summaryrefslogtreecommitdiffstats
path: root/generic/regc_locale.c
diff options
context:
space:
mode:
Diffstat (limited to 'generic/regc_locale.c')
-rw-r--r--generic/regc_locale.c1610
1 files changed, 1011 insertions, 599 deletions
diff --git a/generic/regc_locale.c b/generic/regc_locale.c
index 98ea392..0f8d1b2 100644
--- a/generic/regc_locale.c
+++ b/generic/regc_locale.c
@@ -1,4 +1,4 @@
-/*
+/*
* regc_locale.c --
*
* This file contains the Unicode locale specific regexp routines.
@@ -6,564 +6,913 @@
*
* 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 1.3 1999/06/02 01:53:29 stanton Exp $
+ * See the file "license.terms" for information on usage and redistribution of
+ * this file, and for a DISCLAIMER OF ALL WARRANTIES.
*/
-
+
/* ASCII character-name table */
-static struct cname {
- char *name;
- char code;
+static const struct cname {
+ const char *name;
+ const char code;
} cnames[] = {
- {"NUL", '\0'},
- {"SOH", '\001'},
- {"STX", '\002'},
- {"ETX", '\003'},
- {"EOT", '\004'},
- {"ENQ", '\005'},
- {"ACK", '\006'},
- {"BEL", '\007'},
- {"alert", '\007'},
- {"BS", '\010'},
- {"backspace", '\b'},
- {"HT", '\011'},
- {"tab", '\t'},
- {"LF", '\012'},
- {"newline", '\n'},
- {"VT", '\013'},
- {"vertical-tab", '\v'},
- {"FF", '\014'},
- {"form-feed", '\f'},
- {"CR", '\015'},
- {"carriage-return", '\r'},
- {"SO", '\016'},
- {"SI", '\017'},
- {"DLE", '\020'},
- {"DC1", '\021'},
- {"DC2", '\022'},
- {"DC3", '\023'},
- {"DC4", '\024'},
- {"NAK", '\025'},
- {"SYN", '\026'},
- {"ETB", '\027'},
- {"CAN", '\030'},
- {"EM", '\031'},
- {"SUB", '\032'},
- {"ESC", '\033'},
- {"IS4", '\034'},
- {"FS", '\034'},
- {"IS3", '\035'},
- {"GS", '\035'},
- {"IS2", '\036'},
- {"RS", '\036'},
- {"IS1", '\037'},
- {"US", '\037'},
- {"space", ' '},
- {"exclamation-mark", '!'},
- {"quotation-mark", '"'},
- {"number-sign", '#'},
- {"dollar-sign", '$'},
- {"percent-sign", '%'},
- {"ampersand", '&'},
- {"apostrophe", '\''},
- {"left-parenthesis", '('},
- {"right-parenthesis", ')'},
- {"asterisk", '*'},
- {"plus-sign", '+'},
- {"comma", ','},
- {"hyphen", '-'},
- {"hyphen-minus", '-'},
- {"period", '.'},
- {"full-stop", '.'},
- {"slash", '/'},
- {"solidus", '/'},
- {"zero", '0'},
- {"one", '1'},
- {"two", '2'},
- {"three", '3'},
- {"four", '4'},
- {"five", '5'},
- {"six", '6'},
- {"seven", '7'},
- {"eight", '8'},
- {"nine", '9'},
- {"colon", ':'},
- {"semicolon", ';'},
- {"less-than-sign", '<'},
- {"equals-sign", '='},
- {"greater-than-sign", '>'},
- {"question-mark", '?'},
- {"commercial-at", '@'},
- {"left-square-bracket", '['},
- {"backslash", '\\'},
- {"reverse-solidus", '\\'},
- {"right-square-bracket", ']'},
- {"circumflex", '^'},
- {"circumflex-accent", '^'},
- {"underscore", '_'},
- {"low-line", '_'},
- {"grave-accent", '`'},
- {"left-brace", '{'},
- {"left-curly-bracket", '{'},
- {"vertical-line", '|'},
- {"right-brace", '}'},
- {"right-curly-bracket", '}'},
- {"tilde", '~'},
- {"DEL", '\177'},
- {NULL, 0}
+ {"NUL", '\0'},
+ {"SOH", '\001'},
+ {"STX", '\002'},
+ {"ETX", '\003'},
+ {"EOT", '\004'},
+ {"ENQ", '\005'},
+ {"ACK", '\006'},
+ {"BEL", '\007'},
+ {"alert", '\007'},
+ {"BS", '\010'},
+ {"backspace", '\b'},
+ {"HT", '\011'},
+ {"tab", '\t'},
+ {"LF", '\012'},
+ {"newline", '\n'},
+ {"VT", '\013'},
+ {"vertical-tab", '\v'},
+ {"FF", '\014'},
+ {"form-feed", '\f'},
+ {"CR", '\015'},
+ {"carriage-return", '\r'},
+ {"SO", '\016'},
+ {"SI", '\017'},
+ {"DLE", '\020'},
+ {"DC1", '\021'},
+ {"DC2", '\022'},
+ {"DC3", '\023'},
+ {"DC4", '\024'},
+ {"NAK", '\025'},
+ {"SYN", '\026'},
+ {"ETB", '\027'},
+ {"CAN", '\030'},
+ {"EM", '\031'},
+ {"SUB", '\032'},
+ {"ESC", '\033'},
+ {"IS4", '\034'},
+ {"FS", '\034'},
+ {"IS3", '\035'},
+ {"GS", '\035'},
+ {"IS2", '\036'},
+ {"RS", '\036'},
+ {"IS1", '\037'},
+ {"US", '\037'},
+ {"space", ' '},
+ {"exclamation-mark",'!'},
+ {"quotation-mark", '"'},
+ {"number-sign", '#'},
+ {"dollar-sign", '$'},
+ {"percent-sign", '%'},
+ {"ampersand", '&'},
+ {"apostrophe", '\''},
+ {"left-parenthesis",'('},
+ {"right-parenthesis", ')'},
+ {"asterisk", '*'},
+ {"plus-sign", '+'},
+ {"comma", ','},
+ {"hyphen", '-'},
+ {"hyphen-minus", '-'},
+ {"period", '.'},
+ {"full-stop", '.'},
+ {"slash", '/'},
+ {"solidus", '/'},
+ {"zero", '0'},
+ {"one", '1'},
+ {"two", '2'},
+ {"three", '3'},
+ {"four", '4'},
+ {"five", '5'},
+ {"six", '6'},
+ {"seven", '7'},
+ {"eight", '8'},
+ {"nine", '9'},
+ {"colon", ':'},
+ {"semicolon", ';'},
+ {"less-than-sign", '<'},
+ {"equals-sign", '='},
+ {"greater-than-sign", '>'},
+ {"question-mark", '?'},
+ {"commercial-at", '@'},
+ {"left-square-bracket", '['},
+ {"backslash", '\\'},
+ {"reverse-solidus", '\\'},
+ {"right-square-bracket", ']'},
+ {"circumflex", '^'},
+ {"circumflex-accent", '^'},
+ {"underscore", '_'},
+ {"low-line", '_'},
+ {"grave-accent", '`'},
+ {"left-brace", '{'},
+ {"left-curly-bracket", '{'},
+ {"vertical-line", '|'},
+ {"right-brace", '}'},
+ {"right-curly-bracket", '}'},
+ {"tilde", '~'},
+ {"DEL", '\177'},
+ {NULL, 0}
};
+
+/*
+ * Unicode character-class tables.
+ */
-/* Unicode character-class tables */
-
-typedef struct crange {
+typedef struct {
chr start;
chr end;
} crange;
-static crange alphaTable[] = {
- {0X0041, 0X005A}, {0X0061, 0X007A}, {0X00AA, 0X00AA}, {0X00B5, 0X00B5},
- {0X00BA, 0X00BA}, {0X00C0, 0X00D6}, {0X00D8, 0X00F6}, {0X00F8, 0X01F5},
- {0X01FA, 0X0217}, {0X0250, 0X02A8}, {0X02B0, 0X02B8}, {0X02BB, 0X02C1},
- {0X02E0, 0X02E4}, {0X037A, 0X037A}, {0x0386, 0x0386}, {0X0388, 0X038A},
- {0X038C, 0X038C}, {0X038E, 0X03A1}, {0X03A3, 0X03CE}, {0X03D0, 0X03D6},
- {0X03DA, 0X03DA}, {0X03DC, 0X03DC}, {0X03DE, 0X03DE}, {0X03E0, 0X03E0},
- {0X03E2, 0X03F3}, {0X0401, 0X040C}, {0X040E, 0X044F}, {0X0451, 0X045C},
- {0X045E, 0X0481}, {0X0490, 0X04C4}, {0X04C7, 0X04C8}, {0X04CB, 0X04CC},
- {0X04D0, 0X04EB}, {0X04EE, 0X04F5}, {0X04F8, 0X04F9}, {0x0531, 0x0556},
- {0x0559, 0x0559}, {0x0561, 0x0587}, {0X05D0, 0X05EA}, {0X05F0, 0X05F2},
- {0X0621, 0X063A}, {0x0641, 0x0652}, {0X0670, 0X06B7}, {0X06BA, 0X06BE},
- {0X06C0, 0X06CE}, {0X06D0, 0X06D3}, {0X06D5, 0X06DC}, {0X06E1, 0X06E8},
- {0X06ED, 0X06ED}, {0x0901, 0x0903}, {0x0905, 0x0939}, {0X093D, 0X094C},
- {0x0958, 0x0963}, {0x0981, 0x0983}, {0X0985, 0X098C}, {0X098F, 0X0990},
- {0X0993, 0X09A8}, {0X09AA, 0X09B0}, {0X09B2, 0X09B2}, {0X09B6, 0X09B9},
- {0X09BE, 0X09C4}, {0X09C7, 0X09C8}, {0X09CB, 0X09CC}, {0X09D7, 0X09D7},
- {0X09DC, 0X09DD}, {0X09DF, 0X09E3}, {0X09F0, 0X09F1}, {0X0A02, 0X0A02},
- {0X0A05, 0X0A0A}, {0X0A0F, 0X0A10}, {0X0A13, 0X0A28}, {0X0A2A, 0X0A30},
- {0X0A32, 0X0A33}, {0X0A35, 0X0A36}, {0X0A38, 0X0A39}, {0X0A3E, 0X0A42},
- {0X0A47, 0X0A48}, {0X0A4B, 0X0A4C}, {0X0A59, 0X0A5C}, {0X0A5E, 0X0A5E},
- {0X0A70, 0X0A74}, {0X0A81, 0X0A83}, {0X0A85, 0X0A8B}, {0X0A8D, 0X0A8D},
- {0X0A8F, 0X0A91}, {0X0A93, 0X0AA8}, {0X0AAA, 0X0AB0}, {0X0AB2, 0X0AB3},
- {0X0AB5, 0X0AB9}, {0X0ABD, 0X0AC5}, {0X0AC7, 0X0AC9}, {0X0ACB, 0X0ACC},
- {0X0AE0, 0X0AE0}, {0X0B01, 0X0B03}, {0X0B05, 0X0B0C}, {0X0B0F, 0X0B10},
- {0X0B13, 0X0B28}, {0X0B2A, 0X0B30}, {0X0B32, 0X0B33}, {0X0B36, 0X0B39},
- {0X0B3D, 0X0B43}, {0X0B47, 0X0B48}, {0X0B4B, 0X0B4C}, {0X0B56, 0X0B57},
- {0X0B5C, 0X0B5D}, {0X0B5F, 0X0B61}, {0X0B82, 0X0B83}, {0X0B85, 0X0B8A},
- {0X0B8E, 0X0B90}, {0X0B92, 0X0B95}, {0X0B99, 0X0B9A}, {0X0B9C, 0X0B9C},
- {0X0B9E, 0X0B9F}, {0X0BA3, 0X0BA4}, {0X0BA8, 0X0BAA}, {0X0BAE, 0X0BB5},
- {0X0BB7, 0X0BB9}, {0X0BBE, 0X0BC2}, {0X0BC6, 0X0BC8}, {0X0BCA, 0X0BCC},
- {0X0BD7, 0X0BD7}, {0X0C01, 0X0C03}, {0X0C05, 0X0C0C}, {0X0C0E, 0X0C10},
- {0X0C12, 0X0C28}, {0X0C2A, 0X0C33}, {0X0C35, 0X0C39}, {0X0C3E, 0X0C44},
- {0X0C46, 0X0C48}, {0X0C4A, 0X0C4C}, {0X0C55, 0X0C56}, {0X0C60, 0X0C61},
- {0X0C82, 0X0C83}, {0X0C85, 0X0C8C}, {0X0C8E, 0X0C90}, {0X0C92, 0X0CA8},
- {0X0CAA, 0X0CB3}, {0X0CB5, 0X0CB9}, {0X0CBE, 0X0CC4}, {0X0CC6, 0X0CC8},
- {0X0CCA, 0X0CCC}, {0X0CD5, 0X0CD6}, {0X0CDE, 0X0CDE}, {0X0CE0, 0X0CE1},
- {0X0D02, 0X0D03}, {0X0D05, 0X0D0C}, {0X0D0E, 0X0D10}, {0X0D12, 0X0D28},
- {0X0D2A, 0X0D39}, {0X0D3E, 0X0D43}, {0X0D46, 0X0D48}, {0X0D4A, 0X0D4C},
- {0X0D57, 0X0D57}, {0X0D60, 0X0D61}, {0X0E01, 0X0E2E}, {0X0E30, 0X0E3A},
- {0X0E40, 0X0E45}, {0X0E47, 0X0E47}, {0X0E4D, 0X0E4D}, {0X0E81, 0X0E82},
- {0X0E84, 0X0E84}, {0X0E87, 0X0E88}, {0X0E8A, 0X0E8A}, {0X0E8D, 0X0E8D},
- {0X0E94, 0X0E97}, {0X0E99, 0X0E9F}, {0X0EA1, 0X0EA3}, {0X0EA5, 0X0EA5},
- {0X0EA7, 0X0EA7}, {0X0EAA, 0X0EAB}, {0X0EAD, 0X0EAE}, {0X0EB0, 0X0EB9},
- {0X0EBB, 0X0EBD}, {0X0EC0, 0X0EC4}, {0X0ECD, 0X0ECD}, {0X0EDC, 0X0EDD},
- {0X0F40, 0X0F47}, {0X0F49, 0X0F69}, {0X0F71, 0X0F81}, {0X0F90, 0X0F95},
- {0X0F97, 0X0F97}, {0X0F99, 0X0FAD}, {0X0FB1, 0X0FB7}, {0X0FB9, 0X0FB9},
- {0X10A0, 0X10C5}, {0X10D0, 0X10F6}, {0x1100, 0x1159}, {0X115F, 0X11A2},
- {0X11A8, 0X11F9}, {0X1E00, 0X1E9B}, {0X1EA0, 0X1EF9}, {0X1F00, 0X1F15},
- {0X1F18, 0X1F1D}, {0X1F20, 0X1F45}, {0X1F48, 0X1F4D}, {0X1F50, 0X1F57},
- {0X1F59, 0X1F59}, {0X1F5B, 0X1F5B}, {0X1F5D, 0X1F5D}, {0X1F5F, 0X1F7D},
- {0X1F80, 0X1FB4}, {0X1FB6, 0X1FBC}, {0X1FBE, 0X1FBE}, {0X1FC2, 0X1FC4},
- {0X1FC6, 0X1FCC}, {0X1FD0, 0X1FD3}, {0X1FD6, 0X1FDB}, {0X1FE0, 0X1FEC},
- {0X1FF2, 0X1FF4}, {0X1FF6, 0X1FFC}, {0X207F, 0X207F}, {0x2102, 0x2102},
- {0x2107, 0x2107}, {0X210A, 0X2113}, {0x2115, 0x2115}, {0X2118, 0X211D},
- {0x2124, 0x2124}, {0x2126, 0x2126}, {0x2128, 0x2128}, {0X212A, 0X2131},
- {0x2133, 0x2138}, {0x2160, 0x2182}, {0x3041, 0x3094}, {0X30A1, 0X30FA},
- {0X3105, 0X312C}, {0X3131, 0X318E}, {0XAC00, 0XD7A3}, {0XFB00, 0XFB06},
- {0XFB13, 0XFB17}, {0XFB1F, 0XFB28}, {0XFB2A, 0XFB36}, {0XFB38, 0XFB3C},
- {0XFB3E, 0XFB3E}, {0XFB40, 0XFB41}, {0XFB43, 0XFB44}, {0XFB46, 0XFBB1},
- {0XFBD3, 0XFD3D}, {0XFD50, 0XFD8F}, {0XFD92, 0XFDC7}, {0XFDF0, 0XFDFB},
