>
 867 # The following test is for #10571
 868 <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>
 869
 870 # UBreakIteratorType UBRK_SENTENCE, Locale "el"
 871 # Add break after Greek question mark (cldrbug #2069).
 872 # "\u0391\u03B2, \u03B3\u03B4; \u0395 \u03B6\u03B7\u037E \u0398 \u03B9\u03BA. "
 873 # "\u039B\u03BC \u03BD\u03BE! \u039F\u03C0, \u03A1\u03C2? \u03A3"
 874 # which is "Αβ, γδ; Ε ζη; Θ ικ. Λμ νξ! Οπ, Ρς? Σ"
 875
 876 <locale root>
 877 <sent>
 878 <data>•Αβ, γδ; Ε ζη; Θ ικ. •Λμ νξ! •Οπ, Ρς? •Σ<100></data>
 879
 880 <locale el>
 881 <sent>
 882 <data>•Αβ, γδ; •Ε ζη; •Θ ικ. •Λμ νξ! •Οπ, Ρς? •Σ<100></data>
 883
 884 #  UBreakIteratorType UBRK_WORD, Locale "en_US_POSIX"
 885 #  Words don't include colon or period (cldrbug #1969).
 886
 887 <locale sv>
 888 <word>
 889 <data>•Can't<200> •have<200> •breaks<200> •in<200> •xx:yy<200> •or<200> •struct.field<200> \
 890 •for<200> •CS<200>-•types<200>.•</data>
 891 <data>•\uFF92\uFF76\uFF9E<400> •</data>
 892
 893 <locale en_US>
 894 <word>
 895 <data>•Can't<200> •have<200> •breaks<200> •in<200> •xx<200>:•yy<200> •or<200> •struct.field<200> \
 896 •for<200> •CS<200>-•types<200>.•</data>
 897 <data>•\uFF92\uFF76\uFF9E<400> •</data>
 898
 899 <locale en_US_POSIX>
 900 <word>
 901 <data>•Can't<200> •have<200> •breaks<200> •in<200> •xx<200>:•yy<200> •or<200> •struct<200>.•field<200> \
 902 •for<200> •CS<200>-•types<200>.•</data>
 903 <data>•\u06c9<200>\uc799\ufffa•</data>
 904 <data>•\uFF92\uFF76\uFF9E<400> •</data>
 905
 906
 907 # UBreakIteratorType UBRK_CHARACTER, Locale "th"
 908 # Clusters should not include spacing Thai/Lao vowels (prefix or postfix), except for [SARA] AM (cldrbug #2161).
 909 # Update:  As of Unicode 6.1 root has same behavior as th for this.
 910 #
 911 # "\u0E01\u0E23\u0E30\u0E17\u0E48\u0E2D\u0E21\u0E23\u0E08\u0E19\u0E32 "
 912 # "(\u0E2A\u0E38\u0E0A\u0E32\u0E15\u0E34-\u0E08\u0E38\u0E11\u0E32\u0E21\u0E32\u0E28) "
 913 # "\u0E40\u0E14\u0E47\u0E01\u0E21\u0E35\u0E1B\u0E31\u0E0D\u0E2B\u0E32 "
 914 # which is "กระท่อมรจนา (สุชาติ-จุฑามาศ) เด็กมีปัญหา "
 915
 916 <locale th>
 917 <char>
 918 <data>•\u0E01•\u0E23•\u0E30•\u0E17\u0E48•\u0E2D•\u0E21•\u0E23•\u0E08•\u0E19•\u0E32• •\
 919 (•\u0E2A\u0E38•\u0E0A•\u0E32•\u0E15\u0E34•-•\u0E08\u0E38•\u0E11•\u0E32•\u0E21•\u0E32•\u0E28•)• •\
 920 \u0E40•\u0E14\u0E47•\u0E01•\u0E21\u0E35•\u0E1B\u0E31•\u0E0D•\u0E2B•\u0E32• •</data>
 921
 922 # Finnish line breaking
 923 #
 924 # These rules deal with hyphens when there is a space on the leading side.
 925 # There should be a break opportunity between the space and the hyphen, and not after the hyphen.
 926 # See CLDR ticket 3029.
 927 # See ICU ticket 8151
 928
 929 <locale root>
 930 <line>
 931 <data>•abc •- •def    •abc •-•def    •abc- •def   •abc-•def•</data>   # With ASCII hyphen
 932 <data>•abc •‐ •def    •abc •‐•def    •abc‐ •def   •abc‐•def•</data>   # With Unicode u2010 hyphen
 933
 934 <locale fi>
 935 <line>
 936 # TODO: problems with Finnish line break rules cause these two lines to fail.
