]> git.saurik.com Git - apple/icu.git/blame - icuSources/data/brkitr/rules/line_loose.txt
ICU-59180.0.1.tar.gz
[apple/icu.git] / icuSources / data / brkitr / rules / line_loose.txt
CommitLineData
f3c0d7a5
A
1# Copyright (C) 2016 and later: Unicode, Inc. and others.
2# License & terms of use: http://www.unicode.org/copyright.html
2ca993e8 3# Copyright (c) 2002-2016 International Business Machines Corporation and
08b89b0a
A
4# others. All Rights Reserved.
5#
6# file: line_loose.txt
7#
8# Line Breaking Rules
2ca993e8 9# Implement default line breaking as defined by
f3c0d7a5 10# Unicode Standard Annex #14 Revision 37 for Unicode 9.0
08b89b0a 11# http://www.unicode.org/reports/tr14/
2ca993e8 12#
f3c0d7a5
A
13# Includes extensions to the handling of emoji ZWJ sequences from
14# https://goo.gl/cluFCn
2ca993e8
A
15#
16# tailored as noted in 2nd paragraph below.
08b89b0a
A
17#
18# TODO: Rule LB 8 remains as it was in Unicode 5.2
19# This is only because of a limitation of ICU break engine implementation,
20# not because the older behavior is desirable.
21#
22# This tailors the line break behavior to correspond to CSS
f3c0d7a5 23# line-break=loose (BCP47 -u-lb-loose) as defined for languages other than
08b89b0a
A
24# Chinese & Japanese.
25# It sets characters of class CJ to behave like ID.
26# In addition, it allows breaks:
27# * before iteration marks 3005, 303B, 309D, 309E, 30FD, 30FE (all NS)
28# * between characters of LineBreak class IN
29
30#
31# Character Classes defined by TR 14.
32#
33
34!!chain;
08b89b0a 35
f3c0d7a5
A
36$AI = [:LineBreak = Ambiguous:];
37$AL = [:LineBreak = Alphabetic:];
08b89b0a
A
38$BA = [:LineBreak = Break_After:];
39$BB = [:LineBreak = Break_Before:];
40$BK = [:LineBreak = Mandatory_Break:];
41$B2 = [:LineBreak = Break_Both:];
42$CB = [:LineBreak = Contingent_Break:];
43$CJ = [:LineBreak = Conditional_Japanese_Starter:];
44$CL = [:LineBreak = Close_Punctuation:];
f3c0d7a5 45# $CM = [:LineBreak = Combining_Mark:];
08b89b0a
A
46$CP = [:LineBreak = Close_Parenthesis:];
47$CR = [:LineBreak = Carriage_Return:];
f3c0d7a5
A
48$EB = [[:LineBreak = EB:] \U0001F46A-\U0001F46D\U0001F46F\U0001F91D\U0001F93C];
49$EM = [:LineBreak = EM:];
08b89b0a
A
50$EX = [:LineBreak = Exclamation:];
51$GL = [:LineBreak = Glue:];
52$HL = [:LineBreak = Hebrew_Letter:];
53$HY = [:LineBreak = Hyphen:];
54$H2 = [:LineBreak = H2:];
55$H3 = [:LineBreak = H3:];
f3c0d7a5
A
56# CSS Loose tailoring: CJ resolves to ID
57$ID = [[:LineBreak = Ideographic:] $CJ];
08b89b0a
A
58$IN = [:LineBreak = Inseperable:];
59$IS = [:LineBreak = Infix_Numeric:];
60$JL = [:LineBreak = JL:];
61$JV = [:LineBreak = JV:];
62$JT = [:LineBreak = JT:];
63$LF = [:LineBreak = Line_Feed:];
64$NL = [:LineBreak = Next_Line:];
65$NSX = [\u3005 \u303B \u309D \u309E \u30FD \u30FE];
66$NS = [[:LineBreak = Nonstarter:] - $NSX];
67$NU = [:LineBreak = Numeric:];
68$OP = [:LineBreak = Open_Punctuation:];
69$PO = [:LineBreak = Postfix_Numeric:];
70$PR = [:LineBreak = Prefix_Numeric:];
71$QU = [:LineBreak = Quotation:];