- {0XFE70, 0XFE72}, {0XFE74, 0XFE74}, {0XFE76, 0XFEFC}, {0XFF21, 0XFF3A},
- {0XFF41, 0XFF5A}, {0XFF66, 0XFF6F}, {0XFF71, 0XFF9D}, {0XFFA0, 0XFFBE},
- {0XFFC2, 0XFFC7}, {0XFFCA, 0XFFCF}, {0XFFD2, 0XFFD7}, {0XFFDA, 0XFFDC}
+/*
+ * 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 const crange alphaRangeTable[] = {
+ {0x41, 0x5a}, {0x61, 0x7a}, {0xc0, 0xd6}, {0xd8, 0xf6},
+ {0xf8, 0x2c1}, {0x2c6, 0x2d1}, {0x2e0, 0x2e4}, {0x370, 0x374},
+ {0x37a, 0x37d}, {0x388, 0x38a}, {0x38e, 0x3a1}, {0x3a3, 0x3f5},
+ {0x3f7, 0x481}, {0x48a, 0x527}, {0x531, 0x556}, {0x561, 0x587},
+ {0x5d0, 0x5ea}, {0x5f0, 0x5f2}, {0x620, 0x64a}, {0x671, 0x6d3},
+ {0x6fa, 0x6fc}, {0x712, 0x72f}, {0x74d, 0x7a5}, {0x7ca, 0x7ea},
+ {0x800, 0x815}, {0x840, 0x858}, {0x8a2, 0x8ac}, {0x904, 0x939},
+ {0x958, 0x961}, {0x971, 0x977}, {0x979, 0x97f}, {0x985, 0x98c},
+ {0x993, 0x9a8}, {0x9aa, 0x9b0}, {0x9b6, 0x9b9}, {0x9df, 0x9e1},
+ {0xa05, 0xa0a}, {0xa13, 0xa28}, {0xa2a, 0xa30}, {0xa59, 0xa5c},
+ {0xa72, 0xa74}, {0xa85, 0xa8d}, {0xa8f, 0xa91}, {0xa93, 0xaa8},
+ {0xaaa, 0xab0}, {0xab5, 0xab9}, {0xb05, 0xb0c}, {0xb13, 0xb28},
+ {0xb2a, 0xb30}, {0xb35, 0xb39}, {0xb5f, 0xb61}, {0xb85, 0xb8a},
+ {0xb8e, 0xb90}, {0xb92, 0xb95}, {0xba8, 0xbaa}, {0xbae, 0xbb9},
+ {0xc05, 0xc0c}, {0xc0e, 0xc10}, {0xc12, 0xc28}, {0xc2a, 0xc33},
+ {0xc35, 0xc39}, {0xc85, 0xc8c}, {0xc8e, 0xc90}, {0xc92, 0xca8},
+ {0xcaa, 0xcb3}, {0xcb5, 0xcb9}, {0xd05, 0xd0c}, {0xd0e, 0xd10},
+ {0xd12, 0xd3a}, {0xd7a, 0xd7f}, {0xd85, 0xd96}, {0xd9a, 0xdb1},
+ {0xdb3, 0xdbb}, {0xdc0, 0xdc6}, {0xe01, 0xe30}, {0xe40, 0xe46},
+ {0xe94, 0xe97}, {0xe99, 0xe9f}, {0xea1, 0xea3}, {0xead, 0xeb0},
+ {0xec0, 0xec4}, {0xedc, 0xedf}, {0xf40, 0xf47}, {0xf49, 0xf6c},
+ {0xf88, 0xf8c}, {0x1000, 0x102a}, {0x1050, 0x1055}, {0x105a, 0x105d},
+ {0x106e, 0x1070}, {0x1075, 0x1081}, {0x10a0, 0x10c5}, {0x10d0, 0x10fa},
+ {0x10fc, 0x1248}, {0x124a, 0x124d}, {0x1250, 0x1256}, {0x125a, 0x125d},
+ {0x1260, 0x1288}, {0x128a, 0x128d}, {0x1290, 0x12b0}, {0x12b2, 0x12b5},
+ {0x12b8, 0x12be}, {0x12c2, 0x12c5}, {0x12c8, 0x12d6}, {0x12d8, 0x1310},
+ {0x1312, 0x1315}, {0x1318, 0x135a}, {0x1380, 0x138f}, {0x13a0, 0x13f4},
+ {0x1401, 0x166c}, {0x166f, 0x167f}, {0x1681, 0x169a}, {0x16a0, 0x16ea},
+ {0x1700, 0x170c}, {0x170e, 0x1711}, {0x1720, 0x1731}, {0x1740, 0x1751},
+ {0x1760, 0x176c}, {0x176e, 0x1770}, {0x1780, 0x17b3}, {0x1820, 0x1877},
+ {0x1880, 0x18a8}, {0x18b0, 0x18f5}, {0x1900, 0x191c}, {0x1950, 0x196d},
+ {0x1970, 0x1974}, {0x1980, 0x19ab}, {0x19c1, 0x19c7}, {0x1a00, 0x1a16},
+ {0x1a20, 0x1a54}, {0x1b05, 0x1b33}, {0x1b45, 0x1b4b}, {0x1b83, 0x1ba0},
+ {0x1bba, 0x1be5}, {0x1c00, 0x1c23}, {0x1c4d, 0x1c4f}, {0x1c5a, 0x1c7d},
+ {0x1ce9, 0x1cec}, {0x1cee, 0x1cf1}, {0x1d00, 0x1dbf}, {0x1e00, 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}, {0x2090, 0x209c}, {0x210a, 0x2113},
+ {0x2119, 0x211d}, {0x212a, 0x212d}, {0x212f, 0x2139}, {0x213c, 0x213f},
+ {0x2145, 0x2149}, {0x2c00, 0x2c2e}, {0x2c30, 0x2c5e}, {0x2c60, 0x2ce4},
+ {0x2ceb, 0x2cee}, {0x2d00, 0x2d25}, {0x2d30, 0x2d67}, {0x2d80, 0x2d96},
+ {0x2da0, 0x2da6}, {0x2da8, 0x2dae}, {0x2db0, 0x2db6}, {0x2db8, 0x2dbe},
+ {0x2dc0, 0x2dc6}, {0x2dc8, 0x2dce}, {0x2dd0, 0x2dd6}, {0x2dd8, 0x2dde},
+ {0x3031, 0x3035}, {0x3041, 0x3096}, {0x309d, 0x309f}, {0x30a1, 0x30fa},
+ {0x30fc, 0x30ff}, {0x3105, 0x312d}, {0x3131, 0x318e}, {0x31a0, 0x31ba},
+ {0x31f0, 0x31ff}, {0x3400, 0x4db5}, {0x4e00, 0x9fcc}, {0xa000, 0xa48c},
+ {0xa4d0, 0xa4fd}, {0xa500, 0xa60c}, {0xa610, 0xa61f}, {0xa640, 0xa66e},
+ {0xa67f, 0xa697}, {0xa6a0, 0xa6e5}, {0xa717, 0xa71f}, {0xa722, 0xa788},
+ {0xa78b, 0xa78e}, {0xa790, 0xa793}, {0xa7a0, 0xa7aa}, {0xa7f8, 0xa801},
+ {0xa803, 0xa805}, {0xa807, 0xa80a}, {0xa80c, 0xa822}, {0xa840, 0xa873},
+ {0xa882, 0xa8b3}, {0xa8f2, 0xa8f7}, {0xa90a, 0xa925}, {0xa930, 0xa946},
+ {0xa960, 0xa97c}, {0xa984, 0xa9b2}, {0xaa00, 0xaa28}, {0xaa40, 0xaa42},
+ {0xaa44, 0xaa4b}, {0xaa60, 0xaa76}, {0xaa80, 0xaaaf}, {0xaab9, 0xaabd},
+ {0xaadb, 0xaadd}, {0xaae0, 0xaaea}, {0xaaf2, 0xaaf4}, {0xab01, 0xab06},
+ {0xab09, 0xab0e}, {0xab11, 0xab16}, {0xab20, 0xab26}, {0xab28, 0xab2e},
+ {0xabc0, 0xabe2}, {0xac00, 0xd7a3}, {0xd7b0, 0xd7c6}, {0xd7cb, 0xd7fb},
+ {0xf900, 0xfa6d}, {0xfa70, 0xfad9}, {0xfb00, 0xfb06}, {0xfb13, 0xfb17},
+ {0xfb1f, 0xfb28}, {0xfb2a, 0xfb36}, {0xfb38, 0xfb3c}, {0xfb46, 0xfbb1},
+ {0xfbd3, 0xfd3d}, {0xfd50, 0xfd8f}, {0xfd92, 0xfdc7}, {0xfdf0, 0xfdfb},
+ {0xfe70, 0xfe74}, {0xfe76, 0xfefc}, {0xff21, 0xff3a}, {0xff41, 0xff5a},
+ {0xff66, 0xffbe}, {0xffc2, 0xffc7}, {0xffca, 0xffcf}, {0xffd2, 0xffd7},
+ {0xffda, 0xffdc}
+#if TCL_UTF_MAX > 4
+ ,{0x10000, 0x1000b}, {0x1000d, 0x10026}, {0x10028, 0x1003a}, {0x1003f, 0x1004d},
+ {0x10050, 0x1005d}, {0x10080, 0x100fa}, {0x10280, 0x1029c}, {0x102a0, 0x102d0},
+ {0x10300, 0x1031e}, {0x10330, 0x10340}, {0x10342, 0x10349}, {0x10380, 0x1039d},
+ {0x103a0, 0x103c3}, {0x103c8, 0x103cf}, {0x10400, 0x1049d}, {0x10800, 0x10805},
+ {0x1080a, 0x10835}, {0x1083f, 0x10855}, {0x10900, 0x10915}, {0x10920, 0x10939},
+ {0x10980, 0x109b7}, {0x10a10, 0x10a13}, {0x10a15, 0x10a17}, {0x10a19, 0x10a33},
+ {0x10a60, 0x10a7c}, {0x10b00, 0x10b35}, {0x10b40, 0x10b55}, {0x10b60, 0x10b72},
+ {0x10c00, 0x10c48}, {0x11003, 0x11037}, {0x11083, 0x110af}, {0x110d0, 0x110e8},
+ {0x11103, 0x11126}, {0x11183, 0x111b2}, {0x111c1, 0x111c4}, {0x11680, 0x116aa},
+ {0x12000, 0x1236e}, {0x13000, 0x1342e}, {0x16800, 0x16a38}, {0x16f00, 0x16f44},
+ {0x16f93, 0x16f9f}, {0x1d400, 0x1d454}, {0x1d456, 0x1d49c}, {0x1d4a9, 0x1d4ac},
+ {0x1d4ae, 0x1d4b9}, {0x1d4bd, 0x1d4c3}, {0x1d4c5, 0x1d505}, {0x1d507, 0x1d50a},
+ {0x1d50d, 0x1d514}, {0x1d516, 0x1d51c}, {0x1d51e, 0x1d539}, {0x1d53b, 0x1d53e},
+ {0x1d540, 0x1d544}, {0x1d54a, 0x1d550}, {0x1d552, 0x1d6a5}, {0x1d6a8, 0x1d6c0},
+ {0x1d6c2, 0x1d6da}, {0x1d6dc, 0x1d6fa}, {0x1d6fc, 0x1d714}, {0x1d716, 0x1d734},
+ {0x1d736, 0x1d74e}, {0x1d750, 0x1d76e}, {0x1d770, 0x1d788}, {0x1d78a, 0x1d7a8},
+ {0x1d7aa, 0x1d7c2}, {0x1d7c4, 0x1d7cb}, {0x1ee00, 0x1ee03}, {0x1ee05, 0x1ee1f},
+ {0x1ee29, 0x1ee32}, {0x1ee34, 0x1ee37}, {0x1ee4d, 0x1ee4f}, {0x1ee67, 0x1ee6a},
+ {0x1ee6c, 0x1ee72}, {0x1ee74, 0x1ee77}, {0x1ee79, 0x1ee7c}, {0x1ee80, 0x1ee89},
+ {0x1ee8b, 0x1ee9b}, {0x1eea1, 0x1eea3}, {0x1eea5, 0x1eea9}, {0x1eeab, 0x1eebb},
+ {0x20000, 0x2a6d6}, {0x2a700, 0x2b734}, {0x2b740, 0x2b81d}, {0x2f800, 0x2fa1d}
+#endif
};
-#define NUM_ALPHA (sizeof(alphaTable)/sizeof(crange))
+#define NUM_ALPHA_RANGE (sizeof(alphaRangeTable)/sizeof(crange))
-static crange digitTable[] = {
- {0x0030, 0x0039}
+static const chr alphaCharTable[] = {
+ 0xaa, 0xb5, 0xba, 0x2ec, 0x2ee, 0x376, 0x377, 0x386, 0x38c,
+ 0x559, 0x66e, 0x66f, 0x6d5, 0x6e5, 0x6e6, 0x6ee, 0x6ef, 0x6ff,
+ 0x710, 0x7b1, 0x7f4, 0x7f5, 0x7fa, 0x81a, 0x824, 0x828, 0x8a0,
+ 0x93d, 0x950, 0x98f, 0x990, 0x9b2, 0x9bd, 0x9ce, 0x9dc, 0x9dd,
+ 0x9f0, 0x9f1, 0xa0f, 0xa10, 0xa32, 0xa33, 0xa35, 0xa36, 0xa38,
+ 0xa39, 0xa5e, 0xab2, 0xab3, 0xabd, 0xad0, 0xae0, 0xae1, 0xb0f,
+ 0xb10, 0xb32, 0xb33, 0xb3d, 0xb5c, 0xb5d, 0xb71, 0xb83, 0xb99,
+ 0xb9a, 0xb9c, 0xb9e, 0xb9f, 0xba3, 0xba4, 0xbd0, 0xc3d, 0xc58,
+ 0xc59, 0xc60, 0xc61, 0xcbd, 0xcde, 0xce0, 0xce1, 0xcf1, 0xcf2,
+ 0xd3d, 0xd4e, 0xd60, 0xd61, 0xdbd, 0xe32, 0xe33, 0xe81, 0xe82,
+ 0xe84, 0xe87, 0xe88, 0xe8a, 0xe8d, 0xea5, 0xea7, 0xeaa, 0xeab,
+ 0xeb2, 0xeb3, 0xebd, 0xec6, 0xf00, 0x103f, 0x1061, 0x1065, 0x1066,
+ 0x108e, 0x10c7, 0x10cd, 0x1258, 0x12c0, 0x17d7, 0x17dc, 0x18aa, 0x1aa7,
+ 0x1bae, 0x1baf, 0x1cf5, 0x1cf6, 0x1f59, 0x1f5b, 0x1f5d, 0x1fbe, 0x2071,
+ 0x207f, 0x2102, 0x2107, 0x2115, 0x2124, 0x2126, 0x2128, 0x214e, 0x2183,
+ 0x2184, 0x2cf2, 0x2cf3, 0x2d27, 0x2d2d, 0x2d6f, 0x2e2f, 0x3005, 0x3006,
+ 0x303b, 0x303c, 0xa62a, 0xa62b, 0xa8fb, 0xa9cf, 0xaa7a, 0xaab1, 0xaab5,
+ 0xaab6, 0xaac0, 0xaac2, 0xfb1d, 0xfb3e, 0xfb40, 0xfb41, 0xfb43, 0xfb44
+#if TCL_UTF_MAX > 4
+ ,0x1003c, 0x1003d, 0x10808, 0x10837, 0x10838, 0x1083c, 0x109be, 0x109bf, 0x10a00,
+ 0x16f50, 0x1b000, 0x1b001, 0x1d49e, 0x1d49f, 0x1d4a2, 0x1d4a5, 0x1d4a6, 0x1d4bb,
+ 0x1d546, 0x1ee21, 0x1ee22, 0x1ee24, 0x1ee27, 0x1ee39, 0x1ee3b, 0x1ee42, 0x1ee47,
+ 0x1ee49, 0x1ee4b, 0x1ee51, 0x1ee52, 0x1ee54, 0x1ee57, 0x1ee59, 0x1ee5b, 0x1ee5d,
+ 0x1ee5f, 0x1ee61, 0x1ee62, 0x1ee64, 0x1ee7e
+#endif
};
-#define NUM_DIGIT (sizeof(digitTable)/sizeof(crange))
-
-static crange punctTable[] = {
- {0x0021, 0x0023}, {0X0025, 0X002A}, {0X002C, 0X002F}, {0X003A, 0X003B},
- {0X003F, 0X0040}, {0X005B, 0X005D}, {0X005F, 0X005F}, {0X007B, 0X007B},
- {0X007D, 0X007D}, {0X00A1, 0X00A1}, {0X00AB, 0X00AB}, {0X00AD, 0X00AD},
- {0X00BB, 0X00BB}, {0X00BF, 0X00BF}, {0X02BC, 0X02BC}, {0x0374, 0x0375},
- {0X037E, 0X037E}, {0x0387, 0x0387}, {0X055A, 0X055F}, {0x0589, 0x0589},
- {0X05BE, 0X05BE}, {0X05C0, 0X05C0}, {0X05C3, 0X05C3}, {0X05F3, 0X05F4},
- {0X060C, 0X060C}, {0X061B, 0X061B}, {0X061F, 0X061F}, {0X066A, 0X066D},
- {0X06D4, 0X06D4}, {0x0964, 0x0965}, {0x0970, 0x0970}, {0X0E2F, 0X0E2F},
- {0X0E5A, 0X0E5B}, {0X0EAF, 0X0EAF}, {0X0F04, 0X0F12}, {0X0F3A, 0X0F3F},
- {0X0F85, 0X0F85}, {0X10FB, 0X10FB}, {0x2010, 0x2027}, {0x2030, 0x2043},
- {0x2045, 0x2046}, {0X207D, 0X207E}, {0X208D, 0X208E}, {0X2329, 0X232A},
- {0x3001, 0x3003}, {0x3006, 0x3006}, {0x3008, 0x3011}, {0X3014, 0X301F},
- {0x3030, 0x3030}, {0X30FB, 0X30FB}, {0XFD3E, 0XFD3F}, {0XFE30, 0XFE44},
- {0XFE49, 0XFE52}, {0XFE54, 0XFE61}, {0XFE63, 0XFE63}, {0XFE68, 0XFE68},
- {0XFE6A, 0XFE6B}, {0XFF01, 0XFF03}, {0XFF05, 0XFF0A}, {0XFF0C, 0XFF0F},
- {0XFF1A, 0XFF1B}, {0XFF1F, 0XFF20}, {0XFF3B, 0XFF3D}, {0XFF3F, 0XFF3F},
- {0XFF5B, 0XFF5B}, {0XFF5D, 0XFF5D}, {0XFF61, 0XFF65}
+#define NUM_ALPHA_CHAR (sizeof(alphaCharTable)/sizeof(chr))
+
+/*
+ * Unicode: control characters.
