line_loose.txt (14572B)
1 # Copyright (C) 2016 and later: Unicode, Inc. and others. 2 # License & terms of use: http://www.unicode.org/copyright.html 3 # 4 # Copyright (c) 2002-2016 International Business Machines Corporation and 5 # others. All Rights Reserved. 6 # 7 # file: line_loose.txt 8 # 9 # Line Breaking Rules 10 # Implement default line breaking as defined by 11 # Unicode Standard Annex #14 (https://www.unicode.org/reports/tr14/) 12 # for Unicode 14.0, with the following modification: 13 # 14 # Boundaries between hyphens and following letters are suppressed when 15 # there is a boundary preceding the hyphen. See rule 20.9 16 # 17 # This tailors the line break behavior to correspond to CSS 18 # line-break=loose (BCP47 -u-lb-loose) as defined for languages other than 19 # Chinese & Japanese. 20 # It sets characters of class CJ to behave like ID. 21 # In addition, it allows breaks: 22 # * before iteration marks 3005, 303B, 309D, 309E, 30FD, 30FE (all NS) 23 # * between characters of LineBreak class IN 24 25 # 26 # Character Classes defined by TR 14. 27 # 28 29 !!chain; 30 !!quoted_literals_only; 31 32 $AI = [:LineBreak = Ambiguous:]; 33 $AK = [:LineBreak = Aksara:]; 34 $AL = [:LineBreak = Alphabetic:]; 35 $AP = [:LineBreak = Aksara_Prebase:]; 36 $AS = [:LineBreak = Aksara_Start:]; 37 $BA = [:LineBreak = Break_After:]; 38 $HH = [:LineBreak = Unambiguous_Hyphen:]; 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:]; 45 # $CM = [:LineBreak = Combining_Mark:]; 46 $CP = [:LineBreak = Close_Parenthesis:]; 47 $CR = [:LineBreak = Carriage_Return:]; 48 $EB = [:LineBreak = EB:]; 49 $EM = [:LineBreak = EM:]; 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:]; 56 # CSS Loose tailoring: CJ resolves to ID 57 $ID = [[:LineBreak = Ideographic:] $CJ]; 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 $VF = [:LineBreak = Virama_Final:]; 78 $VI = [:LineBreak = Virama:]; 79 $WJ = [:LineBreak = Word_Joiner:]; 80 $XX = [:LineBreak = Unknown:]; 81 $ZW = [:LineBreak = ZWSpace:]; 82 $ZWJ = [:LineBreak = ZWJ:]; 83 84 $EastAsian = [\p{ea=F}\p{ea=W}\p{ea=H}]; 85 86 $ExtPictUnassigned = [\p{Extended_Pictographic} & \p{Cn}]; 87 88 # By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly 89 # list it in the numerous rules that use CM. 90 # By LB1, SA characters with general categor of Mn or Mc also resolve to CM. 91 92 $CM = [[:LineBreak = Combining_Mark:] $ZWJ [$SA & [[:Mn:][:Mc:]]]]; 93 $CMX = [[$CM] - [$ZWJ]]; 94 95 # Dictionary character set, for triggering language-based break engines. Currently 96 # limited to LineBreak=Complex_Context (SA). 97 98 $dictionary = [$SA]; 99 100 # 101 # Rule LB1. By default, treat AI (characters with ambiguous east Asian width), 102 # SA (Dictionary chars, excluding Mn and Mc) 103 # SG (Unpaired Surrogates) 104 # XX (Unknown, unassigned) 105 # as $AL (Alphabetic) 106 # 107 $ALPlus = [$AL $AI $SG $XX [$SA-[[:Mn:][:Mc:]]]]; 108 109 110 ## ------------------------------------------------- 111 112 # 113 # CAN_CM is the set of characters that may combine with CM combining chars. 114 # Note that Linebreak UAX 14's concept of a combining char and the rules 115 # for what they can combine with are _very_ different from the rest of Unicode. 