72$RI = [:LineBreak = Regional_Indicator:];
73$SA = [:LineBreak = Complex_Context:];
74$SG = [:LineBreak = Surrogate:];
75$SP = [:LineBreak = Space:];
76$SY = [:LineBreak = Break_Symbols:];
77$WJ = [:LineBreak = Word_Joiner:];
78$XX = [:LineBreak = Unknown:];
79$ZW = [:LineBreak = ZWSpace:];
f3c0d7a5
A
80$ZWJ = [:LineBreak = ZWJ:];
81
82$EmojiNRK = [[\p{Emoji}] - [$RI \u002a\u00230-9©®™〰〽]];
83# Data for Extended Pictographic scraped from CLDR common/properties/ExtendedPictographic.txt, r13267
84$Extended_Pict = [\U0001F774-\U0001F77F\U00002700-\U00002701\U00002703-\U00002704\U0000270E\U00002710-\U00002711\U00002765-\U00002767\U0001F030-\U0001F093\U0001F094-\U0001F09F\U0001F10D-\U0001F10F\U0001F12F\U0001F16C-\U0001F16F\U0001F1AD-\U0001F1E5\U0001F260-\U0001F265\U0001F203-\U0001F20F\U0001F23C-\U0001F23F\U0001F249-\U0001F24F\U0001F252-\U0001F25F\U0001F266-\U0001F2FF\U0001F7D5-\U0001F7FF\U0001F000-\U0001F003\U0001F005-\U0001F02B\U0001F02C-\U0001F02F\U0001F322-\U0001F323\U0001F394-\U0001F395\U0001F398\U0001F39C-\U0001F39D\U0001F3F1-\U0001F3F2\U0001F3F6\U0001F4FE\U0001F53E-\U0001F548\U0001F54F\U0001F568-\U0001F56E\U0001F571-\U0001F572\U0001F57B-\U0001F586\U0001F588-\U0001F589\U0001F58E-\U0001F58F\U0001F591-\U0001F594\U0001F597-\U0001F5A3\U0001F5A6-\U0001F5A7\U0001F5A9-\U0001F5B0\U0001F5B3-\U0001F5BB\U0001F5BD-\U0001F5C1\U0001F5C5-\U0001F5D0\U0001F5D4-\U0001F5DB\U0001F5DF-\U0001F5E0\U0001F5E2\U0001F5E4-\U0001F5E7\U0001F5E9-\U0001F5EE\U0001F5F0-\U0001F5F2\U0001F5F4-\U0001F5F9\U00002605\U00002607-\U0000260D\U0000260F-\U00002610\U00002612\U00002616-\U00002617\U00002619-\U0000261C\U0000261E-\U0000261F\U00002621\U00002624-\U00002625\U00002627-\U00002629\U0000262B-\U0000262D\U00002630-\U00002637\U0000263B-\U00002647\U00002654-\U0000265F\U00002661-\U00002662\U00002664\U00002667\U00002669-\U0000267A\U0000267C-\U0000267E\U00002680-\U00002691\U00002695\U00002698\U0000269A\U0000269D-\U0000269F\U000026A2-\U000026A9\U000026AC-\U000026AF\U000026B2-\U000026BC\U000026BF-\U000026C3\U000026C6-\U000026C7\U000026C9-\U000026CD\U000026D0\U000026D2\U000026D5-\U000026E8\U000026EB-\U000026EF\U000026F6\U000026FB-\U000026FC\U000026FE-\U000026FF\U00002388\U0001FA00-\U0001FFFD\U0001F0A0-\U0001F0AE\U0001F0B1-\U0001F0BF\U0001F0C1-\U0001F0CF\U0001F0D1-\U0001F0F5\U0001F0AF-\U0001F0B0\U0001F0C0\U0001F0D0\U0001F0F6-\U0001F0FF\U0001F80C-\U0001F80F\U0001F848-\U0001F84F\U0001F85A-\U0001F85F\U0001F888-\U0001F88F\U0001F8AE-\U0001F8FF\U0001F900-\U0001F90B\U0001F91F\U0001F928-\U0001F92F\U0001F931-\U0001F932\U0001F94C\U0001F95F-\U0001F96B\U0001F992-\U0001F997\U0001F9D0-\U0001F9E6\U0001F90C-\U0001F90F\U0001F93F\U0001F94D-\U0001F94F\U0001F96C-\U0001F97F\U0001F998-\U0001F9BF\U0001F9C1-\U0001F9CF\U0001F9E7-\U0001F9FF\U0001F6C6-\U0001F6CA\U0001F6D3-\U0001F6D4\U0001F6E6-\U0001F6E8\U0001F6EA\U0001F6F1-\U0001F6F2\U0001F6F7-\U0001F6F8\U0001F6D5-\U0001F6DF\U0001F6ED-\U0001F6EF\U0001F6F9-\U0001F6FF];
85
86# By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly
87# list it in the numerous rules that use CM.
