line_normal_cj.txt (9558B)
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_normal_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=normal (BCP47 -u-lb-normal) as defined for Chinese & Japanese. 21 # It sets characters of class CJ to behave like ID. 22 # In addition, it allows breaks: 23 # * before 301C, 30A0 (both NS) 24 # It allows breaking before 201C and after 201D, for zh_Hans, zh_Hant, and ja. 25 26 type = line; 27 locale = ja@lb=normal; 28 29 AI = [:LineBreak = Ambiguous:]; 30 AK = [:LineBreak = Aksara:]; 31 AL = [:LineBreak = Alphabetic:]; 32 AP = [:LineBreak = Aksara_Prebase:]; 33 AS = [:LineBreak = Aksara_Start:]; 34 BA = [:LineBreak = Break_After:]; 35 HH = [:LineBreak = Unambiguous_Hyphen:]; 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:] \u201d]; 42 CMS = [:LineBreak = Combining_Mark:]; 43 CP = [:LineBreak = Close_Parenthesis:]; 44 CR = [:LineBreak = Carriage_Return:]; 45 EB = [:LineBreak = EB:]; 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 ID = [[:LineBreak = Ideographic:] CJ]; # CSS Normal tailoring: CJ resolves to ID 54 IN = [:LineBreak = Inseperable:]; 55 IS = [:LineBreak = Infix_Numeric:]; 56 JL = [:LineBreak = JL:]; 57 JV = [:LineBreak = JV:]; 58 JT = [:LineBreak = JT:]; 59 LF = [:LineBreak = Line_Feed:]; 60 NL = [:LineBreak = Next_Line:]; 61 NSX = [\u301C \u30A0]; 62 NS = [[:LineBreak = Nonstarter:] - NSX]; 63 NU = [:LineBreak = Numeric:]; 64 OP = [[:LineBreak = Open_Punctuation:] \u201c]; 65 PO = [:LineBreak = Postfix_Numeric:]; 66 PR = [:LineBreak = Prefix_Numeric:]; 67 QU = [[:LineBreak = Quotation:] - [\u201c\u201d]]; 68 RI = [:LineBreak = Regional_Indicator:]; 69 SA = [:LineBreak = Complex_Context:]; 70 SG = [:LineBreak = Surrogate:]; 71 SP = [:LineBreak = Space:]; 72 SY = [:LineBreak = Break_Symbols:]; 73 VF = [:LineBreak = Virama_Final:]; 74 VI = [:LineBreak = Virama:]; 75 WJ = [:LineBreak = Word_Joiner:]; 76 XX = [:LineBreak = Unknown:]; 77 ZW = [:LineBreak = ZWSpace:]; 78 ZWJ = [:LineBreak = ZWJ:]; 79 80 # OP30 and CP30 are variants of OP and CP that appear in rule LB30 from UAX 14. 81 # Limitations of this monkey test rule parser require that these definitions be pulled out 82 # rather than appearing in-line in LB 30. 83 84 OP30 = [OP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 85 CP30 = [CP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 86 87 eaFWH = [\p{ea=F}\p{ea=W}\p{ea=H}]; 88 eaFWHminusCMOPGL = [ eaFWH - [CMS OP GL] ]; 89 eaFWHminusCM = [ eaFWH - CMS ]; 90 eaFWHBreakableAtLB19 = [ eaFWH - [NS BA EX CL IN IS GL CMS] ]; 91 BAminuseaFWH = [BA - eaFWH ]; 92 93 PiQU = [\p{Pi}&QU]; 94 PfQU = [\p{Pf}&QU]; 95 96 # The redundant-looking inner brackets are required for the current parser in the test code. 97 ExtPictUnassigned = [[\p{Extended_Pictographic}]&[\p{Cn}]]; 98 99 # LB1 - Resolve AI, CB, CJ, SA, SG, and XX into other line breaking classes 100 AL = [AL AI SG XX ]; 101 dictionary = SA; 102 103 # By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly 104 # list it in the numerous rules that use CM. 105 CM = [CMS ZWJ]; 106 107 LB4: BK ÷; 108 LB5: CR LF; 109 LB5.1: CR ÷; 110 LB5.2: LF ÷; 111 LB5.3: NL ÷; 112 113 LB6: . (BK | CR | LF | NL); 114 LB6.1: [^BK CR LF NL SP ZW] CM* (BK | CR | LF | NL); 115 116 # LB8 break after ZW SP*. 