char.txt   [plain text]

#
#   Copyright (C) 2002-2003, International Business Machines Corporation and others.
#       All Rights Reserved.
#
#   file:  char.txt 
#
#   ICU Character Break Rules, also known as Grapheme Cluster Boundaries
#      See Unicode Standard Annex #29.
#      These rules are based on TR29 Version 4.0.0
#

#
#  Character Class Definitions.
#    The names are those from TR29.
#
$CR = \r;
$LF = \n;
$Control    = [[:Zl:] [:Zp:] [:Cc:] [:Cf:]];

$Extend     = [[:Grapheme_Extend = TRUE:]]; 

#
# Korean Syllable Definitions
#
$L   = [:Hangul_Syllable_Type = L:];
$V   = [:Hangul_Syllable_Type = V:];
$T   = [:Hangul_Syllable_Type = T:];

$LV  = [:Hangul_Syllable_Type = LV:];
$LVT = [:Hangul_Syllable_Type = LVT:];

$HangulSyllable = $L+ | ($L* ($LV? $V+ | $LV | $LVT) $T*) | $T+;

#
#  Forward Break Rules
#
$CR $LF;
([^$Control] | $HangulSyllable) $Extend*;
.;


#
#  Reverse Rule, back up to the beginning of some preceding grapheme cluster.
#
! ($Extend | $V | $T )*   ($LF $CR | ($LV | $LVT)?$L* | .);