88# By LB1, SA characters with general categor of Mn or Mc also resolve to CM.
89
90$CM = [[:LineBreak = Combining_Mark:] $ZWJ [$SA & [[:Mn:][:Mc:]]]];
08b89b0a
A
91
92# Dictionary character set, for triggering language-based break engines. Currently
f3c0d7a5 93# limited to LineBreak=Complex_Context (SA).
08b89b0a 94
f3c0d7a5 95$dictionary = [$SA];
08b89b0a
A
96
97#
98# Rule LB1. By default, treat AI (characters with ambiguous east Asian width),
f3c0d7a5 99# SA (Dictionary chars, excluding Mn and Mc)
08b89b0a
A
100# SG (Unpaired Surrogates)
101# XX (Unknown, unassigned)
102# as $AL (Alphabetic)
103#
f3c0d7a5
A
104$ALPlus = [$AL $AI $SG $XX [$SA-[[:Mn:][:Mc:]]]];
105
08b89b0a
A
106
107## -------------------------------------------------
108
109!!forward;
110
08b89b0a
A
111#
112# CAN_CM is the set of characters that may combine with CM combining chars.
113# Note that Linebreak UAX 14's concept of a combining char and the rules
114# for what they can combine with are _very_ different from the rest of Unicode.
115#
116# Note that $CM itself is left out of this set. If CM is needed as a base
117# it must be listed separately in the rule.
118#
119$CAN_CM = [^$SP $BK $CR $LF $NL $ZW $CM]; # Bases that can take CMs
120$CANT_CM = [ $SP $BK $CR $LF $NL $ZW $CM]; # Bases that can't take CMs
121
122#
123# AL_FOLLOW set of chars that can unconditionally follow an AL
124# Needed in rules where stand-alone $CM s are treated as AL.
08b89b0a 125#
f3c0d7a5 126$AL_FOLLOW = [$BK $CR $LF $NL $ZW $SP $CL $CP $EX $HL $IS $SY $WJ $GL $OP $QU $BA $HY $NS $IN $NU $PR $PO $ALPlus];
08b89b0a
A
127
128
129#
130# Rule LB 4, 5 Mandatory (Hard) breaks.
131#
132$LB4Breaks = [$BK $CR $LF $NL];
2ca993e8 133$LB4NonBreaks = [^$BK $CR $LF $NL $CM];
08b89b0a
A
134$CR $LF {100};
135
136#
137# LB 6 Do not break before hard line breaks.
138#
139$LB4NonBreaks? $LB4Breaks {100}; # LB 5 do not break before hard breaks.
140$CAN_CM $CM* $LB4Breaks {100};
2ca993e8 141^$CM+ $LB4Breaks {100};
08b89b0a
A
142
143# LB 7 x SP
144# x ZW
145$LB4NonBreaks [$SP $ZW];
146$CAN_CM $CM* [$SP $ZW];
2ca993e8 147^$CM+ [$SP $ZW];
08b89b0a
A
148
149#
150# LB 8 Break after zero width space
151# TODO: ZW SP* <break>
152# An engine change is required to write the reverse rule for this.
