icuSources/i18n/regexcst.h

   1 //---------------------------------------------------------------------------------
   2 //
   3 // Generated Header File.  Do not edit by hand.
   4 //    This file contains the state table for the ICU Regular Expression Pattern Parser
   5 //    It is generated by the Perl script "regexcst.pl" from
   6 //    the rule parser state definitions file "regexcst.txt".
   7 //
   8 //   Copyright (C) 2002-2003 International Business Machines Corporation
   9 //   and others. All rights reserved.
  10 //
  11 //---------------------------------------------------------------------------------
  12 #ifndef RBBIRPT_H
  13 #define RBBIRPT_H
  14
  15 U_NAMESPACE_BEGIN
  16 //
  17 // Character classes for regex pattern scanning.
  18 //
  19     static const uint8_t kRuleSet_digit_char = 128;
  20     static const uint8_t kRuleSet_white_space = 129;
  21     static const uint8_t kRuleSet_rule_char = 130;
  22
  23
  24 enum Regex_PatternParseAction {
  25     doPossessivePlus,
  26     doCloseParen,
  27     doProperty,
  28     doBeginMatchMode,
  29     doOrOperator,
  30     doOpenCaptureParen,
  31     doBadOpenParenType,
  32     doRuleError,
  33     doIntevalLowerDigit,
  34     doBackslashs,
  35     doNGOpt,
  36     doBackslashw,
  37     doMismatchedParenErr,
  38     doOpenLookBehind,
  39     doBackslashz,
  40     doIntervalError,
  41     doStar,
  42     doCaret,
  43     doEnterQuoteMode,
  44     doNGStar,
  45     doMatchMode,
  46     doIntervalUpperDigit,
  47     doOpenLookAheadNeg,
  48     doPlus,
  49     doOpenNonCaptureParen,
  50     doBackslashA,
  51     doBackslashB,
  52     doNGPlus,
  53     doSetMatchMode,
  54     doPatFinish,
  55     doBackslashD,
  56     doPossessiveInterval,
  57     doEscapeError,
  58     doBackslashG,
  59     doSuppressComments,
  60     doMatchModeParen,
  61     doOpt,
  62     doInterval,
  63     doLiteralChar,
  64     doIntervalInit,
  65     doOpenAtomicParen,
  66     doBackslashS,
  67     doOpenLookAhead,
  68     doBackRef,
  69     doDollar,
  70     doDotAny,
  71     doBackslashW,
  72     doBackslashX,
  73     doScanUnicodeSet,
  74     doBackslashZ,
  75     doPerlInline,
  76     doPossessiveOpt,
  77     doNOP,
  78     doConditionalExpr,
  79     doExit,
  80     doNGInterval,
  81     doPatStart,
  82     doBadModeFlag,
  83     doBackslashb,
  84     doPossessiveStar,
  85     doBackslashd,
  86     doIntervalSame,
  87     doOpenLookBehindNeg,
  88     rbbiLastAction};
  89
  90 //-------------------------------------------------------------------------------
  91 //
  92 //  RegexTableEl       represents the structure of a row in the transition table
  93 //                     for the pattern parser state machine.
  94 //-------------------------------------------------------------------------------
  95 struct RegexTableEl {
  96     Regex_PatternParseAction      fAction;
  97     uint8_t                       fCharClass;       // 0-127:    an individual ASCII character
  98                                                     // 128-255:  character class index
  99     uint8_t                       fNextState;       // 0-250:    normal next-state numbers
 100                                                     // 255:      pop next-state from stack.
