regc_locale.c

来自「A*算法 A*算法 A*算法 A*算法A*算法A*算法」· C语言 代码 · 共 1,179 行 · 第 1/4 页

C
1,179
字号
/* 
 * regc_locale.c --
 *
 *	This file contains the Unicode locale specific regexp routines.
 *	This file is #included by regcomp.c.
 *
 * Copyright (c) 1998 by Scriptics Corporation.
 *
 * See the file "license.terms" for information on usage and redistribution
 * of this file, and for a DISCLAIMER OF ALL WARRANTIES.
 *
 * RCS: @(#) $Id: regc_locale.c,v 1.7 2004/05/23 14:56:16 JS Exp $
 *
 * wxWidgets:
 *  The Scriptics license can be found in the file COPYRIGHT. Modifications
 *  for wxWidgets are under the wxWidgets licence, see README for details.
 */

#define CONST const

/* ASCII character-name table */

static struct cname {
    chr *name;
    char code;
} cnames[] = {
    {_T("NUL"),                 '\0'},
    {_T("SOH"),                 '\001'},
    {_T("STX"),                 '\002'},
    {_T("ETX"),                 '\003'},
    {_T("EOT"),                 '\004'},
    {_T("ENQ"),                 '\005'},
    {_T("ACK"),                 '\006'},
    {_T("BEL"),                 '\007'},
    {_T("alert"),               '\007'},
    {_T("BS"),                  '\010'},
    {_T("backspace"),           '\b'},
    {_T("HT"),                  '\011'},
    {_T("tab"),                 '\t'},
    {_T("LF"),                  '\012'},
    {_T("newline"),             '\n'},
    {_T("VT"),                  '\013'},
    {_T("vertical-tab"),        '\v'},
    {_T("FF"),                  '\014'},
    {_T("form-feed"),           '\f'},
    {_T("CR"),                  '\015'},
    {_T("carriage-return"),     '\r'},
    {_T("SO"),                  '\016'},
    {_T("SI"),                  '\017'},
    {_T("DLE"),                 '\020'},
    {_T("DC1"),                 '\021'},
    {_T("DC2"),                 '\022'},
    {_T("DC3"),                 '\023'},
    {_T("DC4"),                 '\024'},
    {_T("NAK"),                 '\025'},
    {_T("SYN"),                 '\026'},
    {_T("ETB"),                 '\027'},
    {_T("CAN"),                 '\030'},
    {_T("EM"),                  '\031'},
    {_T("SUB"),                 '\032'},
    {_T("ESC"),                 '\033'},
    {_T("IS4"),                 '\034'},
    {_T("FS"),                  '\034'},
    {_T("IS3"),                 '\035'},
    {_T("GS"),                  '\035'},
    {_T("IS2"),                 '\036'},
    {_T("RS"),                  '\036'},
    {_T("IS1"),                 '\037'},
    {_T("US"),                  '\037'},
    {_T("space"),               ' '},
    {_T("exclamation-mark"),    '!'},
    {_T("quotation-mark"),      '"'},
    {_T("number-sign"),         '#'},
    {_T("dollar-sign"),         '$'},
    {_T("percent-sign"),        '%'},
    {_T("ampersand"),           '&'},
    {_T("apostrophe"),          '\''},
    {_T("left-parenthesis"),    '('},
    {_T("right-parenthesis"),   ')'},
    {_T("asterisk"),            '*'},
    {_T("plus-sign"),           '+'},
    {_T("comma"),               ','},
    {_T("hyphen"),              '-'},
    {_T("hyphen-minus"),        '-'},
    {_T("period"),              '.'},
    {_T("full-stop"),           '.'},
    {_T("slash"),               '/'},
    {_T("solidus"),             '/'},
    {_T("zero"),                '0'},
    {_T("one"),                 '1'},
    {_T("two"),                 '2'},
    {_T("three"),               '3'},
    {_T("four"),                '4'},
    {_T("five"),                '5'},
    {_T("six"),                 '6'},
    {_T("seven"),               '7'},
    {_T("eight"),               '8'},
    {_T("nine"),                '9'},
    {_T("colon"),               ':'},
    {_T("semicolon"),           ';'},
    {_T("less-than-sign"),      '<'},
    {_T("equals-sign"),         '='},
    {_T("greater-than-sign"),   '>'},
    {_T("question-mark"),       '?'},
    {_T("commercial-at"),       '@'},
    {_T("left-square-bracket"), '['},
    {_T("backslash"),           '\\'},
    {_T("reverse-solidus"),     '\\'},
    {_T("right-square-bracket"), ']'},
    {_T("circumflex"),          '^'},
    {_T("circumflex-accent"),   '^'},
    {_T("underscore"),          '_'},
    {_T("low-line"),            '_'},
    {_T("grave-accent"),        '`'},
    {_T("left-brace"),          '{'},
    {_T("left-curly-bracket"),  '{'},
    {_T("vertical-line"),       '|'},
    {_T("right-brace"),         '}'},
    {_T("right-curly-bracket"), '}'},
    {_T("tilde"),               '~'},
    {_T("DEL"),                 '\177'},
    {NULL,                      0}
};

