diff --git a/genucd.py b/genucd.py new file mode 100644 index 0000000..c29d422 --- /dev/null +++ b/genucd.py @@ -0,0 +1,77 @@ +# Create utfdata.h from UnicodeData.txt + +tolower = [] +toupper = [] +isalpha = [] + +for line in open("UnicodeData.txt").readlines(): + line = line.split(";") + code = int(line[0],16) + # if code > 65535: continue # skip non-BMP codepoints + if line[2][0] == 'L': + isalpha.append(code) + if line[12]: + toupper.append((code,int(line[12],16))) + if line[13]: + tolower.append((code,int(line[13],16))) + +def dumpalpha(): + table = [] + prev = 0 + start = 0 + for code in isalpha: + if code != prev+1: + if start: + table.append((start,prev)) + start = code + prev = code + table.append((start,prev)) + + print("") + print("static const Rune ucd_alpha2[] = {") + for a, b in table: + if b - a > 0: + print(hex(a)+","+hex(b)+",") + print("};"); + + print("") + print("static const Rune ucd_alpha1[] = {") + for a, b in table: + if b - a == 0: + print(hex(a)+",") + print("};"); + +def dumpmap(name, input): + table = [] + prev_a = 0 + prev_b = 0 + start_a = 0 + start_b = 0 + for a, b in input: + if a != prev_a+1 or b != prev_b+1: + if start_a: + table.append((start_a,prev_a,start_b)) + start_a = a + start_b = b + prev_a = a + prev_b = b + table.append((start_a,prev_a,start_b)) + + print("") + print("static const Rune " + name + "2[] = {") + for a, b, n in table: + if b - a > 0: + print(hex(a)+","+hex(b)+","+str(n-a)+",") + print("};"); + + print("") + print("static const Rune " + name + "1[] = {") + for a, b, n in table: + if b - a == 0: + print(hex(a)+","+str(n-a)+",") + print("};"); + +print("/* This file was automatically created from UnicodeData.txt */") +dumpalpha() +dumpmap("ucd_tolower", tolower) +dumpmap("ucd_toupper", toupper) diff --git a/one.c b/one.c index 3eebe08..11ae034 100644 --- a/one.c +++ b/one.c @@ -24,4 +24,3 @@ #include "jsvalue.c" #include "regexp.c" #include "utf.c" -#include "utftype.c" diff --git a/utf.c b/utf.c index 3c97ce3..567790a 100644 --- a/utf.c +++ b/utf.c @@ -15,6 +15,9 @@ #include #include "utf.h" +#include "utfdata.h" + +#define nelem(a) (int)(sizeof (a) / sizeof (a)[0]) typedef unsigned char uchar; @@ -210,3 +213,93 @@ utflen(const char *s) n++; } } + +static const Rune * +ucd_bsearch(Rune c, const Rune *t, int n, int ne) +{ + const Rune *p; + int m; + + while(n > 1) { + m = n/2; + p = t + m*ne; + if(c >= p[0]) { + t = p; + n = n-m; + } else + n = m; + } + if(n && c >= t[0]) + return t; + return 0; +} + +Rune +tolowerrune(Rune c) +{ + const Rune *p; + + p = ucd_bsearch(c, ucd_tolower2, nelem(ucd_tolower2)/3, 3); + if(p && c >= p[0] && c <= p[1]) + return c + p[2]; + p = ucd_bsearch(c, ucd_tolower1, nelem(ucd_tolower1)/2, 2); + if(p && c == p[0]) + return c + p[1]; + return c; +} + +Rune +toupperrune(Rune c) +{ + const Rune *p; + + p = ucd_bsearch(c, ucd_toupper2, nelem(ucd_toupper2)/3, 3); + if(p && c >= p[0] && c <= p[1]) + return c + p[2]; + p = ucd_bsearch(c, ucd_toupper1, nelem(ucd_toupper1)/2, 2); + if(p && c == p[0]) + return c + p[1]; + return c; +} + +int +islowerrune(Rune c) +{ + const Rune *p; + + p = ucd_bsearch(c, ucd_toupper2, nelem(ucd_toupper2)/3, 3); + if(p && c >= p[0] && c <= p[1]) + return 1; + p = ucd_bsearch(c, ucd_toupper1, nelem(ucd_toupper1)/2, 2); + if(p && c == p[0]) + return 1; + return 0; +} + +int +isupperrune(Rune c) +{ + const Rune *p; + + p = ucd_bsearch(c, ucd_tolower2, nelem(ucd_tolower2)/3, 3); + if(p && c >= p[0] && c <= p[1]) + return 1; + p = ucd_bsearch(c, ucd_tolower1, nelem(ucd_tolower1)/2, 2); + if(p && c == p[0]) + return 1; + return 0; +} + +int +isalpharune(Rune c) +{ + const Rune *p; + + p = ucd_bsearch(c, ucd_alpha2, nelem(ucd_alpha2)/2, 2); + if(p && c >= p[0] && c <= p[1]) + return 1; + p = ucd_bsearch(c, ucd_alpha1, nelem(ucd_alpha1), 1); + if(p && c == p[0]) + return 1; + return 0; +} diff --git a/utf.h b/utf.h index e612918..638ba6e 100644 --- a/utf.h +++ b/utf.h @@ -23,11 +23,8 @@ typedef int Rune; /* 32 bits */ #define isalpharune jsU_isalpharune #define islowerrune jsU_islowerrune -#define isspacerune jsU_isspacerune -#define istitlerune jsU_istitlerune #define isupperrune jsU_isupperrune #define tolowerrune jsU_tolowerrune -#define totitlerune jsU_totitlerune #define toupperrune jsU_toupperrune enum @@ -46,11 +43,8 @@ int utflen(const char *s); int isalpharune(Rune c); int islowerrune(Rune c); -int isspacerune(Rune c); -int istitlerune(Rune c); int isupperrune(Rune c); Rune tolowerrune(Rune c); -Rune totitlerune(Rune c); Rune toupperrune(Rune c); #endif diff --git a/utfdata.h b/utfdata.h new file mode 100644 index 0000000..ba2d74e --- /dev/null +++ b/utfdata.h @@ -0,0 +1,1988 @@ +/* This file was automatically created from UnicodeData.txt */ + +static const Rune