 937 #<data>•abc •- •def    •abc •-def    •abc- •def   •abc-•def•</data>   # With ASCII hyphen
 938 #<data>•abc •‐ •def    •abc •‐def    •abc‐ •def   •abc‐•def•</data>   # With Unicode u2010 hyphen
 939
 940 <data>•abc •- •def    •abc •-def    •abc- •def   •</data>   # With ASCII hyphen
 941 <data>•abc •‐ •def    •abc •‐def    •abc‐ •def   •</data>   # With Unicode u2010 hyphen
 942
 943 # Test for #10176 (in fi)
 944 <line>
 945 <data>•abc/•s •def•</data>
 946 <data>•abc/\u05D9 •def•</data>
 947 <data>•\u05E7\u05D7/\u05D9 •\u05DE\u05E2\u05D9\u05DC•</data>
 948 <data>•\u05D3\u05E8\u05D5\u05E9\u05D9\u05DD •\u05E9\u05D7\u05E7\u05E0\u05D9\u05DD/\u05D9\u05D5\u05EA•</data>
 949
 950 ####################################################################################
 951 #
 952 #  Test CSS line break variants: strict, normal, loose
 953 #
 954 ####################################################################################
 955
 956 <locale ja@lb=strict>
 957 <line>
 958 #     •no brk before 3063              •no brk before 301C•no brk btw 2026   •no brk before FF01•
 959 <data>•\u3084\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
 960
 961 <locale ja@lb=normal>
 962 <line>
 963 #     •brk OK before 3063               •brk OK before 301C •no brk btw 2026   •no brk before FF01•
 964 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031•\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
 965
 966 <locale ja@lb=loose>
 967 <line>
 968 #     •brk OK before 3063               •brk OK before 301C •brk OK btw 2026    •brk OK before FF01•
 969 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031•\u301C\u0020•\u2026•\u2026\u0020•u30A2•\uFF01\u0020•</data>
 970
 971 <locale en@lb=strict>
 972 <line>
 973 #     •no brk before 3063              •no brk before 301C•no brk btw 2026   •no brk before FF01•
 974 <data>•\u3084\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
 975
 976 <locale en@lb=normal>
 977 <line>
 978 #     •brk OK before 3063               •no brk before 301C •no brk btw 2026  •no brk before FF01•
 979 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026\u2026\u0020•\u30A2\uFF01\u0020•</data>
 980
 981 <locale en@lb=loose>
 982 <line>
 983 #     •brk OK before 3063               •no brk before 301C •brk OK btw 2026   •no brk before FF01•
 984 <data>•\u3084•\u3063•\u3071•\u308A\u0020•\u0031\u301C\u0020•\u2026•\u2026\u0020•u30A2\uFF01\u0020•</data>
 985
 986 ####################################################################################
 987 #
 988 #  Test Apple early change of lb class for 22EF
 989 #
 990 ####################################################################################
 991
 992 <locale en>
 993 <line>
 994 <data>•\u4E00\u2026\u2026•\u4E00\u22EF\u22EF•\u4E00\u0020•</data>
 995
 996 ####################################################################################
 997 #
 998 #  Test Apple early change of cjdict
 999 #
1000 ####################################################################################
1001
1002 <locale en>
1003 <word>
1004 <data>•ジョージア<400> •</data>
1005 <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>
1006
1007 ####################################################################################
1008 #
1009 #  Test Apple early change of thaidict
1010 #
1011 ####################################################################################
1012
