1# Copyright (C) 2016 and later: Unicode, Inc. and others. 2# License & terms of use: http://www.unicode.org/copyright.html 3# Copyright (c) 2002-2016 International Business Machines Corporation and 4# others. All Rights Reserved. 5# 6# file: line_loose_fi.txt 7# 8# Line Breaking Rules 9# Implement default line breaking as defined by 10# Unicode Standard Annex #14 Revision 37 for Unicode 9.0 11# http://www.unicode.org/reports/tr14/ 12# tailored as noted in 3rd paragraph below. 13# 14# Includes extensions to the handling of emoji ZWJ sequences from 15# https://goo.gl/cluFCn 16# 17# TODO: Rule LB 8 remains as it was in Unicode 5.2 18# This is only because of a limitation of ICU break engine implementation, 19# not because the older behavior is desirable. 20# 21# This tailors the line break behavior both for Finnish and to correpond to CSS 22# line-break=loose (BCP47 -u-lb-loose) as defined for languages other than 23# Chinese & Japanese. 24# It sets characters of class CJ to behave like ID. 25# In addition, it allows breaks before 3005, 303B, 309D, 309E, 30FD, 30FE (all NS). 26# 27# Character Classes defined by TR 14. 28# 29 30!!chain; 31 32$AI = [:LineBreak = Ambiguous:]; 33$AL = [:LineBreak = Alphabetic:]; 34$BA = [:LineBreak = Break_After:]; 35$HH = [\u2010]; # \u2010 is HYPHEN, default line break is BA. 36$BB = [:LineBreak = Break_Before:]; 37$BK = [:LineBreak = Mandatory_Break:]; 38$B2 = [:LineBreak = Break_Both:]; 39$CB = [:LineBreak = Contingent_Break:]; 40$CJ = [:LineBreak = Conditional_Japanese_Starter:]; 41$CL = [:LineBreak = Close_Punctuation:]; 42# $CM = [:LineBreak = Combining_Mark:]; 43$CP = [:LineBreak = Close_Parenthesis:]; 44$CR = [:LineBreak = Carriage_Return:]; 45$EB = [[:LineBreak = EB:]\U0001F3C2\U0001F3C7\U0001F3CC\U0001F46A-\U0001F46D\U0001F46F\U0001F574\U0001F6CC]; 46$EM = [:LineBreak = EM:]; 47$EX = [:LineBreak = Exclamation:]; 48$GL = [:LineBreak = Glue:]; 49$HL = [:LineBreak = Hebrew_Letter:]; 50$HY = [:LineBreak = Hyphen:]; 51$H2 = [:LineBreak = H2:]; 52$H3 = [:LineBreak = H3:]; 53# CSS Loose tailoring: CJ resolves to ID 54$ID = [[:LineBreak = Ideographic:] $CJ]; 55$IN = [:LineBreak = Inseperable:]; 56$IS = [:LineBreak = Infix_Numeric:]; 57$JL = [:LineBreak = JL:]; 58$JV = [:LineBreak = JV:]; 59$JT = [:LineBreak = JT:]; 60$LF = [:LineBreak = Line_Feed:]; 61$NL = [:LineBreak = Next_Line:]; 62$NSX = [\u3005 \u303B \u309D \u309E \u30FD \u30FE]; 63$NS = [[:LineBreak = Nonstarter:] - $NSX]; 64$NU = [:LineBreak = Numeric:]; 65$OP = [:LineBreak = Open_Punctuation:]; 66$PO = [:LineBreak = Postfix_Numeric:]; 67$PR = [:LineBreak = Prefix_Numeric:]; 68$QU = [:LineBreak = Quotation:]; 69$RI = [:LineBreak = Regional_Indicator:]; 70$SA = [:LineBreak = Complex_Context:]; 71$SG = [:LineBreak = Surrogate:]; 72$SP = [:LineBreak = Space:]; 73$SY = [:LineBreak = Break_Symbols:]; 74$WJ = [:LineBreak = Word_Joiner:]; 75$XX = [:LineBreak = Unknown:]; 76$ZW = [:LineBreak = ZWSpace:]; 77$ZWJ = [:LineBreak = ZWJ:]; 78 79$EmojiNRK = [[\p{Emoji}] - [$RI \u002a\u00230-9©®™〰〽]]; 80# Data for Extended Pictographic scraped from CLDR common/properties/ExtendedPictographic.txt, r12773 81$Extended_Pict = [\U0001F774-\U0001F77F\u2700-\u2701\u2703-\u2704\u270E\u2710-\u2711\u2765-\u2767\U0001F030-\U0001F093\U0001F094-\U0001F09F\U0001F10D-\U0001F10F\U0001F12F\U0001F16C-\U0001F16F\U0001F1AD-\U0001F1E5\U0001F203-\U0001F20F\U0001F23C-\U0001F23F\U0001F249-\U0001F24F\U0001F252-\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\u2605\u2607-\u260D\u260F-\u2610\u2612\u2616-\u2617\u2619-\u261C\u261E-\u261F\u2621\u2624-\u2625\u2627-\u2629\u262B-\u262D\u2630-\u2637\u263B-\u2647\u2654-\u265F\u2661-\u2662\u2664\u2667\u2669-\u267A\u267C-\u267E\u2680-\u2691\u2695\u2698\u269A\u269D-\u269F\u26A2-\u26A9\u26AC-\u26AF\u26B2-\u26BC\u26BF-\u26C3\u26C6-\u26C7\u26C9-\u26CD\u26D0\u26D2\u26D5-\u26E8\u26EB-\u26EF\u26F6\u26FB-\u26FC\u26FE-\u26FF\u2388\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-\U0001F90F\U0001F91F\U0001F928-\U0001F92F\U0001F931-\U0001F932\U0001F93F\U0001F94C-\U0001F94F\U0001F95F-\U0001F97F\U0001F992-\U0001F9BF\U0001F9C1-\U0001F9FF\U0001F6C6-\U0001F6CA\U0001F6E6-\U0001F6E8\U0001F6EA\U0001F6F1-\U0001F6F2\U0001F6D3-\U0001F6DF\U0001F6ED-\U0001F6EF\U0001F6F7-\U0001F6FF]; 82 83# By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly 84# list it in the numerous rules that use CM. 85# By LB1, SA characters with general categor of Mn or Mc also resolve to CM. 86 87$CM = [[:LineBreak = Combining_Mark:] $ZWJ [$SA & [[:Mn:][:Mc:]]]]; 88 89# Dictionary character set, for triggering language-based break engines. Currently 90# limited to LineBreak=Complex_Context (SA). 91 92$dictionary = [$SA]; 93 94# 95# Rule LB1. By default, treat AI (characters with ambiguous east Asian width), 96# SA (Dictionary chars, excluding Mn and Mc) 97# SG (Unpaired Surrogates) 98# XX (Unknown, unassigned) 99# as $AL (Alphabetic) 100# 101$ALPlus = [$AL $AI $SG $XX [$SA-[[:Mn:][:Mc:]]]]; 102 103 104## ------------------------------------------------- 105 106!!forward; 107 108# 109# CAN_CM is the set of characters that may combine with CM combining chars. 110# Note that Linebreak UAX 14's concept of a combining char and the rules 111# for what they can combine with are _very_ different from the rest of Unicode. 112# 113# Note that $CM itself is left out of this set. If CM is needed as a base 114# it must be listed separately in the rule. 115# 116$CAN_CM = [^$SP $BK $CR $LF $NL $ZW $CM]; # Bases that can take CMs 117$CANT_CM = [ $SP $BK $CR $LF $NL $ZW $CM]; # Bases that can't take CMs 118 119# 120# AL_FOLLOW set of chars that can unconditionally follow an AL 121# Needed in rules where stand-alone $CM s are treated as AL. 122# 123$AL_FOLLOW = [$BK $CR $LF $NL $ZW $SP $CL $CP $EX $HL $IS $SY $WJ $GL $OP $QU $BA $HH $HY $NS $IN $NU $PR $PO $ALPlus]; 124 125 126# 127# Rule LB 4, 5 Mandatory (Hard) breaks. 128# 129$LB4Breaks = [$BK $CR $LF $NL]; 130$LB4NonBreaks = [^$BK $CR $LF $NL $CM]; 131$CR $LF {100}; 132 133# 134# LB 6 Do not break before hard line breaks. 135# 136$LB4NonBreaks? $LB4Breaks {100}; # LB 5 do not break before hard breaks. 137$CAN_CM $CM* $LB4Breaks {100}; 138^$CM+ $LB4Breaks {100}; 139 140# LB 7 x SP 141# x ZW 142$LB4NonBreaks [$SP $ZW]; 143$CAN_CM $CM* [$SP $ZW]; 144^$CM+ [$SP $ZW]; 145 146# 147# LB 8 Break after zero width space 148# TODO: ZW SP* <break> 149# An engine change is required to write the reverse rule for this. 150# For now, leave the Unicode 5.2 rule, ZW <break> 151# 152$LB8Breaks = [$LB4Breaks $ZW]; 153$LB8NonBreaks = [[$LB4NonBreaks] - [$ZW]]; 154 155# LB 8a ZWJ x (ID | Extended_Pict | EmojiNRK) Emoji ZWJ sequences. 