+ */
+
+static const crange controlRangeTable[] = {
+ {0x0, 0x1f}, {0x7f, 0x9f}, {0x600, 0x604}, {0x200b, 0x200f},
+ {0x202a, 0x202e}, {0x2060, 0x2064}, {0x2066, 0x206f}, {0xe000, 0xf8ff},
+ {0xfff9, 0xfffb}
+#if TCL_UTF_MAX > 4
+ ,{0x1d173, 0x1d17a}, {0xe0020, 0xe007f}, {0xf0000, 0xffffd}, {0x100000, 0x10fffd}
+#endif
};
-#define NUM_PUNCT (sizeof(punctTable)/sizeof(crange))
-
-static crange spaceTable[] = {
- {0x0000, 0x0000}, {0x0009, 0x000D}, {0x0020, 0x0020}, {0x00A0, 0x00A0},
- {0x2000, 0x200F}, {0x2028, 0x202E}, {0X206A, 0X206F}, {0x3000, 0x3000},
- {0xFEFF, 0xFEFF}
-};
-
-#define NUM_SPACE (sizeof(spaceTable)/sizeof(crange))
-
-static crange upperRangeTable[] = {
- {0x0041, 0x005a}, {0x00c0, 0x00d6}, {0x00d8, 0x00de}, {0x0189, 0x018b},
- {0x018e, 0x0191}, {0x0388, 0x038a}, {0x0391, 0x03a1}, {0x03a3, 0x03ab},
- {0x03d2, 0x03d4}, {0x0401, 0x040c}, {0x040e, 0x042f}, {0x0531, 0x0556},
- {0x10a0, 0x10c5}, {0x1f08, 0x1f0f}, {0x1f18, 0x1f1d}, {0x1f28, 0x1f2f},
- {0x1f38, 0x1f3f}, {0x1f48, 0x1f4d}, {0x1f68, 0x1f6f}, {0x1f88, 0x1f8f},
- {0x1f98, 0x1f9f}, {0x1fa8, 0x1faf}, {0x1fb8, 0x1fbc}, {0x1fc8, 0x1fcc},
- {0x1fd8, 0x1fdb}, {0x1fe8, 0x1fec}, {0x1ff8, 0x1ffc}, {0x210b, 0x210d},
- {0x2110, 0x2112}, {0x2118, 0x211d}, {0x212a, 0x212d}, {0x2130, 0x2131},
- {0xff21, 0xff3a}
+#define NUM_CONTROL_RANGE (sizeof(controlRangeTable)/sizeof(crange))
+
+static const chr controlCharTable[] = {
+ 0xad, 0x61c, 0x6dd, 0x70f, 0x180e, 0xfeff
+#if TCL_UTF_MAX > 4
+ ,0x110bd, 0xe0001
+#endif
};
-#define NUM_UPPER_RANGE (sizeof(upperRangeTable)/sizeof(crange))
+#define NUM_CONTROL_CHAR (sizeof(controlCharTable)/sizeof(chr))
-static chr upperCharTable[] = {
- 0x0100, 0x0102, 0x0104, 0x0106, 0x0108, 0x010a, 0x010c, 0x010e, 0x0110,
- 0x0112, 0x0114, 0x0116, 0x0118, 0x011a, 0x011c, 0x011e, 0x0120, 0x0122,
- 0x0124, 0x0126, 0x0128, 0x012a, 0x012c, 0x012e, 0x0130, 0x0132, 0x0134,
- 0x0136, 0x0139, 0x013b, 0x013d, 0x013f, 0x0141, 0x0143, 0x0145, 0x0147,
- 0x014a, 0x014c, 0x014e, 0x0150, 0x0152, 0x0154, 0x0156, 0x0158, 0x015a,
- 0x015c, 0x015e, 0x0160, 0x0162, 0x0164, 0x0166, 0x0168, 0x016a, 0x016c,
- 0x016e, 0x0170, 0x0172, 0x0174, 0x0176, 0x0178, 0x0179, 0x017b, 0x017d,
- 0x0181, 0x0182, 0x0184, 0x0186, 0x0187, 0x0193, 0x0194, 0x0196, 0x0197,
- 0x0198, 0x019c, 0x019d, 0x019f, 0x01a0, 0x01a2, 0x01a4, 0x01a6, 0x01a7,
- 0x01a9, 0x01ac, 0x01ae, 0x01af, 0x01b1, 0x01b2, 0x01b3, 0x01b5, 0x01b7,
- 0x01b8, 0x01bc, 0x01c4, 0x01c7, 0x01ca, 0x01cd, 0x01cf, 0x01d1, 0x01d3,
- 0x01d5, 0x01d7, 0x01d9, 0x01db, 0x01de, 0x01e0, 0x01e2, 0x01e4, 0x01e6,
- 0x01e8, 0x01ea, 0x01ec, 0x01ee, 0x01f1, 0x01f4, 0x01fa, 0x01fc, 0x01fe,
- 0x0200, 0x0202, 0x0204, 0x0206, 0x0208, 0x020a, 0x020c, 0x020e, 0x0210,
- 0x0212, 0x0214, 0x0216, 0x0386, 0x038c, 0x038e, 0x038f, 0x03da, 0x03dc,
- 0x03de, 0x03e0, 0x03e2, 0x03e4, 0x03e6, 0x03e8, 0x03ea, 0x03ec, 0x03ee,
- 0x0460, 0x0462, 0x0464, 0x0466, 0x0468, 0x046a, 0x046c, 0x046e, 0x0470,
- 0x0472, 0x0474, 0x0476, 0x0478, 0x047a, 0x047c, 0x047e, 0x0480, 0x0490,
- 0x0492, 0x0494, 0x0496, 0x0498, 0x049a, 0x049c, 0x049e, 0x04a0, 0x04a2,
- 0x04a4, 0x04a6, 0x04a8, 0x04aa, 0x04ac, 0x04ae, 0x04b0, 0x04b2, 0x04b4,
- 0x04b6, 0x04b8, 0x04ba, 0x04bc, 0x04be, 0x04c1, 0x04c3, 0x04c7, 0x04cb,
- 0x04d0, 0x04d2, 0x04d4, 0x04d6, 0x04d8, 0x04da, 0x04dc, 0x04de, 0x04e0,
- 0x04e2, 0x04e4, 0x04e6, 0x04e8, 0x04ea, 0x04ee, 0x04f0, 0x04f2, 0x04f4,
- 0x04f8, 0x1e00, 0x1e02, 0x1e04, 0x1e06, 0x1e08, 0x1e0a, 0x1e0c, 0x1e0e,
- 0x1e10, 0x1e12, 0x1e14, 0x1e16, 0x1e18, 0x1e1a, 0x1e1c, 0x1e1e, 0x1e20,
- 0x1e22, 0x1e24, 0x1e26, 0x1e28, 0x1e2a, 0x1e2c, 0x1e2e, 0x1e30, 0x1e32,
- 0x1e34, 0x1e36, 0x1e38, 0x1e3a, 0x1e3c, 0x1e3e, 0x1e40, 0x1e42, 0x1e44,
- 0x1e46, 0x1e48, 0x1e4a, 0x1e4c, 0x1e4e, 0x1e50, 0x1e52, 0x1e54, 0x1e56,
- 0x1e58, 0x1e5a, 0x1e5c, 0x1e5e, 0x1e60, 0x1e62, 0x1e64, 0x1e66, 0x1e68,
- 0x1e6a, 0x1e6c, 0x1e6e, 0x1e70, 0x1e72, 0x1e74, 0x1e76, 0x1e78, 0x1e7a,
- 0x1e7c, 0x1e7e, 0x1e80, 0x1e82, 0x1e84, 0x1e86, 0x1e88, 0x1e8a, 0x1e8c,
- 0x1e8e, 0x1e90, 0x1e92, 0x1e94, 0x1ea0, 0x1ea2, 0x1ea4, 0x1ea6, 0x1ea8,
- 0x1eaa, 0x1eac, 0x1eae, 0x1eb0, 0x1eb2, 0x1eb4, 0x1eb6, 0x1eb8, 0x1eba,
- 0x1ebc, 0x1ebe, 0x1ec0, 0x1ec2, 0x1ec4, 0x1ec6, 0x1ec8, 0x1eca, 0x1ecc,
- 0x1ece, 0x1ed0, 0x1ed2, 0x1ed4, 0x1ed6, 0x1ed8, 0x1eda, 0x1edc, 0x1ede,
- 0x1ee0, 0x1ee2, 0x1ee4, 0x1ee6, 0x1ee8, 0x1eea, 0x1eec, 0x1eee, 0x1ef0,
- 0x1ef2, 0x1ef4, 0x1ef6, 0x1ef8, 0x1f59, 0x1f5b, 0x1f5d, 0x1f5f, 0x1fbe,
- 0x2102, 0x2107, 0x2115, 0x2124, 0x2126, 0x2128, 0x2133
+/*
+ * Unicode: decimal digit characters.
+ */
+
+static const crange digitRangeTable[] = {
+ {0x30, 0x39}, {0x660, 0x669}, {0x6f0, 0x6f9}, {0x7c0, 0x7c9},
+ {0x966, 0x96f}, {0x9e6, 0x9ef}, {0xa66, 0xa6f}, {0xae6, 0xaef},
+ {0xb66, 0xb6f}, {0xbe6, 0xbef}, {0xc66, 0xc6f}, {0xce6, 0xcef},
+ {0xd66, 0xd6f}, {0xe50, 0xe59}, {0xed0, 0xed9}, {0xf20, 0xf29},
+ {0x1040, 0x1049}, {0x1090, 0x1099}, {0x17e0, 0x17e9}, {0x1810, 0x1819},
+ {0x1946, 0x194f}, {0x19d0, 0x19d9}, {0x1a80, 0x1a89}, {0x1a90, 0x1a99},
+ {0x1b50, 0x1b59}, {0x1bb0, 0x1bb9}, {0x1c40, 0x1c49}, {0x1c50, 0x1c59},
+ {0xa620, 0xa629}, {0xa8d0, 0xa8d9}, {0xa900, 0xa909}, {0xa9d0, 0xa9d9},
+ {0xaa50, 0xaa59}, {0xabf0, 0xabf9}, {0xff10, 0xff19}
+#if TCL_UTF_MAX > 4
+ ,{0x104a0, 0x104a9}, {0x11066, 0x1106f}, {0x110f0, 0x110f9}, {0x11136, 0x1113f},
+ {0x111d0, 0x111d9}, {0x116c0, 0x116c9}, {0x1d7ce, 0x1d7ff}
+#endif
};
-#define NUM_UPPER_CHAR (sizeof(upperCharTable)/sizeof(chr))
+#define NUM_DIGIT_RANGE (sizeof(digitRangeTable)/sizeof(crange))
+
+/*
+ * no singletons of digit characters.
+ */
+
+/*
+ * Unicode: punctuation characters.
