line_loose_cj.txt (10602B)
1 # 2 # Copyright (C) 2016 and later: Unicode, Inc. and others. 3 # License & terms of use: http://www.unicode.org/copyright.html 4 # Copyright (c) 2016, International Business Machines Corporation and others. All Rights Reserved. 5 # 6 # file: line_loose_cj.txt 7 # 8 # Reference Line Break rules for intltest rbbi/RBBIMonkeyTest. 9 # Rules derived from Unicode Standard Annex #14 for Unicode 16.0. 10 # 11 # Note: Rule syntax and the monkey test itself are still a work in progress. 12 # They are expected to change with review and the addition of support for rule tailoring. 13 # 14 # Line Breaking Rules 15 # Implement default line breaking as defined by 16 # Unicode Standard Annex #14 17 # http://www.unicode.org/reports/tr14/, tailored as noted below. 18 # 19 # This tailors the line break behavior to correspond to CSS 20 # line-break=loose (BCP47 -u-lb-loose) as defined for Chinese & Japanese. 21 # It sets characters of class CJ to behave like ID. 22 # In addition, it allows breaks: 23 # * between ID and hyphens 2010 & 2013 (as well as the rest of the HH class) 24 # * before 301C, 30A0 (both NS) 25 # * before iteration marks 3005, 303B, 309D, 309E, 30FD, 30FE (all NS) 26 # * between characters of LineBreak class IN such as 2026 27 # * before some centered punct 203C, 2047, 2048, 2049, 30FB, FF1A, FF1B, 28 # FF65 (all NS) and FF01, FF1F (both EX). 29 # * before suffix characters with LineBreak class PO and EastAsianWidth A,F,W; 30 # this includes: 00B0 2030 2032 2033 2035 2103 2109 FE6A FF05 FFE0 31 # * after prefix characters with LineBreak class PR and EastAsianWidth A,F,W; 32 # this includes: 00A4 00B1 20AC 2116 FE69 FF04 FFE1 FFE5 FFE6 33 # It allows breaking before 201C and after 201D, for zh_Hans, zh_Hant, and ja. 34 35 36 type = line; 37 locale = ja@lb=loose; 38 39 40 AI = [:LineBreak = Ambiguous:]; 41 AK = [:LineBreak = Aksara:]; 42 AL = [:LineBreak = Alphabetic:]; 43 AP = [:LineBreak = Aksara_Prebase:]; 44 AS = [:LineBreak = Aksara_Start:]; 45 BA = [:LineBreak = Break_After:]; 46 HH = [:LineBreak = Unambiguous_Hyphen:]; 47 BB = [:LineBreak = Break_Before:]; 48 BK = [:LineBreak = Mandatory_Break:]; 49 B2 = [:LineBreak = Break_Both:]; 50 CB = [:LineBreak = Contingent_Break:]; 51 CJ = [:LineBreak = Conditional_Japanese_Starter:]; 52 CL = [[:LineBreak = Close_Punctuation:] \u201d]; 53 CMS = [:LineBreak = Combining_Mark:]; 54 CP = [:LineBreak = Close_Parenthesis:]; 55 CR = [:LineBreak = Carriage_Return:]; 56 EB = [:LineBreak = EB:]; 57 EM = [:LineBreak = EM:]; 58 EXX = [\uFF01 \uFF1F]; 59 EX = [[:LineBreak = Exclamation:] - EXX]; 60 GL = [:LineBreak = Glue:]; 61 HL = [:LineBreak = Hebrew_Letter:]; 62 HY = [:LineBreak = Hyphen:]; 63 H2 = [:LineBreak = H2:]; 64 H3 = [:LineBreak = H3:]; 65 ID = [[:LineBreak = Ideographic:] CJ]; # CSS Loose tailoring: CJ resolves to ID 66 IN = [:LineBreak = Inseperable:]; 67 IS = [:LineBreak = Infix_Numeric:]; 68 JL = [:LineBreak = JL:]; 69 JV = [:LineBreak = JV:]; 70 JT = [:LineBreak = JT:]; 71 LF = [:LineBreak = Line_Feed:]; 72 NL = [:LineBreak = Next_Line:]; 73 NSX = [\u301C \u30A0 \u3005 \u303B \u309D \u309E \u30FD \u30FE \u203C \u2047 \u2048 \u2049 \u30FB \uFF1A \uFF1B \uFF65]; 74 NS = [[:LineBreak = Nonstarter:] - NSX]; 75 NU = [:LineBreak = Numeric:]; 76 OP = [[:LineBreak = Open_Punctuation:] \u201c]; 77 POX = [\u00B0 \u2030 \u2032 \u2033 \u2035 \u2103 \u2109 \uFE6A \uFF05 \uFFE0]; 78 PO = [[:LineBreak = Postfix_Numeric:] - POX]; 79 PRX = [\u00A4 \u00B1 \u20AC \u2116 \uFE69 \uFF04 \uFFE1 \uFFE5 \uFFE6]; 80 PR = [[:LineBreak = Prefix_Numeric:] - PRX]; 81 QU = [[:LineBreak = Quotation:] - [\u201c\u201d]]; 82 RI = [:LineBreak = Regional_Indicator:]; 83 SA = [:LineBreak = Complex_Context:]; 84 SG = [:LineBreak = Surrogate:]; 85 SP = [:LineBreak = Space:]; 86 SY = [:LineBreak = Break_Symbols:]; 87 VF = [:LineBreak = Virama_Final:]; 88 VI = [:LineBreak = Virama:]; 89 WJ = [:LineBreak = Word_Joiner:]; 90 XX = [:LineBreak = Unknown:]; 91 ZW = [:LineBreak = ZWSpace:]; 92 ZWJ = [:LineBreak = ZWJ:]; 93 94 # OP30 and CP30 are variants of OP and CP that appear in rule LB30 from UAX 14. 