156# 157$ZWJ ($ID | $Extended_Pict | $EmojiNRK); 158 159# LB 9 Combining marks. X $CM needs to behave like X, where X is not $SP, $BK $CR $LF $NL 160# $CM not covered by the above needs to behave like $AL 161# See definition of $CAN_CM. 162 163$CAN_CM $CM+; # Stick together any combining sequences that don't match other rules. 164^$CM+; 165 166# 167# LB 11 Do not break before or after WORD JOINER & related characters. 168# 169$CAN_CM $CM* $WJ; 170$LB8NonBreaks $WJ; 171^$CM+ $WJ; 172 173$WJ $CM* .; 174 175# 176# LB 12 Do not break after NBSP and related characters. 177# GL x 178# 179$GL $CM* .; 180 181# 182# LB 12a Do not break before NBSP and related characters ... 183# [^SP BA HY] x GL 184# 185[[$LB8NonBreaks] - [$SP $BA $HH $HY]] $CM* $GL; 186^$CM+ $GL; 187 188 189 190# 191# LB 13 Don't break before ']' or '!' or ';' or '/', even after spaces. 192# 193$LB8NonBreaks $CL; 194$CAN_CM $CM* $CL; 195^$CM+ $CL; # by rule 10, stand-alone CM behaves as AL 196 197$LB8NonBreaks $CP; 198$CAN_CM $CM* $CP; 199^$CM+ $CP; # by rule 10, stand-alone CM behaves as AL 200 201$LB8NonBreaks $EX; 202$CAN_CM $CM* $EX; 203^$CM+ $EX; # by rule 10, stand-alone CM behaves as AL 204 205$LB8NonBreaks $IS; 206$CAN_CM $CM* $IS; 207^$CM+ $IS; # by rule 10, stand-alone CM behaves as AL 208 209$LB8NonBreaks $SY; 210$CAN_CM $CM* $SY; 211^$CM+ $SY; # by rule 10, stand-alone CM behaves as AL 212 213 214# 215# LB 14 Do not break after OP, even after spaces 216# 217$OP $CM* $SP* .; 218 219$OP $CM* $SP+ $CM+ $AL_FOLLOW?; # by rule 10, stand-alone CM behaves as AL 220 # by rule 8, CM following a SP is stand-alone. 221 222# LB 15 223$QU $CM* $SP* $OP; 224 225# LB 16 226# Do not break between closing punctuation and $NS, even with intervening spaces 227# But DO allow a break between closing punctuation and $NSX, don't include it here 228($CL | $CP) $CM* $SP* $NS; 229 230# LB 17 231$B2 $CM* $SP* $B2; 232 233# 234# LB 18 Break after spaces. 235# 236$LB18NonBreaks = [$LB8NonBreaks - [$SP]]; 237$LB18Breaks = [$LB8Breaks $SP]; 238 239 240# LB 19 241# x QU 242$LB18NonBreaks $CM* $QU; 243^$CM+ $QU; 244 245# QU x 246$QU $CM* .; 247 248 249# LB 20 250# <break> $CB 251# $CB <break> 252 253$LB20NonBreaks = [$LB18NonBreaks - $CB]; 254 255# LB 20.09 added rule for Finnish tailoring 256# LB 21 x (BA | HY | NS) 257# BB x 258# 259# DO allow breaks here before NSX, so don't include it 260$LB20NonBreaks $CM* ($BA | $HH | $HY | $NS) $CM* / $AL; 261$LB20NonBreaks $CM* ($BA | $HH | $HY | $NS); 262($HY | $HH) $AL; 263^$CM+ ($BA | $HY | $HH | $NS); 264 265$BB $CM* [^$CB]; # $BB x 266$BB $CM* $LB20NonBreaks; 267 268# LB 21a Don't break after Hebrew + Hyphen 269# HL (HY | BA) x 270# 271$HL $CM* ($HY | $BA | $HH) $CM* [^$CB]?; 272 273# LB 21b (forward) Don't break between SY and HL 274# (break between HL and SY already disallowed by LB 13 above) 275$SY $CM* $HL; 276 277# LB 22 278($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; 284 285 286# $LB 23 287# 288($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* $PO; 296 297 298# 299# LB 24 300# 301($PR | $PO) $CM* ($ALPlus | $HL); 302($ALPlus | $HL) $CM* ($PR | $PO); 303^$CM+ ($PR | $PO); # Rule 10, any otherwise unattached CM behaves as AL 304 305# 306# LB 25 Numbers. 