153# For now, leave the Unicode 5.2 rule, ZW <break>
154#
155$LB8Breaks = [$LB4Breaks $ZW];
156$LB8NonBreaks = [[$LB4NonBreaks] - [$ZW]];
157
f3c0d7a5 158# LB 8a ZWJ x (ID | Extended_Pict | EmojiNRK) Emoji ZWJ sequences.
2ca993e8 159#
f3c0d7a5 160$ZWJ ($ID | $Extended_Pict | $EmojiNRK);
08b89b0a 161
2ca993e8
A
162# LB 9 Combining marks. X $CM needs to behave like X, where X is not $SP, $BK $CR $LF $NL
163# $CM not covered by the above needs to behave like $AL
08b89b0a
A
164# See definition of $CAN_CM.
165
166$CAN_CM $CM+; # Stick together any combining sequences that don't match other rules.
2ca993e8 167^$CM+;
08b89b0a
A
168
169#
170# LB 11 Do not break before or after WORD JOINER & related characters.
171#
f3c0d7a5
A
172$CAN_CM $CM* $WJ;
173$LB8NonBreaks $WJ;
174^$CM+ $WJ;
08b89b0a 175
f3c0d7a5 176$WJ $CM* .;
08b89b0a
A
177
178#
179# LB 12 Do not break after NBSP and related characters.
180# GL x
181#
f3c0d7a5 182$GL $CM* .;
2ca993e8 183
08b89b0a
A
184#
185# LB 12a Do not break before NBSP and related characters ...
186# [^SP BA HY] x GL
187#
f3c0d7a5
A
188[[$LB8NonBreaks] - [$SP $BA $HY]] $CM* $GL;
189^$CM+ $GL;
08b89b0a
A
190
191
192
193#
194# LB 13 Don't break before ']' or '!' or ';' or '/', even after spaces.
195#
196$LB8NonBreaks $CL;
197$CAN_CM $CM* $CL;
2ca993e8 198^$CM+ $CL; # by rule 10, stand-alone CM behaves as AL
08b89b0a
A
199
200$LB8NonBreaks $CP;
201$CAN_CM $CM* $CP;
2ca993e8 202^$CM+ $CP; # by rule 10, stand-alone CM behaves as AL
08b89b0a
A
203
204$LB8NonBreaks $EX;
205$CAN_CM $CM* $EX;
2ca993e8 206^$CM+ $EX; # by rule 10, stand-alone CM behaves as AL
08b89b0a
A
207
208$LB8NonBreaks $IS;
209$CAN_CM $CM* $IS;
2ca993e8 210^$CM+ $IS; # by rule 10, stand-alone CM behaves as AL
08b89b0a
A
211
212$LB8NonBreaks $SY;
213$CAN_CM $CM* $SY;
2ca993e8 214^$CM+ $SY; # by rule 10, stand-alone CM behaves as AL
08b89b0a
A
215
216
217#
218# LB 14 Do not break after OP, even after spaces
219#
f3c0d7a5 220$OP $CM* $SP* .;
08b89b0a 221
f3c0d7a5
A
222$OP $CM* $SP+ $CM+ $AL_FOLLOW?; # by rule 10, stand-alone CM behaves as AL
223 # by rule 8, CM following a SP is stand-alone.
08b89b0a
A
224
225# LB 15
f3c0d7a5 226$QU $CM* $SP* $OP;
08b89b0a
A
227
228# LB 16
229# Do not break between closing punctuation and $NS, even with intervening spaces
230# But DO allow a break between closing punctuation and $NSX, don't include it here
f3c0d7a5 231($CL | $CP) $CM* $SP* $NS;
08b89b0a
A
232
233# LB 17
f3c0d7a5 234$B2 $CM* $SP* $B2;
08b89b0a
A
235
236#
237# LB 18 Break after spaces.