117 # Precedes LB7 because both rules will match the sequences like ZW SP, 118 # and LB8 must take precedence. 119 120 LB8: ZW SP* ÷ [^ZW SP BK CR LF NL]; 121 122 # Numbers. Equivalent to Tailoring example 8 from UAX 14. 123 # Moved up, before LB14, because it can match longer sequences which must take precedence. 124 LB25: ((PR | PO)CM*)? ((OP | HY)CM*)? (IS CM*)? NU (CM*(NU | SY | IS))* (CM*(CL | CP))? (CM*(PR | PO))?; 125 126 # Rules LB14 - LB17. 127 128 # Moved before LB14, because it matches a supersequence. 129 LB20a.3: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 130 LB15a.1: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ .; 131 132 # Moved before LB14. These are really the cases where LB19a does not apply, but 133 # the old LB19 would. This is to avoid many instances of chaining over two code 134 # points. 135 LB19a.1: eaFWHminusCMOPGL ÷ PiQU CM* eaFWHminusCM; 136 LB19a.2: eaFWHminusCMOPGL CM* CMS ÷ PiQU CM* eaFWHminusCM; 137 LB19a.5: eaFWHminusCM CM* PfQU ÷ eaFWHBreakableAtLB19; 138 LB19a.6: eaFWHminusCM CM* PfQU CM* CMS ÷ eaFWHBreakableAtLB19; 139 140 # Moved up, before LB7, because they can match a longer sequence that would also match LB7. 141 # For example, the sequence "OP CM SP AL" matches LB14 142 # while the prefix of it, "OP CM SP" matches LB7.1 143 LB20a.7: OP CM* SP+ (HY | HH) CM* (AL | HL); 144 LB14: OP CM* SP* .; 145 146 LB20a.4: ^ (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 147 LB15a.2: ^ (PiQU CM* SP*)+ .; 148 # LB15b/LB15a chaining. 149 LB20a.5: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 150 LB15b.1: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ .; 151 LB15b.2: ([^SP] CM* | SP) PfQU CM* ( SP | GL | WJ | CL | QU | CP | EX | IS | SY | BK | CR | LF | NL | ZW | $ ); 152 153 # LB 15c Break before an IS that begins a number and follows a space. 154 LB15c: SP ÷ IS CM* NU; 155 156 # LB15d × IS 157 LB15d.1: [^SP] CM* IS; 158 LB15d.2: SP IS; 159 160 161 # Do not break between closing punctuation and $NS, even with intervening spaces 162 # But DO allow a break between closing punctuation and $NSX, don't include it here 163 LB16: (CL | CP)CM* SP* NS; 164 LB17: B2 CM* SP* B2; 165 166 # LB7 Do not break before spaces or zero width space. 