#
-# Copyright (C) 2002-2008, International Business Machines Corporation and others.
+# Copyright (C) 2002-2011, 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 Revision 12, for Unicode Version 5.1
+# These rules are based on UAX #29 Revision 19 for Unicode Version 6.1
#
#
$CR = [\p{Grapheme_Cluster_Break = CR}];
$LF = [\p{Grapheme_Cluster_Break = LF}];
$Control = [\p{Grapheme_Cluster_Break = Control}];
-$Prepend = [\p{Grapheme_Cluster_Break = Prepend}];
+# TODO: Restore if the Prepend set becomes non-empty again: $Prepend = [\p{Grapheme_Cluster_Break = Prepend}];
$Extend = [\p{Grapheme_Cluster_Break = Extend}];
$SpacingMark = [\p{Grapheme_Cluster_Break = SpacingMark}];
$LV = [\p{Grapheme_Cluster_Break = LV}];
$LVT = [\p{Grapheme_Cluster_Break = LVT}];
+#
+# Emoji regional indicators
+#
+$RI = [\U0001F1E6-\U0001F1FF];
## -------------------------------------------------
!!chain;
[^$Control $CR $LF] $Extend;
[^$Control $CR $LF] $SpacingMark;
-$Prepend [^$Control $CR $LF];
+# TODO: Restore if the Prepend set becomes non-empty again: $Prepend [^$Control $CR $LF];
+$RI $RI / $RI;
+$RI $RI;
## -------------------------------------------------
$Extend [^$Control $CR $LF];
$SpacingMark [^$Control $CR $LF];
-[^$Control $CR $LF] $Prepend;
+# TODO: Restore if the Prepend set becomes non-empty again: [^$Control $CR $LF] $Prepend;
+$RI $RI / $RI $RI;
+$RI $RI;
## -------------------------------------------------