1013 <locale th>
1014 <line>
1015 <data>•อัปเดต•อีเวนต์•</data>
1016
1017 ####################################################################################
1018 #
1019 #  Apple tests for currency break per <rdar://problem/24359173>
1020 #
1021 ####################################################################################
1022
1023 <locale en>
1024 <line>
1025 <data>•CN¥522.5•</data>
1026 <locale zh>
1027 <line>
1028 <data>•￥522.5•</data>
1029
1030 ####################################################################################
1031 #
1032 #  Apple tests for @lw=keepall per <rdar://problem/11679343&16501490>
1033 #
1034 ####################################################################################
1035
1036 <locale ko>
1037 <line>
1038 <data>•\uBAA8•\uB4E0 •\uC778•\uB958 •\uAD6C•\uC131•\uC6D0•\uC758 •\uCC9C•\uBD80•\uC758 •\uC874•\uC5C4•\uC131•\uACFC •\uB3D9•\uB4F1•\uD558•\uACE0 •\uC591•\uB3C4•\uD560•</data>
1039
1040 <locale ko@lw=keepall>
1041 <line>
1042 <data>•\uBAA8\uB4E0 •\uC778\uB958 •\uAD6C\uC131\uC6D0\uC758 •\uCC9C\uBD80\uC758 •\uC874\uC5C4\uC131\uACFC •\uB3D9\uB4F1\uD558\uACE0 •\uC591\uB3C4\uD560•</data>
1043
1044 ####################################################################################
1045 #
1046 #  Test Apple breaks for emoji clusters (same for all locales and break types)
1047 #
1048 ####################################################################################
1049
1050 <locale root>
1051
1052 <char>
1053 # woman zwj woman zwj girl zwj girl, woman/fitz-1-2 zwj woman/fitz-4 zwj boy/fitz-6
1054 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•\U0001F469\U0001F3FB\u200D\U0001F469\U0001F3FD\u200D\U0001F466\U0001F3FF•</data>
1055 # woman zwj, baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1056 <data>•\U0001F469\u200D•\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1057 # man zwj  hvy_blk_heart zwj  man, woman, man zwj hvy_blk_heart esel zwj  man, woman
1058 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1059 # woman zwj hvy_blk_heart/esel zwj kiss_mark zwj woman, sleuth/fitz-4, horseRacing/fitz-5
1060 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1061 # victory_hand esel, victory_hand/esel/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5, space,
1062 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE•\u0020•</data>
1063 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1064 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8•\u0020•</data>
1065 # family (man, boy, boy), family (woman, girl), space
1066 <data>•\U0001F468\u200D\U0001F466\u200D\U0001F466•\U0001F469\u200D\U0001F467•\u0020•</data>
1067 # rainbow flag using FE0F, rainbow flag no FE0F, space
1068 <data>•\U0001F3F3\uFE0F\u200D\U0001F308•\U0001F3F3\u200D\U0001F308•\u0020•</data>
1069 # woman with ball using FE0F, woman with ball no FE0F, woman with ball fitz 4 using FE0F, woman with ball fitz 4 no FE0F, space
1070 <data>•\u26F9\uFE0F\u200D\u2640\uFE0F•\u26F9\u200D\u2640•\u26F9\U0001F3FD\u200D\u2640\uFE0F•\u26F9\U0001F3FD\u200D\u2640•\u0020•</data>
1071 # female runner using FE0F, female runner no FE0F, female runner fitz 4 using FE0F, female runner fitz 4 no FE0F, space
1072 <data>•\U0001F3C3\u200D\u2640\uFE0F•\U0001F3C3\u200D\u2640•\U0001F3C3\U0001F3FD\u200D\u2640\uFE0F•\U0001F3C3\U0001F3FD\u200D\u2640•\u0020•</data>
1073 # 9.0 + professions