#if wxUSE_UNICODE

/* Unicode character-class tables */

typedef struct crange {
    chr start;
    chr end;
} crange;

/*
 *	Declarations of Unicode character ranges.  This code
 *	is automatically generated by the tools/uniClass.tcl script
 *	and used in generic/regc_locale.c.  Do not modify by hand.
 */

/* Unicode: alphabetic characters */

static crange alphaRangeTable[] = {
    {0x0041, 0x005a}, {0x0061, 0x007a}, {0x00c0, 0x00d6}, {0x00d8, 0x00f6}, 
    {0x00f8, 0x021f}, {0x0222, 0x0233}, {0x0250, 0x02ad}, {0x02b0, 0x02b8}, 
    {0x02bb, 0x02c1}, {0x02e0, 0x02e4}, {0x0388, 0x038a}, {0x038e, 0x03a1}, 
    {0x03a3, 0x03ce}, {0x03d0, 0x03d7}, {0x03da, 0x03f5}, {0x0400, 0x0481}, 
    {0x048c, 0x04c4}, {0x04d0, 0x04f5}, {0x0531, 0x0556}, {0x0561, 0x0587}, 
    {0x05d0, 0x05ea}, {0x05f0, 0x05f2}, {0x0621, 0x063a}, {0x0640, 0x064a}, 
    {0x0671, 0x06d3}, {0x06fa, 0x06fc}, {0x0712, 0x072c}, {0x0780, 0x07a5}, 
    {0x0905, 0x0939}, {0x0958, 0x0961}, {0x0985, 0x098c}, {0x0993, 0x09a8}, 
    {0x09aa, 0x09b0}, {0x09b6, 0x09b9}, {0x09df, 0x09e1}, {0x0a05, 0x0a0a}, 
    {0x0a13, 0x0a28}, {0x0a2a, 0x0a30}, {0x0a59, 0x0a5c}, {0x0a72, 0x0a74}, 
    {0x0a85, 0x0a8b}, {0x0a8f, 0x0a91}, {0x0a93, 0x0aa8}, {0x0aaa, 0x0ab0}, 
    {0x0ab5, 0x0ab9}, {0x0b05, 0x0b0c}, {0x0b13, 0x0b28}, {0x0b2a, 0x0b30}, 
    {0x0b36, 0x0b39}, {0x0b5f, 0x0b61}, {0x0b85, 0x0b8a}, {0x0b8e, 0x0b90}, 
    {0x0b92, 0x0b95}, {0x0ba8, 0x0baa}, {0x0bae, 0x0bb5}, {0x0bb7, 0x0bb9}, 
    {0x0c05, 0x0c0c}, {0x0c0e, 0x0c10}, {0x0c12, 0x0c28}, {0x0c2a, 0x0c33}, 
    {0x0c35, 0x0c39}, {0x0c85, 0x0c8c}, {0x0c8e, 0x0c90}, {0x0c92, 0x0ca8}, 
    {0x0caa, 0x0cb3}, {0x0cb5, 0x0cb9}, {0x0d05, 0x0d0c}, {0x0d0e, 0x0d10}, 
    {0x0d12, 0x0d28}, {0x0d2a, 0x0d39}, {0x0d85, 0x0d96}, {0x0d9a, 0x0db1}, 
    {0x0db3, 0x0dbb}, {0x0dc0, 0x0dc6}, {0x0e01, 0x0e30}, {0x0e40, 0x0e46}, 
    {0x0e94, 0x0e97}, {0x0e99, 0x0e9f}, {0x0ea1, 0x0ea3}, {0x0ead, 0x0eb0}, 
    {0x0ec0, 0x0ec4}, {0x0f40, 0x0f47}, {0x0f49, 0x0f6a}, {0x0f88, 0x0f8b}, 
    {0x1000, 0x1021}, {0x1023, 0x1027}, {0x1050, 0x1055}, {0x10a0, 0x10c5}, 
    {0x10d0, 0x10f6}, {0x1100, 0x1159}, {0x115f, 0x11a2}, {0x11a8, 0x11f9}, 
    {0x1200, 0x1206}, {0x1208, 0x1246}, {0x124a, 0x124d}, {0x1250, 0x1256}, 
    {0x125a, 0x125d}, {0x1260, 0x1286}, {0x128a, 0x128d}, {0x1290, 0x12ae}, 
    {0x12b2, 0x12b5}, {0x12b8, 0x12be}, {0x12c2, 0x12c5}, {0x12c8, 0x12ce}, 
    {0x12d0, 0x12d6}, {0x12d8, 0x12ee}, {0x12f0, 0x130e}, {0x1312, 0x1315}, 
    {0x1318, 0x131e}, {0x1320, 0x1346}, {0x1348, 0x135a}, {0x13a0, 0x13f4}, 
    {0x1401, 0x166c}, {0x166f, 0x1676}, {0x1681, 0x169a}, {0x16a0, 0x16ea}, 
    {0x1780, 0x17b3}, {0x1820, 0x1877}, {0x1880, 0x18a8}, {0x1e00, 0x1e9b}, 
    {0x1ea0, 0x1ef9}, {0x1f00, 0x1f15}, {0x1f18, 0x1f1d}, {0x1f20, 0x1f45}, 
    {0x1f48, 0x1f4d}, {0x1f50, 0x1f57}, {0x1f5f, 0x1f7d}, {0x1f80, 0x1fb4}, 
    {0x1fb6, 0x1fbc}, {0x1fc2, 0x1fc4}, {0x1fc6, 0x1fcc}, {0x1fd0, 0x1fd3}, 
    {0x1fd6, 0x1fdb}, {0x1fe0, 0x1fec}, {0x1ff2, 0x1ff4}, {0x1ff6, 0x1ffc}, 
    {0x210a, 0x2113}, {0x2119, 0x211d}, {0x212a, 0x212d}, {0x212f, 0x2131}, 
    {0x2133, 0x2139}, {0x3031, 0x3035}, {0x3041, 0x3094}, {0x30a1, 0x30fa}, 
    {0x30fc, 0x30fe}, {0x3105, 0x312c}, {0x3131, 0x318e}, {0x31a0, 0x31b7}, 
    {0x3400, 0x4db5}, {0x4e00, 0x9fa5}, {0xa000, 0xa48c}, {0xac00, 0xd7a3}, 
    {0xf900, 0xfa2d}, {0xfb00, 0xfb06}, {0xfb13, 0xfb17}, {0xfb1f, 0xfb28}, 
    {0xfb2a, 0xfb36}, {0xfb38, 0xfb3c}, {0xfb46, 0xfbb1}, {0xfbd3, 0xfd3d}, 
    {0xfd50, 0xfd8f}, {0xfd92, 0xfdc7}, {0xfdf0, 0xfdfb}, {0xfe70, 0xfe72}, 
    {0xfe76, 0xfefc}, {0xff21, 0xff3a}, {0xff41, 0xff5a}, {0xff66, 0xffbe}, 
    {0xffc2, 0xffc7}, {0xffca, 0xffcf}, {0xffd2, 0xffd7}, {0xffda, 0xffdc}
};

