line.txt (13975B)
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.txt 7 # 8 # Line Breaking Rules 9 # Implement default line breaking as defined by 10 # Unicode Standard Annex #14 (https://www.unicode.org/reports/tr14/) 11 # for Unicode 14.0, with the following modification: 12 # 13 # Boundaries between hyphens and following letters are suppressed when 14 # there is a boundary preceding the hyphen. See rule 20.9 15 # 16 # This corresponds to CSS line-break=strict (BCP47 -u-lb-strict). 17 # It sets characters of class CJ to behave like NS. 18 19 # 20 # Character Classes defined by TR 14. 21 # 22 23 !!chain; 24 !!quoted_literals_only; 25 26 $AI = [:LineBreak = Ambiguous:]; 27 $AK = [:LineBreak = Aksara:]; 28 $AL = [:LineBreak = Alphabetic:]; 29 $AP = [:LineBreak = Aksara_Prebase:]; 30 $AS = [:LineBreak = Aksara_Start:]; 31 $BA = [:LineBreak = Break_After:]; 32 $HH = [:LineBreak = Unambiguous_Hyphen:]; 33 $BB = [:LineBreak = Break_Before:]; 34 $BK = [:LineBreak = Mandatory_Break:]; 35 $B2 = [:LineBreak = Break_Both:]; 36 $CB = [:LineBreak = Contingent_Break:]; 37 $CJ = [:LineBreak = Conditional_Japanese_Starter:]; 38 $CL = [:LineBreak = Close_Punctuation:]; 39 # $CM = [:LineBreak = Combining_Mark:]; 40 $CP = [:LineBreak = Close_Parenthesis:]; 41 $CR = [:LineBreak = Carriage_Return:]; 42 $EB = [:LineBreak = EB:]; 43 $EM = [:LineBreak = EM:]; 44 $EX = [:LineBreak = Exclamation:]; 45 $GL = [:LineBreak = Glue:]; 46 $HL = [:LineBreak = Hebrew_Letter:]; 47 $HY = [:LineBreak = Hyphen:]; 48 $H2 = [:LineBreak = H2:]; 49 $H3 = [:LineBreak = H3:]; 50 $ID = [:LineBreak = Ideographic:]; 51 $IN = [:LineBreak = Inseperable:]; 52 $IS = [:LineBreak = Infix_Numeric:]; 53 $JL = [:LineBreak = JL:]; 54 $JV = [:LineBreak = JV:]; 55 $JT = [:LineBreak = JT:]; 56 $LF = [:LineBreak = Line_Feed:]; 57 $NL = [:LineBreak = Next_Line:]; 58 # NS includes CJ for CSS strict line breaking. 59 $NS = [[:LineBreak = Nonstarter:] $CJ]; 60 $NU = [:LineBreak = Numeric:]; 61 $OP = [:LineBreak = Open_Punctuation:]; 62 $PO = [:LineBreak = Postfix_Numeric:]; 63 $PR = [:LineBreak = Prefix_Numeric:]; 64 $QU = [:LineBreak = Quotation:]; 65 $RI = [:LineBreak = Regional_Indicator:]; 66 $SA = [:LineBreak = Complex_Context:]; 67 $SG = [:LineBreak = Surrogate:]; 68 $SP = [:LineBreak = Space:]; 69 $SY = [:LineBreak = Break_Symbols:]; 70 $VF = [:LineBreak = Virama_Final:]; 71 $VI = [:LineBreak = Virama:]; 72 $WJ = [:LineBreak = Word_Joiner:]; 73 $XX = [:LineBreak = Unknown:]; 74 $ZW = [:LineBreak = ZWSpace:]; 75 $ZWJ = [:LineBreak = ZWJ:]; 76 77 $EastAsian = [\p{ea=F}\p{ea=W}\p{ea=H}]; 78 79 $ExtPictUnassigned = [\p{Extended_Pictographic} & \p{Cn}]; 80 81 # By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly 82 # list it in the numerous rules that use CM. 83 # By LB1, SA characters with general categor of Mn or Mc also resolve to CM. 84 85 $CM = [[:LineBreak = Combining_Mark:] $ZWJ [$SA & [[:Mn:][:Mc:]]]]; 86 $CMX = [[$CM] - [$ZWJ]]; 87 88 # Dictionary character set, for triggering language-based break engines. Currently 89 # limited to LineBreak=Complex_Context (SA). 