icuSources/test/testdata/break_rules/grapheme.txt

   1 #
   2 # Copyright (C) 2016 and later: Unicode, Inc. and others.
   3 # License & terms of use: http://www.unicode.org/copyright.html
   4 # Copyright (c) 2016, International Business Machines Corporation and others. All Rights Reserved.
   5
   6 # file: grapheme.txt
   7 #
   8 # Reference Grapheme Break rules for intltest rbbi/RBBIMonkeyTest
   9 #
  10 #
  11 # Note: Rule syntax and the monkey test itself are still a work in progress.
  12 #       They are expected to change with review and the addition of support for rule tailoring.
  13
  14 type = grapheme;      # one of grapheme | word | line | sentence
  15 locale = en;
  16
  17 CR                 = [\p{Grapheme_Cluster_Break = CR}];
  18 LF                 = [\p{Grapheme_Cluster_Break = LF}];
  19
  20 Control            = [[\p{Grapheme_Cluster_Break = Control}]];
  21 Extend             = [[\p{Grapheme_Cluster_Break = Extend}]];
  22 ZWJ                = [\p{Grapheme_Cluster_Break = ZWJ}];
  23 Regional_Indicator = [\p{Grapheme_Cluster_Break = Regional_Indicator}];
  24 Prepend            = [\p{Grapheme_Cluster_Break = Prepend}];
  25 SpacingMark        = [\p{Grapheme_Cluster_Break = SpacingMark}];
  26
  27 #
  28 # Korean Syllable Definitions
  29 #
  30 L                  = [\p{Grapheme_Cluster_Break = L}];
  31 V                  = [\p{Grapheme_Cluster_Break = V}];
  32 T                  = [\p{Grapheme_Cluster_Break = T}];
  33 LV                 = [\p{Grapheme_Cluster_Break = LV}];
  34 LVT                = [\p{Grapheme_Cluster_Break = LVT}];
  35
  36 # Emoji defintions
  37
  38 Extended_Pict      = [:ExtPict:];
  39
  40 # Indic Sequences
  41 Virama_            = [[\p{Gujr}\p{sc=Telu}\p{sc=Mlym}\p{sc=Orya}\p{sc=Beng}\p{sc=Deva}] & [\p{Indic_Syllabic_Category=Virama}]];
  42
  43 LinkingConsonant   = [[\p{Gujr}\p{sc=Telu}\p{sc=Mlym}\p{sc=Orya}\p{sc=Beng}\p{sc=Deva}] & [\p{Indic_Syllabic_Category=Consonant}]];
  44
  45 ExtCccZwj          = [[Extend-[\p{ccc=0}]] ZWJ];
  46
  47 GB3:     CR LF;
  48 GB4:     (Control | CR | LF) ÷;
  49 GB5:     . ÷ (Control | CR | LF);
  50
  51 GB6:     L (L | V | LV | LVT);
  52 GB7:     (LV | V) (V | T);
  53 GB8:     (LVT | T) T;
  54
  55 GB11:    Extended_Pict Extend* ZWJ Extended_Pict;
  56 GB9c:    LinkingConsonant ExtCccZwj* Virama_ ExtCccZwj* LinkingConsonant;
  57 GB9:     . (Extend | ZWJ);
  58
  59 GB9a:    . SpacingMark;
  60 GB9b:    Prepend .;
  61
  62 # Regional Indicators, split into pairs.
  63 #      Note that a pair of RIs that is not followed by a third RI will fall into
  64 #      the normal rules for Extend, etc.
  65 #
  66 GB12:  Regional_Indicator Regional_Indicator ÷ Regional_Indicator;
  67 GB13:  Regional_Indicator Regional_Indicator;
  68
  69 GB999:     . ÷;