ucd_alpha2[] = { +0x41,0x5a, +0x61,0x7a, +0xc0,0xd6, +0xd8,0xf6, +0xf8,0x2c1, +0x2c6,0x2d1, +0x2e0,0x2e4, +0x370,0x374, +0x376,0x377, +0x37a,0x37d, +0x388,0x38a, +0x38e,0x3a1, +0x3a3,0x3f5, +0x3f7,0x481, +0x48a,0x52f, +0x531,0x556, +0x560,0x588, +0x5d0,0x5ea, +0x5ef,0x5f2, +0x620,0x64a, +0x66e,0x66f, +0x671,0x6d3, +0x6e5,0x6e6, +0x6ee,0x6ef, +0x6fa,0x6fc, +0x712,0x72f, +0x74d,0x7a5, +0x7ca,0x7ea, +0x7f4,0x7f5, +0x800,0x815, +0x840,0x858, +0x860,0x86a, +0x8a0,0x8b4, +0x8b6,0x8c7, +0x904,0x939, +0x958,0x961, +0x971,0x980, +0x985,0x98c, +0x98f,0x990, +0x993,0x9a8, +0x9aa,0x9b0, +0x9b6,0x9b9, +0x9dc,0x9dd, +0x9df,0x9e1, +0x9f0,0x9f1, +0xa05,0xa0a, +0xa0f,0xa10, +0xa13,0xa28, +0xa2a,0xa30, +0xa32,0xa33, +0xa35,0xa36, +0xa38,0xa39, +0xa59,0xa5c, +0xa72,0xa74, +0xa85,0xa8d, +0xa8f,0xa91, +0xa93,0xaa8, +0xaaa,0xab0, +0xab2,0xab3, +0xab5,0xab9, +0xae0,0xae1, +0xb05,0xb0c, +0xb0f,0xb10, +0xb13,0xb28, +0xb2a,0xb30, +0xb32,0xb33, +0xb35,0xb39, +0xb5c,0xb5d, +0xb5f,0xb61, +0xb85,0xb8a, +0xb8e,0xb90, +0xb92,0xb95, +0xb99,0xb9a, +0xb9e,0xb9f, +0xba3,0xba4, +0xba8,0xbaa, +0xbae,0xbb9, +0xc05,0xc0c, +0xc0e,0xc10, +0xc12,0xc28, +0xc2a,0xc39, +0xc58,0xc5a, +0xc60,0xc61, +0xc85,0xc8c, +0xc8e,0xc90, +0xc92,0xca8, +0xcaa,0xcb3, +0xcb5,0xcb9, +0xce0,0xce1, +0xcf1,0xcf2, +0xd04,0xd0c, +0xd0e,0xd10, +0xd12,0xd3a, +0xd54,0xd56, +0xd5f,0xd61, +0xd7a,0xd7f, +0xd85,0xd96, +0xd9a,0xdb1, +0xdb3,0xdbb, +0xdc0,0xdc6, +0xe01,0xe30, +0xe32,0xe33, +0xe40,0xe46, +0xe81,0xe82, +0xe86,0xe8a, +0xe8c,0xea3, +0xea7,0xeb0, +0xeb2,0xeb3, +0xec0,0xec4, +0xedc,0xedf, +0xf40,0xf47, +0xf49,0xf6c, +0xf88,0xf8c, +0x1000,0x102a, +0x1050,0x1055, +0x105a,0x105d, +0x1065,0x1066, +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, +0x13a0,0x13f5, +0x13f8,0x13fd, +0x1401,0x166c, +0x166f,0x167f, +0x1681,0x169a, +0x16a0,0x16ea, +0x16f1,0x16f8, +0x1700,0x170c, +0x170e,0x1711, +0x1720,0x1731, +0x1740,0x1751, +0x1760,0x176c, +0x176e,0x1770, +0x1780,0x17b3, +0x1820,0x1878, +0x1880,0x1884, +0x1887,0x18a8, +0x18b0,0x18f5, +0x1900,0x191e, +0x1950,0x196d, +0x1970,0x1974, +0x1980,0x19ab, +0x19b0,0x19c9, +0x1a00,0x1a16, +0x1a20,0x1a54, +0x1b05,0x1b33, +0x1b45,0x1b4b, +0x1b83,0x1ba0, +0x1bae,0x1baf, +0x1bba,0x1be5, +0x1c00,0x1c23, +0x1c4d,0x1c4f, +0x1c5a,0x1c7d, +0x1c80,0x1c88, +0x1c90,0x1cba, +0x1cbd,0x1cbf, +0x1ce9,0x1cec, +0x1cee,0x1cf3, +0x1cf5,0x1cf6, +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, +0x2183,0x2184, +0x2c00,0x2c2e, +0x2c30,0x2c5e, +0x2c60,0x2ce4, +0x2ceb,0x2cee, +0x2cf2,0x2cf3, +0x2d00,0x2d25, +0x2d30,0x2d67, +0x2d80,0x2d96, +0x2da0,0x2da6, +0x2da8,0x2dae, +0x2db0,0x2db6, +0x2db8,0x2dbe, +0x2