95 # Limitations of this monkey test rule parser require that these definitions be pulled out 96 # rather than appearing in-line in LB 30. 97 98 OP30 = [OP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 99 CP30 = [CP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 100 101 eaFWH = [\p{ea=F}\p{ea=W}\p{ea=H}]; 102 eaFWHminusCMOPGL = [ eaFWH - [CMS OP GL] ]; 103 eaFWHminusCM = [ eaFWH - CMS ]; 104 eaFWHBreakableAtLB19 = [ eaFWH - [NS BA EX CL IN IS GL CMS] ]; 105 BAminuseaFWH = [BA - eaFWH ]; 106 107 PiQU = [\p{Pi}&QU]; 108 PfQU = [\p{Pf}&QU]; 109 110 # The redundant-looking inner brackets are required for the current parser in the test code. 111 ExtPictUnassigned = [[\p{Extended_Pictographic}]&[\p{Cn}]]; 112 113 # LB1 - Resolve AI, CB, CJ, SA, SG, and XX into other line breaking classes 114 AL = [AL AI SG XX ]; 115 dictionary = SA; 116 117 # By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly 118 # list it in the numerous rules that use CM. 119 CM = [CMS ZWJ]; 120 121 LB4: BK ÷; 122 LB5: CR LF; 123 LB5.1: CR ÷; 124 LB5.2: LF ÷; 125 LB5.3: NL ÷; 126 127 LB6: . (BK | CR | LF | NL); 128 LB6.1: [^BK CR LF NL SP ZW] CM* (BK | CR | LF | NL); 129 130 # LB8 break after ZW SP*. 131 # Precedes LB7 because both rules will match the sequences like ZW SP, 132 # and LB8 must take precedence. 133 134 LB8: ZW SP* ÷ [^ZW SP BK CR LF NL]; 135 136 # Numbers. Equivalent to Tailoring example 8 from UAX 14. 137 # Moved up, before LB14, because it can match longer sequences which must take precedence. 138 # Loose_cj tailoring: do not include $PRX at the beginning or $POX at the end. 139 LB25: ((PR | PO | POX)CM*)? ((OP | HY)CM*)? (IS CM*)? NU (CM*(NU | SY | IS))* (CM*(CL | CP))? (CM*(PR | PRX | PO))?; 140 141 # Rules LB14 - LB17. 142 143 # Moved before LB14, because it matches a supersequence. 144 LB20a.3: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 145 LB15a.1: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ .; 146 147 # Moved before LB14. These are really the cases where LB19a does not apply, but 148 # the old LB19 would. This is to avoid many instances of chaining over two code 149 # points. 150 LB19a.1: eaFWHminusCMOPGL ÷ PiQU CM* eaFWHminusCM; 151 LB19a.2: eaFWHminusCMOPGL CM* CMS ÷ PiQU CM* eaFWHminusCM; 152 LB19a.5: eaFWHminusCM CM* PfQU ÷ eaFWHBreakableAtLB19; 153 LB19a.6: eaFWHminusCM CM* PfQU CM* CMS ÷ eaFWHBreakableAtLB19; 154 155 # Moved up, before LB7, because they can match a longer sequence that would also match LB7. 156 # For example, the sequence "OP CM SP AL" matches LB14 157 # while the prefix of it, "OP CM SP" matches LB7.1 158 LB20a.7: OP CM* SP+ (HY | HH) CM* (AL | HL); 159 LB14: OP CM* SP* .; 160 161 LB20a.4: ^ (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 162 LB15a.2: ^ (PiQU CM* SP*)+ .; 163 # LB15b/LB15a chaining. 164 LB20a.5: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 165 LB15b.1: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ .; 166 LB15b.2: ([^SP] CM* | SP) PfQU CM* ( SP | GL | WJ | CL | QU | CP | EX | IS | SY | BK | CR | LF | NL | ZW | $ ); 167 168 # LB 15c Break before an IS that begins a number and follows a space. 