git.saurik.com Git - apple/icu.git/blob - icuSources/test/testdata/break_rules/line_loose

1 #

3 # License & terms of use: http://www.unicode.org/copyright.html

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 Revision 44 for Unicode 13.0

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 # * before hyphens 2010 & 2013 (both BA) and 301C, 30A0 (both NS)

24 # * before iteration marks 3005, 303B, 309D, 309E, 30FD, 30FE (all NS)

25 # * between characters of LineBreak class IN such as 2026

26 # * before some centered punct 203C, 2047, 2048, 2049, 30FB, FF1A, FF1B,

27 # FF65 (all NS) and FF01, FF1F (both EX).

28 # * before suffix characters with LineBreak class PO and EastAsianWidth A,F,W;

29 # this includes: 00B0 2030 2032 2033 2035 2103 2109 FE6A FF05 FFE0

30 # * after prefix characters with LineBreak class PR and EastAsianWidth A,F,W;

31 # this includes: 00A4 00B1 20AC 2116 FE69 FF04 FFE1 FFE5 FFE6

32 # It allows breaking before 201C and after 201D, for zh_Hans, zh_Hant, and ja.

35 type = line;

36 locale = ja@lb=loose;

39 AI = [:LineBreak = Ambiguous:];

40 AL = [:LineBreak = Alphabetic:];

41 BAX = [\u2010 \u2013];

42 BA = [[:LineBreak = Break_After:] - BAX];

43 HH = [\u2010]; # \u2010 is HYPHEN, default line break is BA.

44 BB = [:LineBreak = Break_Before:];

45 BK = [:LineBreak = Mandatory_Break:];

46 B2 = [:LineBreak = Break_Both:];

47 CB = [:LineBreak = Contingent_Break:];

48 CJ = [:LineBreak = Conditional_Japanese_Starter:];

49 CL = [[:LineBreak = Close_Punctuation:] \u201d];

50 CMS = [:LineBreak = Combining_Mark:];

51 CP = [:LineBreak = Close_Parenthesis:];

52 CR = [:LineBreak = Carriage_Return:];

53 EB = [[:LineBreak = EB:] \U0001F46A-\U0001F46D\U0001F46F\U0001F91D\U0001F93C];

54 EM = [:LineBreak = EM:];

55 EXX = [\uFF01 \uFF1F];

56 EX = [[:LineBreak = Exclamation:] - EXX];

57 GL = [:LineBreak = Glue:];

58 HL = [:LineBreak = Hebrew_Letter:];

59 HY = [:LineBreak = Hyphen:];

60 H2 = [:LineBreak = H2:];

61 H3 = [:LineBreak = H3:];

62 ID = [[:LineBreak = Ideographic:] CJ]; # CSS Loose tailoring: CJ resolves to ID

63 IN = [:LineBreak = Inseperable:];

64 IS = [:LineBreak = Infix_Numeric:];

65 JL = [:LineBreak = JL:];

66 JV = [:LineBreak = JV:];

67 JT = [:LineBreak = JT:];

68 LF = [:LineBreak = Line_Feed:];

69 NL = [:LineBreak = Next_Line:];

70 NSX = [\u301C \u30A0 \u3005 \u303B \u309D \u309E \u30FD \u30FE \u203C \u2047 \u2048 \u2049 \u30FB \uFF1A \uFF1B \uFF65];

71 NS = [[:LineBreak = Nonstarter:] - NSX];

72 NU = [:LineBreak = Numeric:];

73 OP = [[:LineBreak = Open_Punctuation:] \u201c];

74 POX = [\u00B0 \u2030 \u2032 \u2033 \u2035 \u2103 \u2109 \uFE6A \uFF05 \uFFE0];

75 PO = [[:LineBreak = Postfix_Numeric:] - POX];

76 PRX = [\u00A4 \u00B1 \u20AC \u2116 \uFE69 \uFF04 \uFFE1 \uFFE5 \uFFE6];

77 PR = [[:LineBreak = Prefix_Numeric:] - PRX];

78 QU = [[:LineBreak = Quotation:] - [\u201c\u201d]];

79 RI = [:LineBreak = Regional_Indicator:];

80 SA = [:LineBreak = Complex_Context:];

81 SG = [:LineBreak = Surrogate:];

82 SP = [:LineBreak = Space:];

83 SY = [:LineBreak = Break_Symbols:];

84 WJ = [:LineBreak = Word_Joiner:];

85 XX = [:LineBreak = Unknown:];

86 ZW = [:LineBreak = ZWSpace:];

87 ZWJ = [:LineBreak = ZWJ:];

89 # OP30 and CP30 are variants of OP and CP that appear in rule LB30 from UAX 14.

90 # Limitations of this monkey test rule parser require that these definitions be pulled out

91 # rather than appearing in-line in LB 30.

93 OP30 = [OP - [\p{ea=F}\p{ea=W}\p{ea=H}]];

94 CP30 = [CP - [\p{ea=F}\p{ea=W}\p{ea=H}]];

96 # LB1 - Resolve AI, CB, CJ, SA, SG, and XX into other line breaking classes

97 AL = [AL AI SG XX ];

98 dictionary = SA;

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 # Loose_cj tailoring: do not include $PRX at the beginning or $POX at the end.

122 LB25: ((PR | PO | POX)CM*)? ((OP | HY)CM*)? (IS CM*)? NU (CM*(NU | SY | IS))* (CM*(CL | CP))? (CM*(PR | PRX | PO))?;

123

124 # Rules LB14 - LB17.