dc0,0x2dc6, +0x2dc8,0x2dce, +0x2dd0,0x2dd6, +0x2dd8,0x2dde, +0x3005,0x3006, +0x3031,0x3035, +0x303b,0x303c, +0x3041,0x3096, +0x309d,0x309f, +0x30a1,0x30fa, +0x30fc,0x30ff, +0x3105,0x312f, +0x3131,0x318e, +0x31a0,0x31bf, +0x31f0,0x31ff, +0xa000,0xa48c, +0xa4d0,0xa4fd, +0xa500,0xa60c, +0xa610,0xa61f, +0xa62a,0xa62b, +0xa640,0xa66e, +0xa67f,0xa69d, +0xa6a0,0xa6e5, +0xa717,0xa71f, +0xa722,0xa788, +0xa78b,0xa7bf, +0xa7c2,0xa7ca, +0xa7f5,0xa801, +0xa803,0xa805, +0xa807,0xa80a, +0xa80c,0xa822, +0xa840,0xa873, +0xa882,0xa8b3, +0xa8f2,0xa8f7, +0xa8fd,0xa8fe, +0xa90a,0xa925, +0xa930,0xa946, +0xa960,0xa97c, +0xa984,0xa9b2, +0xa9e0,0xa9e4, +0xa9e6,0xa9ef, +0xa9fa,0xa9fe, +0xaa00,0xaa28, +0xaa40,0xaa42, +0xaa44,0xaa4b, +0xaa60,0xaa76, +0xaa7e,0xaaaf, +0xaab5,0xaab6, +0xaab9,0xaabd, +0xaadb,0xaadd, +0xaae0,0xaaea, +0xaaf2,0xaaf4, +0xab01,0xab06, +0xab09,0xab0e, +0xab11,0xab16, +0xab20,0xab26, +0xab28,0xab2e, +0xab30,0xab5a, +0xab5c,0xab69, +0xab70,0xabe2, +0xd7b0,0xd7c6, +0xd7cb,0xd7fb, +0xf900,0xfa6d, +0xfa70,0xfad9, +0xfb00,0xfb06, +0xfb13,0xfb17, +0xfb1f,0xfb28, +0xfb2a,0xfb36, +0xfb38,0xfb3c, +0xfb40,0xfb41, +0xfb43,0xfb44, +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, +0x10000,0x1000b, +0x1000d,0x10026, +0x10028,0x1003a, +0x1003c,0x1003d, +0x1003f,0x1004d, +0x10050,0x1005d, +0x10080,0x100fa, +0x10280,0x1029c, +0x102a0,0x102d0, +0x10300,0x1031f, +0x1032d,0x10340, +0x10342,0x10349, +0x10350,0x10375, +0x10380,0x1039d, +0x103a0,0x103c3, +0x103c8,0x103cf, +0x10400,0x1049d, +0x104b0,0x104d3, +0x104d8,0x104fb, +0x10500,0x10527, +0x10530,0x10563, +0x10600,0x10736, +0x10740,0x10755, +0x10760,0x10767, +0x10800,0x10805, +0x1080a,0x10835, +0x10837,0x10838, +0x1083f,0x10855, +0x10860,0x10876, +0x10880,0x1089e, +0x108e0,0x108f2, +0x108f4,0x108f5, +0x10900,0x10915, +0x10920,0x10939, +0x10980,0x109b7, +0x109be,0x109bf, +0x10a10,0x10a13, +0x10a15,0x10a17, +0x10a19,0x10a35, +0x10a60,0x10a7c, +0x10a80,0x10a9c, +0x10ac0,0x10ac7, +0x10ac9,0x10ae4, +0x10b00,0x10b35, +0x10b40,0x10b55, +0x10b60,0x10b72, +0x10b80,0x10b91, +0x10c00,0x10c48, +0x10c80,0x10cb2, +0x10cc0,0x10cf2, +0x10d00,0x10d23, +0x10e80,0x10ea9, +0x10eb0,0x10eb1, +0x10f00,0x10f1c, +0x10f30,0x10f45, +0x10fb0,0x10fc4, +0x10fe0,0x10ff6, +0x11003,0x11037, +0x11083,0x110af, +0x110d0,0x110e8, +0x11103,0x11126, +0x11150,0x11172, +0x11183,0x111b2, +0x111c1,0x111c4, +0x11200,0x11211, +0x11213,0x1122b, +0x11280,0x11286, +0x1128a,0x1128d, +0x1128f,0x1129d, +0x1129f,0x112a8, +0x112