#define NUM_ALPHA_RANGE (sizeof(alphaRangeTable)/sizeof(crange))

static chr alphaCharTable[] = {
    0x00aa, 0x00b5, 0x00ba, 0x02d0, 0x02d1, 0x02ee, 0x037a, 0x0386, 0x038c, 
    0x04c7, 0x04c8, 0x04cb, 0x04cc, 0x04f8, 0x04f9, 0x0559, 0x06d5, 0x06e5, 
    0x06e6, 0x0710, 0x093d, 0x0950, 0x098f, 0x0990, 0x09b2, 0x09dc, 0x09dd, 
    0x09f0, 0x09f1, 0x0a0f, 0x0a10, 0x0a32, 0x0a33, 0x0a35, 0x0a36, 0x0a38, 
    0x0a39, 0x0a5e, 0x0a8d, 0x0ab2, 0x0ab3, 0x0abd, 0x0ad0, 0x0ae0, 0x0b0f, 
    0x0b10, 0x0b32, 0x0b33, 0x0b3d, 0x0b5c, 0x0b5d, 0x0b99, 0x0b9a, 0x0b9c, 
    0x0b9e, 0x0b9f, 0x0ba3, 0x0ba4, 0x0c60, 0x0c61, 0x0cde, 0x0ce0, 0x0ce1, 
    0x0d60, 0x0d61, 0x0dbd, 0x0e32, 0x0e33, 0x0e81, 0x0e82, 0x0e84, 0x0e87, 
    0x0e88, 0x0e8a, 0x0e8d, 0x0ea5, 0x0ea7, 0x0eaa, 0x0eab, 0x0eb2, 0x0eb3, 
    0x0ebd, 0x0ec6, 0x0edc, 0x0edd, 0x0f00, 0x1029, 0x102a, 0x1248, 0x1258, 
    0x1288, 0x12b0, 0x12c0, 0x1310, 0x1f59, 0x1f5b, 0x1f5d, 0x1fbe, 0x207f, 
    0x2102, 0x2107, 0x2115, 0x2124, 0x2126, 0x2128, 0x3005, 0x3006, 0x309d, 
    0x309e, 0xfb1d, 0xfb3e, 0xfb40, 0xfb41, 0xfb43, 0xfb44, 0xfe74, 0xfffe
};