125 # Moved up, before LB7, because they can match a longer sequence that would also match LB7.

126 # For example, the sequence "OP CM SP AL" matches LB14

127 # while the prefix of it, "OP CM SP" matches LB7.1

128 LB14: OP CM* SP* .;

129

130 # LB 14a Break before an IS that begins a number and follows a space.

131 LB14a: SP ÷ IS CM* NU;

132

133 # LB14b × IS

134 LB14b.1: [^SP] CM* IS;

135 LB14b.2: SP IS;

136

137 LB15: QU CM* SP* OP;

138 LB16: (CL | CP)CM* SP* NS;

139 LB17: B2 CM* SP* B2;

140

141

142 # LB7 Do not break before spaces or zero width space.

143

144 LB7.1: [^ZW SP] CM* [SP ZW];

145 LB7.2: [ZW SP] [SP ZW];

146

147 # LB8a

148 # ZWJ x

149 # Don't match a CM on the right - let other rules pick up CM sequences, where

150 # the ZWJ behaves as just another generic CM.

151 LB8a: ZWJ [^CM];

152

153

154 # LB9: X CM -> X

155 # LB10: Unattached CM -> AL

156

157 #LB11: × WJ;

158 # WJ ×

159

160 LB11.1: [^SP] CM* WJ;

161 LB11.2: SP WJ;

162 LB11.3: WJ CM* [^CM];

163

164 LB12: GL CM* [^CM];

165

166 LB12a: [^SP BA BAX HY] CM* GL;

167

168 # LB 13 Do not break before ‘]’ or ‘!’ or ‘/’, even after spaces.

169 LB13.1: [^SP] CM* [CL CP EX SY];

170 LB13.2: SP [CL CP EX SY];

171

172

173 # LB 14-17 are moved above LB 7.

174

175 LB18: SP ÷;

176

177 LB19: . CM* QU;

178 LB19.1: QU CM* [^CM];

179

180 # LB 20 Break before and after CB.

181 # Interaction with LB8a: ZWJ x . is tricky because CM includes ZWJ.

182 # ZWJ acts like a CM to the left, combining with CB.

183 # ZWJ acts independently to the right, no break after by LB8a.

184 LB20.1: . CM* ZWJ CB;

185 LB20.2: . CM* ÷ CB;

186

187 LB20.3: CB CM* ZWJ [^CM];

188 LB20.4: CB CM* ÷;

189

190 # LB 20.09 Don't break between Hyphens and Letters when there is a break preceding the hyphen.

191 LB20.09: ^(HY | HH) CM* AL;

192

193 # Note: Rule 21a must come before 21 to prevent 21.1 from matching HL BA, then

194 # not picking up the continuing match after the BA from 21a.

195 # LB 21a Don't break after Hebrew + Hyphen

196 # HL (HY | BA) x

197

198 LB21a: HL CM* (HY | BA | BAX) CM* [^CM CB]?;

199

200 LB21.1: . CM* [BA HY NS];

201 LB21.2: BB CM* [^CM CB];

202

203 LB21b: SY CM* HL;

204

205

206 LB22: [^IN] CM* IN; # For CSS Loose, allow breaks between adjacent ellipses characters.

207

208 LB23.1: (AL | HL | CM) CM* NU;

209 LB23.2: NU CM* (AL | HL);

210

211 LB23a.1: PR CM* (ID | EB | EM);

212 LB23a.2: (ID | EB | EM) CM* PO;

213

214 LB24.2: (PR | PO | POX) CM* (AL | HL);

215 LB24.3: (AL | HL | CM) CM* (PR | PO | POX);

216

217

218 LB26.1: JL CM* (JL | JV | H2 | H3);

219 LB26.2: (JV | H2) CM* (JV | JT);

220 LB26.3: (JT | H3) CM* JT;

221

222 LB27.1: (JL | JV | JT | H2 | H3) CM* IN;

223 LB27.2: (JL | JV | JT | H2 | H3) CM* PO;

224 LB27.3: PR CM* (JL | JV | JT | H2 | H3);

225

226 # LB28 Do not break between Alphabetics.

227 # Unattached (leading) CM treated as AL.

228 LB28: (AL | HL | CM)CM* (AL | HL);

229

230 LB29: IS CM* (AL | HL);

231

232 # LB30 is adjusted for unattached leading CM being treated as AL.

233 LB30.1: (AL | CM | HL | NU) CM* OP30;

234 LB30.2: CP30 CM* (AL | HL | NU);

235

236 # LB30a keep pairs of RI together.

237 LB30a.1: RI CM* RI ÷ [^BK CR LF NL SP ZW WJ CL CP EX IS SY GL QU BA HY NS IN CM];

238 LB30a.2: RI CM* RI CM* CMS ÷ [^BK CR LF NL SP ZW WJ CL CP EX IS SY GL QU BA HY NS IN CM];

239 LB30a.3: RI CM* RI CM* [BK CR LF NL SP ZW WJ GL CL CP EX IS SY QU BA HY NS IN ZWJ]?;

240

241 # LB30b Do not break between Emoji Base and Emoji Modifier

242 LB30b: EB CM* EM;

243

244 # LB31 Break Everywhere Else.

245 # Include combining marks

246 LB31.1: . CM* ZWJ [^CM];

247 LB31.2: . CM* ÷;