b0,0x112de, +0x11305,0x1130c, +0x1130f,0x11310, +0x11313,0x11328, +0x1132a,0x11330, +0x11332,0x11333, +0x11335,0x11339, +0x1135d,0x11361, +0x11400,0x11434, +0x11447,0x1144a, +0x1145f,0x11461, +0x11480,0x114af, +0x114c4,0x114c5, +0x11580,0x115ae, +0x115d8,0x115db, +0x11600,0x1162f, +0x11680,0x116aa, +0x11700,0x1171a, +0x11800,0x1182b, +0x118a0,0x118df, +0x118ff,0x11906, +0x1190c,0x11913, +0x11915,0x11916, +0x11918,0x1192f, +0x119a0,0x119a7, +0x119aa,0x119d0, +0x11a0b,0x11a32, +0x11a5c,0x11a89, +0x11ac0,0x11af8, +0x11c00,0x11c08, +0x11c0a,0x11c2e, +0x11c72,0x11c8f, +0x11d00,0x11d06, +0x11d08,0x11d09, +0x11d0b,0x11d30, +0x11d60,0x11d65, +0x11d67,0x11d68, +0x11d6a,0x11d89, +0x11ee0,0x11ef2, +0x12000,0x12399, +0x12480,0x12543, +0x13000,0x1342e, +0x14400,0x14646, +0x16800,0x16a38, +0x16a40,0x16a5e, +0x16ad0,0x16aed, +0x16b00,0x16b2f, +0x16b40,0x16b43, +0x16b63,0x16b77, +0x16b7d,0x16b8f, +0x16e40,0x16e7f, +0x16f00,0x16f4a, +0x16f93,0x16f9f, +0x16fe0,0x16fe1, +0x18800,0x18cd5, +0x1b000,0x1b11e, +0x1b150,0x1b152, +0x1b164,0x1b167, +0x1b170,0x1b2fb, +0x1bc00,0x1bc6a, +0x1bc70,0x1bc7c, +0x1bc80,0x1bc88, +0x1bc90,0x1bc99, +0x1d400,0x1d454, +0x1d456,0x1d49c, +0x1d49e,0x1d49f, +0x1d4a5,0x1d4a6, +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, +0x1e100,0x1e12c, +0x1e137,0x1e13d, +0x1e2c0,0x1e2eb, +0x1e800,0x1e8c4, +0x1e900,0x1e943, +0x1ee00,0x1ee03, +0x1ee05,0x1ee1f, +0x1ee21,0x1ee22, +0x1ee29,0x1ee32, +0x1ee34,0x1ee37, +0x1ee4d,0x1ee4f, +0x1ee51,0x1ee52, +0x1ee61,0x1ee62, +0x1ee67,0x1ee6a, +0x1ee6c,0x1ee72, +0x1ee74,0x1ee77, +0x1ee79,0x1ee7c, +0x1ee80,0x1ee89, +0x1ee8b,0x1ee9b, +0x1eea1,0x1eea3, +0x1eea5,0x1eea9, +0x1eeab,0x1eebb, +0x2f800,0x2fa1d, +}; + +static const Rune ucd_alpha1[] = { +0xaa, +0xb5, +0xba, +0x2ec, +0x2ee, +0x37f, +0x386, +0x38c, +0x559, +0x6d5, +0x6ff, +0x710, +0x7b1, +0x7fa, +0x81a, +0x824, +0x828, +0x93d, +0x950, +0x9b2, +0x9bd, +0x9ce, +0x9fc, +0xa5e, +0xabd, +0xad0, +0xaf9, +0xb3d, +0xb71, +0xb83, +0xb9c, +0xbd0, +0xc3d, +0xc80, +0xcbd, +0xcde, +0xd3d, +0xd4e, +0xdbd, +0xe84, +0xea5, +0xebd, +0xec6, +0xf00, +0x103f, +0x1061, +0x108e, +0x10c7, +0x10cd, +0x1258, +0x12c0, +0x17d7, +0x17dc, +0x18aa, +0x1aa7, +0x1cfa, +0x1f59, +0x1f5b, +0x1f5d, +0x1fbe, +0x2071, +0x207f, +0x2102, +0x2107, +0x2115, +0x2124, +0x2126, +0x2128, +0x214e, +0x2d27, +0x2d2d, +0x2d6f, +0x2e2f, +0x3400, +0x4dbf, +0x4e00, +0x9ffc, +0xa8fb, +0xa9cf, +0xaa7a, +0xaab1, +0xaac0, +0xaac2, +0xac00, +0xd7a3, +0xfb1d, +0xfb3e, +0