line.txt (8799B)
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.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 # This corresponds to CSS line-break=strict (BCP47 -u-lb-strict). 12 # It sets characters of class CJ to behave like NS. 13 # 14 # Note: Rule syntax and the monkey test itself are still a work in progress. 15 # They are expected to change with review and the addition of support for rule tailoring. 16 17 18 type = line; 19 locale = en; 20 21 22 AI = [:LineBreak = Ambiguous:]; 23 AK = [:LineBreak = Aksara:]; 24 AL = [:LineBreak = Alphabetic:]; 25 AP = [:LineBreak = Aksara_Prebase:]; 26 AS = [:LineBreak = Aksara_Start:]; 27 BA = [:LineBreak = Break_After:]; 28 HH = [:LineBreak = Unambiguous_Hyphen:]; 29 BB = [:LineBreak = Break_Before:]; 30 BK = [:LineBreak = Mandatory_Break:]; 31 B2 = [:LineBreak = Break_Both:]; 32 CB = [:LineBreak = Contingent_Break:]; 33 CJ = [:LineBreak = Conditional_Japanese_Starter:]; 34 CL = [:LineBreak = Close_Punctuation:]; 35 CMS = [:LineBreak = Combining_Mark:]; 36 CP = [:LineBreak = Close_Parenthesis:]; 37 CR = [:LineBreak = Carriage_Return:]; 38 EB = [:LineBreak = EB:]; 39 EM = [:LineBreak = EM:]; 40 EX = [:LineBreak = Exclamation:]; 41 GL = [:LineBreak = Glue:]; 42 HL = [:LineBreak = Hebrew_Letter:]; 43 HY = [:LineBreak = Hyphen:]; 44 H2 = [:LineBreak = H2:]; 45 H3 = [:LineBreak = H3:]; 46 ID = [:LineBreak = Ideographic:]; 47 IN = [:LineBreak = Inseperable:]; 48 IS = [:LineBreak = Infix_Numeric:]; 49 JL = [:LineBreak = JL:]; 50 JV = [:LineBreak = JV:]; 51 JT = [:LineBreak = JT:]; 52 LF = [:LineBreak = Line_Feed:]; 53 NL = [:LineBreak = Next_Line:]; 54 NS = [[:LineBreak = Nonstarter:] CJ]; # CSS Strict tailoring: CJ resolves to NS. 55 NU = [:LineBreak = Numeric:]; 56 OP = [:LineBreak = Open_Punctuation:]; 57 PO = [:LineBreak = Postfix_Numeric:]; 58 PR = [:LineBreak = Prefix_Numeric:]; 59 QU = [:LineBreak = Quotation:]; 60 RI = [:LineBreak = Regional_Indicator:]; 61 SA = [:LineBreak = Complex_Context:]; 62 SG = [:LineBreak = Surrogate:]; 63 SP = [:LineBreak = Space:]; 64 SY = [:LineBreak = Break_Symbols:]; 65 VF = [:LineBreak = Virama_Final:]; 66 VI = [:LineBreak = Virama:]; 67 WJ = [:LineBreak = Word_Joiner:]; 68 XX = [:LineBreak = Unknown:]; 69 ZW = [:LineBreak = ZWSpace:]; 70 ZWJ = [:LineBreak = ZWJ:]; 71 72 # OP30 and CP30 are variants of OP and CP that appear in rule LB30 from UAX 14. 73 # Limitations of this monkey test rule parser require that these definitions be pulled out 74 # rather than appearing in-line in LB 30. 