#define NUM_ALPHA_CHAR (sizeof(alphaCharTable)/sizeof(chr))

/* Unicode: decimal digit characters */

static crange digitRangeTable[] = {
    {0x0030, 0x0039}, {0x0660, 0x0669}, {0x06f0, 0x06f9}, {0x0966, 0x096f}, 
    {0x09e6, 0x09ef}, {0x0a66, 0x0a6f}, {0x0ae6, 0x0aef}, {0x0b66, 0x0b6f}, 
    {0x0be7, 0x0bef}, {0x0c66, 0x0c6f}, {0x0ce6, 0x0cef}, {0x0d66, 0x0d6f}, 
    {0x0e50, 0x0e59}, {0x0ed0, 0x0ed9}, {0x0f20, 0x0f29}, {0x1040, 0x1049}, 
    {0x1369, 0x1371}, {0x17e0, 0x17e9}, {0x1810, 0x1819}, {0xff10, 0xff19}
};

#define NUM_DIGIT_RANGE (sizeof(digitRangeTable)/sizeof(crange))

/* no singletons of digit characters */

/* Unicode: punctuation characters */

static crange punctRangeTable[] = {
    {0x0021, 0x0023}, {0x0025, 0x002a}, {0x002c, 0x002f}, {0x005b, 0x005d}, 
    {0x055a, 0x055f}, {0x066a, 0x066d}, {0x0700, 0x070d}, {0x0f04, 0x0f12}, 
    {0x0f3a, 0x0f3d}, {0x104a, 0x104f}, {0x1361, 0x1368}, {0x16eb, 0x16ed}, 
    {0x17d4, 0x17da}, {0x1800, 0x180a}, {0x2010, 0x2027}, {0x2030, 0x2043}, 
    {0x2048, 0x204d}, {0x3001, 0x3003}, {0x3008, 0x3011}, {0x3014, 0x301f}, 
    {0xfe30, 0xfe44}, {0xfe49, 0xfe52}, {0xfe54, 0xfe61}, {0xff01, 0xff03}, 
    {0xff05, 0xff0a}, {0xff0c, 0xff0f}, {0xff3b, 0xff3d}, {0xff61, 0xff65}
};

#define NUM_PUNCT_RANGE (sizeof(punctRangeTable)/sizeof(crange))

static chr punctCharTable[] = {
    0x003a, 0x003b, 0x003f, 0x0040, 0x005f, 0x007b, 0x007d, 0x00a1, 0x00ab, 
    0x00ad, 0x00b7, 0x00bb, 0x00bf, 0x037e, 0x0387, 0x0589, 0x058a, 0x05be, 
    0x05c0, 0x05c3, 0x05f3, 0x05f4, 0x060c, 0x061b, 0x061f, 0x06d4, 0x0964, 
    0x0965, 0x0970, 0x0df4, 0x0e4f, 0x0e5a, 0x0e5b, 0x0f85, 0x10fb, 0x166d, 
    0x166e, 0x169b, 0x169c, 0x17dc, 0x2045, 0x2046, 0x207d, 0x207e, 0x208d, 
    0x208e, 0x2329, 0x232a, 0x3030, 0x30fb, 0xfd3e, 0xfd3f, 0xfe63, 0xfe68, 
    0xfe6a, 0xfe6b, 0xff1a, 0xff1b, 0xff1f, 0xff20, 0xff3f, 0xff5b, 0xff5d
};

#define NUM_PUNCT_CHAR (sizeof(punctCharTable)/sizeof(chr))

/* Unicode: white space characters */

static crange spaceRangeTable[] = {
    {0x0009, 0x000d}, {0x2000, 0x200b}
};