 101     uint8_t                       fPushState;
 102     UBool                         fNextChar;
 103 };
 104
 105 static const struct RegexTableEl gRuleParseStateTable[] = {
 106     {doNOP, 0, 0, 0, TRUE}
 107     , {doPatStart, 255, 2,0,  FALSE}     //  1      start
 108     , {doLiteralChar, 254, 14,0,  TRUE}     //  2      term
 109     , {doLiteralChar, 130, 14,0,  TRUE}     //  3
 110     , {doScanUnicodeSet, 91 /* [ */, 14,0,  TRUE}     //  4
 111     , {doNOP, 40 /* ( */, 27,0,  TRUE}     //  5
 112     , {doDotAny, 46 /* . */, 14,0,  TRUE}     //  6
 113     , {doCaret, 94 /* ^ */, 2,0,  TRUE}     //  7
 114     , {doDollar, 36 /* $ */, 2,0,  TRUE}     //  8
 115     , {doNOP, 92 /* \ */, 81,0,  TRUE}     //  9
 116     , {doOrOperator, 124 /* | */, 2,0,  TRUE}     //  10
 117     , {doCloseParen, 41 /* ) */, 255,0,  TRUE}     //  11
 118     , {doPatFinish, 253, 2,0,  FALSE}     //  12
 119     , {doRuleError, 255, 101,0,  FALSE}     //  13
 120     , {doNOP, 42 /* * */, 59,0,  TRUE}     //  14      expr-quant
 121     , {doNOP, 43 /* + */, 62,0,  TRUE}     //  15
 122     , {doNOP, 63 /* ? */, 65,0,  TRUE}     //  16
 123     , {doIntervalInit, 123 /* { */, 68,0,  TRUE}     //  17
 124     , {doNOP, 40 /* ( */, 23,0,  TRUE}     //  18
 125     , {doNOP, 255, 20,0,  FALSE}     //  19
 126     , {doOrOperator, 124 /* | */, 2,0,  TRUE}     //  20      expr-cont
 127     , {doCloseParen, 41 /* ) */, 255,0,  TRUE}     //  21
 128     , {doNOP, 255, 2,0,  FALSE}     //  22
 129     , {doSuppressComments, 63 /* ? */, 25,0,  TRUE}     //  23      open-paren-quant
 130     , {doNOP, 255, 27,0,  FALSE}     //  24
 131     , {doNOP, 35 /* # */, 47, 14, TRUE}     //  25      open-paren-quant2
 132     , {doNOP, 255, 29,0,  FALSE}     //  26
 133     , {doSuppressComments, 63 /* ? */, 29,0,  TRUE}     //  27      open-paren
 134     , {doOpenCaptureParen, 255, 2, 14, FALSE}     //  28
 135     , {doOpenNonCaptureParen, 58 /* : */, 2, 14, TRUE}     //  29      open-paren-extended
 136     , {doOpenAtomicParen, 62 /* > */, 2, 14, TRUE}     //  30
 137     , {doOpenLookAhead, 61 /* = */, 2, 20, TRUE}     //  31
 138     , {doOpenLookAheadNeg, 33 /* ! */, 2, 20, TRUE}     //  32
 139     , {doNOP, 60 /* < */, 44,0,  TRUE}     //  33
 140     , {doNOP, 35 /* # */, 47, 2, TRUE}     //  34
 141     , {doBeginMatchMode, 105 /* i */, 50,0,  FALSE}     //  35
 142     , {doBeginMatchMode, 109 /* m */, 50,0,  FALSE}     //  36
 143     , {doBeginMatchMode, 115 /* s */, 50,0,  FALSE}     //  37
 144     , {doBeginMatchMode, 119 /* w */, 50,0,  FALSE}     //  38
 145     , {doBeginMatchMode, 120 /* x */, 50,0,  FALSE}     //  39
 146     , {doBeginMatchMode, 45 /* - */, 50,0,  FALSE}     //  40
 147     , {doConditionalExpr, 40 /* ( */, 101,0,  TRUE}     //  41
 148     , {doPerlInline, 123 /* { */, 101,0,  TRUE}     //  42
 149     , {doBadOpenParenType, 255, 101,0,  FALSE}     //  43
 150     , {doOpenLookBehind, 61 /* = */, 2, 20, TRUE}     //  44      open-paren-lookbehind
 151     , {doOpenLookBehindNeg, 33 /* ! */, 2, 20, TRUE}     //  45
 152     , {doBadOpenParenType, 255, 101,0,  FALSE}     //  46
 153     , {doNOP, 41 /* ) */, 255,0,  TRUE}     //  47      paren-comment
 154     , {doMismatchedParenErr, 253, 101,0,  FALSE}     //  48
 155     , {doNOP, 255, 47,0,  TRUE}     //  49
 156     , {doMatchMode, 105 /* i */, 50,0,  TRUE}     //  50      paren-flag
 157     , {doMatchMode, 109 /* m */, 50,0,  TRUE}     //  51
 158     , {doMatchMode, 115 /* s */, 50,0,  TRUE}     //  52
 159     , {doMatchMode, 119 /* w */, 50,0,  TRUE}     //  53
 160     , {doMatchMode, 120 /* x */, 50,0,  TRUE}     //  54
 161     , {doMatchMode, 45 /* - */, 50,0,  TRUE}     //  55
 162     , {doSetMatchMode, 41 /* ) */, 2,0,  TRUE}     //  56
 163     , {doMatchModeParen, 58 /* : */, 2, 14, TRUE}     //  57
 164     , {doBadModeFlag, 255, 101,0,  FALSE}     //  58
 165     , {doNGStar, 63 /* ? */, 20,0,  TRUE}     //  59      quant-star
 166     , {doPossessiveStar, 43 /* + */, 20,0,  TRUE}     //  60
 167     , {doStar, 255, 20,0,  FALSE}     //  61