307# 308(($PR | $PO) $CM*)? (($OP | $HY) $CM*)? $NU ($CM* ($NU | $SY | $IS))* 309 ($CM* ($CL | $CP))? ($CM* ($PR | $PO))?; 310 311# LB 26 Do not break a Korean syllable 312# 313$JL $CM* ($JL | $JV | $H2 | $H3); 314($JV | $H2) $CM* ($JV | $JT); 315($JT | $H3) $CM* $JT; 316 317# LB 27 Treat korean Syllable Block the same as ID (don't break it) 318($JL | $JV | $JT | $H2 | $H3) $CM* $IN; 319($JL | $JV | $JT | $H2 | $H3) $CM* $PO; 320$PR $CM* ($JL | $JV | $JT | $H2 | $H3); 321 322 323# LB 28 Do not break between alphabetics 324# 325($ALPlus | $HL) $CM* ($ALPlus | $HL); 326^$CM+ ($ALPlus | $HL); # The $CM+ is from rule 10, an unattached CM is treated as AL 327 328# LB 29 329$IS $CM* ($ALPlus | $HL); 330 331# LB 30 332($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 | $NU); 335 336# LB 30a Do not break between regional indicators. Break after pairs of them. 337# Tricky interaction with LB8a: ZWJ x ID 338$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 $CM] {eof}]; 341 342$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 | $EmojiNRK); 344 345# LB 30b Do not break between an Emoji Base and an Emoji Modifier 346$EB $CM* $EM; 347 348# 349# Reverse Rules. 350# 351## ------------------------------------------------- 352 353!!reverse; 354 355# LB 9 Combining Marks. 356# Stick together any combining sequences that don't match other rules. 357 358^$CM+ $CAN_CM?; 359 360# 361# Sequences of the form (shown forwards) 362# [CANT_CM] <break> [CM] [whatever] 363# The CM needs to behave as an AL 364# 365$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. 369 # LB14 says OP SP* x . 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. 376 377 378# LB 4, 5, 6 379 380$LB4Breaks [$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* $ZW 394 395# LB 8a ZWJ x (ID | Extended_Pict | EmojiNRK) 396# 397($ID | $Extended_Pict | $EmojiNRK) $ZWJ $CM* $CAN_CM?; 398 399 400# 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 $AL 403# Stick together any combining sequences that don't match other rules. 404^$CM+ $CAN_CM; 405 406 407# LB 11 408# 409$WJ $CM* $CAN_CM; 410$WJ [$LB8NonBreaks-$CM]; 411 412 $CANT_CM $CM* $WJ; 413$CAN_CM $CM* $WJ; 414 415# LB 12a 416# [^SP BA HY] x GL 417# 418$GL $CM* [$LB8NonBreaks-[$CM $SP $BA $HH $HY]]; 419 420# LB 12 421# GL x 422# 423$CANT_CM $CM* $GL; 424$CAN_CM $CM* $GL; 425 426 427# LB 13 428$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]; 439 440 441# LB 14 OP SP* x 442# 443. $SP* $CM* $OP; 444$AL_FOLLOW? $CM+ $SP+ $CM* $OP; # by LB 10, behaves like $AL_FOLLOW? $AL $SP+ $CM* $OP 445 446 447# LB 15 448$OP $SP* $CM* $QU; 449 450# LB 16 451# Don't include $NSX here 452$NS $SP* $CM* ($CL | $CP); 453 454# LB 17 455$B2 $SP* $CM* $B2; 456 457# LB 18 break after spaces 458# Nothing explicit needed here. 459 460 461# 462# LB 19 463# 464$QU $CM* $CAN_CM; # . x QU 465$QU $LB18NonBreaks; 466 467 468$CAN_CM $CM* $QU; # QU x . 469 $CANT_CM $CM* $QU; 470 471# 472# LB 20 Break before and after CB. 473# nothing needed here. 474# 475 476# LB 20.09 added rule for Finnish tailoring 477$AL ($HY | $HH) / $SP; 478 479# LB 21 480# Don't include $NSX here 481($BA | $HH | $HY | $NS) $CM* [$LB20NonBreaks-$CM]; # . x (BA | HY | NS) 482 483[$LB20NonBreaks-$CM] $CM* $BB; # BB x . 