116 # 117 # Note that $CM itself is left out of this set. If CM is needed as a base 118 # it must be listed separately in the rule. 119 # 120 $CAN_CM = [^$SP $BK $CR $LF $NL $ZW $CM]; # Bases that can take CMs 121 $CANT_CM = [ $SP $BK $CR $LF $NL $ZW $CM]; # Bases that can't take CMs 122 123 # 124 # AL_FOLLOW set of chars that can unconditionally follow an AL 125 # Needed in rules where stand-alone $CM s are treated as AL. 126 # 127 $AL_FOLLOW = [$BK $CR $LF $NL $ZW $SP $CL $CP $EX $HL $IS $SY $WJ $GL [$OP - $EastAsian] $QU $BA $HH $HY $NS $IN $NU $PR $PO $ALPlus]; 128 129 130 # 131 # Rule LB 4, 5 Mandatory (Hard) breaks. 132 # 133 $LB4Breaks = [$BK $CR $LF $NL]; 134 $LB4NonBreaks = [^$BK $CR $LF $NL $CM]; 135 $CR $LF {100}; 136 137 # 138 # LB 6 Do not break before hard line breaks. 139 # 140 $LB4NonBreaks? $LB4Breaks {100}; # LB 5 do not break before hard breaks. 141 $CAN_CM $CM* $LB4Breaks {100}; 142 ^$CM+ $LB4Breaks {100}; 143 144 # LB 7 x SP 145 # x ZW 146 $LB4NonBreaks [$SP $ZW]; 147 $CAN_CM $CM* [$SP $ZW]; 148 ^$CM+ [$SP $ZW]; 149 150 # 151 # LB 8 Break after zero width space 152 # ZW SP* ÷ 153 # 154 $LB8Breaks = [$LB4Breaks $ZW]; 155 $LB8NonBreaks = [[$LB4NonBreaks] - [$ZW]]; 156 $ZW $SP* / [^$SP $ZW $LB4Breaks]; 157 158 # LB 8a ZWJ x Do not break Emoji ZWJ sequences. 159 # 160 $ZWJ [^$CM]; 161 162 # 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 164 # See definition of $CAN_CM. 165 166 $CAN_CM $CM+; # Stick together any combining sequences that don't match other rules. 167 ^$CM+; 168 169 # 170 # LB 11 Do not break before or after WORD JOINER & related characters. 171 # 172 $CAN_CM $CM* $WJ; 173 $LB8NonBreaks $WJ; 174 ^$CM+ $WJ; 175 176 $WJ $CM* .; 177 178 # 179 # LB 12 Do not break after NBSP and related characters. 180 # GL x 181 # 182 $GL $CM* .; 183 184 # 185 # LB 12a Do not break before NBSP and related characters ... 186 # [^SP BA HY HH] x GL 187 # 188 [[$LB8NonBreaks] - [$SP $BA $HY $HH]] $CM* $GL; 189 ^$CM+ $GL; 190 191 192 193 194 # LB 13 Don't break before ']' or '!' or '/', even after spaces. 195 # 196 $LB8NonBreaks $CL; 197 $CAN_CM $CM* $CL; 198 ^$CM+ $CL; # by rule 10, stand-alone CM behaves as AL 199 200 $LB8NonBreaks $CP; 201 $CAN_CM $CM* $CP; 202 ^$CM+ $CP; # by rule 10, stand-alone CM behaves as AL 203 204 $LB8NonBreaks $EX; 205 $CAN_CM $CM* $EX; 206 ^$CM+ $EX; # by rule 10, stand-alone CM behaves as AL 207 208 $LB8NonBreaks $SY; 209 $CAN_CM $CM* $SY; 210 ^$CM+ $SY; # by rule 10, stand-alone CM behaves as AL 211 212 213 # 214 # LB 14 Do not break after OP, even after spaces 215 # Note subtle interaction with "SP IS /" rules in LB14a. 216 # This rule consumes the SP, chaining happens on the IS, effectivley overriding the SP IS rules, 217 # which is the desired behavior. 218 # 219 $OP $CM* $SP* .; 220 221 $OP $CM* $SP+ $CM+ $AL_FOLLOW?; # by rule 10, stand-alone CM behaves as AL 222 # by rule 8, CM following a SP is stand-alone. 223 224 225 # LB 15a 226 ($OP $CM* $SP+ | [$OP $QU $GL] $CM*) ([\p{Pi} & $QU] $CM* $SP*)+ .; 227 ($OP $CM* $SP+ | [$OP $QU $GL] $CM*) ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 228 ^([\p{Pi} & $QU] $CM* $SP*)+ .; 229 ^([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 230 231 # LB 15b 232 $LB8NonBreaks [\p{Pf} & $QU] $CM* [$SP $GL $WJ $CL $QU $CP $EX $IS $SY $BK $CR $LF $NL $ZW {eof}]; 233 $CAN_CM $CM* [\p{Pf} & $QU] $CM* [$SP $GL $WJ $CL $QU $CP $EX $IS $SY $BK $CR $LF $NL $ZW {eof}]; 234 ^$CM+ [\p{Pf} & $QU] $CM* [$SP $GL $WJ $CL $QU $CP $EX $IS $SY $BK $CR $LF $NL $ZW {eof}]; 235 236 # Messy interaction: manually chain between LB 15b and LB 15a on Pf Pi. 237 $LB8NonBreaks [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ .; 238 $LB8NonBreaks [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 239 $CAN_CM $CM* [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ .; 240 $CAN_CM $CM* [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 241 ^$CM+ [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ .; 242 ^$CM+ [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 243 244 245 # LB 15c Force a break before start of a number with a leading decimal pt, e.g. " .23" 246 # Note: would be simpler to express as "$SP / $IS $CM* $NU;", but ICU rules have limitations. 247 # See issue ICU-20303 248 249 250 $CanFollowIS = [$BK $CR $LF $NL $SP $ZW $WJ $GL $CL $CP $EX $IS $SY $QU $BA $HH $HY $NS $ALPlus $HL $IN]; 251 $SP $IS / [^ $CanFollowIS $NU $CM]; 252 $SP $IS $CM* $CMX / [^ $CanFollowIS $NU $CM]; 253 254 # 255 # LB 15d Do not break before numeric separators (IS), even after spaces. 256 # SP IS QU is handled below as part of LB 19. 257 258 [$LB8NonBreaks - $SP] $IS; 259 $SP $IS $CM* [$CanFollowIS {eof}]; 260 $SP $IS $CM* $ZWJ [^$CM $NU]; 261 262 $CAN_CM $CM* $IS; 263 ^$CM+ $IS; # by rule 10, stand-alone CM behaves as AL 264 265 266 # LB 16 267 # Do not break between closing punctuation and $NS, even with intervening spaces 268 # But DO allow a break between closing punctuation and $NSX, don't include it here 269 ($CL | $CP) $CM* $SP* $NS; 270 271 # LB 17 272 $B2 $CM* $SP* $B2; 273 274 # 275 # LB 18 Break after spaces. 276 # 277 $LB18NonBreaks = [$LB8NonBreaks - [$SP]]; 278 $LB18Breaks = [$LB8Breaks $SP]; 279 280 281 # LB 19 and LB 19a. 282 # Instead of implementing both as keep-together rules as in UAX #14, we have an 283 # East_Asian_Width and General_Category-insensitive keep-together rule 284 # equivalent to the old LB19 × QU and QU ×, and then we poke holes into it based 285 # on context. This avoids having to do manual chaining over multiple characters 286 # with many other rules over multiple characters, as a keep-together LB19a would 287 # overlap in context with at least LB14, LB15a, LB15a, LB15d, LB30a, and itself. 288 $LB18NonBreaks $CM* $QU; 289 ^$CM+ $QU; 290 291 [$LB18NonBreaks & $EastAsian - [$OP $GL]] / [\p{Pi} & $QU] $CM* [ $EastAsian - $CM]; 292 [$LB18NonBreaks & $EastAsian - [$OP $GL]] $CM* $CMX / [\p{Pi} & $QU] $CM* [ $EastAsian - $CM]; 293 294 $QU $CM* .; 295 [$LB18NonBreaks & $EastAsian] $CM* [\p{Pf} & $QU] / [ $EastAsian - [$NS $BA $EX $CL $IN $IS $GL $CM]]; 296 [$LB18NonBreaks & $EastAsian] $CM* [\p{Pf} & $QU] $CM* $CMX / [ $EastAsian - [$NS $BA $EX $CL $IN $IS $GL $CM]]; 297 298 # LB 20 299 # <break> $CB 300 # $CB <break> 301 # 302 $LB20NonBreaks = [$LB18NonBreaks - $CB]; 303 304 # LB 20a Don't break between Hyphens and Letters when there is a break preceding the hyphen. 305 # Originally added as a Finnish tailoring, promoted to default ICU behavior (ICU-8151), 306 # and then to default UAX #14 behaviour (UTC-179-C32). 307 # 308 ^($HY | $HH) $CM* ($ALPlus | $HL); 309 $GL $CM* ($HY | $HH) $CM* ($ALPlus | $HL ); 310 # Non-breaking CB from LB8a: 311 $CB $CM* $ZWJ ($HY | $HH) $CM* ($ALPlus | $HL ); 312 # Non-breaking SP from LB14: 313 $OP $CM* $SP+ ($HY | $HH) $CM* ($ALPlus | $HL ); 314 # Non-breaking SP from LB15a: 315 ($OP $CM* $SP+ | [$OP $QU $GL] $CM*) ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 316 ^([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 317 # Non-breaking SP from LB15a following LB15b: 318 $LB8NonBreaks [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 319 $CAN_CM $CM* [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 320 ^$CM+ [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 321 322 # LB 21 x (BA | HH | HY | NS) 323 # BB x 324 # 325 # DO allow breaks here before NSX, so don't include it 326 $LB20NonBreaks $CM* ($BA | $HH | $HY | $NS); 327 328 329 ^$CM+ ($BA | $HH | $HY | $NS); 330 331 $BB $CM* [^$CB]; # $BB x 332 $BB $CM* $LB20NonBreaks; 333 334 # LB 21a Do not break after the hyphen in Hebrew + Hyphen + non-Hebrew 335 # HL (HY | HH) x [^HL] 336 # 337 $HL $CM* ($HY | $HH) $CM* [^$CB $HL]?; 338 339 # LB 21b (forward) Don't break between SY and HL 340 # (break between HL and SY already disallowed by LB 13 above) 341 $SY $CM* $HL; 342 343 344 # LB 22 Do not break before ellipses 345 # 346 [$LB20NonBreaks - $IN] $CM* $IN; # line_loose tailoring 347 ^$CM+ $IN; 348 349 350 # LB 23 351 # 352 ($ALPlus | $HL) $CM* $NU; 353 ^$CM+ $NU; # Rule 10, any otherwise unattached CM behaves as AL 354 $NU $CM* ($ALPlus | $HL); 355 356 # LB 23a 357 # 358 $PR $CM* ($ID | $EB | $EM); 359 ($ID | $EB | $EM) $CM* $PO; 360 361 362 # 363 # LB 24 364 # 365 ($PR | $PO) $CM* ($ALPlus | $HL); 366 ($ALPlus | $HL) $CM* ($PR | $PO); 367 ^$CM+ ($PR | $PO); # Rule 10, any otherwise unattached CM behaves as AL 368 369 # 370 # LB 25 Numbers. 371 # 372 (($PR | $PO) $CM*)? (($OP | $HY) $CM*)? ($IS $CM*)? $NU ($CM* ($NU | $SY | $IS))* 373 ($CM* ($CL | $CP))? ($CM* ($PR | $PO))?; 374 375 # LB 26 Do not break a Korean syllable 376 # 377 $JL $CM* ($JL | $JV | $H2 | $H3); 378 ($JV | $H2) $CM* ($JV | $JT); 379 ($JT | $H3) $CM* $JT; 380 381 # LB 27 Treat korean Syllable Block the same as ID (don't break it) 382 ($JL | $JV | $JT | $H2 | $H3) $CM* $PO; 383 $PR $CM* ($JL | $JV | $JT | $H2 | $H3); 384 385 386 # LB 28 Do not break between alphabetics 387 # 388 ($ALPlus | $HL) $CM* ($ALPlus | $HL); 389 ^$CM+ ($ALPlus | $HL); # The $CM+ is from rule 10, an unattached CM is treated as AL 390 391 #LB 28a Do not break Orthographic syllables 392 ($AP $CM*)? ($AS | $AK | [◌] ) ($CM* $VI $CM* ($AK | [◌] ))* ($CM* $VI | (($CM* ($AS | $AK | [◌] ) )? $CM* $VF))?; 393 394 # LB 29 395 $IS $CM* ($ALPlus | $HL); 396 397 # LB 30 398 ($ALPlus | $HL | $NU) $CM* [$OP - $EastAsian]; 399 ^$CM+ [$OP - $EastAsian]; # The $CM+ is from rule 10, an unattached CM is treated as AL. 400 [$CP - $EastAsian] $CM* ($ALPlus | $HL | $NU); 401 402 # LB 30a Do not break between regional indicators. Break after pairs of them. 403 # Tricky interaction with LB8a: ZWJ x . together with ZWJ acting like a CM. 404 $RI $CM* $RI / [[^$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HH $HY $NS $IN $CM]]; 405 $RI $CM* $RI $CM* [$CM-$ZWJ] / [[^$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HH $HY $NS $IN $CM]]; 406 $RI $CM* $RI $CM* [$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HH $HY $NS $IN $ZWJ {eof}]; 407 # note: the preceding rule includes {eof} rather than having the last [set] term qualified with '?' 408 # because of the chain-out behavior difference. The rule must chain out only from the [set characters], 409 # not from the preceding $RI or $CM, which it would be able to do if the set were optional. 410 411 # LB30b Do not break between an emoji base (or potential emoji) and an emoji modifier. 412 $EB $CM* $EM; 413 $ExtPictUnassigned $CM* $EM; 414 415 # LB 31 Break everywhere else. 416 # Match a single code point if no other rule applies. 417 .;