+ */
-static crange lowerRangeTable[] = {
- {0x0061, 0x007a}, {0x00df, 0x00f6}, {0x00f8, 0x00ff}, {0x0199, 0x019b},
- {0x0250, 0x02a8}, {0x03ac, 0x03ce}, {0x03ef, 0x03f2}, {0x0430, 0x044f},
- {0x0451, 0x045c}, {0x0561, 0x0587}, {0x10d0, 0x10f6}, {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},
- {0x1fd0, 0x1fd3}, {0x1fe0, 0x1fe7}, {0x1ff2, 0x1ff4}, {0xfb00, 0xfb06},
- {0xfb13, 0xfb17}, {0xff41, 0xff5a}
+static const crange punctRangeTable[] = {
+ {0x21, 0x23}, {0x25, 0x2a}, {0x2c, 0x2f}, {0x5b, 0x5d},
+ {0x55a, 0x55f}, {0x66a, 0x66d}, {0x700, 0x70d}, {0x7f7, 0x7f9},
+ {0x830, 0x83e}, {0xf04, 0xf12}, {0xf3a, 0xf3d}, {0xfd0, 0xfd4},
+ {0x104a, 0x104f}, {0x1360, 0x1368}, {0x16eb, 0x16ed}, {0x17d4, 0x17d6},
+ {0x17d8, 0x17da}, {0x1800, 0x180a}, {0x1aa0, 0x1aa6}, {0x1aa8, 0x1aad},
+ {0x1b5a, 0x1b60}, {0x1bfc, 0x1bff}, {0x1c3b, 0x1c3f}, {0x1cc0, 0x1cc7},
+ {0x2010, 0x2027}, {0x2030, 0x2043}, {0x2045, 0x2051}, {0x2053, 0x205e},
+ {0x2308, 0x230b}, {0x2768, 0x2775}, {0x27e6, 0x27ef}, {0x2983, 0x2998},
+ {0x29d8, 0x29db}, {0x2cf9, 0x2cfc}, {0x2e00, 0x2e2e}, {0x2e30, 0x2e3b},
+ {0x3001, 0x3003}, {0x3008, 0x3011}, {0x3014, 0x301f}, {0xa60d, 0xa60f},
+ {0xa6f2, 0xa6f7}, {0xa874, 0xa877}, {0xa8f8, 0xa8fa}, {0xa9c1, 0xa9cd},
+ {0xaa5c, 0xaa5f}, {0xfe10, 0xfe19}, {0xfe30, 0xfe52}, {0xfe54, 0xfe61},
+ {0xff01, 0xff03}, {0xff05, 0xff0a}, {0xff0c, 0xff0f}, {0xff3b, 0xff3d},
+ {0xff5f, 0xff65}
+#if TCL_UTF_MAX > 4
+ ,{0x10100, 0x10102}, {0x10a50, 0x10a58}, {0x10b39, 0x10b3f}, {0x11047, 0x1104d},
+ {0x110be, 0x110c1}, {0x11140, 0x11143}, {0x111c5, 0x111c8}, {0x12470, 0x12473}
+#endif
};
-#define NUM_LOWER_RANGE (sizeof(lowerRangeTable)/sizeof(crange))
+#define NUM_PUNCT_RANGE (sizeof(punctRangeTable)/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, 0x0138, 0x013a, 0x013c, 0x013e, 0x0140, 0x0142,
- 0x0144, 0x0146, 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, 0x017e, 0x017f, 0x0180, 0x0183, 0x0185, 0x0188, 0x018c, 0x018d,
- 0x0192, 0x0195, 0x019e, 0x01a1, 0x01a3, 0x01a5, 0x01a8, 0x01ab, 0x01ad,
- 0x01b0, 0x01b4, 0x01b6, 0x01b9, 0x01ba, 0x01bd, 0x01c6, 0x01c9, 0x01cc,
- 0x01ce, 0x01d0, 0x01d2, 0x01d4, 0x01d6, 0x01d8, 0x01da, 0x01dd, 0x01df,
- 0x01e1, 0x01e3, 0x01e5, 0x01e7, 0x01e9, 0x01eb, 0x01ed, 0x01f0, 0x01f3,
- 0x01f5, 0x01fb, 0x01fd, 0x01ff, 0x0201, 0x0203, 0x0205, 0x0207, 0x0209,
- 0x020b, 0x020d, 0x020f, 0x0211, 0x0213, 0x0215, 0x0217, 0x0390, 0x03d0,
- 0x03d1, 0x03d5, 0x03d6, 0x03e3, 0x03e5, 0x03e7, 0x03e9, 0x03eb, 0x03ed,
- 0x045e, 0x045f, 0x0461, 0x0463, 0x0465, 0x0467, 0x0469, 0x046b, 0x046d,
- 0x046f, 0x0471, 0x0473, 0x0475, 0x0477, 0x0479, 0x047b, 0x047d, 0x047f,
- 0x0481, 0x0491, 0x0493, 0x0495, 0x0497, 0x0499, 0x049b, 0x049d, 0x049f,
- 0x04a1, 0x04a3, 0x04a5, 0x04a7, 0x04a9, 0x04ab, 0x04ad, 0x04af, 0x04b1,
- 0x04b3, 0x04b5, 0x04b7, 0x04b9, 0x04bb, 0x04bd, 0x04bf, 0x04c2, 0x04c4,
- 0x04c8, 0x04cc, 0x04d1, 0x04d3, 0x04d5, 0x04d7, 0x04d9, 0x04db, 0x04dd,
- 0x04df, 0x04e1, 0x04e3, 0x04e5, 0x04e7, 0x04e9, 0x04eb, 0x04ef, 0x04f1,
- 0x04f3, 0x04f5, 0x04f9, 0x1e01, 0x1e03, 0x1e05, 0x1e07, 0x1e09, 0x1e0b,
- 0x1e0d, 0x1e0f, 0x1e11, 0x1e13, 0x1e15, 0x1e17, 0x1e19, 0x1e1b, 0x1e1d,
- 0x1e1f, 0x1e21, 0x1e23, 0x1e25, 0x1e27, 0x1e29, 0x1e2b, 0x1e2d, 0x1e2f,
- 0x1e31, 0x1e33, 0x1e35, 0x1e37, 0x1e39, 0x1e3b, 0x1e3d, 0x1e3f, 0x1e41,
- 0x1e43, 0x1e45, 0x1e47, 0x1e49, 0x1e4b, 0x1e4d, 0x1e4f, 0x1e51, 0x1e53,
- 0x1e55, 0x1e57, 0x1e59, 0x1e5b, 0x1e5d, 0x1e5f, 0x1e61, 0x1e63, 0x1e65,
- 0x1e67, 0x1e69, 0x1e6b, 0x1e6d, 0x1e6f, 0x1e71, 0x1e73, 0x1e75, 0x1e77,
- 0x1e79, 0x1e7b, 0x1e7d, 0x1e7f, 0x1e81, 0x1e83, 0x1e85, 0x1e87, 0x1e89,
- 0x1e8b, 0x1e8d, 0x1e8f, 0x1e91, 0x1e93, 0x1ea1, 0x1ea3, 0x1ea5, 0x1ea7,
- 0x1ea9, 0x1eab, 0x1ead, 0x1eaf, 0x1eb1, 0x1eb3, 0x1eb5, 0x1eb7, 0x1eb9,
- 0x1ebb, 0x1ebd, 0x1ebf, 0x1ec1, 0x1ec3, 0x1ec5, 0x1ec7, 0x1ec9, 0x1ecb,
- 0x1ecd, 0x1ecf, 0x1ed1, 0x1ed3, 0x1ed5, 0x1ed7, 0x1ed9, 0x1edb, 0x1edd,
- 0x1edf, 0x1ee1, 0x1ee3, 0x1ee5, 0x1ee7, 0x1ee9, 0x1eeb, 0x1eed, 0x1eef,
- 0x1ef1, 0x1ef3, 0x1ef5, 0x1ef7, 0x1ef9, 0x1fb6, 0x1fb7, 0x1fc2, 0x1fc3,
- 0x1fc4, 0x1fc6, 0x1fc7, 0x1fd6, 0x1fd7, 0x1ff6, 0x1ff7, 0x207f, 0x210a,
- 0x210e, 0x210f, 0x2113, 0x212e, 0x212f, 0x2134
+static const chr punctCharTable[] = {
+ 0x3a, 0x3b, 0x3f, 0x40, 0x5f, 0x7b, 0x7d, 0xa1, 0xa7,
+ 0xab, 0xb6, 0xb7, 0xbb, 0xbf, 0x37e, 0x387, 0x589, 0x58a,
+ 0x5be, 0x5c0, 0x5c3, 0x5c6, 0x5f3, 0x5f4, 0x609, 0x60a, 0x60c,
+ 0x60d, 0x61b, 0x61e, 0x61f, 0x6d4, 0x85e, 0x964, 0x965, 0x970,
+ 0xaf0, 0xdf4, 0xe4f, 0xe5a, 0xe5b, 0xf14, 0xf85, 0xfd9, 0xfda,
+ 0x10fb, 0x1400, 0x166d, 0x166e, 0x169b, 0x169c, 0x1735, 0x1736, 0x1944,
+ 0x1945, 0x1a1e, 0x1a1f, 0x1c7e, 0x1c7f, 0x1cd3, 0x207d, 0x207e, 0x208d,
+ 0x208e, 0x2329, 0x232a, 0x27c5, 0x27c6, 0x29fc, 0x29fd, 0x2cfe, 0x2cff,
+ 0x2d70, 0x3030, 0x303d, 0x30a0, 0x30fb, 0xa4fe, 0xa4ff, 0xa673, 0xa67e,
+ 0xa8ce, 0xa8cf, 0xa92e, 0xa92f, 0xa95f, 0xa9de, 0xa9df, 0xaade, 0xaadf,
+ 0xaaf0, 0xaaf1, 0xabeb, 0xfd3e, 0xfd3f, 0xfe63, 0xfe68, 0xfe6a, 0xfe6b,
+ 0xff1a, 0xff1b, 0xff1f, 0xff20, 0xff3f, 0xff5b, 0xff5d
+#if TCL_UTF_MAX > 4
+ ,0x1039f, 0x103d0, 0x10857, 0x1091f, 0x1093f, 0x10a7f, 0x110bb, 0x110bc
+#endif
};
-#define NUM_LOWER_CHAR (sizeof(lowerCharTable)/sizeof(chr))
+#define NUM_PUNCT_CHAR (sizeof(punctCharTable)/sizeof(chr))
/*
- * The graph table includes the set of characters that are neither ISO control
- * characters nor in the space table.
+ * Unicode: white space characters.
*/
-static crange graphTable[] = {
- {0x0021, 0x007e}, {0x00a1, 0x1fff}, {0x2010, 0x2027}, {0x202f, 0x2069},
- {0x2070, 0x2fff}, {0x3001, 0xfefe}, {0xff00, 0xffff}
+static const crange spaceRangeTable[] = {
+ {0x9, 0xd}, {0x2000, 0x200b}
};
-#define NUM_GRAPH (sizeof(graphTable)/sizeof(crange))
-#define CH NOCELT
+#define NUM_SPACE_RANGE (sizeof(spaceRangeTable)/sizeof(crange))
+
+static const chr spaceCharTable[] = {
+ 0x20, 0x85, 0xa0, 0x1680, 0x180e, 0x2028, 0x2029, 0x202f, 0x205f,
+ 0x2060, 0x3000, 0xfeff
+};
+
+#define NUM_SPACE_CHAR (sizeof(spaceCharTable)/sizeof(chr))
/*
- - nmcces - how many distinct MCCEs are there?
- ^ static int nmcces(struct vars *);
+ * Unicode: lowercase characters.
*/
-static int
-nmcces(v)
-struct vars *v;
-{
- return 0;
-}
+
+static const crange lowerRangeTable[] = {
+ {0x61, 0x7a}, {0xdf, 0xf6}, {0xf8, 0xff}, {0x17e, 0x180},
+ {0x199, 0x19b}, {0x1bd, 0x1bf}, {0x233, 0x239}, {0x24f, 0x293},
+ {0x295, 0x2af}, {0x37b, 0x37d}, {0x3ac, 0x3ce}, {0x3d5, 0x3d7},
+ {0x3ef, 0x3f3}, {0x430, 0x45f}, {0x561, 0x587}, {0x1d00, 0x1d2b},
+ {0x1d6b, 0x1d77}, {0x1d79, 0x1d9a}, {0x1e95, 0x1e9d}, {0x1eff, 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}, {0x2146, 0x2149},
+ {0x2c30, 0x2c5e}, {0x2c76, 0x2c7b}, {0x2d00, 0x2d25}, {0xa72f, 0xa731},
+ {0xa771, 0xa778}, {0xfb00, 0xfb06}, {0xfb13, 0xfb17}, {0xff41, 0xff5a}
+#if TCL_UTF_MAX > 4
+ ,{0x10428, 0x1044f}, {0x1d41a, 0x1d433}, {0x1d44e, 0x1d454}, {0x1d456, 0x1d467},
+ {0x1d482, 0x1d49b}, {0x1d4b6, 0x1d4b9}, {0x1d4bd, 0x1d4c3}, {0x1d4c5, 0x1d4cf},
+ {0x1d4ea, 0x1d503}, {0x1d51e, 0x1d537}, {0x1d552, 0x1d56b}, {0x1d586, 0x1d59f},
+ {0x1d5ba, 0x1d5d3}, {0x1d5ee, 0x1d607}, {0x1d622, 0x1d63b}, {0x1d656, 0x1d66f},
+ {0x1d68a, 0x1d6a5}, {0x1d6c2, 0x1d6da}, {0x1d6dc, 0x1d6e1}, {0x1d6fc, 0x1d714},
+ {0x1d716, 0x1d71b}, {0x1d736, 0x1d74e}, {0x1d750, 0x1d755}, {0x1d770, 0x1d788},
+ {0x1d78a, 0x1d78f}, {0x1d7aa, 0x1d7c2}, {0x1d7c4, 0x1d7c9}
+#endif
+};
+
+#define NUM_LOWER_RANGE (sizeof(lowerRangeTable)/sizeof(crange))
+
+static const chr lowerCharTable[] = {
+ 0xb5, 0x101, 0x103, 0x105, 0x107, 0x109, 0x10b, 0x10d, 0x10f,
+ 0x111, 0x113, 0x115, 0x117, 0x119, 0x11b, 0x11d, 0x11f, 0x121,
+ 0x123, 0x125, 0x127, 0x129, 0x12b, 0x12d, 0x12f, 0x131, 0x133,
+ 0x135, 0x137, 0x138, 0x13a, 0x13c, 0x13e, 0x140, 0x142, 0x144,
+ 0x146, 0x148, 0x149, 0x14b, 0x14d, 0x14f, 0x151, 0x153, 0x155,
+ 0x157, 0x159, 0x15b, 0x15d, 0x15f, 0x161, 0x163, 0x165, 0x167,
+ 0x169, 0x16b, 0x16d, 0x16f, 0x171, 0x173, 0x175, 0x177, 0x17a,
+ 0x17c, 0x183, 0x185, 0x188, 0x18c, 0x18d, 0x192, 0x195, 0x19e,
+ 0x1a1, 0x1a3, 0x1a5, 0x1a8, 0x1aa, 0x1ab, 0x1ad, 0x1b0, 0x1b4,
+ 0x1b6, 0x1b9, 0x1ba, 0x1c6, 0x1c9, 0x1cc, 0x1ce, 0x1d0, 0x1d2,
+ 0x1d4, 0x1d6, 0x1d8, 0x1da, 0x1dc, 0x1dd, 0x1df, 0x1e1, 0x1e3,
+ 0x1e5, 0x1e7, 0x1e9, 0x1eb, 0x1ed, 0x1ef, 0x1f0, 0x1f3, 0x1f5,
+ 0x1f9, 0x1fb, 0x1fd, 0x1ff, 0x201, 0x203, 0x205, 0x207, 0x209,
+ 0x20b, 0x20d, 0x20f, 0x211, 0x213, 0x215, 0x217, 0x219, 0x21b,
+ 0x21d, 0x21f, 0x221, 0x223, 0x225, 0x227, 0x229, 0x22b, 0x22d,
+ 0x22f, 0x231, 0x23c, 0x23f, 0x240, 0x242, 0x247, 0x249, 0x24b,
+ 0x24d, 0x371, 0x373, 0x377, 0x390, 0x3d0, 0x3d1, 0x3d9, 0x3db,
+ 0x3dd, 0x3df, 0x3e1, 0x3e3, 0x3e5, 0x3e7, 0x3e9, 0x3eb, 0x3ed,
+ 0x3f5, 0x3f8, 0x3fb, 0x3fc, 0x461, 0x463, 0x465, 0x467, 0x469,
+ 0x46b, 0x46d, 0x46f, 0x471, 0x473, 0x475, 0x477, 0x479, 0x47b,
+ 0x47d, 0x47f, 0x481, 0x48b, 0x48d, 0x48f, 0x491, 0x493, 0x495,
+ 0x497, 0x499, 0x49b, 0x49d, 0x49f, 0x4a1, 0x4a3, 0x4a5, 0x4a7,
+ 0x4a9, 0x4ab, 0x4ad, 0x4af, 0x4b1, 0x4b3, 0x4b5, 0x4b7, 0x4b9,
+ 0x4bb, 0x4bd, 0x4bf, 0x4c2, 0x4c4, 0x4c6, 0x4c8, 0x4ca, 0x4cc,
+ 0x4ce, 0x4cf, 0x4d1, 0x4d3, 0x4d5, 0x4d7, 0x4d9, 0x4db, 0x4dd,