#define NUM_SPACE_RANGE (sizeof(spaceRangeTable)/sizeof(crange))

static chr spaceCharTable[] = {
    0x0020, 0x00a0, 0x1680, 0x2028, 0x2029, 0x202f, 0x3000
};

#define NUM_SPACE_CHAR (sizeof(spaceCharTable)/sizeof(chr))

/* Unicode: lowercase characters */

static crange lowerRangeTable[] = {
    {0x0061, 0x007a}, {0x00df, 0x00f6}, {0x00f8, 0x00ff}, {0x017e, 0x0180}, 
    {0x0199, 0x019b}, {0x01bd, 0x01bf}, {0x0250, 0x02ad}, {0x03ac, 0x03ce}, 
    {0x03d5, 0x03d7}, {0x03ef, 0x03f3}, {0x0430, 0x045f}, {0x0561, 0x0587}, 
    {0x1e95, 0x1e9b}, {0x1f00, 0x1f07}, {0x1f10, 0x1f15}, {0x1f20, 0x1f27}, 
    {0x1f30, 0x1f37}, {0x1f40, 0x1f45}, {0x1f50, 0x1f57}, {0x1f60, 0x1f67}, 
    {0x1f70, 0x1f7d}, {0x1f80, 0x1f87}, {0x1f90, 0x1f97}, {0x1fa0, 0x1fa7}, 
    {0x1fb0, 0x1fb4}, {0x1fc2, 0x1fc4}, {0x1fd0, 0x1fd3}, {0x1fe0, 0x1fe7}, 
    {0x1ff2, 0x1ff4}, {0xfb00, 0xfb06}, {0xfb13, 0xfb17}, {0xff41, 0xff5a}
};

#define NUM_LOWER_RANGE (sizeof(lowerRangeTable)/sizeof(crange))

static chr lowerCharTable[] = {
    0x00aa, 0x00b5, 0x00ba, 0x0101, 0x0103, 0x0105, 0x0107, 0x0109, 0x010b, 
    0x010d, 0x010f, 0x0111, 0x0113, 0x0115, 0x0117, 0x0119, 0x011b, 0x011d, 
    0x011f, 0x0121, 0x0123, 0x0125, 0x0127, 0x0129, 0x012b, 0x012d, 0x012f, 
    0x0131, 0x0133, 0x0135, 0x0137, 0x0138, 0x013a, 0x013c, 0x013e, 0x0140, 
    0x0142, 0x0144, 0x0146, 0x0148, 0x0149, 0x014b, 0x014d, 0x014f, 0x0151, 
    0x0153, 0x0155, 0x0157, 0x0159, 0x015b, 0x015d, 0x015f, 0x0161, 0x0163, 
    0x0165, 0x0167, 0x0169, 0x016b, 0x016d, 0x016f, 0x0171, 0x0173, 0x0175, 
    0x0177, 0x017a, 0x017c, 0x0183, 0x0185, 0x0188, 0x018c, 0x018d, 0x0192, 
    0x0195, 0x019e, 0x01a1, 0x01a3, 0x01a5, 0x01a8, 0x01aa, 0x01ab, 0x01ad, 
    0x01b0, 0x01b4, 0x01b6, 0x01b9, 0x01ba, 0x01c6, 0x01c9, 0x01cc, 0x01ce, 
    0x01d0, 0x01d2, 0x01d4, 0x01d6, 0x01d8, 0x01da, 0x01dc, 0x01dd, 0x01df, 
    0x01e1, 0x01e3, 0x01e5, 0x01e7, 0x01e9, 0x01eb, 0x01ed, 0x01ef, 0x01f0, 
    0x01f3, 0x01f5, 0x01f9, 0x01fb, 0x01fd, 0x01ff, 0x0201, 0x0203, 0x0205, 
    0x0207, 0x0209, 0x020b, 0x020d, 0x020f, 0x0211, 0x0213, 0x0215, 0x0217, 
    0x0219, 0x021b, 0x021d, 0x021f, 0x0223, 0x0225, 0x0227, 0x0229, 0x022b, 
    0x022d, 0x022f, 0x0231, 0x0233, 0x0390, 0x03d0, 0x03d1, 0x03db, 0x03dd, 
    0x03df, 0x03e1, 0x03e3, 0x03e5, 0x03e7, 0x03e9, 0x03eb, 0x03ed, 0x03f5, 
    0x0461, 0x0463, 0x0465, 0x0467, 0x0469, 0x046b, 0x046d, 0x046f, 0x0471, 

⌨️ 快捷键说明

复制代码Ctrl + C
搜索代码Ctrl + F
全屏模式F11
增大字号Ctrl + =
减小字号Ctrl + -
显示快捷键?