167 168 LB7.1: [^ZW SP] CM* [SP ZW]; 169 LB7.2: [ZW SP] [SP ZW]; 170 171 # LB8a 172 # ZWJ x 173 # Don't match a CM on the right - let other rules pick up CM sequences, where 174 # the ZWJ behaves as just another generic CM. 175 LB8a: ZWJ [^CM]; 176 177 178 # LB9: X CM -> X 179 # LB10: Unattached CM -> AL 180 181 #LB11: × WJ; 182 # WJ × 183 184 LB11.1: [^SP] CM* WJ; 185 LB11.2: SP WJ; 186 LB11.3: WJ CM* [^CM]; 187 188 # Needs to apply before LB12, because the new monkeys are not greedy. 189 LB20a.2: GL CM* (HY | HH) CM* (AL | HL); 190 LB12: GL CM* [^CM]; 191 192 LB12a: [^SP BA HY HH] CM* GL; 193 194 # LB 13 Do not break before ‘]’ or ‘!’ or ‘/’, even after spaces. 195 LB13.1: [^SP] CM* [CL CP EX SY]; 196 LB13.2: SP [CL CP EX SY]; 197 198 # LB 14-17 are moved above LB 7. 199 200 LB18: SP ÷; 201 202 LB19: . CM* QU; 203 LB19.1: QU CM* [^CM]; 204 205 # LB 20 Break before and after CB. 206 # Interaction with LB8a: ZWJ x . is tricky because CM includes ZWJ. 207 # ZWJ acts like a CM to the left, combining with CB. 208 # ZWJ acts independently to the right, no break after by LB8a. 209 LB20.1: . CM* ZWJ CB; 210 LB20.2: . CM* ÷ CB; 211 212 LB20a.6: CB CM* ZWJ (HY | HH) CM* (AL | HL); 213 LB20.3: CB CM* ZWJ [^CM]; 214 LB20.4: CB CM* ÷; 215 216 # LB 20a Do not break after a word-initial hyphen. 217 LB20a.1: ^(HY | HH) CM* (AL | HL); 218 219 # Note: Rule 21a must come before 21 to prevent 21.1 from matching HL BA, then 220 # not picking up the continuing match after the BA from 21a. 221 LB21a: HL CM* (HY | HH) CM* [^CM CB HL]; 222 223 # DO allow breaks here before $NSXcm, so don't include it 224 LB21.1: . CM* [BA HY HH NS]; 225 LB21.2: BB CM* [^CM CB]; 226 227 LB21b: SY CM* HL; 228 229 LB22: . CM* IN; 230 231 LB23.1: (AL | HL | CM) CM* NU; 232 LB23.2: NU CM* (AL | HL); 233 234 LB23a.1: PR CM* (ID | EB | EM); 235 LB23a.2: (ID | EB | EM) CM* PO; 236 237 LB24.2: (PR | PO) CM* (AL | HL); 238 LB24.3: (AL | HL | CM) CM* (PR | PO); 239 240 LB26.1: JL CM* (JL | JV | H2 | H3); 241 LB26.2: (JV | H2) CM* (JV | JT); 242 LB26.3: (JT | H3) CM* JT; 243 244 LB27.1: (JL | JV | JT | H2 | H3) CM* PO; 245 LB27.2: PR CM* (JL | JV | JT | H2 | H3); 246 247 # LB28 Do not break between Alphabetics. 248 # Unattached (leading) CM treated as AL. 249 LB28: (AL | HL | CM)CM* (AL | HL); 250 251 LB28: (AP CM*)? (AS | AK | [◌] ) (CM* VI CM* (AK | [◌] ))* (CM* VI | ((CM* (AS | AK | [◌] ) )? CM* VF))?; 252 253 LB29: IS CM* (AL | HL); 254 255 # LB30 is adjusted for unattached leading CM being treated as AL. 256 LB30.1: (AL | CM | HL | NU) CM* OP30; 257 LB30.2: CP30 CM* (AL | HL | NU); 258 259 # LB30a keep pairs of RI together. 260 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]; 261 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]; 262 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]?; 263 264 # LB30b Do not break between Emoji Base (or potential emoji) and Emoji Modifier 265 LB30b.1: EB CM* EM; 266 LB30b.2: ExtPictUnassigned CM* EM; 267 268 # LB31 Break Everywhere Else. 269 # Include combining marks 270 LB31.1: . CM* ZWJ [^CM]; 271 LB31.2: . CM* ÷;