1074 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1075 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD•\u0020•</data>
1076 # facepalm, facepalm / fitz4, facepalm + woman +var, facepalm + woman -var, facepalm/fitz4 + woman +var, facepalm/fitz4 + woman -var
1077 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640•\u0020•</data>
1078 # handball, handball / fitz4, handball + man +var, handball + man -var, handball/fitz4 + man +var, handball/fitz4 + man -var
1079 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642•\u0020•</data>
1080 # male health worker +var, male health worker -var, male health worker / fitz4 +var, male health worker / fitz4 -var
1081 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695•\u0020•</data>
1082 # female astronaut, female astronaut / fitz4
1083 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680•\u0020•</data>
1084 # modifiers we do not support, but other platforms do
1085 # family/fitz-3, wrestlers/fitz-4, space
1086 <data>•\U0001F46A\U0001F3FC•\U0001F93C\U0001F3FD•\u0020•</data>
1087
1088 # flags1    AE AF AL AM AO AR AT
1089 <data>•\U0001F1E6\U0001F1EA•\U0001F1E6\U0001F1EB•\U0001F1E6\U0001F1F1•\U0001F1E6\U0001F1F2•\U0001F1E6\U0001F1F4•\U0001F1E6\U0001F1F7•\U0001F1E6\U0001F1F9•</data>
1090 # flags2    AU AZ BA BD BE BF BG
1091 <data>•\U0001F1E6\U0001F1FA•\U0001F1E6\U0001F1FF•\U0001F1E7\U0001F1E6•\U0001F1E7\U0001F1E9•\U0001F1E7\U0001F1EA•\U0001F1E7\U0001F1EB•\U0001F1E7\U0001F1EC•</data>
1092 # flags3    BH BJ BN BO BR BS BT
1093 <data>•\U0001F1E7\U0001F1ED•\U0001F1E7\U0001F1EF•\U0001F1E7\U0001F1F3•\U0001F1E7\U0001F1F4•\U0001F1E7\U0001F1F7•\U0001F1E7\U0001F1F8•\U0001F1E7\U0001F1F9•</data>
1094 # flags4    BW BY BZ CA CD CF CG
1095 <data>•\U0001F1E7\U0001F1FC•\U0001F1E7\U0001F1FE•\U0001F1E7\U0001F1FF•\U0001F1E8\U0001F1E6•\U0001F1E8\U0001F1E9•\U0001F1E8\U0001F1EB•\U0001F1E8\U0001F1EC•</data>
1096 # flags5    CH CI CL CM CN CO CR
1097 <data>•\U0001F1E8\U0001F1ED•\U0001F1E8\U0001F1EE•\U0001F1E8\U0001F1F1•\U0001F1E8\U0001F1F2•\U0001F1E8\U0001F1F3•\U0001F1E8\U0001F1F4•\U0001F1E8\U0001F1F7•</data>
1098 # flags6    CU CV CY CZ DE DJ DK
1099 <data>•\U0001F1E8\U0001F1FA•\U0001F1E8\U0001F1FB•\U0001F1E8\U0001F1FE•\U0001F1E8\U0001F1FF•\U0001F1E9\U0001F1EA•\U0001F1E9\U0001F1EF•\U0001F1E9\U0001F1F0•</data>
1100 # flags7    DM DO DZ EC EE EG ER
1101 <data>•\U0001F1E9\U0001F1F2•\U0001F1E9\U0001F1F4•\U0001F1E9\U0001F1FF•\U0001F1EA\U0001F1E8•\U0001F1EA\U0001F1EA•\U0001F1EA\U0001F1EC•\U0001F1EA\U0001F1F7•</data>
1102 # flags8    ES ET FI FJ FR GA GB
1103 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F9•\U0001F1EB\U0001F1EE•\U0001F1EB\U0001F1EF•\U0001F1EB\U0001F1F7•\U0001F1EC\U0001F1E6•\U0001F1EC\U0001F1E7•</data>
1104 # flags9    GE GH GM GN GR GT GW
1105 <data>•\U0001F1EC\U0001F1EA•\U0001F1EC\U0001F1ED•\U0001F1EC\U0001F1F2•\U0001F1EC\U0001F1F3•\U0001F1EC\U0001F1F7•\U0001F1EC\U0001F1F9•\U0001F1EC\U0001F1FC•</data>
1106 # flags10   GY HK HN HR HT HU ID
1107 <data>•\U0001F1EC\U0001F1FE•\U0001F1ED\U0001F1F0•\U0001F1ED\U0001F1F3•\U0001F1ED\U0001F1F7•\U0001F1ED\U0001F1F9•\U0001F1ED\U0001F1FA•\U0001F1EE\U0001F1E9•</data>
1108 # flags11   IE IL IN IQ IR IS IT
1109 <data>•\U0001F1EE\U0001F1EA•\U0001F1EE\U0001F1F1•\U0001F1EE\U0001F1F3•\U0001F1EE\U0001F1F6•\U0001F1EE\U0001F1F7•\U0001F1EE\U0001F1F8•\U0001F1EE\U0001F1F9•</data>
1110 # flags12   JM JO JP KE KG KH KR
