[apple/icu.git] / icuSources / data / translit / Han_Spacedhan.txt

# © 2016 and later: Unicode, Inc. and others.
# License & terms of use: http://www.unicode.org/copyright.html#License
#
# File: Han_Spacedhan.txt
# Generated from CLDR
#

# Only intended for internal use
# Make sure Han are normalized, including characters that contain them.
# The first set in the filter is computed with http://unicode.org/cldr/utility/list-unicodeset.jsp?a=[:tonfkd:/XXX/:]-[:ideographic:]-[:sc=han:]
# Where XXX is the resolved [:ideographic:][:sc=han:]. It needs updating with each Unicode release!
:: [[㆒-㆟㈠-㉇㊀-㊰㋀-㋋㍘-㍰㍻-㍿㏠-㏾ 🈐-🈒🈔-🈺🉀-🉈🉐🉑][:ideographic:][:sc=han:]] nfkc;
:: fullwidth-halfwidth;
｡ → '.';
$terminalPunct = [\.\,\:\;\?\!．，：？！｡、；[:Pe:][:Pf:]];
$initialPunct = [:Ps:][:Pi:];
# add space between any Han or terminal punctuation and letters, and
# between letters and Han or initial punct
[[:Ideographic:] $terminalPunct] {} [:Letter:] → ' ' ;
[:Letter:] [:Mark:]* {} [[:Ideographic:] $initialPunct] → ' ' ;
# remove spacing between ideographs and other letters
← [:Ideographic:] { ' ' } [:Letter:] ;
← [:Letter:] [:Mark:]* { ' ' } [:Ideographic:] ;
Commit	Line	Data
f3c0d7a5 A	1	# © 2016 and later: Unicode, Inc. and others.
	2	# License & terms of use: http://www.unicode.org/copyright.html#License
	3	#
73c04bcf	4	# File: Han_Spacedhan.txt
f3c0d7a5	5	# Generated from CLDR
73c04bcf	6	#
2ca993e8 A	7
	8	# Only intended for internal use
	9	# Make sure Han are normalized, including characters that contain them.
	10	# The first set in the filter is computed with http://unicode.org/cldr/utility/list-unicodeset.jsp?a=[:tonfkd:/XXX/:]-[:ideographic:]-[:sc=han:]
	11	# Where XXX is the resolved [:ideographic:][:sc=han:]. It needs updating with each Unicode release!
	12	:: [[㆒-㆟㈠-㉇㊀-㊰㋀-㋋㍘-㍰㍻-㍿㏠-㏾ 🈐-🈒🈔-🈺🉀-🉈🉐🉑][:ideographic:][:sc=han:]] nfkc;
374ca955	13	:: fullwidth-halfwidth;
729e4ab9	14	｡ → '.';
374ca955 A	15	$terminalPunct = [\.\,\:\;\?\!．，：？！｡、；[:Pe:][:Pf:]];
374ca955 A	16	$initialPunct = [:Ps:][:Pi:];
2ca993e8 A	17	# add space between any Han or terminal punctuation and letters, and
2ca993e8 A	18	# between letters and Han or initial punct
729e4ab9 A	19	[[:Ideographic:] $terminalPunct] {} [:Letter:] → ' ' ;
729e4ab9 A	20	[:Letter:] [:Mark:]* {} [[:Ideographic:] $initialPunct] → ' ' ;
2ca993e8	21	# remove spacing between ideographs and other letters
729e4ab9 A	22	← [:Ideographic:] { ' ' } [:Letter:] ;
729e4ab9 A	23	← [:Letter:] [:Mark:]* { ' ' } [:Ideographic:] ;
2ca993e8	24