+ 0x4df, 0x4e1, 0x4e3, 0x4e5, 0x4e7, 0x4e9, 0x4eb, 0x4ed, 0x4ef,
+ 0x4f1, 0x4f3, 0x4f5, 0x4f7, 0x4f9, 0x4fb, 0x4fd, 0x4ff, 0x501,
+ 0x503, 0x505, 0x507, 0x509, 0x50b, 0x50d, 0x50f, 0x511, 0x513,
+ 0x515, 0x517, 0x519, 0x51b, 0x51d, 0x51f, 0x521, 0x523, 0x525,
+ 0x527, 0x1e01, 0x1e03, 0x1e05, 0x1e07, 0x1e09, 0x1e0b, 0x1e0d, 0x1e0f,
+ 0x1e11, 0x1e13, 0x1e15, 0x1e17, 0x1e19, 0x1e1b, 0x1e1d, 0x1e1f, 0x1e21,
+ 0x1e23, 0x1e25, 0x1e27, 0x1e29, 0x1e2b, 0x1e2d, 0x1e2f, 0x1e31, 0x1e33,
+ 0x1e35, 0x1e37, 0x1e39, 0x1e3b, 0x1e3d, 0x1e3f, 0x1e41, 0x1e43, 0x1e45,
+ 0x1e47, 0x1e49, 0x1e4b, 0x1e4d, 0x1e4f, 0x1e51, 0x1e53, 0x1e55, 0x1e57,
+ 0x1e59, 0x1e5b, 0x1e5d, 0x1e5f, 0x1e61, 0x1e63, 0x1e65, 0x1e67, 0x1e69,
+ 0x1e6b, 0x1e6d, 0x1e6f, 0x1e71, 0x1e73, 0x1e75, 0x1e77, 0x1e79, 0x1e7b,
+ 0x1e7d, 0x1e7f, 0x1e81, 0x1e83, 0x1e85, 0x1e87, 0x1e89, 0x1e8b, 0x1e8d,
+ 0x1e8f, 0x1e91, 0x1e93, 0x1e9f, 0x1ea1, 0x1ea3, 0x1ea5, 0x1ea7, 0x1ea9,
+ 0x1eab, 0x1ead, 0x1eaf, 0x1eb1, 0x1eb3, 0x1eb5, 0x1eb7, 0x1eb9, 0x1ebb,
+ 0x1ebd, 0x1ebf, 0x1ec1, 0x1ec3, 0x1ec5, 0x1ec7, 0x1ec9, 0x1ecb, 0x1ecd,
+ 0x1ecf, 0x1ed1, 0x1ed3, 0x1ed5, 0x1ed7, 0x1ed9, 0x1edb, 0x1edd, 0x1edf,
+ 0x1ee1, 0x1ee3, 0x1ee5, 0x1ee7, 0x1ee9, 0x1eeb, 0x1eed, 0x1eef, 0x1ef1,
+ 0x1ef3, 0x1ef5, 0x1ef7, 0x1ef9, 0x1efb, 0x1efd, 0x1fb6, 0x1fb7, 0x1fbe,
+ 0x1fc6, 0x1fc7, 0x1fd6, 0x1fd7, 0x1ff6, 0x1ff7, 0x210a, 0x210e, 0x210f,
+ 0x2113, 0x212f, 0x2134, 0x2139, 0x213c, 0x213d, 0x214e, 0x2184, 0x2c61,
+ 0x2c65, 0x2c66, 0x2c68, 0x2c6a, 0x2c6c, 0x2c71, 0x2c73, 0x2c74, 0x2c81,
+ 0x2c83, 0x2c85, 0x2c87, 0x2c89, 0x2c8b, 0x2c8d, 0x2c8f, 0x2c91, 0x2c93,
+ 0x2c95, 0x2c97, 0x2c99, 0x2c9b, 0x2c9d, 0x2c9f, 0x2ca1, 0x2ca3, 0x2ca5,
+ 0x2ca7, 0x2ca9, 0x2cab, 0x2cad, 0x2caf, 0x2cb1, 0x2cb3, 0x2cb5, 0x2cb7,
+ 0x2cb9, 0x2cbb, 0x2cbd, 0x2cbf, 0x2cc1, 0x2cc3, 0x2cc5, 0x2cc7, 0x2cc9,
+ 0x2ccb, 0x2ccd, 0x2ccf, 0x2cd1, 0x2cd3, 0x2cd5, 0x2cd7, 0x2cd9, 0x2cdb,
+ 0x2cdd, 0x2cdf, 0x2ce1, 0x2ce3, 0x2ce4, 0x2cec, 0x2cee, 0x2cf3, 0x2d27,
+ 0x2d2d, 0xa641, 0xa643, 0xa645, 0xa647, 0xa649, 0xa64b, 0xa64d, 0xa64f,
+ 0xa651, 0xa653, 0xa655, 0xa657, 0xa659, 0xa65b, 0xa65d, 0xa65f, 0xa661,
+ 0xa663, 0xa665, 0xa667, 0xa669, 0xa66b, 0xa66d, 0xa681, 0xa683, 0xa685,
+ 0xa687, 0xa689, 0xa68b, 0xa68d, 0xa68f, 0xa691, 0xa693, 0xa695, 0xa697,
+ 0xa723, 0xa725, 0xa727, 0xa729, 0xa72b, 0xa72d, 0xa733, 0xa735, 0xa737,
+ 0xa739, 0xa73b, 0xa73d, 0xa73f, 0xa741, 0xa743, 0xa745, 0xa747, 0xa749,
+ 0xa74b, 0xa74d, 0xa74f, 0xa751, 0xa753, 0xa755, 0xa757, 0xa759, 0xa75b,
+ 0xa75d, 0xa75f, 0xa761, 0xa763, 0xa765, 0xa767, 0xa769, 0xa76b, 0xa76d,
+ 0xa76f, 0xa77a, 0xa77c, 0xa77f, 0xa781, 0xa783, 0xa785, 0xa787, 0xa78c,
+ 0xa78e, 0xa791, 0xa793, 0xa7a1, 0xa7a3, 0xa7a5, 0xa7a7, 0xa7a9, 0xa7fa
+#if TCL_UTF_MAX > 4
+ ,0x1d4bb, 0x1d7cb
+#endif
+};
+
+#define NUM_LOWER_CHAR (sizeof(lowerCharTable)/sizeof(chr))
/*
- - nleaders - how many chrs can be first chrs of MCCEs?
- ^ static int nleaders(struct vars *);
+ * Unicode: uppercase characters.
*/
-static int
-nleaders(v)
-struct vars *v;
-{
- return 0;
-}
+
+static const crange upperRangeTable[] = {
+ {0x41, 0x5a}, {0xc0, 0xd6}, {0xd8, 0xde}, {0x189, 0x18b},
+ {0x18e, 0x191}, {0x196, 0x198}, {0x1b1, 0x1b3}, {0x1f6, 0x1f8},
+ {0x243, 0x246}, {0x388, 0x38a}, {0x391, 0x3a1}, {0x3a3, 0x3ab},
+ {0x3d2, 0x3d4}, {0x3fd, 0x42f}, {0x531, 0x556}, {0x10a0, 0x10c5},
+ {0x1f08, 0x1f0f}, {0x1f18, 0x1f1d}, {0x1f28, 0x1f2f}, {0x1f38, 0x1f3f},
+ {0x1f48, 0x1f4d}, {0x1f68, 0x1f6f}, {0x1fb8, 0x1fbb}, {0x1fc8, 0x1fcb},
+ {0x1fd8, 0x1fdb}, {0x1fe8, 0x1fec}, {0x1ff8, 0x1ffb}, {0x210b, 0x210d},
+ {0x2110, 0x2112}, {0x2119, 0x211d}, {0x212a, 0x212d}, {0x2130, 0x2133},
+ {0x2c00, 0x2c2e}, {0x2c62, 0x2c64}, {0x2c6d, 0x2c70}, {0x2c7e, 0x2c80},
+ {0xff21, 0xff3a}
+#if TCL_UTF_MAX > 4
+ ,{0x10400, 0x10427}, {0x1d400, 0x1d419}, {0x1d434, 0x1d44d}, {0x1d468, 0x1d481},
+ {0x1d4a9, 0x1d4ac}, {0x1d4ae, 0x1d4b5}, {0x1d4d0, 0x1d4e9}, {0x1d507, 0x1d50a},
+ {0x1d50d, 0x1d514}, {0x1d516, 0x1d51c}, {0x1d53b, 0x1d53e}, {0x1d540, 0x1d544},
+ {0x1d54a, 0x1d550}, {0x1d56c, 0x1d585}, {0x1d5a0, 0x1d5b9}, {0x1d5d4, 0x1d5ed},
+ {0x1d608, 0x1d621}, {0x1d63c, 0x1d655}, {0x1d670, 0x1d689}, {0x1d6a8, 0x1d6c0},
+ {0x1d6e2, 0x1d6fa}, {0x1d71c, 0x1d734}, {0x1d756, 0x1d76e}, {0x1d790, 0x1d7a8}
+#endif
+};
+
+#define NUM_UPPER_RANGE (sizeof(upperRangeTable)/sizeof(crange))
+
+static const chr upperCharTable[] = {
+ 0x100, 0x102, 0x104, 0x106, 0x108, 0x10a, 0x10c, 0x10e, 0x110,
+ 0x112, 0x114, 0x116, 0x118, 0x11a, 0x11c, 0x11e, 0x120, 0x122,
+ 0x124, 0x126, 0x128, 0x12a, 0x12c, 0x12e, 0x130, 0x132, 0x134,
+ 0x136, 0x139, 0x13b, 0x13d, 0x13f, 0x141, 0x143, 0x145, 0x147,
+ 0x14a, 0x14c, 0x14e, 0x150, 0x152, 0x154, 0x156, 0x158, 0x15a,
+ 0x15c, 0x15e, 0x160, 0x162, 0x164, 0x166, 0x168, 0x16a, 0x16c,
+ 0x16e, 0x170, 0x172, 0x174, 0x176, 0x178, 0x179, 0x17b, 0x17d,
+ 0x181, 0x182, 0x184, 0x186, 0x187, 0x193, 0x194, 0x19c, 0x19d,
+ 0x19f, 0x1a0, 0x1a2, 0x1a4, 0x1a6, 0x1a7, 0x1a9, 0x1ac, 0x1ae,
+ 0x1af, 0x1b5, 0x1b7, 0x1b8, 0x1bc, 0x1c4, 0x1c7, 0x1ca, 0x1cd,
+ 0x1cf, 0x1d1, 0x1d3, 0x1d5, 0x1d7, 0x1d9, 0x1db, 0x1de, 0x1e0,
+ 0x1e2, 0x1e4, 0x1e6, 0x1e8, 0x1ea, 0x1ec, 0x1ee, 0x1f1, 0x1f4,
+ 0x1fa, 0x1fc, 0x1fe, 0x200, 0x202, 0x204, 0x206, 0x208, 0x20a,
+ 0x20c, 0x20e, 0x210, 0x212, 0x214, 0x216, 0x218, 0x21a, 0x21c,
+ 0x21e, 0x220, 0x222, 0x224, 0x226, 0x228, 0x22a, 0x22c, 0x22e,
+ 0x230, 0x232, 0x23a, 0x23b, 0x23d, 0x23e, 0x241, 0x248, 0x24a,
+ 0x24c, 0x24e, 0x370, 0x372, 0x376, 0x386, 0x38c, 0x38e, 0x38f,
+ 0x3cf, 0x3d8, 0x3da, 0x3dc, 0x3de, 0x3e0, 0x3e2, 0x3e4, 0x3e6,
+ 0x3e8, 0x3ea, 0x3ec, 0x3ee, 0x3f4, 0x3f7, 0x3f9, 0x3fa, 0x460,
+ 0x462, 0x464, 0x466, 0x468, 0x46a, 0x46c, 0x46e, 0x470, 0x472,
+ 0x474, 0x476, 0x478, 0x47a, 0x47c, 0x47e, 0x480, 0x48a, 0x48c,
+ 0x48e, 0x490, 0x492, 0x494, 0x496, 0x498, 0x49a, 0x49c, 0x49e,
+ 0x4a0, 0x4a2, 0x4a4, 0x4a6, 0x4a8, 0x4aa, 0x4ac, 0x4ae, 0x4b0,
+ 0x4b2, 0x4b4, 0x4b6, 0x4b8, 0x4ba, 0x4bc, 0x4be, 0x4c0, 0x4c1,
+ 0x4c3, 0x4c5, 0x4c7, 0x4c9, 0x4cb, 0x4cd, 0x4d0, 0x4d2, 0x4d4,
+ 0x4d6, 0x4d8, 0x4da, 0x4dc, 0x4de, 0x4e0, 0x4e2, 0x4e4, 0x4e6,
+ 0x4e8, 0x4ea, 0x4ec, 0x4ee, 0x4f0, 0x4f2, 0x4f4, 0x4f6, 0x4f8,
+ 0x4fa, 0x4fc, 0x4fe, 0x500, 0x502, 0x504, 0x506, 0x508, 0x50a,
+ 0x50c, 0x50e, 0x510, 0x512, 0x514, 0x516, 0x518, 0x51a, 0x51c,
+ 0x51e, 0x520, 0x522, 0x524, 0x526, 0x10c7, 0x10cd, 0x1e00, 0x1e02,
+ 0x1e04, 0x1e06, 0x1e08, 0x1e0a, 0x1e0c, 0x1e0e, 0x1e10, 0x1e12, 0x1e14,
+ 0x1e16, 0x1e18, 0x1e1a, 0x1e1c, 0x1e1e, 0x1e20, 0x1e22, 0x1e24, 0x1e26,
+ 0x1e28, 0x1e2a, 0x1e2c, 0x1e2e, 0x1e30, 0x1e32, 0x1e34, 0x1e36, 0x1e38,
+ 0x1e3a, 0x1e3c, 0x1e3e, 0x1e40, 0x1e42, 0x1e44, 0x1e46, 0x1e48, 0x1e4a,
+ 0x1e4c, 0x1e4e, 0x1e50, 0x1e52, 0x1e54, 0x1e56, 0x1e58, 0x1e5a, 0x1e5c,
+ 0x1e5e, 0x1e60, 0x1e62, 0x1e64, 0x1e66, 0x1e68, 0x1e6a, 0x1e6c, 0x1e6e,
+ 0x1e70, 0x1e72, 0x1e74, 0x1e76, 0x1e78, 0x1e7a, 0x1e7c, 0x1e7e, 0x1e80,
+ 0x1e82, 0x1e84, 0x1e86, 0x1e88, 0x1e8a, 0x1e8c, 0x1e8e, 0x1e90, 0x1e92,
+ 0x1e94, 0x1e9e, 0x1ea0, 0x1ea2, 0x1ea4, 0x1ea6, 0x1ea8, 0x1eaa, 0x1eac,
+ 0x1eae, 0x1eb0, 0x1eb2, 0x1eb4, 0x1eb6, 0x1eb8, 0x1eba, 0x1ebc, 0x1ebe,
+ 0x1ec0, 0x1ec2, 0x1ec4, 0x1ec6, 0x1ec8, 0x1eca, 0x1ecc, 0x1ece, 0x1ed0,
+ 0x1ed2, 0x1ed4, 0x1ed6, 0x1ed8, 0x1eda, 0x1edc, 0x1ede, 0x1ee0, 0x1ee2,
+ 0x1ee4, 0x1ee6, 0x1ee8, 0x1eea, 0x1eec, 0x1eee, 0x1ef0, 0x1ef2, 0x1ef4,
+ 0x1ef6, 0x1ef8, 0x1efa, 0x1efc, 0x1efe, 0x1f59, 0x1f5b, 0x1f5d, 0x1f5f,
+ 0x2102, 0x2107, 0x2115, 0x2124, 0x2126, 0x2128, 0x213e, 0x213f, 0x2145,
+ 0x2183, 0x2c60, 0x2c67, 0x2c69, 0x2c6b, 0x2c72, 0x2c75, 0x2c82, 0x2c84,
+ 0x2c86, 0x2c88, 0x2c8a, 0x2c8c, 0x2c8e, 0x2c90, 0x2c92, 0x2c94, 0x2c96,
+ 0x2c98, 0x2c9a, 0x2c9c, 0x2c9e, 0x2ca0, 0x2ca2, 0x2ca4, 0x2ca6, 0x2ca8,
+ 0x2caa, 0x2cac, 0x2cae, 0x2cb0, 0x2cb2, 0x2cb4, 0x2cb6, 0x2cb8, 0x2cba,
+ 0x2cbc, 0x2cbe, 0x2cc0, 0x2cc2, 0x2cc4, 0x2cc6, 0x2cc8, 0x2cca, 0x2ccc,
+ 0x2cce, 0x2cd0, 0x2cd2, 0x2cd4, 0x2cd6, 0x2cd8, 0x2cda, 0x2cdc, 0x2cde,
+ 0x2ce0, 0x2ce2, 0x2ceb, 0x2ced, 0x2cf2, 0xa640, 0xa642, 0xa644, 0xa646,
+ 0xa648, 0xa64a, 0xa64c, 0xa64e, 0xa650, 0xa652, 0xa654, 0xa656, 0xa658,
+ 0xa65a, 0xa65c, 0xa65e, 0xa660, 0xa662, 0xa664, 0xa666, 0xa668, 0xa66a,
+ 0xa66c, 0xa680, 0xa682, 0xa684, 0xa686, 0xa688, 0xa68a, 0xa68c, 0xa68e,
+ 0xa690, 0xa692, 0xa694, 0xa696, 0xa722, 0xa724, 0xa726, 0xa728, 0xa72a,
+ 0xa72c, 0xa72e, 0xa732, 0xa734, 0xa736, 0xa738, 0xa73a, 0xa73c, 0xa73e,
+ 0xa740, 0xa742, 0xa744, 0xa746, 0xa748, 0xa74a, 0xa74c, 0xa74e, 0xa750,
+ 0xa752, 0xa754, 0xa756, 0xa758, 0xa75a, 0xa75c, 0xa75e, 0xa760, 0xa762,
+ 0xa764, 0xa766, 0xa768, 0xa76a, 0xa76c, 0xa76e, 0xa779, 0xa77b, 0xa77d,
+ 0xa77e, 0xa780, 0xa782, 0xa784, 0xa786, 0xa78b, 0xa78d, 0xa790, 0xa792,
+ 0xa7a0, 0xa7a2, 0xa7a4, 0xa7a6, 0xa7a8, 0xa7aa
+#if TCL_UTF_MAX > 4
+ ,0x1d49c, 0x1d49e, 0x1d49f, 0x1d4a2, 0x1d4a5, 0x1d4a6, 0x1d504, 0x1d505, 0x1d538,
+ 0x1d539, 0x1d546, 0x1d7ca
+#endif
+};
+
+#define NUM_UPPER_CHAR (sizeof(upperCharTable)/sizeof(chr))
/*
- - allmcces - return a cvec with all the MCCEs of the locale
- ^ static struct cvec *allmcces(struct vars *, struct cvec *);
+ * Unicode: unicode print characters excluding space.