1111 <data>•\U0001F1EF\U0001F1F2•\U0001F1EF\U0001F1F4•\U0001F1EF\U0001F1F5•\U0001F1F0\U0001F1EA•\U0001F1F0\U0001F1EC•\U0001F1F0\U0001F1ED•\U0001F1F0\U0001F1F7•</data>
1112 # flags13   MX MY NL NO PL PT
1113 <data>•\U0001F1F2\U0001F1FD•\U0001F1F2\U0001F1FE•\U0001F1F3\U0001F1F1•\U0001F1F3\U0001F1F4•\U0001F1F5\U0001F1F1•\U0001F1F5\U0001F1F9•</data>
1114 # flags14   RO RU SA SE SK TH TR
1115 <data>•\U0001F1F7\U0001F1F4•\U0001F1F7\U0001F1FA•\U0001F1F8\U0001F1E6•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1F0•\U0001F1F9\U0001F1ED•\U0001F1F9\U0001F1F7•</data>
1116 # flags15   UA US VN XK ZW
1117 <data>•\U0001F1FA\U0001F1E6•\U0001F1FA\U0001F1F8•\U0001F1FB\U0001F1F3•\U0001F1FD\U0001F1F0•\U0001F1FF\U0001F1FC•</data>
1118 # flagsX1   ES ES ES SE SE SE
1119 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•</data>
1120 # flagsX2   GB GB GB BG BG BG
1121 <data>•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•</data>
1122 # flagsXtnd AE AF AL AM AO AR
1123 <data>•\U0001F1E6\U0001F1EA\u200C•\U0001F1E6\U0001F1EB\u200C•\U0001F1E6\U0001F1F1\u200C•\U0001F1E6\U0001F1F2\u0300•\U0001F1E6\U0001F1F4\u20DE•\U0001F1E6\U0001F1F7\u200C•</data>
1124
1125 <word>
1126 # woman zwj woman zwj girl zwj girl, woman/fitz-1-2 zwj woman/fitz-4 zwj boy/fitz-6
1127 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•\U0001F469\U0001F3FB\u200D\U0001F469\U0001F3FD\u200D\U0001F466\U0001F3FF•</data>
1128 # woman zwj, baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1129 <data>•\U0001F469\u200D•\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1130 # man zwj  hvy_blk_heart zwj  man, woman, man zwj hvy_blk_heart esel zwj  man, woman
1131 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1132 # woman zwj hvy_blk_heart/esel zwj kiss_mark zwj woman, sleuth/fitz-4, horseRacing/fitz-5
1133 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1134 # victory_hand esel, victory_hand/esel/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5, space,
1135 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE•\u0020•</data>
1136 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1137 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8•\u0020•</data>
1138 # family (man, boy, boy), family (woman, girl), space
1139 <data>•\U0001F468\u200D\U0001F466\u200D\U0001F466•\U0001F469\u200D\U0001F467•\u0020•</data>
1140 # rainbow flag using FE0F, rainbow flag no FE0F, space
1141 <data>•\U0001F3F3\uFE0F\u200D\U0001F308•\U0001F3F3\u200D\U0001F308•\u0020•</data>
1142 # woman with ball using FE0F, woman with ball no FE0F, woman with ball fitz 4 using FE0F, woman with ball fitz 4 no FE0F, space
1143 <data>•\u26F9\uFE0F\u200D\u2640\uFE0F•\u26F9\u200D\u2640•\u26F9\U0001F3FD\u200D\u2640\uFE0F•\u26F9\U0001F3FD\u200D\u2640•\u0020•</data>
1144 # female runner using FE0F, female runner no FE0F, female runner fitz 4 using FE0F, female runner fitz 4 no FE0F, space
1145 <data>•\U0001F3C3\u200D\u2640\uFE0F•\U0001F3C3\u200D\u2640•\U0001F3C3\U0001F3FD\u200D\u2640\uFE0F•\U0001F3C3\U0001F3FD\u200D\u2640•\u0020•</data>
1146 # 9.0 + professions
1147 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1148 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD•\u0020•</data>
1149 # facepalm, facepalm / fitz4, facepalm + woman +var, facepalm + woman -var, facepalm/fitz4 + woman +var, facepalm/fitz4 + woman -var
1150 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640•\u0020•</data>
1151 # handball, handball / fitz4, handball + man +var, handball + man -var, handball/fitz4 + man +var, handball/fitz4 + man -var
