line_loose.txt (9235B)
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.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 # This tailors the line break behavior to correspond to CSS 15 # line-break=loose (BCP47 -u-lb-loose) as defined for languages other than 16 # Chinese & Japanese. 17 # It sets characters of class CJ to behave like ID. 18 # In addition, it allows breaks: 19 # * before iteration marks 3005, 303B, 309D, 309E, 30FD, 30FE (all NS) 20 # * between characters of LineBreak class IN 21 22 type = line; 23 locale = en@lb=loose; 24 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 CMS = [: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:] CJ]; # CSS Normal tailoring: CJ resolves to ID 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 NSX = [\u3005 \u303B \u309D \u309E \u30FD \u30FE]; 59 NS = [[:LineBreak = Nonstarter:] - NSX]; 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 # OP30 and CP30 are variants of OP and CP that appear in rule LB30 from UAX 14. 78 # Limitations of this monkey test rule parser require that these definitions be pulled out 79 # rather than appearing in-line in LB 30. 80 81 OP30 = [OP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 82 CP30 = [CP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 83 84 eaFWH = [\p{ea=F}\p{ea=W}\p{ea=H}]; 85 eaFWHminusCMOPGL = [ eaFWH - [CMS OP GL] ]; 86 eaFWHminusCM = [ eaFWH - CMS ]; 87 eaFWHBreakableAtLB19 = [ eaFWH - [NS BA EX CL IN IS GL CMS] ]; 88 BAminuseaFWH = [BA - eaFWH ]; 89 90 PiQU = [\p{Pi}&QU]; 91 PfQU = [\p{Pf}&QU]; 92 93 # The redundant-looking inner brackets are required for the current parser in the test code. 94 ExtPictUnassigned = [[\p{Extended_Pictographic}]&[\p{Cn}]]; 95 96 # LB1 - Resolve AI, CB, CJ, SA, SG, and XX into other line breaking classes 97 AL = [AL AI SG XX ]; 98 dictionary = SA; 99 100 # By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly 101 # list it in the numerous rules that use CM. 102 CM = [CMS ZWJ]; 103 104 LB4: BK ÷; 105 LB5: CR LF; 106 LB5.1: CR ÷; 107 LB5.2: LF ÷; 108 LB5.3: NL ÷; 109 110 LB6: . (BK | CR | LF | NL); 111 LB6.1: [^BK CR LF NL SP ZW] CM* (BK | CR | LF | NL); 112 113 # LB8 break after ZW SP*. 114 # Precedes LB7 because both rules will match the sequences like ZW SP, 115 # and LB8 must take precedence. 116 117 LB8: ZW SP* ÷ [^ZW SP BK CR LF NL]; 118 119 # Numbers. Equivalent to Tailoring example 8 from UAX 14. 120 # Moved up, before LB14, because it can match longer sequences which must take precedence. 121 LB25: ((PR | PO)CM*)? ((OP | HY)CM*)? (IS CM*)? NU (CM*(NU | SY | IS))* (CM*(CL | CP))? (CM*(PR | PO))?; 122 123 # Rules LB14 - LB17. 124 125 # Moved before LB14, because it matches a supersequence. 126 LB20a.3: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 127 LB15a.1: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ .; 128 129 # Moved before LB14. These are really the cases where LB19a does not apply, but 130 # the old LB19 would. This is to avoid many instances of chaining over two code 131 # points. 132 LB19a.1: eaFWHminusCMOPGL ÷ PiQU CM* eaFWHminusCM; 133 LB19a.2: eaFWHminusCMOPGL CM* CMS ÷ PiQU CM* eaFWHminusCM; 134 LB19a.5: eaFWHminusCM CM* PfQU ÷ eaFWHBreakableAtLB19; 135 LB19a.6: eaFWHminusCM CM* PfQU CM* CMS ÷ eaFWHBreakableAtLB19; 136 137 # Moved up, before LB7, because they can match a longer sequence that would also match LB7. 138 # For example, the sequence "OP CM SP AL" matches LB14 139 # while the prefix of it, "OP CM SP" matches LB7.1 140 LB20a.7: OP CM* SP+ (HY | HH) CM* (AL | HL); 141 LB14: OP CM* SP* .; 142 143 LB20a.4: ^ (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 144 LB15a.2: ^ (PiQU CM* SP*)+ .; 145 # LB15b/LB15a chaining. 