90 91 $dictionary = [$SA]; 92 93 # 94 # Rule LB1. By default, treat AI (characters with ambiguous east Asian width), 95 # SA (Dictionary chars, excluding Mn and Mc) 96 # SG (Unpaired Surrogates) 97 # XX (Unknown, unassigned) 98 # as $AL (Alphabetic) 99 # 100 $ALPlus = [$AL $AI $SG $XX [$SA-[[:Mn:][:Mc:]]]]; 101 102 103 ## ------------------------------------------------- 104 105 # 106 # CAN_CM is the set of characters that may combine with CM combining chars. 107 # Note that Linebreak UAX 14's concept of a combining char and the rules 108 # for what they can combine with are _very_ different from the rest of Unicode. 109 # 110 # Note that $CM itself is left out of this set. If CM is needed as a base 111 # it must be listed separately in the rule. 112 # 113 $CAN_CM = [^$SP $BK $CR $LF $NL $ZW $CM]; # Bases that can take CMs 114 $CANT_CM = [ $SP $BK $CR $LF $NL $ZW $CM]; # Bases that can't take CMs 115 116 # 117 # AL_FOLLOW set of chars that can unconditionally follow an AL 118 # Needed in rules where stand-alone $CM s are treated as AL. 119 # 120 $AL_FOLLOW = [$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]; 121 122 123 # 124 # Rule LB 4, 5 Mandatory (Hard) breaks. 125 # 126 $LB4Breaks = [$BK $CR $LF $NL]; 127 $LB4NonBreaks = [^$BK $CR $LF $NL $CM]; 128 $CR $LF {100}; 129 130 # 131 # LB 6 Do not break before hard line breaks. 132 # 133 $LB4NonBreaks? $LB4Breaks {100}; # LB 5 do not break before hard breaks. 134 $CAN_CM $CM* $LB4Breaks {100}; 135 ^$CM+ $LB4Breaks {100}; 136 137 # LB 7 x SP 138 # x ZW 139 $LB4NonBreaks [$SP $ZW]; 140 $CAN_CM $CM* [$SP $ZW]; 141 ^$CM+ [$SP $ZW]; 142 143 # 144 # LB 8 Break after zero width space 145 # ZW SP* ÷ 146 # 147 $LB8Breaks = [$LB4Breaks $ZW]; 148 $LB8NonBreaks = [[$LB4NonBreaks] - [$ZW]]; 149 $ZW $SP* / [^$SP $ZW $LB4Breaks]; 150 151 # LB 8a ZWJ x Do not break Emoji ZWJ sequences. 152 # 153 $ZWJ [^$CM]; 154 155 # LB 9 Combining marks. X $CM needs to behave like X, where X is not $SP, $BK $CR $LF $NL 156 # $CM not covered by the above needs to behave like $AL 157 # See definition of $CAN_CM. 158 159 $CAN_CM $CM+; # Stick together any combining sequences that don't match other rules. 160 ^$CM+; 161 162 # 163 # LB 11 Do not break before or after WORD JOINER & related characters. 164 # 165 $CAN_CM $CM* $WJ; 166 $LB8NonBreaks $WJ; 167 ^$CM+ $WJ; 168 169 $WJ $CM* .; 170 171 # 172 # LB 12 Do not break after NBSP and related characters. 173 # GL x 174 # 175 $GL $CM* .; 176 177 # 178 # LB 12a Do not break before NBSP and related characters ... 