ICU-461.12.tar.gz

[apple/icu.git] / icuSources / data / brkitr / char.txt

diff --git a/icuSources/data/brkitr/char.txt b/icuSources/data/brkitr/char.txt
index 58170ddf698ac12110091b4b08a477de66569b6a..5f2417c15a4483d30cbbc2b02d5cca177056e522 100644 (file)
--- a/icuSources/data/brkitr/char.txt
+++ b/icuSources/data/brkitr/char.txt
@@ -1,26 +1,23 @@
 #
-#   Copyright (C) 2002-2006, International Business Machines Corporation and others.
+#   Copyright (C) 2002-2010, 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 5.0.0
-#        Includes post-5.0 change to treat Japanese half width voicing marks
-#        as Grapheme Extend.
+#      These rules are based on TR29 Revision 16, for Unicode Version 6.0
 #
 
 #
 #  Character Class Definitions.
 #
-$CR      = [\p{Grapheme_Cluster_Break = CR}];
-$LF      = [\p{Grapheme_Cluster_Break = LF}];
-$Control = [\p{Grapheme_Cluster_Break = Control}];
-
-# add Japanese Half Width voicing marks to $Extend
-$VoiceMarks = [\uff9e\uff9f]; 
-$Extend  = [\p{Grapheme_Cluster_Break = Extend} $VoiceMarks];
+$CR          = [\p{Grapheme_Cluster_Break = CR}];
+$LF          = [\p{Grapheme_Cluster_Break = LF}];
+$Control     = [\p{Grapheme_Cluster_Break = Control}];
+$Prepend     = [\p{Grapheme_Cluster_Break = Prepend}];
+$Extend      = [\p{Grapheme_Cluster_Break = Extend}];
+$SpacingMark = [\p{Grapheme_Cluster_Break = SpacingMark}];
 
 #
 # Korean Syllable Definitions
@@ -32,33 +29,43 @@ $T       = [\p{Grapheme_Cluster_Break = T}];
 $LV      = [\p{Grapheme_Cluster_Break = LV}];
 $LVT     = [\p{Grapheme_Cluster_Break = LVT}];
 
-$HangulSyllable = $L+ | ($L* ($LV? $V+ | $LV | $LVT) $T*) | $T+;
 
 ## -------------------------------------------------
+!!chain;
 
 !!forward;
 
 $CR $LF;
-([^$Control $CR $LF] | $HangulSyllable) $Extend*;
+
+$L ($L | $V | $LV | $LVT);
+($LV | $V) ($V | $T);
+($LVT | $T) $T;
+
+[^$Control $CR $LF] $Extend;
+
+[^$Control $CR $LF] $SpacingMark;
+$Prepend [^$Control $CR $LF];
+
 
 ## -------------------------------------------------
 
 !!reverse;
+$LF $CR;
+($L | $V | $LV | $LVT) $L;
+($V | $T) ($LV | $V);
+$T ($LVT | $T);
+
+$Extend      [^$Control $CR $LF];
+$SpacingMark [^$Control $CR $LF];
+[^$Control $CR $LF] $Prepend;
 
-$BackHangulSyllable = $L+ | ($T* ($V+$LV? | $LV | $LVT) $L*) | $T+;
-$BackOneCluster = ($LF $CR) | ($Extend* ([^$Control $CR $LF] | $BackHangulSyllable));
-$BackOneCluster;
 
 ## -------------------------------------------------
 
 !!safe_reverse;
 
-# rule 6, 7, 8
-$V+ $L;
 
 ## -------------------------------------------------
 
 !!safe_forward;
 
-# rule 6, 7, 8
-$V+ $T;