x10808, +0x1083c, +0x10a00, +0x10f27, +0x11144, +0x11147, +0x11176, +0x111da, +0x111dc, +0x11288, +0x1133d, +0x11350, +0x114c7, +0x11644, +0x116b8, +0x11909, +0x1193f, +0x11941, +0x119e1, +0x119e3, +0x11a00, +0x11a3a, +0x11a50, +0x11a9d, +0x11c40, +0x11d46, +0x11d98, +0x11fb0, +0x16f50, +0x16fe3, +0x17000, +0x187f7, +0x18d00, +0x18d08, +0x1d4a2, +0x1d4bb, +0x1d546, +0x1e14e, +0x1e94b, +0x1ee24, +0x1ee27, +0x1ee39, +0x1ee3b, +0x1ee42, +0x1ee47, +0x1ee49, +0x1ee4b, +0x1ee54, +0x1ee57, +0x1ee59, +0x1ee5b, +0x1ee5d, +0x1ee5f, +0x1ee64, +0x1ee7e, +0x20000, +0x2a6dd, +0x2a700, +0x2b734, +0x2b740, +0x2b81d, +0x2b820, +0x2cea1, +0x2ceb0, +0x2ebe0, +0x30000, +0x3134a, +}; + +static const Rune ucd_tolower2[] = { +0x41,0x5a,32, +0xc0,0xd6,32, +0xd8,0xde,32, +0x189,0x18a,205, +0x1b1,0x1b2,217, +0x388,0x38a,37, +0x38e,0x38f,63, +0x391,0x3a1,32, +0x3a3,0x3ab,32, +0x3fd,0x3ff,-130, +0x400,0x40f,80, +0x410,0x42f,32, +0x531,0x556,48, +0x10a0,0x10c5,7264, +0x13a0,0x13ef,38864, +0x13f0,0x13f5,8, +0x1c90,0x1cba,-3008, +0x1cbd,0x1cbf,-3008, +0x1f08,0x1f0f,-8, +0x1f18,0x1f1d,-8, +0x1f28,0x1f2f,-8, +0x1f38,0x1f3f,-8, +0x1f48,0x1f4d,-8, +0x1f68,0x1f6f,-8, +0x1f88,0x1f8f,-8, +0x1f98,0x1f9f,-8, +0x1fa8,0x1faf,-8, +0x1fb8,0x1fb9,-8, +0x1fba,0x1fbb,-74, +0x1fc8,0x1fcb,-86, +0x1fd8,0x1fd9,-8, +0x1fda,0x1fdb,-100, +0x1fe8,0x1fe9,-8, +0x1fea,0x1feb,-112, +0x1ff8,0x1ff9,-128, +0x1ffa,0x1ffb,-126, +0x2160,0x216f,16, +0x24b6,0x24cf,26, +0x2c00,0x2c2e,48, +0x2c7e,0x2c7f,-10815, +0xff21,0xff3a,32, +0x10400,0x10427,40, +0x104b0,0x104d3,40, +0x10c80,0x10cb2,64, +0x118a0,0x118bf,32, +0x16e40,0x16e5f,32, +0x1e900,0x1e921,34, +}; + +static const Rune ucd_tolower1[] = { +0x100,1, +0x102,1, +0x104,1, +0x106,1, +0x108,1, +0x10a,1, +0x10c,1, +0x10e,1, +0x110,1, +0x112,1, +0x114,1, +0x116,1, +0x118,1, +0x11a,1, +0x11c,1, +0x11e,1, +0x120,1, +0x122,1, +0x124,1, +0x126,1, +0x128,1, +0x12a,1, +0x12c,1, +0x12e,1, +0x130,-199, +0x132,1, +0x134,1, +0x136,1, +0x139,1, +0x13b,1, +0x13d,1, +0x13f,1, +0x141,1, +0x143,1, +0x145,1, +0x147,1, +0x14a,1, +0x14c,1, +0x14e,1, +0x150,1, +0x152,1, +0x154,1, +0x156,1, +0x158,1, +0x15a,1, +0x15c,1, +0x15e,1, +0x160,1, +0x162,1, +0x164,1, +0x166,1, +0x168,1, +0x16a,1, +0x16c,1, +0x16e,1, +0x170,1, +0x172,1, +0x174,1, +0x176,1, +0x178,-121, +0x179,1, +0x17b,1, +0x17d,1, +0x181,210, +0x182,1, +0x184,1, +0x186,206, +0x187,1, +0x18b,1, +0x18e,79, +0x18f,202, +0x190,203, +0x191,1, +0x193,205, +0x194,207, +0x196,211, +0x197,209, +0x