179 # [^SP BA HY HH] x GL 180 # 181 [[$LB8NonBreaks] - [$SP $BA $HY $HH]] $CM* $GL; 182 ^$CM+ $GL; 183 184 185 186 187 # LB 13 Don't break before ']' or '!' or '/', even after spaces. 188 # 189 $LB8NonBreaks $CL; 190 $CAN_CM $CM* $CL; 191 ^$CM+ $CL; # by rule 10, stand-alone CM behaves as AL 192 193 $LB8NonBreaks $CP; 194 $CAN_CM $CM* $CP; 195 ^$CM+ $CP; # by rule 10, stand-alone CM behaves as AL 196 197 $LB8NonBreaks $EX; 198 $CAN_CM $CM* $EX; 199 ^$CM+ $EX; # by rule 10, stand-alone CM behaves as AL 200 201 $LB8NonBreaks $SY; 202 $CAN_CM $CM* $SY; 203 ^$CM+ $SY; # by rule 10, stand-alone CM behaves as AL 204 205 206 # 207 # LB 14 Do not break after OP, even after spaces 208 # Note subtle interaction with "SP IS /" rules in LB14a. 209 # This rule consumes the SP, chaining happens on the IS, effectivley overriding the SP IS rules, 210 # which is the desired behavior. 211 # 212 $OP $CM* $SP* .; 213 214 $OP $CM* $SP+ $CM+ $AL_FOLLOW?; # by rule 10, stand-alone CM behaves as AL 215 # by rule 8, CM following a SP is stand-alone. 216 217 218 # LB 15a 219 ($OP $CM* $SP+ | [$OP $QU $GL] $CM*) ([\p{Pi} & $QU] $CM* $SP*)+ .; 220 ($OP $CM* $SP+ | [$OP $QU $GL] $CM*) ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 221 ^([\p{Pi} & $QU] $CM* $SP*)+ .; 222 ^([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 223 224 # LB 15b 225 $LB8NonBreaks [\p{Pf} & $QU] $CM* [$SP $GL $WJ $CL $QU $CP $EX $IS $SY $BK $CR $LF $NL $ZW {eof}]; 226 $CAN_CM $CM* [\p{Pf} & $QU] $CM* [$SP $GL $WJ $CL $QU $CP $EX $IS $SY $BK $CR $LF $NL $ZW {eof}]; 227 ^$CM+ [\p{Pf} & $QU] $CM* [$SP $GL $WJ $CL $QU $CP $EX $IS $SY $BK $CR $LF $NL $ZW {eof}]; 228 229 # Messy interaction: manually chain between LB 15b and LB 15a on Pf Pi. 230 $LB8NonBreaks [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ .; 231 $LB8NonBreaks [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 232 $CAN_CM $CM* [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ .; 233 $CAN_CM $CM* [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 234 ^$CM+ [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ .; 235 ^$CM+ [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP $CM+ $AL_FOLLOW?; 236 237 238 # LB 15c Force a break before start of a number with a leading decimal pt, e.g. " .23" 239 # Note: would be simpler to express as "$SP / $IS $CM* $NU;", but ICU rules have limitations. 240 # See issue ICU-20303 241 242 243 $CanFollowIS = [$BK $CR $LF $NL $SP $ZW $WJ $GL $CL $CP $EX $IS $SY $QU $BA $HH $HY $NS $ALPlus $HL $IN]; 244 $SP $IS / [^ $CanFollowIS $NU $CM]; 245 $SP $IS $CM* $CMX / [^ $CanFollowIS $NU $CM]; 246 247 # 248 # LB 15d Do not break before numeric separators (IS), even after spaces. 249 # SP IS QU is handled below as part of LB 19. 