146 LB20a.5: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 147 LB15b.1: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ .; 148 LB15b.2: ([^SP] CM* | SP) PfQU CM* ( SP | GL | WJ | CL | QU | CP | EX | IS | SY | BK | CR | LF | NL | ZW | $ ); 149 150 # LB 15c Break before an IS that begins a number and follows a space. 151 LB15c: SP ÷ IS CM* NU; 152 153 # LB15d × IS 154 LB15d.1: [^SP] CM* IS; 155 LB15d.2: SP IS; 156 157 LB16: (CL | CP)CM* SP* NS; 158 LB17: B2 CM* SP* B2; 159 160 161 # LB7 Do not break before spaces or zero width space. 162 163 LB7.1: [^ZW SP] CM* [SP ZW]; 164 LB7.2: [ZW SP] [SP ZW]; 165 166 # LB8a 167 # ZWJ x 168 # Don't match a CM on the right - let other rules pick up CM sequences, where 169 # the ZWJ behaves as just another generic CM. 170 LB8a: ZWJ [^CM]; 171 172 173 # LB9: X CM -> X 174 # LB10: Unattached CM -> AL 175 176 #LB11: × WJ; 177 # WJ × 178 179 LB11.1: [^SP] CM* WJ; 180 LB11.2: SP WJ; 181 LB11.3: WJ CM* [^CM]; 182 183 # Needs to apply before LB12, because the new monkeys are not greedy. 184 LB20a.2: GL CM* (HY | HH) CM* (AL | HL); 185 LB12: GL CM* [^CM]; 186 187 LB12a: [^SP BA HY HH] CM* GL; 188 189 # LB 13 Do not break before ‘]’ or ‘!’ or ‘/’, even after spaces. 190 LB13.1: [^SP] CM* [CL CP EX SY]; 191 LB13.2: SP [CL CP EX SY]; 192 193 # LB 14-17 are moved above LB 7. 194 195 LB18: SP ÷; 196 197 LB19: . CM* QU; 198 LB19.1: QU CM* [^CM]; 199 200 # LB 20 Break before and after CB. 201 # Interaction with LB8a: ZWJ x . is tricky because CM includes ZWJ. 202 # ZWJ acts like a CM to the left, combining with CB. 203 # ZWJ acts independently to the right, no break after by LB8a. 204 LB20.1: . CM* ZWJ CB; 205 LB20.2: . CM* ÷ CB; 206 207 LB20a.6: CB CM* ZWJ (HY | HH) CM* (AL | HL); 208 LB20.3: CB CM* ZWJ [^CM]; 209 LB20.4: CB CM* ÷; 210 211 # LB 20a Do not break after a word-initial hyphen. 212 LB20a.1: ^(HY | HH) CM* (AL | HL); 213 214 # Note: Rule 21a must come before 21 to prevent 21.1 from matching HL BA, then 215 # not picking up the continuing match after the BA from 21a. 216 LB21a: HL CM* (HY | HH) CM* [^CM CB HL]; 217 218 LB21.1: . CM* [BA HY HH NS]; 219 LB21.2: BB CM* [^CM CB]; 220 221 LB21b: SY CM* HL; 222 223 224 LB22: [^IN] CM* IN; # For CSS Loose, allow breaks between adjacent ellipses characters. 225 226 LB23.1: (AL | HL | CM) CM* NU; 227 LB23.2: NU CM* (AL | HL); 228 229 LB23a.1: PR CM* (ID | EB | EM); 230 LB23a.2: (ID | EB | EM) CM* PO; 231 232 LB24.2: (PR | PO) CM* (AL | HL); 233 LB24.3: (AL | HL | CM) CM* (PR | PO); 234 235 236 LB26.1: JL CM* (JL | JV | H2 | H3); 237 LB26.2: (JV | H2) CM* (JV | JT); 238 LB26.3: (JT | H3) CM* JT; 239 240 LB27.1: (JL | JV | JT | H2 | H3) CM* PO; 241 LB27.2: PR CM* (JL | JV | JT | H2 | H3); 242 243 # LB28 Do not break between Alphabetics. 244 # Unattached (leading) CM treated as AL. 245 LB28: (AL | HL | CM)CM* (AL | HL); 246 247 LB28: (AP CM*)? (AS | AK | [◌] ) (CM* VI CM* (AK | [◌] ))* (CM* VI | ((CM* (AS | AK | [◌] ) )? CM* VF))?; 248 249 LB29: IS CM* (AL | HL); 250 251 # LB30 is adjusted for unattached leading CM being treated as AL. 252 LB30.1: (AL | CM | HL | NU) CM* OP30; 253 LB30.2: CP30 CM* (AL | HL | NU); 254 255 # LB30a keep pairs of RI together. 256 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]; 257 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]; 258 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]?; 259 260 # LB30b Do not break between Emoji Base (or potential emoji) and Emoji Modifier 261 LB30b.1: EB CM* EM; 262 LB30b.2: ExtPictUnassigned CM* EM; 263 264 # LB31 Break Everywhere Else. 265 # Include combining marks 266 LB31.1: . CM* ZWJ [^CM]; 267 LB31.2: . CM* ÷;