198,1, +0x19c,211, +0x19d,213, +0x19f,214, +0x1a0,1, +0x1a2,1, +0x1a4,1, +0x1a6,218, +0x1a7,1, +0x1a9,218, +0x1ac,1, +0x1ae,218, +0x1af,1, +0x1b3,1, +0x1b5,1, +0x1b7,219, +0x1b8,1, +0x1bc,1, +0x1c4,2, +0x1c5,1, +0x1c7,2, +0x1c8,1, +0x1ca,2, +0x1cb,1, +0x1cd,1, +0x1cf,1, +0x1d1,1, +0x1d3,1, +0x1d5,1, +0x1d7,1, +0x1d9,1, +0x1db,1, +0x1de,1, +0x1e0,1, +0x1e2,1, +0x1e4,1, +0x1e6,1, +0x1e8,1, +0x1ea,1, +0x1ec,1, +0x1ee,1, +0x1f1,2, +0x1f2,1, +0x1f4,1, +0x1f6,-97, +0x1f7,-56, +0x1f8,1, +0x1fa,1, +0x1fc,1, +0x1fe,1, +0x200,1, +0x202,1, +0x204,1, +0x206,1, +0x208,1, +0x20a,1, +0x20c,1, +0x20e,1, +0x210,1, +0x212,1, +0x214,1, +0x216,1, +0x218,1, +0x21a,1, +0x21c,1, +0x21e,1, +0x220,-130, +0x222,1, +0x224,1, +0x226,1, +0x228,1, +0x22a,1, +0x22c,1, +0x22e,1, +0x230,1, +0x232,1, +0x23a,10795, +0x23b,1, +0x23d,-163, +0x23e,10792, +0x241,1, +0x243,-195, +0x244,69, +0x245,71, +0x246,1, +0x248,1, +0x24a,1, +0x24c,1, +0x24e,1, +0x370,1, +0x372,1, +0x376,1, +0x37f,116, +0x386,38, +0x38c,64, +0x3cf,8, +0x3d8,1, +0x3da,1, +0x3dc,1, +0x3de,1, +0x3e0,1, +0x3e2,1, +0x3e4,1, +0x3e6,1, +0x3e8,1, +0x3ea,1, +0x3ec,1, +0x3ee,1, +0x3f4,-60, +0x3f7,1, +0x3f9,-7, +0x3fa,1, +0x460,1, +0x462,1, +0x464,1, +0x466,1, +0x468,1, +0x46a,1, +0x46c,1, +0x46e,1, +0x470,1, +0x472,1, +0x474,1, +0x476,1, +0x478,1, +0x47a,1, +0x47c,1, +0x47e,1, +0x480,1, +0x48a,1, +0x48c,1, +0x48e,1, +0x490,1, +0x492,1, +0x494,1, +0x496,1, +0x498,1, +0x49a,1, +0x49c,1, +0x49e,1, +0x4a0,1, +0x4a2,1, +0x4a4,1, +0x4a6,1, +0x4a8,1, +0x4aa,1, +0x4ac,1, +0x4ae,1, +0x4b0,1, +0x4b2,1, +0x4b4,1, +0x4b6,1, +0x4b8,1, +0x4ba,1, +0x4bc,1, +0x4be,1, +0x4c0,15, +0x4c1,1, +0x4c3,1, +0x4c5,1, +0x4c7,1, +0x4c9,1, +0x4cb,1, +0x4cd,1, +0x4d0,1, +0x4d2,1, +0x4d4,1, +0x4d6,1, +0x4d8,1, +0x4da,1, +0x4dc,1, +0x4de,1, +0x4e0,1, +0x4e2,1, +0x4e4,1, +0x4e6,1, +0x4e8,1, +0x4ea,1, +0x4ec,1, +0x4ee,1, +0x4f0,1, +0x4f2,1, +0x4f4,1, +0x4f6,1, +0x4f8,1, +0x4fa,1, +0x4fc,1, +0x4fe,1, +0x500,1, +0x502,1, +0x504,1, +0x506,1, +0x508,1, +0x50a,1, +0x50c,1, +0x50e,1, +0x510,1, +0x512,1, +0x514,1, +0x516,1, +0x518,1, +0x51a,1, +0x51c,1, +0x51e,1, +0x520,1, +0x522,1, +0x524,1, +0x526,1, +0x528,1, +0x52a,1, +0x52c,1, +0x52e,1, +0x10c7,7264, +0x10cd,7264, +0x1e00,1, +0x1e02,1, +0x1e04,1, +0x1e06,1, +0x1e08,1, +0x1e0a,1, +0x1e0c,1, +0x1e0e,1, +0x1e10,1, +0x1e12,1, +0x1e14,1, +0x1e16,1, +0x1e18,1, +0x1e1a,1, +0x1e1c,1, +0x1e1e,1, +0x1e20,1, +0x1e22,1, +0x1e24,1, +0x1e26,1, +0x1e28,1, +0x1e2a,1, +0x1e2c,1, +0x1e2e,1