484[^$CB] $CM* $BB; # 485 486# LB21a 487[^$CB] $CM* ($HY | $BA | $HH) $CM* $HL; 488 489# LB21b (reverse) 490$HL $CM* $SY; 491 492# LB 22 493$IN $CM* ($ALPlus | $HL); 494$IN $CM* $EX; 495$IN $CM* ($ID | $EB | $EM); 496# $IN $CM* $IN; # delete this rule for CSS loose 497$IN $CM* $NU; 498 499# LB 23 500$NU $CM* ($ALPlus | $HL); 501($ALPlus | $HL) $CM* $NU; 502 503# LB23a 504($ID | $EB | $EM) $CM* $PR; 505$PO $CM* ($ID | $EB | $EM); 506 507# LB 24 508($ALPlus | $HL) $CM* ($PR | $PO); 509($PR | $PO) $CM* ($ALPlus | $HL); 510 511 512# LB 25 513($CM* ($PR | $PO))? ($CM* ($CL | $CP))? ($CM* ($NU | $IS | $SY))* $CM* $NU ($CM* ($OP | $HY))? ($CM* ($PR | $PO))?; 514 515# LB 26 516($H3 | $H2 | $JV | $JL) $CM* $JL; 517($JT | $JV) $CM* ($H2 | $JV); 518$JT $CM* ($H3 | $JT); 519 520# LB 27 521$IN $CM* ($H3 | $H2 | $JT | $JV | $JL); 522$PO $CM* ($H3 | $H2 | $JT | $JV | $JL); 523 ($H3 | $H2 | $JT | $JV | $JL) $CM* $PR; 524 525# LB 28 526($ALPlus | $HL) $CM* ($ALPlus | $HL); 527 528 529# LB 29 530($ALPlus | $HL) $CM* $IS; 531 532# LB 30 533$OP $CM* ($ALPlus | $HL | $NU); 534($ALPlus | $HL | $NU) $CM* $CP; 535 536# LB 30a 537# Pairs of Regional Indicators. 538# The following two rules are nearly identical. The first matches only sequences with an odd number of adjacent RIs, 539# the second with an even number. Stripping away the cruft they look like 540# [^RI] RI / (RI RI)+ ^RI; 541# [^RI] RI RI / (RI RI)+ ^RI; 542# 543# Line Loose tailoring: Don't include NSX here. 544[{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]]; 545[{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 $CM]]; 546 547# In general, adjacent RIs stay together. The hard-break rules, above, overide this, forcing in the boundaries between pairs. 548$RI $CM* $RI; 549 550# WJ, GL, QU, etc. are classes with rules like "WJ x " which includes "WJ x RI". 551$RI $CM* ([$WJ $GL $QU $BB] | (($HY | $BA)$CM* $HL)); 552 553 554# LB 30b Do not break between an Emoji Base and an Emoji Modifier 555$EM $CM* $EB; 556 557 558## ------------------------------------------------- 559 560!!safe_reverse; 561 562# LB 9 563^$CM+ [^$CM $BK $CR $LF $NL $ZW $SP]; 564^$CM+ $SP / .; 565 566# LB 14 567$SP+ $CM* $OP; 568 569# LB 15 570$SP+ $CM* $QU; 571 572# LB 16 573$SP+ $CM* ($CL | $CP); 574 575# LB 17 576$SP+ $CM* $B2; 577 578# LB 21 579$CM* ($HY | $BA | $HH) $CM* $HL; 580 581# LB 25 582($CM* ($IS | $SY))+ $CM* $NU; 583($CL | $CP) $CM* ($NU | $IS | $SY); 584 585# LB 30 586($CM* $RI)+; 587 588# For dictionary-based break 589$dictionary $dictionary; 590 591## ------------------------------------------------- 592 593!!safe_forward; 594 595# Skip forward over all character classes that are involved in 596# rules containing patterns with possibly more than one char 597# of context. 598# 599# It might be slightly more efficient to have specific rules 600# instead of one generic one, but only if we could 601# turn off rule chaining. We don't want to move more 602# than necessary. 603# 604^[$CM $OP $QU $CL $CP $B2 $PR $HY $BA $SP $RI $ZWJ $dictionary]+ [^$CM $OP $QU $CL $CP $B2 $PR $HY $BA $RI $ZWJ $dictionary]; 605$dictionary $dictionary; 606 607