75 76 OP30 = [OP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 77 CP30 = [CP - [\p{ea=F}\p{ea=W}\p{ea=H}]]; 78 79 eaFWH = [\p{ea=F}\p{ea=W}\p{ea=H}]; 80 eaFWHminusCMOPGL = [ eaFWH - [CMS OP GL] ]; 81 eaFWHminusCM = [ eaFWH - CMS ]; 82 eaFWHBreakableAtLB19 = [ eaFWH - [NS BA EX CL IN IS GL CMS] ]; 83 BAminuseaFWH = [BA - eaFWH ]; 84 85 PiQU = [\p{Pi}&QU]; 86 PfQU = [\p{Pf}&QU]; 87 88 # The redundant-looking inner brackets are required for the current parser in the test code. 89 ExtPictUnassigned = [[\p{Extended_Pictographic}]&[\p{Cn}]]; 90 91 # LB1 - Resolve AI, CB, CJ, SA, SG, and XX into other line breaking classes 92 AL = [AL AI SG XX ]; 93 dictionary = SA; 94 95 # By LB9, a ZWJ also behaves as a CM. Including it in the definition of CM avoids having to explicitly 96 # list it in the numerous rules that use CM. 97 CM = [CMS ZWJ]; 98 99 LB4: BK ÷; 100 LB5: CR LF; 101 LB5.1: CR ÷; 102 LB5.2: LF ÷; 103 LB5.3: NL ÷; 104 105 LB6: . (BK | CR | LF | NL); 106 LB6.1: [^BK CR LF NL SP ZW] CM* (BK | CR | LF | NL); 107 108 # LB8 break after ZW SP*. 109 # Precedes LB7 because both rules will match the sequences like ZW SP, 110 # and LB8 must take precedence. 111 112 LB8: ZW SP* ÷ [^ZW SP BK CR LF NL]; 113 114 # Numbers. Equivalent to Tailoring example 8 from UAX 14. 115 # Moved up, before LB14, because it can match longer sequences which must take precedence. 116 LB25: ((PR | PO)CM*)? ((OP | HY)CM*)? (IS CM*)? NU (CM*(NU | SY | IS))* (CM*(CL | CP))? (CM*(PR | PO))?; 117 118 # Rules LB14 - LB17. 119 120 # Moved before LB14, because it matches a supersequence. 121 LB20a.3: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 122 LB15a.1: ( OP CM* SP* | QU CM* | GL CM* ) (PiQU CM* SP*)+ .; 123 124 # Moved before LB14. These are really the cases where LB19a does not apply, but 125 # the old LB19 would. This is to avoid many instances of chaining over two code 126 # points. 127 LB19a.1: eaFWHminusCMOPGL ÷ PiQU CM* eaFWHminusCM; 128 LB19a.2: eaFWHminusCMOPGL CM* CMS ÷ PiQU CM* eaFWHminusCM; 129 LB19a.5: eaFWHminusCM CM* PfQU ÷ eaFWHBreakableAtLB19; 130 LB19a.6: eaFWHminusCM CM* PfQU CM* CMS ÷ eaFWHBreakableAtLB19; 131 132 # Moved up, before LB7, because they can match a longer sequence that would also match LB7. 133 # For example, the sequence "OP CM SP AL" matches LB14 134 # while the prefix of it, "OP CM SP" matches LB7.1 135 LB20a.7: OP CM* SP+ (HY | HH) CM* (AL | HL); 136 LB14: OP CM* SP* .; 137 138 LB20a.4: ^ (PiQU CM* SP*)+ SP (HY | HH) CM* (AL | HL); 139 LB15a.2: ^ (PiQU CM* SP*)+ .; 140 # LB15b/LB15a chaining. 141 LB20a.5: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ SP (HY | HH) CM* AL; 142 LB15b.1: ([^SP] CM* | SP) PfQU CM* (PiQU CM* SP*)+ .; 143 LB15b.2: ([^SP] CM* | SP) PfQU CM* ( SP | GL | WJ | CL | QU | CP | EX | IS | SY | BK | CR | LF | NL | ZW | $ ); 144 145 # LB 15c Break before an IS that begins a number and follows a space. 146 LB15c: SP ÷ IS CM* NU; 147 148 # LB15d × IS 149 LB15d.1: [^SP] CM* IS; 150 LB15d.2: SP IS; 151 152 LB16: (CL | CP)CM* SP* NS; 153 LB17: B2 CM* SP* B2; 154 155 156 # LB7 Do not break before spaces or zero width space. 