238#
239$LB18NonBreaks = [$LB8NonBreaks - [$SP]];
240$LB18Breaks = [$LB8Breaks $SP];
241
242
243# LB 19
244# x QU
f3c0d7a5
A
245$LB18NonBreaks $CM* $QU;
246^$CM+ $QU;
08b89b0a
A
247
248# QU x
f3c0d7a5 249$QU $CM* .;
08b89b0a
A
250
251
252# LB 20
253# <break> $CB
254# $CB <break>
255
256$LB20NonBreaks = [$LB18NonBreaks - $CB];
257
258# LB 21 x (BA | HY | NS)
259# BB x
260#
f3c0d7a5
A
261# DO allow breaks here before NSX, so don't include it
262$LB20NonBreaks $CM* ($BA | $HY | $NS);
263^$CM+ ($BA | $HY | $NS);
08b89b0a 264
f3c0d7a5
A
265$BB $CM* [^$CB]; # $BB x
266$BB $CM* $LB20NonBreaks;
08b89b0a
A
267
268# LB 21a Don't break after Hebrew + Hyphen
269# HL (HY | BA) x
2ca993e8 270#
f3c0d7a5 271$HL $CM* ($HY | $BA) $CM* [^$CB]?;
08b89b0a
A
272
273# LB 21b (forward) Don't break between SY and HL
274# (break between HL and SY already disallowed by LB 13 above)
f3c0d7a5 275$SY $CM* $HL;
08b89b0a
A
276
277# LB 22
f3c0d7a5
A
278($ALPlus | $HL) $CM* $IN;
279^$CM+ $IN; # by rule 10, any otherwise unattached CM behaves as AL
280$EX $CM* $IN;
281($ID | $EB | $EM) $CM* $IN;
282# $IN $CM* $IN; # delete this rule for CSS loose
283$NU $CM* $IN;
08b89b0a
A
284
285
286# $LB 23
f3c0d7a5
A
287#
288($ALPlus | $HL) $CM* $NU;
289^$CM+ $NU; # Rule 10, any otherwise unattached CM behaves as AL
290$NU $CM* ($ALPlus | $HL);
291
292# LB 23a
293#
294$PR $CM* ($ID | $EB | $EM);
295($ID | $EB | $EM) $CM* $PO;
296
08b89b0a
A
297
298#
299# LB 24
300#
f3c0d7a5
A
301($PR | $PO) $CM* ($ALPlus | $HL);
302($ALPlus | $HL) $CM* ($PR | $PO);
303^$CM+ ($PR | $PO); # Rule 10, any otherwise unattached CM behaves as AL
08b89b0a
A
304
305#
306# LB 25 Numbers.
307#
f3c0d7a5
A
308(($PR | $PO) $CM*)? (($OP | $HY) $CM*)? $NU ($CM* ($NU | $SY | $IS))*
309 ($CM* ($CL | $CP))? ($CM* ($PR | $PO))?;
08b89b0a
A
310
311# LB 26 Do not break a Korean syllable
312#
f3c0d7a5
A
313$JL $CM* ($JL | $JV | $H2 | $H3);
314($JV | $H2) $CM* ($JV | $JT);
315($JT | $H3) $CM* $JT;
08b89b0a
A
316
317# LB 27 Treat korean Syllable Block the same as ID (don't break it)
f3c0d7a5
A
318($JL | $JV | $JT | $H2 | $H3) $CM* $IN;
319($JL | $JV | $JT | $H2 | $H3) $CM* $PO;
320$PR $CM* ($JL | $JV | $JT | $H2 | $H3);
08b89b0a
A
321
322
323# LB 28 Do not break between alphabetics
324#
f3c0d7a5
A
325($ALPlus | $HL) $CM* ($ALPlus | $HL);
326^$CM+ ($ALPlus | $HL); # The $CM+ is from rule 10, an unattached CM is treated as AL
08b89b0a
A
327
328# LB 29
f3c0d7a5 329$IS $CM* ($ALPlus | $HL);
08b89b0a
A
330
331# LB 30
f3c0d7a5
A
332($ALPlus | $HL | $NU) $CM* $OP;
333^$CM+ $OP; # The $CM+ is from rule 10, an unattached CM is treated as AL.