1152 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642•\u0020•</data>
1153 # male health worker +var, male health worker -var, male health worker / fitz4 +var, male health worker / fitz4 -var
1154 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695•\u0020•</data>
1155 # female astronaut, female astronaut / fitz4
1156 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680•\u0020•</data>
1157 # modifiers we do not support, but other platforms do
1158 # family/fitz-3, wrestlers/fitz-4, space
1159 <data>•\U0001F46A\U0001F3FC•\U0001F93C\U0001F3FD•\u0020•</data>
1160
1161 # flags1    AE AF AL AM AO AR AT
1162 <data>•\U0001F1E6\U0001F1EA•\U0001F1E6\U0001F1EB•\U0001F1E6\U0001F1F1•\U0001F1E6\U0001F1F2•\U0001F1E6\U0001F1F4•\U0001F1E6\U0001F1F7•\U0001F1E6\U0001F1F9•</data>
1163 # flags2    AU AZ BA BD BE BF BG
1164 <data>•\U0001F1E6\U0001F1FA•\U0001F1E6\U0001F1FF•\U0001F1E7\U0001F1E6•\U0001F1E7\U0001F1E9•\U0001F1E7\U0001F1EA•\U0001F1E7\U0001F1EB•\U0001F1E7\U0001F1EC•</data>
1165 # flags3    BH BJ BN BO BR BS BT
1166 <data>•\U0001F1E7\U0001F1ED•\U0001F1E7\U0001F1EF•\U0001F1E7\U0001F1F3•\U0001F1E7\U0001F1F4•\U0001F1E7\U0001F1F7•\U0001F1E7\U0001F1F8•\U0001F1E7\U0001F1F9•</data>
1167 # flags4    BW BY BZ CA CD CF CG
1168 <data>•\U0001F1E7\U0001F1FC•\U0001F1E7\U0001F1FE•\U0001F1E7\U0001F1FF•\U0001F1E8\U0001F1E6•\U0001F1E8\U0001F1E9•\U0001F1E8\U0001F1EB•\U0001F1E8\U0001F1EC•</data>
1169 # flags5    CH CI CL CM CN CO CR
1170 <data>•\U0001F1E8\U0001F1ED•\U0001F1E8\U0001F1EE•\U0001F1E8\U0001F1F1•\U0001F1E8\U0001F1F2•\U0001F1E8\U0001F1F3•\U0001F1E8\U0001F1F4•\U0001F1E8\U0001F1F7•</data>
1171 # flags6    CU CV CY CZ DE DJ DK
1172 <data>•\U0001F1E8\U0001F1FA•\U0001F1E8\U0001F1FB•\U0001F1E8\U0001F1FE•\U0001F1E8\U0001F1FF•\U0001F1E9\U0001F1EA•\U0001F1E9\U0001F1EF•\U0001F1E9\U0001F1F0•</data>
1173 # flags7    DM DO DZ EC EE EG ER
1174 <data>•\U0001F1E9\U0001F1F2•\U0001F1E9\U0001F1F4•\U0001F1E9\U0001F1FF•\U0001F1EA\U0001F1E8•\U0001F1EA\U0001F1EA•\U0001F1EA\U0001F1EC•\U0001F1EA\U0001F1F7•</data>
1175 # flags8    ES ET FI FJ FR GA GB
1176 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F9•\U0001F1EB\U0001F1EE•\U0001F1EB\U0001F1EF•\U0001F1EB\U0001F1F7•\U0001F1EC\U0001F1E6•\U0001F1EC\U0001F1E7•</data>
1177 # flags9    GE GH GM GN GR GT GW
1178 <data>•\U0001F1EC\U0001F1EA•\U0001F1EC\U0001F1ED•\U0001F1EC\U0001F1F2•\U0001F1EC\U0001F1F3•\U0001F1EC\U0001F1F7•\U0001F1EC\U0001F1F9•\U0001F1EC\U0001F1FC•</data>
1179 # flags10   GY HK HN HR HT HU ID
1180 <data>•\U0001F1EC\U0001F1FE•\U0001F1ED\U0001F1F0•\U0001F1ED\U0001F1F3•\U0001F1ED\U0001F1F7•\U0001F1ED\U0001F1F9•\U0001F1ED\U0001F1FA•\U0001F1EE\U0001F1E9•</data>
1181 # flags11   IE IL IN IQ IR IS IT
1182 <data>•\U0001F1EE\U0001F1EA•\U0001F1EE\U0001F1F1•\U0001F1EE\U0001F1F3•\U0001F1EE\U0001F1F6•\U0001F1EE\U0001F1F7•\U0001F1EE\U0001F1F8•\U0001F1EE\U0001F1F9•</data>
1183 # flags12   JM JO JP KE KG KH KR
1184 <data>•\U0001F1EF\U0001F1F2•\U0001F1EF\U0001F1F4•\U0001F1EF\U0001F1F5•\U0001F1F0\U0001F1EA•\U0001F1F0\U0001F1EC•\U0001F1F0\U0001F1ED•\U0001F1F0\U0001F1F7•</data>
1185 # flags13   MX MY NL NO PL PT
1186 <data>•\U0001F1F2\U0001F1FD•\U0001F1F2\U0001F1FE•\U0001F1F3\U0001F1F1•\U0001F1F3\U0001F1F4•\U0001F1F5\U0001F1F1•\U0001F1F5\U0001F1F9•</data>
1187 # flags14   RO RU SA SE SK TH TR