169 LB15c: SP ÷ IS CM* NU; 170 171 # LB15d × IS 172 LB15d.1: [^SP] CM* IS; 173 LB15d.2: SP IS; 174 175 LB16: (CL | CP)CM* SP* NS; 176 LB17: B2 CM* SP* B2; 177 178 179 # LB7 Do not break before spaces or zero width space. 180 181 LB7.1: [^ZW SP] CM* [SP ZW]; 182 LB7.2: [ZW SP] [SP ZW]; 183 184 # LB8a 185 # ZWJ x 186 # Don't match a CM on the right - let other rules pick up CM sequences, where 187 # the ZWJ behaves as just another generic CM. 188 LB8a: ZWJ [^CM]; 189 190 191 # LB9: X CM -> X 192 # LB10: Unattached CM -> AL 193 194 #LB11: × WJ; 195 # WJ × 196 197 LB11.1: [^SP] CM* WJ; 198 LB11.2: SP WJ; 199 LB11.3: WJ CM* [^CM]; 200 201 # Needs to apply before LB12, because the new monkeys are not greedy. 202 LB20a.2: GL CM* (HY | HH) CM* (AL | HL); 203 LB12: GL CM* [^CM]; 204 205 LB12a: [^SP BA HY HH] CM* GL; 206 207 # LB 13 Do not break before ‘]’ or ‘!’ or ‘/’, even after spaces. 208 LB13.1: [^SP] CM* [CL CP EX SY]; 209 LB13.2: SP [CL CP EX SY]; 210 211 212 # LB 14-17 are moved above LB 7. 213 214 LB18: SP ÷; 215 216 LB19: . CM* QU; 217 LB19.1: QU CM* [^CM]; 218 219 # LB 20 Break before and after CB. 220 # Interaction with LB8a: ZWJ x . is tricky because CM includes ZWJ. 221 # ZWJ acts like a CM to the left, combining with CB. 222 # ZWJ acts independently to the right, no break after by LB8a. 223 LB20.1: . CM* ZWJ CB; 224 LB20.2: . CM* ÷ CB; 225 226 LB20a.6: CB CM* ZWJ (HY | HH) CM* (AL | HL); 227 LB20.3: CB CM* ZWJ [^CM]; 228 LB20.4: CB CM* ÷; 229 230 # LB 20a Do not break after a word-initial hyphen. 231 LB20a.1: ^(HY | HH) CM* (AL | HL); 232 233 # Note: Rule 21a must come before 21 to prevent 21.1 from matching HL BA, then 234 # not picking up the continuing match after the BA from 21a. 235 # LB 21a Don't break after Hebrew + Hyphen 236 # HL (HY | BA) x 237 238 LB21a: HL CM* (HY | HH) CM* [^CM CB HL]; 239 240 LB21.1: [^ID] CM* [BA HY HH NS]; 241 LB21.2: ID CM* [BA HY NS]; 242 LB21.3: BB CM* [^CM CB]; 243 244 LB21b: SY CM* HL; 245 246 247 LB22: [^IN] CM* IN; # For CSS Loose, allow breaks between adjacent ellipses characters. 248 249 LB23.1: (AL | HL | CM) CM* NU; 250 LB23.2: NU CM* (AL | HL); 251 252 LB23a.1: PR CM* (ID | EB | EM); 253 LB23a.2: (ID | EB | EM) CM* PO; 254 255 LB24.2: (PR | PO | POX) CM* (AL | HL); 256 LB24.3: (AL | HL | CM) CM* (PR | PO | POX); 257 258 259 LB26.1: JL CM* (JL | JV | H2 | H3); 260 LB26.2: (JV | H2) CM* (JV | JT); 261 LB26.3: (JT | H3) CM* JT; 262 263 LB27.1: (JL | JV | JT | H2 | H3) CM* PO; 264 LB27.2: PR CM* (JL | JV | JT | H2 | H3); 265 266 # LB28 Do not break between Alphabetics. 267 # Unattached (leading) CM treated as AL. 268 LB28: (AL | HL | CM)CM* (AL | HL); 269 270 LB28: (AP CM*)? (AS | AK | [◌] ) (CM* VI CM* (AK | [◌] ))* (CM* VI | ((CM* (AS | AK | [◌] ) )? CM* VF))?; 271 272 LB29: IS CM* (AL | HL); 273 274 # LB30 is adjusted for unattached leading CM being treated as AL. 275 LB30.1: (AL | CM | HL | NU) CM* OP30; 276 LB30.2: CP30 CM* (AL | HL | NU); 277 278 # LB30a keep pairs of RI together. 279 LB30a.1: RI CM* RI ÷ [^BK CR LF NL SP ZW WJ CL CP EX IS SY GL QU BA HH HY NS IN CM]; 280 LB30a.2: RI CM* RI CM* CMS ÷ [^BK CR LF NL SP ZW WJ CL CP EX IS SY GL QU BA HH HY NS IN CM]; 281 LB30a.3: RI CM* RI CM* [BK CR LF NL SP ZW WJ GL CL CP EX IS SY QU BA HH HY NS IN ZWJ]?; 282 283 # LB30b Do not break between Emoji Base (or potential emoji) and Emoji Modifier 284 LB30b.1: EB CM* EM; 285 LB30b.2: ExtPictUnassigned CM* EM; 286 287 # LB31 Break Everywhere Else. 288 # Include combining marks 289 LB31.1: . CM* ZWJ [^CM]; 290 LB31.2: . CM* ÷;