334$CP $CM* ($ALPlus | $HL | $NU);
08b89b0a 335
2ca993e8
A
336# LB 30a Do not break between regional indicators. Break after pairs of them.
337# Tricky interaction with LB8a: ZWJ x ID
f3c0d7a5
A
338$RI $CM* $RI / [[^$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HY $NS $CM] {eof}];
339$RI $CM* $RI $CM* $ZWJ / [[^$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HY $NS $CM $ID $Extended_Pict $EmojiNRK] {eof}];
340$RI $CM* $RI $CM* [$CM-$ZWJ] / [[^$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HY $NS $CM] {eof}];
08b89b0a 341
f3c0d7a5
A
342$RI $CM* $RI $CM* [$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HY $NS {eof}];
343$RI $CM* $RI $CM* $ZWJ ($ID | $Extended_Pict | $EmojiNRK);
2ca993e8
A
344
345# LB 30b Do not break between an Emoji Base and an Emoji Modifier
346$EB $CM* $EM;
08b89b0a
A
347
348#
349# Reverse Rules.
350#
351## -------------------------------------------------
352
353!!reverse;
354
f3c0d7a5
A
355# LB 9 Combining Marks.
356# Stick together any combining sequences that don't match other rules.
08b89b0a 357
f3c0d7a5 358^$CM+ $CAN_CM?;
08b89b0a
A
359
360#
361# Sequences of the form (shown forwards)
362# [CANT_CM] <break> [CM] [whatever]
363# The CM needs to behave as an AL
364#
365$AL_FOLLOW $CM+ / (
366 [$BK $CR $LF $NL $ZW {eof}] |
367 $SP+ $CM+ $SP |
368 $SP+ $CM* ([^$OP $CM $SP] | [$AL {eof}])); # if LB 14 will match, need to surpress this break.
2ca993e8 369 # LB14 says OP SP* x .
08b89b0a
A
370 # becomes OP SP* x AL
371 # becomes OP SP* x CM+ AL_FOLLOW
372 #
373 # Further note: the $AL in [$AL {eof}] is only to work around
374 # a rule compiler bug which complains about
375 # empty sets otherwise.
2ca993e8 376
08b89b0a 377
2ca993e8 378# LB 4, 5, 6
08b89b0a
A
379
380$LB4Breaks [$LB4NonBreaks-$CM];
381$LB4Breaks $CM+ $CAN_CM;
382$LF $CR;
383
384
385# LB 7 x SP
386# x ZW
387[$SP $ZW] [$LB4NonBreaks-$CM];
388[$SP $ZW] $CM+ $CAN_CM;
389
390# LB 8 ZW SP* <break>
391# TODO: to implement this, we need more than one look-ahead hard break in play at a time.
392# Requires an engine enhancement.
393# / $SP* $ZW
394
f3c0d7a5 395# LB 8a ZWJ x (ID | Extended_Pict | EmojiNRK)
2ca993e8 396#
f3c0d7a5 397($ID | $Extended_Pict | $EmojiNRK) $ZWJ $CM* $CAN_CM?;
2ca993e8 398
08b89b0a
A
399
400# LB 9,10 Combining marks.
401# X $CM needs to behave like X, where X is not $SP or controls.
402# $CM not covered by the above needs to behave like $AL
403# Stick together any combining sequences that don't match other rules.