1188 <data>•\U0001F1F7\U0001F1F4•\U0001F1F7\U0001F1FA•\U0001F1F8\U0001F1E6•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1F0•\U0001F1F9\U0001F1ED•\U0001F1F9\U0001F1F7•</data>
1189 # flags15   UA US VN XK ZW
1190 <data>•\U0001F1FA\U0001F1E6•\U0001F1FA\U0001F1F8•\U0001F1FB\U0001F1F3•\U0001F1FD\U0001F1F0•\U0001F1FF\U0001F1FC•</data>
1191 # flagsX1   ES ES ES SE SE SE
1192 <data>•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1EA\U0001F1F8•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•\U0001F1F8\U0001F1EA•</data>
1193 # flagsX2   GB GB GB BG BG BG
1194 <data>•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1EC\U0001F1E7•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•\U0001F1E7\U0001F1EC•</data>
1195 # flagsXtnd AE AF AL AM AO AR
1196 <data>•\U0001F1E6\U0001F1EA\u200C•\U0001F1E6\U0001F1EB\u200C•\U0001F1E6\U0001F1F1\u200C•\U0001F1E6\U0001F1F2\u0300•\U0001F1E6\U0001F1F4\u20DE•\U0001F1E6\U0001F1F7\u200C•</data>
1197
1198 <line>
1199 # woman zwj woman zwj girl zwj girl # (line, skip this for now, need safe rules and we don't generate it:) woman/fitz-1-2 zwj woman/fitz-4 zwj boy/fitz-6
1200 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•</data>
1201 # woman zwj baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1202 <data>•\U0001F469\u200D\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1203 # man zwj  hvy_blk_heart zwj  man, woman, man zwj hvy_blk_heart esel zwj  man, woman
1204 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1205 # woman zwj hvy_blk_heart/esel zwj kiss_mark zwj woman, sleuth/fitz-4, horseRacing/fitz-5
1206 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1207 # victory_hand esel, victory_hand/esel/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5 space,
1208 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE\u0020•</data>
1209 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1210 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8\u0020•</data>
1211 # family (man, boy, boy), family (woman, girl), space
1212 <data>•\U0001F468\u200D\U0001F466\u200D\U0001F466•\U0001F469\u200D\U0001F467\u0020•</data>
1213 # rainbow flag using FE0F, rainbow flag no FE0F, space
1214 <data>•\U0001F3F3\uFE0F\u200D\U0001F308•\U0001F3F3\u200D\U0001F308\u0020•</data>
1215 # woman with ball using FE0F, woman with ball no FE0F, woman with ball fitz 4 using FE0F, woman with ball fitz 4 no FE0F, space
1216 <data>•\u26F9\uFE0F\u200D\u2640\uFE0F•\u26F9\u200D\u2640•\u26F9\U0001F3FD\u200D\u2640\uFE0F•\u26F9\U0001F3FD\u200D\u2640\u0020•</data>
1217 # female runner using FE0F, female runner no FE0F, female runner fitz 4 using FE0F, female runner fitz 4 no FE0F, space
1218 <data>•\U0001F3C3\u200D\u2640\uFE0F•\U0001F3C3\u200D\u2640•\U0001F3C3\U0001F3FD\u200D\u2640\uFE0F•\U0001F3C3\U0001F3FD\u200D\u2640\u0020•</data>
1219 # 9.0 + professions
1220 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1221 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD\u0020•</data>
1222 # facepalm, facepalm / fitz4, facepalm + woman +var, facepalm + woman -var, facepalm/fitz4 + woman +var, facepalm/fitz4 + woman -var
1223 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640\u0020•</data>
1224 # handball, handball / fitz4, handball + man +var, handball + man -var, handball/fitz4 + man +var, handball/fitz4 + man -var
1225 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642\u0020•</data>
1226 # male health worker +var, male health worker -var, male health worker / fitz4 +var, male health worker / fitz4 -var