*/
-static struct cvec *
-allmcces(v, cv)
-struct vars *v;
-struct cvec *cv; /* this is supposed to have enough room */
-{
- return clearcvec(cv);
-}
+static const crange graphRangeTable[] = {
+ {0x21, 0x7e}, {0xa1, 0xac}, {0xae, 0x377}, {0x37a, 0x37e},
+ {0x384, 0x38a}, {0x38e, 0x3a1}, {0x3a3, 0x527}, {0x531, 0x556},
+ {0x559, 0x55f}, {0x561, 0x587}, {0x591, 0x5c7}, {0x5d0, 0x5ea},
+ {0x5f0, 0x5f4}, {0x606, 0x61b}, {0x61e, 0x6dc}, {0x6de, 0x70d},
+ {0x710, 0x74a}, {0x74d, 0x7b1}, {0x7c0, 0x7fa}, {0x800, 0x82d},
+ {0x830, 0x83e}, {0x840, 0x85b}, {0x8a2, 0x8ac}, {0x8e4, 0x8fe},
+ {0x900, 0x977}, {0x979, 0x97f}, {0x981, 0x983}, {0x985, 0x98c},
+ {0x993, 0x9a8}, {0x9aa, 0x9b0}, {0x9b6, 0x9b9}, {0x9bc, 0x9c4},
+ {0x9cb, 0x9ce}, {0x9df, 0x9e3}, {0x9e6, 0x9fb}, {0xa01, 0xa03},
+ {0xa05, 0xa0a}, {0xa13, 0xa28}, {0xa2a, 0xa30}, {0xa3e, 0xa42},
+ {0xa4b, 0xa4d}, {0xa59, 0xa5c}, {0xa66, 0xa75}, {0xa81, 0xa83},
+ {0xa85, 0xa8d}, {0xa8f, 0xa91}, {0xa93, 0xaa8}, {0xaaa, 0xab0},
+ {0xab5, 0xab9}, {0xabc, 0xac5}, {0xac7, 0xac9}, {0xacb, 0xacd},
+ {0xae0, 0xae3}, {0xae6, 0xaf1}, {0xb01, 0xb03}, {0xb05, 0xb0c},
+ {0xb13, 0xb28}, {0xb2a, 0xb30}, {0xb35, 0xb39}, {0xb3c, 0xb44},
+ {0xb4b, 0xb4d}, {0xb5f, 0xb63}, {0xb66, 0xb77}, {0xb85, 0xb8a},
+ {0xb8e, 0xb90}, {0xb92, 0xb95}, {0xba8, 0xbaa}, {0xbae, 0xbb9},
+ {0xbbe, 0xbc2}, {0xbc6, 0xbc8}, {0xbca, 0xbcd}, {0xbe6, 0xbfa},
+ {0xc01, 0xc03}, {0xc05, 0xc0c}, {0xc0e, 0xc10}, {0xc12, 0xc28},
+ {0xc2a, 0xc33}, {0xc35, 0xc39}, {0xc3d, 0xc44}, {0xc46, 0xc48},
+ {0xc4a, 0xc4d}, {0xc60, 0xc63}, {0xc66, 0xc6f}, {0xc78, 0xc7f},
+ {0xc85, 0xc8c}, {0xc8e, 0xc90}, {0xc92, 0xca8}, {0xcaa, 0xcb3},
+ {0xcb5, 0xcb9}, {0xcbc, 0xcc4}, {0xcc6, 0xcc8}, {0xcca, 0xccd},
+ {0xce0, 0xce3}, {0xce6, 0xcef}, {0xd05, 0xd0c}, {0xd0e, 0xd10},
+ {0xd12, 0xd3a}, {0xd3d, 0xd44}, {0xd46, 0xd48}, {0xd4a, 0xd4e},
+ {0xd60, 0xd63}, {0xd66, 0xd75}, {0xd79, 0xd7f}, {0xd85, 0xd96},
+ {0xd9a, 0xdb1}, {0xdb3, 0xdbb}, {0xdc0, 0xdc6}, {0xdcf, 0xdd4},
+ {0xdd8, 0xddf}, {0xdf2, 0xdf4}, {0xe01, 0xe3a}, {0xe3f, 0xe5b},
+ {0xe94, 0xe97}, {0xe99, 0xe9f}, {0xea1, 0xea3}, {0xead, 0xeb9},
+ {0xebb, 0xebd}, {0xec0, 0xec4}, {0xec8, 0xecd}, {0xed0, 0xed9},
+ {0xedc, 0xedf}, {0xf00, 0xf47}, {0xf49, 0xf6c}, {0xf71, 0xf97},
+ {0xf99, 0xfbc}, {0xfbe, 0xfcc}, {0xfce, 0xfda}, {0x1000, 0x10c5},
+ {0x10d0, 0x1248}, {0x124a, 0x124d}, {0x1250, 0x1256}, {0x125a, 0x125d},
+ {0x1260, 0x1288}, {0x128a, 0x128d}, {0x1290, 0x12b0}, {0x12b2, 0x12b5},
+ {0x12b8, 0x12be}, {0x12c2, 0x12c5}, {0x12c8, 0x12d6}, {0x12d8, 0x1310},
+ {0x1312, 0x1315}, {0x1318, 0x135a}, {0x135d, 0x137c}, {0x1380, 0x1399},
+ {0x13a0, 0x13f4}, {0x1400, 0x167f}, {0x1681, 0x169c}, {0x16a0, 0x16f0},
+ {0x1700, 0x170c}, {0x170e, 0x1714}, {0x1720, 0x1736}, {0x1740, 0x1753},
+ {0x1760, 0x176c}, {0x176e, 0x1770}, {0x1780, 0x17dd}, {0x17e0, 0x17e9},
+ {0x17f0, 0x17f9}, {0x1800, 0x180d}, {0x1810, 0x1819}, {0x1820, 0x1877},
+ {0x1880, 0x18aa}, {0x18b0, 0x18f5}, {0x1900, 0x191c}, {0x1920, 0x192b},
+ {0x1930, 0x193b}, {0x1944, 0x196d}, {0x1970, 0x1974}, {0x1980, 0x19ab},
+ {0x19b0, 0x19c9}, {0x19d0, 0x19da}, {0x19de, 0x1a1b}, {0x1a1e, 0x1a5e},
+ {0x1a60, 0x1a7c}, {0x1a7f, 0x1a89}, {0x1a90, 0x1a99}, {0x1aa0, 0x1aad},
+ {0x1b00, 0x1b4b}, {0x1b50, 0x1b7c}, {0x1b80, 0x1bf3}, {0x1bfc, 0x1c37},
+ {0x1c3b, 0x1c49}, {0x1c4d, 0x1c7f}, {0x1cc0, 0x1cc7}, {0x1cd0, 0x1cf6},
+ {0x1d00, 0x1de6}, {0x1dfc, 0x1f15}, {0x1f18, 0x1f1d}, {0x1f20, 0x1f45},
+ {0x1f48, 0x1f4d}, {0x1f50, 0x1f57}, {0x1f5f, 0x1f7d}, {0x1f80, 0x1fb4},
+ {0x1fb6, 0x1fc4}, {0x1fc6, 0x1fd3}, {0x1fd6, 0x1fdb}, {0x1fdd, 0x1fef},
+ {0x1ff2, 0x1ff4}, {0x1ff6, 0x1ffe}, {0x2010, 0x2027}, {0x2030, 0x205e},
+ {0x2074, 0x208e}, {0x2090, 0x209c}, {0x20a0, 0x20ba}, {0x20d0, 0x20f0},
+ {0x2100, 0x2189}, {0x2190, 0x23f3}, {0x2400, 0x2426}, {0x2440, 0x244a},
+ {0x2460, 0x26ff}, {0x2701, 0x2b4c}, {0x2b50, 0x2b59}, {0x2c00, 0x2c2e},
+ {0x2c30, 0x2c5e}, {0x2c60, 0x2cf3}, {0x2cf9, 0x2d25}, {0x2d30, 0x2d67},
+ {0x2d7f, 0x2d96}, {0x2da0, 0x2da6}, {0x2da8, 0x2dae}, {0x2db0, 0x2db6},
+ {0x2db8, 0x2dbe}, {0x2dc0, 0x2dc6}, {0x2dc8, 0x2dce}, {0x2dd0, 0x2dd6},
+ {0x2dd8, 0x2dde}, {0x2de0, 0x2e3b}, {0x2e80, 0x2e99}, {0x2e9b, 0x2ef3},
+ {0x2f00, 0x2fd5}, {0x2ff0, 0x2ffb}, {0x3001, 0x303f}, {0x3041, 0x3096},
+ {0x3099, 0x30ff}, {0x3105, 0x312d}, {0x3131, 0x318e}, {0x3190, 0x31ba},
+ {0x31c0, 0x31e3}, {0x31f0, 0x321e}, {0x3220, 0x32fe}, {0x3300, 0x4db5},
+ {0x4dc0, 0x9fcc}, {0xa000, 0xa48c}, {0xa490, 0xa4c6}, {0xa4d0, 0xa62b},
+ {0xa640, 0xa697}, {0xa69f, 0xa6f7}, {0xa700, 0xa78e}, {0xa790, 0xa793},
+ {0xa7a0, 0xa7aa}, {0xa7f8, 0xa82b}, {0xa830, 0xa839}, {0xa840, 0xa877},
+ {0xa880, 0xa8c4}, {0xa8ce, 0xa8d9}, {0xa8e0, 0xa8fb}, {0xa900, 0xa953},
+ {0xa95f, 0xa97c}, {0xa980, 0xa9cd}, {0xa9cf, 0xa9d9}, {0xaa00, 0xaa36},
+ {0xaa40, 0xaa4d}, {0xaa50, 0xaa59}, {0xaa5c, 0xaa7b}, {0xaa80, 0xaac2},
+ {0xaadb, 0xaaf6}, {0xab01, 0xab06}, {0xab09, 0xab0e}, {0xab11, 0xab16},
+ {0xab20, 0xab26}, {0xab28, 0xab2e}, {0xabc0, 0xabed}, {0xabf0, 0xabf9},
+ {0xac00, 0xd7a3}, {0xd7b0, 0xd7c6}, {0xd7cb, 0xd7fb}, {0xf900, 0xfa6d},
+ {0xfa70, 0xfad9}, {0xfb00, 0xfb06}, {0xfb13, 0xfb17}, {0xfb1d, 0xfb36},
+ {0xfb38, 0xfb3c}, {0xfb46, 0xfbc1}, {0xfbd3, 0xfd3f}, {0xfd50, 0xfd8f},
+ {0xfd92, 0xfdc7}, {0xfdf0, 0xfdfd}, {0xfe00, 0xfe19}, {0xfe20, 0xfe26},
+ {0xfe30, 0xfe52}, {0xfe54, 0xfe66}, {0xfe68, 0xfe6b}, {0xfe70, 0xfe74},
+ {0xfe76, 0xfefc}, {0xff01, 0xffbe}, {0xffc2, 0xffc7}, {0xffca, 0xffcf},
+ {0xffd2, 0xffd7}, {0xffda, 0xffdc}, {0xffe0, 0xffe6}, {0xffe8, 0xffee}
+#if TCL_UTF_MAX > 4
+ ,{0x10000, 0x1000b}, {0x1000d, 0x10026}, {0x10028, 0x1003a}, {0x1003f, 0x1004d},
+ {0x10050, 0x1005d}, {0x10080, 0x100fa}, {0x10100, 0x10102}, {0x10107, 0x10133},
+ {0x10137, 0x1018a}, {0x10190, 0x1019b}, {0x101d0, 0x101fd}, {0x10280, 0x1029c},
+ {0x102a0, 0x102d0}, {0x10300, 0x1031e}, {0x10320, 0x10323}, {0x10330, 0x1034a},
+ {0x10380, 0x1039d}, {0x1039f, 0x103c3}, {0x103c8, 0x103d5}, {0x10400, 0x1049d},
+ {0x104a0, 0x104a9}, {0x10800, 0x10805}, {0x1080a, 0x10835}, {0x1083f, 0x10855},
+ {0x10857, 0x1085f}, {0x10900, 0x1091b}, {0x1091f, 0x10939}, {0x10980, 0x109b7},
+ {0x10a00, 0x10a03}, {0x10a0c, 0x10a13}, {0x10a15, 0x10a17}, {0x10a19, 0x10a33},
+ {0x10a38, 0x10a3a}, {0x10a3f, 0x10a47}, {0x10a50, 0x10a58}, {0x10a60, 0x10a7f},
+ {0x10b00, 0x10b35}, {0x10b39, 0x10b55}, {0x10b58, 0x10b72}, {0x10b78, 0x10b7f},
+ {0x10c00, 0x10c48}, {0x10e60, 0x10e7e}, {0x11000, 0x1104d}, {0x11052, 0x1106f},
+ {0x11080, 0x110bc}, {0x110be, 0x110c1}, {0x110d0, 0x110e8}, {0x110f0, 0x110f9},
+ {0x11100, 0x11134}, {0x11136, 0x11143}, {0x11180, 0x111c8}, {0x111d0, 0x111d9},
+ {0x11680, 0x116b7}, {0x116c0, 0x116c9}, {0x12000, 0x1236e}, {0x12400, 0x12462},
+ {0x12470, 0x12473}, {0x13000, 0x1342e}, {0x16800, 0x16a38}, {0x16f00, 0x16f44},
+ {0x16f50, 0x16f7e}, {0x16f8f, 0x16f9f}, {0x1d000, 0x1d0f5}, {0x1d100, 0x1d126},
+ {0x1d129, 0x1d172}, {0x1d17b, 0x1d1dd}, {0x1d200, 0x1d245}, {0x1d300, 0x1d356},
+ {0x1d360, 0x1d371}, {0x1d400, 0x1d454}, {0x1d456, 0x1d49c}, {0x1d4a9, 0x1d4ac},
+ {0x1d4ae, 0x1d4b9}, {0x1d4bd, 0x1d4c3}, {0x1d4c5, 0x1d505}, {0x1d507, 0x1d50a},
+ {0x1d50d, 0x1d514}, {0x1d516, 0x1d51c}, {0x1d51e, 0x1d539}, {0x1d53b, 0x1d53e},
+ {0x1d540, 0x1d544}, {0x1d54a, 0x1d550}, {0x1d552, 0x1d6a5}, {0x1d6a8, 0x1d7cb},
+ {0x1d7ce, 0x1d7ff}, {0x1ee00, 0x1ee03}, {0x1ee05, 0x1ee1f}, {0x1ee29, 0x1ee32},
+ {0x1ee34, 0x1ee37}, {0x1ee4d, 0x1ee4f}, {0x1ee67, 0x1ee6a}, {0x1ee6c, 0x1ee72},
+ {0x1ee74, 0x1ee77}, {0x1ee79, 0x1ee7c}, {0x1ee80, 0x1ee89}, {0x1ee8b, 0x1ee9b},
+ {0x1eea1, 0x1eea3}, {0x1eea5, 0x1eea9}, {0x1eeab, 0x1eebb}, {0x1f000, 0x1f02b},
+ {0x1f030, 0x1f093}, {0x1f0a0, 0x1f0ae}, {0x1f0b1, 0x1f0be}, {0x1f0c1, 0x1f0cf},
+ {0x1f0d1, 0x1f0df}, {0x1f100, 0x1f10a}, {0x1f110, 0x1f12e}, {0x1f130, 0x1f16b},
+ {0x1f170, 0x1f19a}, {0x1f1e6, 0x1f202}, {0x1f210, 0x1f23a}, {0x1f240, 0x1f248},
+ {0x1f300, 0x1f320}, {0x1f330, 0x1f335}, {0x1f337, 0x1f37c}, {0x1f380, 0x1f393},
+ {0x1f3a0, 0x1f3c4}, {0x1f3c6, 0x1f3ca}, {0x1f3e0, 0x1f3f0}, {0x1f400, 0x1f43e},
+ {0x1f442, 0x1f4f7}, {0x1f4f9, 0x1f4fc}, {0x1f500, 0x1f53d}, {0x1f540, 0x1f543},