2ca993e8 404^$CM+ $CAN_CM;
08b89b0a
A
405
406
407# LB 11
2ca993e8
A
408#
409$WJ $CM* $CAN_CM;
410$WJ [$LB8NonBreaks-$CM];
08b89b0a
A
411
412 $CANT_CM $CM* $WJ;
2ca993e8 413$CAN_CM $CM* $WJ;
08b89b0a
A
414
415# LB 12a
416# [^SP BA HY] x GL
417#
2ca993e8 418$GL $CM* [$LB8NonBreaks-[$CM $SP $BA $HY]];
08b89b0a
A
419
420# LB 12
421# GL x
422#
423$CANT_CM $CM* $GL;
2ca993e8 424$CAN_CM $CM* $GL;
08b89b0a
A
425
426
427# LB 13
428$CL $CM+ $CAN_CM;
429$CP $CM+ $CAN_CM;
430$EX $CM+ $CAN_CM;
431$IS $CM+ $CAN_CM;
432$SY $CM+ $CAN_CM;
433
434$CL [$LB8NonBreaks-$CM];
435$CP [$LB8NonBreaks-$CM];
436$EX [$LB8NonBreaks-$CM];
437$IS [$LB8NonBreaks-$CM];
438$SY [$LB8NonBreaks-$CM];
439
08b89b0a
A
440
441# LB 14 OP SP* x
442#
f3c0d7a5
A
443. $SP* $CM* $OP;
444$AL_FOLLOW? $CM+ $SP+ $CM* $OP; # by LB 10, behaves like $AL_FOLLOW? $AL $SP+ $CM* $OP
08b89b0a
A
445
446
447# LB 15
2ca993e8 448$OP $SP* $CM* $QU;
08b89b0a
A
449
450# LB 16
451# Don't include $NSX here
2ca993e8 452$NS $SP* $CM* ($CL | $CP);
08b89b0a
A
453
454# LB 17
2ca993e8 455$B2 $SP* $CM* $B2;
08b89b0a
A
456
457# LB 18 break after spaces
458# Nothing explicit needed here.
459
460
461#
462# LB 19
463#
2ca993e8
A
464$QU $CM* $CAN_CM; # . x QU
465$QU $LB18NonBreaks;
08b89b0a
A
466
467
2ca993e8 468$CAN_CM $CM* $QU; # QU x .
08b89b0a 469 $CANT_CM $CM* $QU;
2ca993e8 470
08b89b0a
A
471#
472# LB 20 Break before and after CB.
473# nothing needed here.
474#
475
476# LB 21
477# Don't include $NSX here
2ca993e8 478($BA | $HY | $NS) $CM* [$LB20NonBreaks-$CM]; # . x (BA | HY | NS)
08b89b0a 479
2ca993e8
A
480[$LB20NonBreaks-$CM] $CM* $BB; # BB x .
481[^$CB] $CM* $BB; #
08b89b0a
A
482
483# LB21a
484[^$CB] $CM* ($HY | $BA) $CM* $HL;
485
486# LB21b (reverse)
2ca993e8 487$HL $CM* $SY;
08b89b0a
A
488
489# LB 22
2ca993e8
A
490$IN $CM* ($ALPlus | $HL);
491$IN $CM* $EX;
492$IN $CM* ($ID | $EB | $EM);
493# $IN $CM* $IN; # delete this rule for CSS loose
494$IN $CM* $NU;
08b89b0a
A
495
496# LB 23
2ca993e8
A
497$NU $CM* ($ALPlus | $HL);
498($ALPlus | $HL) $CM* $NU;
08b89b0a 499
f3c0d7a5 500# LB23a
2ca993e8 501($ID | $EB | $EM) $CM* $PR;
f3c0d7a5
A
502$PO $CM* ($ID | $EB | $EM);
503
504# LB 24
505($ALPlus | $HL) $CM* ($PR | $PO);
506($PR | $PO) $CM* ($ALPlus | $HL);
507
08b89b0a
A
508
509# LB 25
510($CM* ($PR | $PO))? ($CM* ($CL | $CP))? ($CM* ($NU | $IS | $SY))* $CM* $NU ($CM* ($OP | $HY))? ($CM* ($PR | $PO))?;
511
512# LB 26
2ca993e8
A
513($H3 | $H2 | $JV | $JL) $CM* $JL;
514($JT | $JV) $CM* ($H2 | $JV);
515$JT $CM* ($H3 | $JT);
08b89b0a
A
516
517# LB 27
2ca993e8
A
518$IN $CM* ($H3 | $H2 | $JT | $JV | $JL);
519$PO $CM* ($H3 | $H2 | $JT | $JV | $JL);
520 ($H3 | $H2 | $JT | $JV | $JL) $CM* $PR;
08b89b0a
A
521
522# LB 28
2ca993e8 523($ALPlus | $HL) $CM* ($ALPlus | $HL);
08b89b0a
A
524
525
526# LB 29
2ca993e8 527($ALPlus | $HL) $CM* $IS;
08b89b0a
A
528
529# LB 30
2ca993e8
A
530$OP $CM* ($ALPlus | $HL | $NU);
531($ALPlus | $HL | $NU) $CM* $CP;
08b89b0a
A
532
533# LB 30a
2ca993e8
A
534# Pairs of Regional Indicators.