250 251 [$LB8NonBreaks - $SP] $IS; 252 $SP $IS $CM* [$CanFollowIS {eof}]; 253 $SP $IS $CM* $ZWJ [^$CM $NU]; 254 255 $CAN_CM $CM* $IS; 256 ^$CM+ $IS; # by rule 10, stand-alone CM behaves as AL 257 258 259 # LB 16 260 ($CL | $CP) $CM* $SP* $NS; 261 262 # LB 17 263 $B2 $CM* $SP* $B2; 264 265 # 266 # LB 18 Break after spaces. 267 # 268 $LB18NonBreaks = [$LB8NonBreaks - [$SP]]; 269 $LB18Breaks = [$LB8Breaks $SP]; 270 271 272 # LB 19 and LB 19a. 273 # Instead of implementing both as keep-together rules as in UAX #14, we have an 274 # East_Asian_Width and General_Category-insensitive keep-together rule 275 # equivalent to the old LB19 × QU and QU ×, and then we poke holes into it based 276 # on context. This avoids having to do manual chaining over multiple characters 277 # with many other rules over multiple characters, as a keep-together LB19a would 278 # overlap in context with at least LB14, LB15a, LB15a, LB15d, LB30a, and itself. 279 $LB18NonBreaks $CM* $QU; 280 ^$CM+ $QU; 281 282 [$LB18NonBreaks & $EastAsian - [$OP $GL]] / [\p{Pi} & $QU] $CM* [ $EastAsian - $CM]; 283 [$LB18NonBreaks & $EastAsian - [$OP $GL]] $CM* $CMX / [\p{Pi} & $QU] $CM* [ $EastAsian - $CM]; 284 285 $QU $CM* .; 286 [$LB18NonBreaks & $EastAsian] $CM* [\p{Pf} & $QU] / [ $EastAsian - [$NS $BA $EX $CL $IN $IS $GL $CM]]; 287 [$LB18NonBreaks & $EastAsian] $CM* [\p{Pf} & $QU] $CM* $CMX / [ $EastAsian - [$NS $BA $EX $CL $IN $IS $GL $CM]]; 288 289 # LB 20 290 # <break> $CB 291 # $CB <break> 292 # 293 $LB20NonBreaks = [$LB18NonBreaks - $CB]; 294 295 # LB 20a Don't break between Hyphens and Letters when there is a break preceding the hyphen. 296 # Originally added as a Finnish tailoring, promoted to default ICU behavior (ICU-8151), 297 # and then to default UAX #14 behaviour (UTC-179-C32). 298 # 299 ^($HY | $HH) $CM* ($ALPlus | $HL); 300 $GL $CM* ($HY | $HH) $CM* ($ALPlus | $HL ); 301 # Non-breaking CB from LB8a: 302 $CB $CM* $ZWJ ($HY | $HH) $CM* ($ALPlus | $HL ); 303 # Non-breaking SP from LB14: 304 $OP $CM* $SP+ ($HY | $HH) $CM* ($ALPlus | $HL ); 305 # Non-breaking SP from LB15a: 306 ($OP $CM* $SP+ | [$OP $QU $GL] $CM*) ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 307 ^([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 308 # Non-breaking SP from LB15a following LB15b: 309 $LB8NonBreaks [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 310 $CAN_CM $CM* [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 311 ^$CM+ [\p{Pf} & $QU] $CM* ([\p{Pi} & $QU] $CM* $SP*)+ $SP ($HY | $HH) $CM* ($ALPlus | $HL ); 312 313 # LB 21 x (BA | HH | HY | NS) 314 # BB x 315 # 316 $LB20NonBreaks $CM* ($BA | $HH | $HY | $NS); 317 318 319 ^$CM+ ($BA | $HH | $HY | $NS); 320 321 $BB $CM* [^$CB]; # $BB x 322 $BB $CM* $LB20NonBreaks; 323 324 # LB 21a Do not break after the hyphen in Hebrew + Hyphen + non-Hebrew 