+ {0x1f550, 0x1f567}, {0x1f5fb, 0x1f640}, {0x1f645, 0x1f64f}, {0x1f680, 0x1f6c5},
+ {0x1f700, 0x1f773}, {0x20000, 0x2a6d6}, {0x2a700, 0x2b734}, {0x2b740, 0x2b81d},
+ {0x2f800, 0x2fa1d}, {0xe0100, 0xe01ef}
+#endif
+};
+
+#define NUM_GRAPH_RANGE (sizeof(graphRangeTable)/sizeof(crange))
+
+static const chr graphCharTable[] = {
+ 0x38c, 0x589, 0x58a, 0x58f, 0x85e, 0x8a0, 0x98f, 0x990, 0x9b2,
+ 0x9c7, 0x9c8, 0x9d7, 0x9dc, 0x9dd, 0xa0f, 0xa10, 0xa32, 0xa33,
+ 0xa35, 0xa36, 0xa38, 0xa39, 0xa3c, 0xa47, 0xa48, 0xa51, 0xa5e,
+ 0xab2, 0xab3, 0xad0, 0xb0f, 0xb10, 0xb32, 0xb33, 0xb47, 0xb48,
+ 0xb56, 0xb57, 0xb5c, 0xb5d, 0xb82, 0xb83, 0xb99, 0xb9a, 0xb9c,
+ 0xb9e, 0xb9f, 0xba3, 0xba4, 0xbd0, 0xbd7, 0xc55, 0xc56, 0xc58,
+ 0xc59, 0xc82, 0xc83, 0xcd5, 0xcd6, 0xcde, 0xcf1, 0xcf2, 0xd02,
+ 0xd03, 0xd57, 0xd82, 0xd83, 0xdbd, 0xdca, 0xdd6, 0xe81, 0xe82,
+ 0xe84, 0xe87, 0xe88, 0xe8a, 0xe8d, 0xea5, 0xea7, 0xeaa, 0xeab,
+ 0xec6, 0x10c7, 0x10cd, 0x1258, 0x12c0, 0x1772, 0x1773, 0x1940, 0x1f59,
+ 0x1f5b, 0x1f5d, 0x2070, 0x2071, 0x2d27, 0x2d2d, 0x2d6f, 0x2d70, 0xa9de,
+ 0xa9df, 0xfb3e, 0xfb40, 0xfb41, 0xfb43, 0xfb44, 0xfffc, 0xfffd
+#if TCL_UTF_MAX > 4
+ ,0x1003c, 0x1003d, 0x10808, 0x10837, 0x10838, 0x1083c, 0x1093f, 0x109be, 0x109bf,
+ 0x10a05, 0x10a06, 0x1b000, 0x1b001, 0x1d49e, 0x1d49f, 0x1d4a2, 0x1d4a5, 0x1d4a6,
+ 0x1d4bb, 0x1d546, 0x1ee21, 0x1ee22, 0x1ee24, 0x1ee27, 0x1ee39, 0x1ee3b, 0x1ee42,
+ 0x1ee47, 0x1ee49, 0x1ee4b, 0x1ee51, 0x1ee52, 0x1ee54, 0x1ee57, 0x1ee59, 0x1ee5b,
+ 0x1ee5d, 0x1ee5f, 0x1ee61, 0x1ee62, 0x1ee64, 0x1ee7e, 0x1eef0, 0x1eef1, 0x1f250,
+ 0x1f251, 0x1f440
+#endif
+};
+
+#define NUM_GRAPH_CHAR (sizeof(graphCharTable)/sizeof(chr))
+
+/*
+ * End of auto-generated Unicode character ranges declarations.
+ */
+
+#define CH NOCELT
+
/*
- element - map collating-element name to celt
- ^ static celt element(struct vars *, chr *, chr *);
+ ^ static celt element(struct vars *, const chr *, const chr *);
*/
static celt
-element(v, startp, endp)
-struct vars *v;
-chr *startp; /* points to start of name */
-chr *endp; /* points just past end of name */
+element(
+ struct vars *v, /* context */
+ const chr *startp, /* points to start of name */
+ const chr *endp) /* points just past end of name */
{
- struct cname *cn;
- size_t len;
- Tcl_DString ds;
- char *np;
-
- /* generic: one-chr names stand for themselves */
- assert(startp < endp);
- len = endp - startp;
- if (len == 1)
- return *startp;
-
- NOTE(REG_ULOCALE);
-
- /* search table */
- Tcl_DStringInit(&ds);
- np = Tcl_UniCharToUtfDString(startp, (int)len, &ds);
- for (cn = cnames; cn->name != NULL; cn++)
- if (strlen(cn->name) == len && strncmp(cn->name, np, len) == 0)
- break; /* NOTE BREAK OUT */
- Tcl_DStringFree(&ds);
- if (cn->name != NULL)
- return CHR(cn->code);
-
- /* couldn't find it */
- ERR(REG_ECOLLATE);
- return 0;
-}
+ const struct cname *cn;
+ size_t len;
+ Tcl_DString ds;
+ const char *np;
+
+ /*
+ * Generic: one-chr names stand for themselves.
+ */
+ assert(startp < endp);
+ len = endp - startp;
+ if (len == 1) {
+ return *startp;
+ }
+
+ NOTE(REG_ULOCALE);
+
+ /*
+ * Search table.
+ */
+
+ Tcl_DStringInit(&ds);
+ np = Tcl_UniCharToUtfDString(startp, (int)len, &ds);
+ for (cn=cnames; cn->name!=NULL; cn++) {
+ if (strlen(cn->name)==len && strncmp(cn->name, np, len)==0) {
+ break; /* NOTE BREAK OUT */
+ }
+ }
+ Tcl_DStringFree(&ds);
+ if (cn->name != NULL) {
+ return CHR(cn->code);
+ }
+
+ /*
+ * Couldn't find it.
+ */
+
+ ERR(REG_ECOLLATE);
+ return 0;
+}
+
/*
- range - supply cvec for a range, including legality check
^ static struct cvec *range(struct vars *, celt, celt, int);
*/
static struct cvec *
-range(v, a, b, cases)
-struct vars *v;
-celt a;
-celt b; /* might equal a */
-int cases; /* case-independent? */
+range(
+ struct vars *v, /* context */
+ celt a, /* range start */
+ celt b, /* range end, might equal a */
+ int cases) /* case-independent? */
{
- int nchrs;
- struct cvec *cv;
- celt c, lc, uc, tc;
+ int nchrs;
+ struct cvec *cv;
+ celt c, lc, uc, tc;
- if (a != b && !before(a, b)) {
- ERR(REG_ERANGE);
- return NULL;
- }
+ if (a != b && !before(a, b)) {
+ ERR(REG_ERANGE);
+ return NULL;
+ }
- if (!cases) { /* easy version */
- cv = getcvec(v, 0, 1, 0);
- NOERRN();
- addrange(cv, a, b);
- return cv;
- }
+ if (!cases) { /* easy version */
+ cv = getcvec(v, 0, 1);
+ NOERRN();
+ addrange(cv, a, b);
+ return cv;
+ }
- /*
- * When case-independent, it's hard to decide when cvec ranges are
- * usable, so for now at least, we won't try. We allocate enough
- * space for two case variants plus a little extra for the two
- * title case variants.
- */
+ /*
+ * When case-independent, it's hard to decide when cvec ranges are usable,
+ * so for now at least, we won't try. We allocate enough space for two
+ * case variants plus a little extra for the two title case variants.
+ */
- nchrs = (b - a + 1)*2 + 4;
+ nchrs = (b - a + 1)*2 + 4;
- cv = getcvec(v, nchrs, 0, 0);
- NOERRN();
+ cv = getcvec(v, nchrs, 0);
+ NOERRN();
- for (c = a; c <= b; c++) {
- addchr(cv, c);
- lc = Tcl_UniCharToLower((chr)c);
- uc = Tcl_UniCharToUpper((chr)c);
- tc = Tcl_UniCharToTitle((chr)c);
- if (c != lc) {
- addchr(cv, lc);
- }
- if (c != uc) {
- addchr(cv, uc);
- }
- if (c != tc && tc != uc) {
- addchr(cv, tc);
- }
+ for (c=a; c<=b; c++) {
+ addchr(cv, c);
+ lc = Tcl_UniCharToLower((chr)c);
+ uc = Tcl_UniCharToUpper((chr)c);
+ tc = Tcl_UniCharToTitle((chr)c);
+ if (c != lc) {
+ addchr(cv, lc);
+ }
+ if (c != uc) {
+ addchr(cv, uc);
}
+ if (c != tc && tc != uc) {
+ addchr(cv, tc);
+ }
+ }
- return cv;
+ return cv;
}
-
+
/*
- before - is celt x before celt y, for purposes of range legality?
^ static int before(celt, celt);
*/
static int /* predicate */
-before(x, y)
-celt x;
-celt y;
+before(
+ celt x, celt y) /* collating elements */
{
- /* trivial because no MCCEs */
- if (x < y)
- return 1;
- return 0;
+ if (x < y) {
+ return 1;
+ }
+ return 0;
}
-
+
/*
- eclass - supply cvec for an equivalence class
* Must include case counterparts on request.
^ static struct cvec *eclass(struct vars *, celt, int);
*/
static struct cvec *
-eclass(v, c, cases)
-struct vars *v;
-celt c;
-int cases; /* all cases? */
+eclass(
+ struct vars *v, /* context */
+ celt c, /* Collating element representing the
+ * equivalence class. */
+ int cases) /* all cases? */
{
- struct cvec *cv;
-
- /* crude fake equivalence class for testing */
- if ((v->cflags&REG_FAKEEC) && c == 'x') {
- cv = getcvec(v, 4, 0, 0);
- addchr(cv, (chr)'x');
- addchr(cv, (chr)'y');
- if (cases) {
- addchr(cv, (chr)'X');
- addchr(cv, (chr)'Y');
- }
- return cv;
- }
+ struct cvec *cv;
- /* otherwise, none */
- if (cases)
- return allcases(v, c);
- cv = getcvec(v, 1, 0, 0);
- assert(cv != NULL);
- addchr(cv, (chr)c);
+ /*
+ * Crude fake equivalence class for testing.
+ */
+
+ if ((v->cflags&REG_FAKE) && c == 'x') {
+ cv = getcvec(v, 4, 0);
+ addchr(cv, (chr)'x');
+ addchr(cv, (chr)'y');
+ if (cases) {
+ addchr(cv, (chr)'X');
+ addchr(cv, (chr)'Y');
+ }
return cv;
-}
+ }
+ /*
+ * Otherwise, none.
+ */
+
+ if (cases) {
+ return allcases(v, c);
+ }
+ cv = getcvec(v, 1, 0);
+ assert(cv != NULL);
+ addchr(cv, (chr)c);
+ return cv;
+}
+
/*
- cclass - supply cvec for a character class
* Must include case counterparts on request.
- ^ static struct cvec *cclass(struct vars *, chr *, chr *, int);
+ ^ static struct cvec *cclass(struct vars *, const chr *, const chr *, int);
*/
static struct cvec *
-cclass(v, startp, endp, cases)
-struct vars *v;
-chr *startp; /* where the name starts */
-chr *endp; /* just past the end of the name */
-int cases; /* case-independent? */
+cclass(
+ struct vars *v, /* context */
+ const chr *startp, /* where the name starts */
+ const chr *endp, /* just past the end of the name */
+ int cases) /* case-independent? */
{
size_t len;
struct cvec *cv = NULL;
Tcl_DString ds;
- char *np, **namePtr;
+ const char *np;
+ const char *const *namePtr;
int i, index;
/*
* The following arrays define the valid character class names.