 168     , {doNGPlus, 63 /* ? */, 20,0,  TRUE}     //  62      quant-plus
 169     , {doPossessivePlus, 43 /* + */, 20,0,  TRUE}     //  63
 170     , {doPlus, 255, 20,0,  FALSE}     //  64
 171     , {doNGOpt, 63 /* ? */, 20,0,  TRUE}     //  65      quant-opt
 172     , {doPossessiveOpt, 43 /* + */, 20,0,  TRUE}     //  66
 173     , {doOpt, 255, 20,0,  FALSE}     //  67
 174     , {doNOP, 129, 68,0,  TRUE}     //  68      interval-open
 175     , {doNOP, 128, 71,0,  FALSE}     //  69
 176     , {doIntervalError, 255, 101,0,  FALSE}     //  70
 177     , {doIntevalLowerDigit, 128, 71,0,  TRUE}     //  71      interval-lower
 178     , {doNOP, 44 /* , */, 75,0,  TRUE}     //  72
 179     , {doIntervalSame, 125 /* } */, 78,0,  TRUE}     //  73
 180     , {doIntervalError, 255, 101,0,  FALSE}     //  74
 181     , {doIntervalUpperDigit, 128, 75,0,  TRUE}     //  75      interval-upper
 182     , {doNOP, 125 /* } */, 78,0,  TRUE}     //  76
 183     , {doIntervalError, 255, 101,0,  FALSE}     //  77
 184     , {doNGInterval, 63 /* ? */, 20,0,  TRUE}     //  78      interval-type
 185     , {doPossessiveInterval, 43 /* + */, 20,0,  TRUE}     //  79
 186     , {doInterval, 255, 20,0,  FALSE}     //  80
 187     , {doBackslashA, 65 /* A */, 2,0,  TRUE}     //  81      backslash
 188     , {doBackslashB, 66 /* B */, 2,0,  TRUE}     //  82
 189     , {doBackslashb, 98 /* b */, 2,0,  TRUE}     //  83
 190     , {doBackslashd, 100 /* d */, 14,0,  TRUE}     //  84
 191     , {doBackslashD, 68 /* D */, 14,0,  TRUE}     //  85
 192     , {doBackslashG, 71 /* G */, 2,0,  TRUE}     //  86
 193     , {doProperty, 78 /* N */, 14,0,  FALSE}     //  87
 194     , {doProperty, 112 /* p */, 14,0,  FALSE}     //  88
 195     , {doProperty, 80 /* P */, 14,0,  FALSE}     //  89
 196     , {doEnterQuoteMode, 81 /* Q */, 2,0,  TRUE}     //  90
 197     , {doBackslashS, 83 /* S */, 14,0,  TRUE}     //  91
 198     , {doBackslashs, 115 /* s */, 14,0,  TRUE}     //  92
 199     , {doBackslashW, 87 /* W */, 14,0,  TRUE}     //  93
 200     , {doBackslashw, 119 /* w */, 14,0,  TRUE}     //  94
 201     , {doBackslashX, 88 /* X */, 14,0,  TRUE}     //  95
 202     , {doBackslashZ, 90 /* Z */, 2,0,  TRUE}     //  96
 203     , {doBackslashz, 122 /* z */, 2,0,  TRUE}     //  97
 204     , {doBackRef, 128, 14,0,  TRUE}     //  98
 205     , {doEscapeError, 253, 101,0,  FALSE}     //  99
 206     , {doLiteralChar, 255, 14,0,  TRUE}     //  100
 207     , {doExit, 255, 101,0,  TRUE}     //  101      errorDeath
 208  };
 209 static const char * const RegexStateNames[] = {    0,
 210      "start",
 211      "term",
 212     0,
 213     0,
 214     0,
 215     0,
 216     0,
 217     0,
 218     0,
 219     0,
 220     0,
 221     0,
 222     0,
 223      "expr-quant",
 224     0,
 225     0,
 226     0,
 227     0,
 228     0,
 229      "expr-cont",
 230     0,
 231     0,
 232      "open-paren-quant",
 233     0,
 234      "open-paren-quant2",
 235     0,
 236      "open-paren",
 237     0,
 238      "open-paren-extended",
 239     0,
 240     0,
 241     0,
 242     0,
 243     0,
 244     0,
 245     0,
 246     0,
 247     0,
 248     0,
 249     0,
 250     0,
 251     0,
 252     0,
 253      "open-paren-lookbehind",
 254     0,
 255     0,
 256      "paren-comment",
 257     0,
 258     0,
 259      "paren-flag",
 260     0,
 261     0,
 262     0,
 263     0,
 264     0,
 265     0,
 266     0,
 267     0,
 268      "quant-star",
 269     0,
 270     0,
 271      "quant-plus",
 272     0,
 273     0,
 274      "quant-opt",
 275     0,
 276     0,
 277      "interval-open",
 278     0,
 279     0,
 280      "interval-lower",
 281     0,
 282     0,
 283     0,
 284      "interval-upper",
 285     0,
 286     0,
 287      "interval-type",
 288     0,
 289     0,
 290      "backslash",
 291     0,
 292     0,
 293     0,
 294     0,
 295     0,
 296     0,
 297     0,
 298     0,
 299     0,
 300     0,
 301     0,
 302     0,
 303     0,
 304     0,
 305     0,
 306     0,
 307     0,
 308     0,
 309     0,
 310      "errorDeath",
 311     0};
 312
 313 U_NAMESPACE_END
 314 #endif