325 # HL (HY | HH) x [^HL] 326 # 327 $HL $CM* ($HY | $HH) $CM* [^$CB $HL]?; 328 329 # LB 21b (forward) Don't break between SY and HL 330 # (break between HL and SY already disallowed by LB 13 above) 331 $SY $CM* $HL; 332 333 # LB 22 Do not break before ellipses 334 # 335 $LB20NonBreaks $CM* $IN; 336 ^$CM+ $IN; 337 338 339 # LB 23 340 # 341 ($ALPlus | $HL) $CM* $NU; 342 ^$CM+ $NU; # Rule 10, any otherwise unattached CM behaves as AL 343 $NU $CM* ($ALPlus | $HL); 344 345 # LB 23a 346 # 347 $PR $CM* ($ID | $EB | $EM); 348 ($ID | $EB | $EM) $CM* $PO; 349 350 351 # 352 # LB 24 353 # 354 ($PR | $PO) $CM* ($ALPlus | $HL); 355 ($ALPlus | $HL) $CM* ($PR | $PO); 356 ^$CM+ ($PR | $PO); # Rule 10, any otherwise unattached CM behaves as AL 357 358 # 359 # LB 25 Numbers. 360 # 361 (($PR | $PO) $CM*)? (($OP | $HY) $CM*)? ($IS $CM*)? $NU ($CM* ($NU | $SY | $IS))* 362 ($CM* ($CL | $CP))? ($CM* ($PR | $PO))?; 363 364 # LB 26 Do not break a Korean syllable 365 # 366 $JL $CM* ($JL | $JV | $H2 | $H3); 367 ($JV | $H2) $CM* ($JV | $JT); 368 ($JT | $H3) $CM* $JT; 369 370 # LB 27 Treat korean Syllable Block the same as ID (don't break it) 371 ($JL | $JV | $JT | $H2 | $H3) $CM* $PO; 372 $PR $CM* ($JL | $JV | $JT | $H2 | $H3); 373 374 375 # LB 28 Do not break between alphabetics 376 # 377 ($ALPlus | $HL) $CM* ($ALPlus | $HL); 378 ^$CM+ ($ALPlus | $HL); # The $CM+ is from rule 10, an unattached CM is treated as AL 379 380 #LB 28a Do not break Orthographic syllables 381 ($AP $CM*)? ($AS | $AK | [◌] ) ($CM* $VI $CM* ($AK | [◌] ))* ($CM* $VI | (($CM* ($AS | $AK | [◌] ) )? $CM* $VF))?; 382 383 # LB 29 384 $IS $CM* ($ALPlus | $HL); 385 386 # LB 30 387 ($ALPlus | $HL | $NU) $CM* [$OP - $EastAsian]; 388 ^$CM+ [$OP - $EastAsian]; # The $CM+ is from rule 10, an unattached CM is treated as AL. 389 [$CP - $EastAsian] $CM* ($ALPlus | $HL | $NU); 390 391 # LB 30a Do not break between regional indicators. Break after pairs of them. 392 # Tricky interaction with LB8a: ZWJ x . together with ZWJ acting like a CM. 393 $RI $CM* $RI / [[^$BK $CR $LF $NL $SP $ZW $WJ $CL $CP $EX $IS $SY $GL $QU $BA $HH $HY $NS $IN $CM]]; 394 $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]]; 395 $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}]; 396 # note: the preceding rule includes {eof} rather than having the last [set] term qualified with '?' 397 # because of the chain-out behavior difference. The rule must chain out only from the [set characters], 398 # not from the preceding $RI or $CM, which it would be able to do if the set were optional. 399 400 # LB30b Do not break between an emoji base (or potential emoji) and an emoji modifier. 401 $EB $CM* $EM; 402 $ExtPictUnassigned $CM* $EM; 403 404 # LB 31 Break everywhere else. 405 # Match a single code point if no other rule applies. 406 .;