*/
- static char *classNames[] = {
- "alnum", "alpha", "blank", "cntrl", "digit", "graph", "lower",
- "print", "punct", "space", "upper", "xdigit", NULL
+ static const char *const classNames[] = {
+ "alnum", "alpha", "ascii", "blank", "cntrl", "digit", "graph",
+ "lower", "print", "punct", "space", "upper", "xdigit", NULL
};
enum classes {
- CC_ALNUM, CC_ALPHA, CC_BLANK, CC_CNTRL, CC_DIGIT, CC_GRAPH, CC_LOWER,
- CC_PRINT, CC_PUNCT, CC_SPACE, CC_UPPER, CC_XDIGIT
+ CC_ALNUM, CC_ALPHA, CC_ASCII, CC_BLANK, CC_CNTRL, CC_DIGIT, CC_GRAPH,
+ CC_LOWER, CC_PRINT, CC_PUNCT, CC_SPACE, CC_UPPER, CC_XDIGIT
};
-
+
/*
* Extract the class name
@@ -574,139 +923,198 @@ int cases; /* case-independent? */
np = Tcl_UniCharToUtfDString(startp, (int)len, &ds);
/*
- * Remap lower and upper to alpha if the match is case insensitive.
- */
-
- if (cases && len == 5 && (strncmp("lower", np, 5) == 0
- || strncmp("upper", np, 5) == 0)) {
- np = "alpha";
- }
-
- /*
* Map the name to the corresponding enumerated value.
*/
index = -1;
- for (namePtr = classNames, i = 0; *namePtr != NULL; namePtr++, i++) {
+ for (namePtr=classNames,i=0 ; *namePtr!=NULL ; namePtr++,i++) {
if ((strlen(*namePtr) == len) && (strncmp(*namePtr, np, len) == 0)) {
index = i;
break;
}
}
- Tcl_DStringInit(&ds);
+ Tcl_DStringFree(&ds);
if (index == -1) {
ERR(REG_ECTYPE);
return NULL;
}
-
+
+ /*
+ * Remap lower and upper to alpha if the match is case insensitive.
+ */
+
+ if (cases && ((index == CC_LOWER) || (index == CC_UPPER))) {
+ index = CC_ALNUM;
+ }
+
/*
* Now compute the character class contents.
*/
switch((enum classes) index) {
- case CC_PRINT:
- case CC_ALNUM:
- cv = getcvec(v, 0, NUM_DIGIT + NUM_ALPHA, 0);
- if (cv) {
- for (i = 0; i < NUM_ALPHA; i++) {
- addrange(cv, alphaTable[i].start, alphaTable[i].end);
- }
- for (i = 0; i < NUM_DIGIT; i++) {
- addrange(cv, digitTable[i].start, digitTable[i].end);
- }
+ case CC_ALNUM:
+ cv = getcvec(v, NUM_ALPHA_CHAR, NUM_DIGIT_RANGE + NUM_ALPHA_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_ALPHA_CHAR ; i++) {
+ addchr(cv, alphaCharTable[i]);
}
- break;
- case CC_ALPHA:
- cv = getcvec(v, 0, NUM_ALPHA, 0);
- if (cv) {
- for (i = 0; i < NUM_ALPHA; i++) {
- addrange(cv, alphaTable[i].start, alphaTable[i].end);
- }
+ for (i=0 ; (size_t)i<NUM_ALPHA_RANGE ; i++) {
+ addrange(cv, alphaRangeTable[i].start,
+ alphaRangeTable[i].end);
}
- break;
- case CC_BLANK:
- cv = getcvec(v, 2, 0, 0);
- addchr(cv, '\t');
- addchr(cv, ' ');
- break;
- case CC_CNTRL:
- cv = getcvec(v, 0, 2, 0);
- addrange(cv, 0x0, 0x1f);
- addrange(cv, 0x7f, 0x9f);
- break;
- case CC_DIGIT:
- cv = getcvec(v, 0, NUM_DIGIT, 0);
- if (cv) {
- for (i = 0; i < NUM_DIGIT; i++) {
- addrange(cv, digitTable[i].start, digitTable[i].end);
- }
+ for (i=0 ; (size_t)i<NUM_DIGIT_RANGE ; i++) {
+ addrange(cv, digitRangeTable[i].start,
+ digitRangeTable[i].end);
}
- break;
- case CC_PUNCT:
- cv = getcvec(v, 0, NUM_PUNCT, 0);
- if (cv) {
- for (i = 0; i < NUM_PUNCT; i++) {
- addrange(cv, punctTable[i].start, punctTable[i].end);
- }
+ }
+ break;
+ case CC_ALPHA:
+ cv = getcvec(v, NUM_ALPHA_CHAR, NUM_ALPHA_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_ALPHA_RANGE ; i++) {
+ addrange(cv, alphaRangeTable[i].start,
+ alphaRangeTable[i].end);
}
- break;
- case CC_XDIGIT:
- cv = getcvec(v, 0, NUM_DIGIT+2, 0);
- if (cv) {
- for (i = 0; i < NUM_DIGIT; i++) {
- addrange(cv, digitTable[i].start, digitTable[i].end);
- }
- addrange(cv, 'a', 'f');
- addrange(cv, 'A', 'F');
+ for (i=0 ; (size_t)i<NUM_ALPHA_CHAR ; i++) {
+ addchr(cv, alphaCharTable[i]);
}
- break;
- case CC_SPACE:
- cv = getcvec(v, 0, NUM_SPACE, 0);
- if (cv) {
- for (i = 0; i < NUM_SPACE; i++) {
- addrange(cv, spaceTable[i].start, spaceTable[i].end);
- }
+ }
+ break;
+ case CC_ASCII:
+ cv = getcvec(v, 0, 1);
+ if (cv) {
+ addrange(cv, 0, 0x7f);
+ }
+ break;
+ case CC_BLANK:
+ cv = getcvec(v, 2, 0);
+ addchr(cv, '\t');
+ addchr(cv, ' ');
+ break;
+ case CC_CNTRL:
+ cv = getcvec(v, NUM_CONTROL_CHAR, NUM_CONTROL_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_CONTROL_RANGE ; i++) {
+ addrange(cv, controlRangeTable[i].start,
+ controlRangeTable[i].end);
}
- break;
- case CC_LOWER:
- cv = getcvec(v, NUM_LOWER_CHAR, NUM_LOWER_RANGE, 0);
- if (cv) {
- for (i = 0; i < NUM_LOWER_RANGE; i++) {
- addrange(cv, lowerRangeTable[i].start,
- lowerRangeTable[i].end);
- }
- for (i = 0; i < NUM_LOWER_CHAR; i++) {
- addchr(cv, lowerCharTable[i]);
- }
+ for (i=0 ; (size_t)i<NUM_CONTROL_CHAR ; i++) {
+ addchr(cv, controlCharTable[i]);
}
- break;
- case CC_UPPER:
- cv = getcvec(v, NUM_UPPER_CHAR, NUM_UPPER_RANGE, 0);
- if (cv) {
- for (i = 0; i < NUM_UPPER_RANGE; i++) {
- addrange(cv, upperRangeTable[i].start,
- upperRangeTable[i].end);
- }
- for (i = 0; i < NUM_UPPER_CHAR; i++) {
- addchr(cv, upperCharTable[i]);
- }
+ }
+ break;
+ case CC_DIGIT:
+ cv = getcvec(v, 0, NUM_DIGIT_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_DIGIT_RANGE ; i++) {
+ addrange(cv, digitRangeTable[i].start,
+ digitRangeTable[i].end);
}
- break;
- case CC_GRAPH:
- cv = getcvec(v, 0, NUM_GRAPH, 0);
- if (cv) {
- for (i = 0; i < NUM_GRAPH; i++) {
- addrange(cv, graphTable[i].start, graphTable[i].end);
- }
+ }
+ break;
+ case CC_PUNCT:
+ cv = getcvec(v, NUM_PUNCT_CHAR, NUM_PUNCT_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_PUNCT_RANGE ; i++) {
+ addrange(cv, punctRangeTable[i].start,
+ punctRangeTable[i].end);
}
- break;
+ for (i=0 ; (size_t)i<NUM_PUNCT_CHAR ; i++) {
+ addchr(cv, punctCharTable[i]);
+ }
+ }
+ break;
+ case CC_XDIGIT:
+ /*
+ * This is a 3 instead of (NUM_DIGIT_RANGE+2) because I've no idea how
+ * to define the digits 'a' through 'f' in non-western locales. The
+ * concept is quite possibly non portable, or only used in contextx
+ * where the characters used would be the western ones anyway!
+ * Whatever is actually the case, the number of ranges is fixed (until
+ * someone comes up with a better arrangement!)
+ */
+
+ cv = getcvec(v, 0, 3);
+ if (cv) {
+ addrange(cv, '0', '9');
+ addrange(cv, 'a', 'f');
+ addrange(cv, 'A', 'F');
+ }
+ break;
+ case CC_SPACE:
+ cv = getcvec(v, NUM_SPACE_CHAR, NUM_SPACE_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_SPACE_RANGE ; i++) {
+ addrange(cv, spaceRangeTable[i].start,
+ spaceRangeTable[i].end);
+ }
+ for (i=0 ; (size_t)i<NUM_SPACE_CHAR ; i++) {
+ addchr(cv, spaceCharTable[i]);
+ }
+ }
+ break;
+ case CC_LOWER:
+ cv = getcvec(v, NUM_LOWER_CHAR, NUM_LOWER_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_LOWER_RANGE ; i++) {
+ addrange(cv, lowerRangeTable[i].start,
+ lowerRangeTable[i].end);
+ }
+ for (i=0 ; (size_t)i<NUM_LOWER_CHAR ; i++) {
+ addchr(cv, lowerCharTable[i]);
+ }
+ }
+ break;
+ case CC_UPPER:
+ cv = getcvec(v, NUM_UPPER_CHAR, NUM_UPPER_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_UPPER_RANGE ; i++) {
+ addrange(cv, upperRangeTable[i].start,
+ upperRangeTable[i].end);
+ }
+ for (i=0 ; (size_t)i<NUM_UPPER_CHAR ; i++) {
+ addchr(cv, upperCharTable[i]);
+ }
+ }
+ break;
+ case CC_PRINT:
+ cv = getcvec(v, NUM_SPACE_CHAR + NUM_GRAPH_CHAR, NUM_SPACE_RANGE + NUM_GRAPH_RANGE - 1);
+ if (cv) {
+ for (i=1 ; (size_t)i<NUM_SPACE_RANGE ; i++) {
+ addrange(cv, spaceRangeTable[i].start,
+ spaceRangeTable[i].end);
+ }
+ for (i=0 ; (size_t)i<NUM_SPACE_CHAR ; i++) {
+ addchr(cv, spaceCharTable[i]);
+ }
+ for (i=0 ; (size_t)i<NUM_GRAPH_RANGE ; i++) {
+ addrange(cv, graphRangeTable[i].start,
+ graphRangeTable[i].end);
+ }
+ for (i=0 ; (size_t)i<NUM_GRAPH_CHAR ; i++) {
+ addchr(cv, graphCharTable[i]);
+ }
+ }
+ break;
+ case CC_GRAPH:
+ cv = getcvec(v, NUM_GRAPH_CHAR, NUM_GRAPH_RANGE);
+ if (cv) {
+ for (i=0 ; (size_t)i<NUM_GRAPH_RANGE ; i++) {
+ addrange(cv, graphRangeTable[i].start,
+ graphRangeTable[i].end);
+ }
+ for (i=0 ; (size_t)i<NUM_GRAPH_CHAR ; i++) {
+ addchr(cv, graphCharTable[i]);
+ }
+ }
+ break;
}
if (cv == NULL) {
ERR(REG_ESPACE);
}
return cv;
}
-
+
/*
- allcases - supply cvec for all case counterparts of a chr (including itself)
* This is a shortcut, preferably an efficient one, for simple characters;
@@ -714,68 +1122,72 @@ int cases; /* case-independent? */
^ static struct cvec *allcases(struct vars *, pchr);
*/
static struct cvec *
-allcases(v, pc)
-struct vars *v;
-pchr pc;
+allcases(
+ struct vars *v, /* context */
+ pchr pc) /* character to get case equivs of */
{
- struct cvec *cv;
- chr c = (chr)pc;
- chr lc, uc, tc;
+ struct cvec *cv;
+ chr c = (chr)pc;
+ chr lc, uc, tc;
- lc = Tcl_UniCharToLower((chr)c);
- uc = Tcl_UniCharToUpper((chr)c);
- tc = Tcl_UniCharToTitle((chr)c);
+ lc = Tcl_UniCharToLower((chr)c);
+ uc = Tcl_UniCharToUpper((chr)c);
+ tc = Tcl_UniCharToTitle((chr)c);
- if (tc != uc) {
- cv = getcvec(v, 3, 0, 0);
- addchr(cv, tc);
- } else {
- cv = getcvec(v, 2, 0, 0);
- }
- addchr(cv, lc);
- if (lc != uc) {
- addchr(cv, uc);
- }
- return cv;
+ if (tc != uc) {
+ cv = getcvec(v, 3, 0);
+ addchr(cv, tc);
+ } else {
+ cv = getcvec(v, 2, 0);
+ }
+ addchr(cv, lc);
+ if (lc != uc) {
+ addchr(cv, uc);
+ }
+ return cv;
}
-
+
/*
- cmp - chr-substring compare
* Backrefs need this. It should preferably be efficient.
* Note that it does not need to report anything except equal/unequal.
* Note also that the length is exact, and the comparison should not
* stop at embedded NULs!
- ^ static int cmp(CONST chr *, CONST chr *, size_t);
+ ^ static int cmp(const chr *, const chr *, size_t);
*/
static int /* 0 for equal, nonzero for unequal */
-cmp(x, y, len)
-CONST chr *x;
-CONST chr *y;
-size_t len; /* exact length of comparison */
+cmp(
+ const chr *x, const chr *y, /* strings to compare */
+ size_t len) /* exact length of comparison */
{
- return memcmp(VS(x), VS(y), len*sizeof(chr));
+ return memcmp(VS(x), VS(y), len*sizeof(chr));
}
-
+
/*
- casecmp - case-independent chr-substring compare
* REG_ICASE backrefs need this. It should preferably be efficient.
* Note that it does not need to report anything except equal/unequal.
* Note also that the length is exact, and the comparison should not
* stop at embedded NULs!
- ^ static int casecmp(CONST chr *, CONST chr *, size_t);
+ ^ static int casecmp(const chr *, const chr *, size_t);
*/
static int /* 0 for equal, nonzero for unequal */
-casecmp(x, y, len)
-CONST chr *x;
-CONST chr *y;
-size_t len; /* exact length of comparison */
+casecmp(
+ const chr *x, const chr *y, /* strings to compare */
+ size_t len) /* exact length of comparison */
{
- size_t i;
- CONST chr *xp;
- CONST chr *yp;
-
- for (xp = x, yp = y, i = len; i > 0; i--)
- if (Tcl_UniCharToLower(*xp++) != Tcl_UniCharToLower(*yp++))
- return 1;
- return 0;
+ for (; len > 0; len--, x++, y++) {
+ if ((*x!=*y) && (Tcl_UniCharToLower(*x) != Tcl_UniCharToLower(*y))) {
+ return 1;
+ }
+ }
+ return 0;
}
+
+/*
+ * Local Variables:
+ * mode: c
+ * c-basic-offset: 4
+ * fill-column: 78
+ * End:
+ */