535# The following two rules are nearly identical. The first matches only sequences with an odd number of adjacent RIs,
536# the second with an even number. Stripping away the cruft they look like
537# [^RI] RI / (RI RI)+ ^RI;
538# [^RI] RI RI / (RI RI)+ ^RI;
539#
540# Line Loose tailoring: Don't include NSX here.
541[{bof} $NS $HY $BA $QU $CL $CP $EX $IS $SY $WJ $GL $ZW $SP $BK $CR $LF $NL $ZWJ] $CM* $RI / ($CM* $RI $CM* $RI)+ $CM* [{eof}[^$RI $CM]];
542[{bof} $NS $HY $BA $QU $CL $CP $EX $IS $SY $WJ $GL $ZW $SP $BK $CR $LF $NL $ZWJ] $CM* $RI $CM* $RI / ($CM* $RI $CM* $RI)+ $CM* [{eof}[^$RI $CM]];
543
544# In general, adjacent RIs stay together. The hard-break rules, above, overide this, forcing in the boundaries between pairs.
545$RI $CM* $RI;
546
547# WJ, GL, QU, etc. are classes with rules like "WJ x " which includes "WJ x RI".
548$RI $CM* ([$WJ $GL $QU $BB] | (($HY | $BA)$CM* $HL));
549
550
551# LB 30b Do not break between an Emoji Base and an Emoji Modifier
552$EM $CM* $EB;
08b89b0a 553
08b89b0a
A
554
555## -------------------------------------------------
556
557!!safe_reverse;
558
559# LB 9
2ca993e8
A
560^$CM+ [^$CM $BK $CR $LF $NL $ZW $SP];
561^$CM+ $SP / .;
08b89b0a
A
562
563# LB 14
564$SP+ $CM* $OP;
565
566# LB 15
567$SP+ $CM* $QU;
568
569# LB 16
570$SP+ $CM* ($CL | $CP);
571
572# LB 17
573$SP+ $CM* $B2;
574
575# LB 21
576$CM* ($HY | $BA) $CM* $HL;
577
578# LB 25
579($CM* ($IS | $SY))+ $CM* $NU;
580($CL | $CP) $CM* ($NU | $IS | $SY);
581
2ca993e8
A
582# LB 30
583($CM* $RI)+;
584
08b89b0a
A
585# For dictionary-based break
586$dictionary $dictionary;
587
588## -------------------------------------------------
589
590!!safe_forward;
591
592# Skip forward over all character classes that are involved in
593# rules containing patterns with possibly more than one char
594# of context.
595#
596# It might be slightly more efficient to have specific rules
597# instead of one generic one, but only if we could
598# turn off rule chaining. We don't want to move more
599# than necessary.
600#
2ca993e8 601^[$CM $OP $QU $CL $CP $B2 $PR $HY $BA $SP $RI $ZWJ $dictionary]+ [^$CM $OP $QU $CL $CP $B2 $PR $HY $BA $RI $ZWJ $dictionary];
08b89b0a
A
602$dictionary $dictionary;
603