1227 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695\u0020•</data>
1228 # female astronaut, female astronaut / fitz4
1229 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680\u0020•</data>
1230 # modifiers we do not support, but other platforms do
1231 # family/fitz-3, wrestlers/fitz-4, space
1232 <data>•\U0001F46A\U0001F3FC•\U0001F93C\U0001F3FD\u0020•</data>
1233
1234 # no special flags handling for line
1235
1236 <locale ja@lb=loose>
1237 <line>
1238 # woman zwj woman zwj girl zwj girl # (line, skip this for now, need safe rules and we don't generate it:) woman/fitz-1-2 zwj woman/fitz-4 zwj boy/fitz-6
1239 <data>•\U0001F469\u200D\U0001F469\u200D\U0001F467\u200D\U0001F467•</data>
1240 # woman zwj baby/fitz-3, older_woman/fitz-5, runner/fitz-4, raised_fist/fitz-3, fuel_pump, fitz-3
1241 <data>•\U0001F469\u200D\U0001F476\U0001F3FC•\U0001F475\U0001F3FE•\U0001F3C3\U0001F3FD•\u270A\U0001F3FC•\u26FD•\U0001F3FC•</data>
1242 # man zwj  hvy_blk_heart zwj  man, woman, man zwj hvy_blk_heart esel zwj  man, woman
1243 <data>•\U0001F468\u200D\u2764\u200D\U0001F468•\U0001F469•\U0001F468\u200D\u2764\uFE0F\u200D\U0001F468•\U0001F469•</data>
1244 # woman zwj hvy_blk_heart/esel zwj kiss_mark zwj woman, sleuth/fitz-4, horseRacing/fitz-5
1245 <data>•\U0001F469\u200D\u2764\uFE0F\u200D\U0001F48B\u200D\U0001F469•\U0001F575\U0001F3FD•\U0001F3C7\U0001F3FE•</data>
1246 # victory_hand esel, victory_hand/esel/fitz-1-2, victory_hand/fitz-1-2, rowboat/fitz-4, vulcan_salute/fitz-5 space,
1247 <data>•\u270C\uFE0F•\u270C\uFE0F\U0001F3FB•\u270C\U0001F3FB•\U0001F6A3\U0001F3FD•\U0001F596\U0001F3FE\u0020•</data>
1248 # writing_hand fitz-1-2, splayed_hand/fitz-3, middle_finger/fitz-4, sign_of_horns/fitz-5, eye zwj left_speech_bubble, space
1249 <data>•\u270D\U0001F3FB•\U0001F590\U0001F3FC•\U0001F595\U0001F3FD•\U0001F918\U0001F3FE•\U0001F441\u200D\U0001F5E8\u0020•</data>
1250 # 9.0 + professions
1251 # black heart, fitz 4, squid, fitz4, man dancing /fitz4, mother xmas /fitz4
1252 <data>•\U0001F5A4•\U0001F3FD•\U0001F991•\U0001F3FD•\U0001F57A\U0001F3FD•\U0001F936\U0001F3FD\u0020•</data>
1253 # facepalm, facepalm / fitz4, facepalm + woman +var, facepalm + woman -var, facepalm/fitz4 + woman +var, facepalm/fitz4 + woman -var
1254 <data>•\U0001F926•\U0001F926\U0001F3FD•\U0001F926\u200D\u2640\uFE0F•\U0001F926\u200D\u2640•\U0001F926\U0001F3FD\u200D\u2640\uFE0F•\U0001F926\U0001F3FD\u200D\u2640\u0020•</data>
1255 # handball, handball / fitz4, handball + man +var, handball + man -var, handball/fitz4 + man +var, handball/fitz4 + man -var
1256 <data>•\U0001F93E•\U0001F93E\U0001F3FD•\U0001F93E\u200D\u2642\uFE0F•\U0001F93E\u200D\u2642•\U0001F93E\U0001F3FD\u200D\u2642\uFE0F•\U0001F93E\U0001F3FD\u200D\u2642\u0020•</data>
1257 # male health worker +var, male health worker -var, male health worker / fitz4 +var, male health worker / fitz4 -var
1258 <data>•\U0001F468\u200D\u2695\uFE0F•\U0001F468\u200D\u2695•\U0001F468\U0001F3FD\u200D\u2695\uFE0F•\U0001F468\U0001F3FD\u200D\u2695\u0020•</data>
1259 # female astronaut, female astronaut / fitz4
1260 <data>•\U0001F469\u200D\U0001F680•\U0001F469\U0001F3FD\u200D\U0001F680\u0020•</data>
1261 # modifiers we do not support, but other platforms do
1262 # family/fitz-3, wrestlers/fitz-4, space
1263 <data>•\U0001F46A\U0001F3FC•\U0001F93C\U0001F3FD\u0020•</data>
1264
1265 # no special flags handling for line