157 158 LB7.1: [^ZW SP] CM* [SP ZW]; 159 LB7.2: [ZW SP] [SP ZW]; 160 161 # LB8a 162 # ZWJ x 163 # Don't match a CM on the right - let other rules pick up CM sequences, where 164 # the ZWJ behaves as just another generic CM. 165 LB8a: ZWJ [^CM]; 166 167 168 # LB9: X CM -> X 169 # LB10: Unattached CM -> AL 170 171 #LB11: × WJ; 172 # WJ × 173 174 LB11.1: [^SP] CM* WJ; 175 LB11.2: SP WJ; 176 LB11.3: WJ CM* [^CM]; 177 178 # Needs to apply before LB12, because the new monkeys are not greedy. 179 LB20a.2: GL CM* (HY | HH) CM* (AL | HL); 180 LB12: GL CM* [^CM]; 181 182 LB12a: [^SP BA HY HH] CM* GL; 183 184 # LB 13 Do not break before ‘]’ or ‘!’ or ‘/’, even after spaces. 185 LB13.1: [^SP] CM* [CL CP EX SY]; 186 LB13.2: SP [CL CP EX SY]; 187 188 # LB 14-17 are moved above LB 7. 189 190 LB18: SP ÷; 191 192 LB19: . CM* QU; 193 LB19.1: QU CM* [^CM]; 194 195 # LB 20 Break before and after CB. 196 # Interaction with LB8a: ZWJ x . is tricky because CM includes ZWJ. 197 # ZWJ acts like a CM to the left, combining with CB. 198 # ZWJ acts independently to the right, no break after by LB8a. 199 LB20.1: . CM* ZWJ CB; 200 LB20.2: . CM* ÷ CB; 201 202 LB20a.6: CB CM* ZWJ (HY | HH) CM* (AL | HL); 203 LB20.3: CB CM* ZWJ [^CM]; 204 LB20.4: CB CM* ÷; 205 206 # LB 20a Do not break after a word-initial hyphen. 207 LB20a.1: ^(HY | HH) CM* (AL | HL); 208 209 # Note: Rule 21a must come before 21 to prevent 21.1 from matching HL BA, then 210 # not picking up the continuing match after the BA from 21a. 211 LB21a: HL CM* (HY | HH) CM* [^CM CB HL]; 212 213 LB21.1: . CM* [BA HY HH NS]; 214 LB21.2: BB CM* [^CM CB]; 215 216 LB21b: SY CM* HL; 217 218 LB22: . CM* IN; 219 220 LB23.1: (AL | HL | CM) CM* NU; 221 LB23.2: NU CM* (AL | HL); 222 223 LB23a.1: PR CM* (ID | EB | EM); 224 LB23a.2: (ID | EB | EM) CM* PO; 225 226 LB24.2: (PR | PO) CM* (AL | HL); 227 LB24.3: (AL | HL | CM) CM* (PR | PO); 228 229 230 LB26.1: JL CM* (JL | JV | H2 | H3); 231 LB26.2: (JV | H2) CM* (JV | JT); 232 LB26.3: (JT | H3) CM* JT; 233 234 LB27.1: (JL | JV | JT | H2 | H3) CM* PO; 235 LB27.2: PR CM* (JL | JV | JT | H2 | H3); 236 237 # LB28 Do not break between Alphabetics. 238 # Unattached (leading) CM treated as AL. 239 LB28: (AL | HL | CM)CM* (AL | HL); 240 241 LB28: (AP CM*)? (AS | AK | [◌] ) (CM* VI CM* (AK | [◌] ))* (CM* VI | ((CM* (AS | AK | [◌] ) )? CM* VF))?; 242 243 LB29: IS CM* (AL | HL); 244 245 # LB30 is adjusted for unattached leading CM being treated as AL. 246 LB30.1: (AL | CM | HL | NU) CM* OP30; 247 LB30.2: CP30 CM* (AL | HL | NU); 248 249 # LB30a keep pairs of RI together. 250 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]; 251 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]; 252 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]?; 253 254 # LB30b Do not break between Emoji Base (or potential emoji) and Emoji Modifier 255 LB30b.1: EB CM* EM; 256 LB30b.2: ExtPictUnassigned CM* EM; 257 258 # LB31 Break Everywhere Else. 259 # Include combining marks 260 LB31.1: . CM* ZWJ [^CM]; 261 LB31.2: . CM* ÷;