, +0x1e30,1, +0x1e32,1, +0x1e34,1, +0x1e36,1, +0x1e38,1, +0x1e3a,1, +0x1e3c,1, +0x1e3e,1, +0x1e40,1, +0x1e42,1, +0x1e44,1, +0x1e46,1, +0x1e48,1, +0x1e4a,1, +0x1e4c,1, +0x1e4e,1, +0x1e50,1, +0x1e52,1, +0x1e54,1, +0x1e56,1, +0x1e58,1, +0x1e5a,1, +0x1e5c,1, +0x1e5e,1, +0x1e60,1, +0x1e62,1, +0x1e64,1, +0x1e66,1, +0x1e68,1, +0x1e6a,1, +0x1e6c,1, +0x1e6e,1, +0x1e70,1, +0x1e72,1, +0x1e74,1, +0x1e76,1, +0x1e78,1, +0x1e7a,1, +0x1e7c,1, +0x1e7e,1, +0x1e80,1, +0x1e82,1, +0x1e84,1, +0x1e86,1, +0x1e88,1, +0x1e8a,1, +0x1e8c,1, +0x1e8e,1, +0x1e90,1, +0x1e92,1, +0x1e94,1, +0x1e9e,-7615, +0x1ea0,1, +0x1ea2,1, +0x1ea4,1, +0x1ea6,1, +0x1ea8,1, +0x1eaa,1, +0x1eac,1, +0x1eae,1, +0x1eb0,1, +0x1eb2,1, +0x1eb4,1, +0x1eb6,1, +0x1eb8,1, +0x1eba,1, +0x1ebc,1, +0x1ebe,1, +0x1ec0,1, +0x1ec2,1, +0x1ec4,1, +0x1ec6,1, +0x1ec8,1, +0x1eca,1, +0x1ecc,1, +0x1ece,1, +0x1ed0,1, +0x1ed2,1, +0x1ed4,1, +0x1ed6,1, +0x1ed8,1, +0x1eda,1, +0x1edc,1, +0x1ede,1, +0x1ee0,1, +0x1ee2,1, +0x1ee4,1, +0x1ee6,1, +0x1ee8,1, +0x1eea,1, +0x1eec,1, +0x1eee,1, +0x1ef0,1, +0x1ef2,1, +0x1ef4,1, +0x1ef6,1, +0x1ef8,1, +0x1efa,1, +0x1efc,1, +0x1efe,1, +0x1f59,-8, +0x1f5b,-8, +0x1f5d,-8, +0x1f5f,-8, +0x1fbc,-9, +0x1fcc,-9, +0x1fec,-7, +0x1ffc,-9, +0x2126,-7517, +0x212a,-8383, +0x212b,-8262, +0x2132,28, +0x2183,1, +0x2c60,1, +0x2c62,-10743, +0x2c63,-3814, +0x2c64,-10727, +0x2c67,1, +0x2c69,1, +0x2c6b,1, +0x2c6d,-10780, +0x2c6e,-10749, +0x2c6f,-10783, +0x2c70,-10782, +0x2c72,1, +0x2c75,1, +0x2c80,1, +0x2c82,1, +0x2c84,1, +0x2c86,1, +0x2c88,1, +0x2c8a,1, +0x2c8c,1, +0x2c8e,1, +0x2c90,1, +0x2c92,1, +0x2c94,1, +0x2c96,1, +0x2c98,1, +0x2c9a,1, +0x2c9c,1, +0x2c9e,1, +0x2ca0,1, +0x2ca2,1, +0x2ca4,1, +0x2ca6,1, +0x2ca8,1, +0x2caa,1, +0x2cac,1, +0x2cae,1, +0x2cb0,1, +0x2cb2,1, +0x2cb4,1, +0x2cb6,1, +0x2cb8,1, +0x2cba,1, +0x2cbc,1, +0x2cbe,1, +0x2cc0,1, +0x2cc2,1, +0x2cc4,1, +0x2cc6,1, +0x2cc8,1, +0x2cca,1, +0x2ccc,1, +0x2cce,1, +0x2cd0,1, +0x2cd2,1, +0x2cd4,1, +0x2cd6,1, +0x2cd8,1, +0x2cda,1, +0x2cdc,1, +0x2cde,1, +0x2ce0,1, +0x2ce2,1, +0x2ceb,1, +0x2ced,1, +0x2cf2,1, +0xa640,1, +0xa642,1, +0xa644,1, +0xa646,1, +0xa648,1, +0xa64a,1, +0xa64c,1, +0xa64e,1, +0xa650,1, +0xa652,1, +0xa654,1, +0xa656,1, +0xa658,1, +0xa65a,1, +0xa65c,1, +0xa65e,1, +0xa660,1, +0xa662,1, +0xa664,1, +0xa666,1, +0xa668,1, +0xa66a,1, +0xa66c,1, +0xa680,1, +0xa682,1, +0xa684,1, +0xa686,1, +0xa688,1, +0xa68a,1, +0xa68c,1, +0xa68e,1, +0xa690,1, +0xa692,1, +0xa694,1, +0xa696,1, +0xa698,1, +0xa69a,1, +0xa722,1, +0xa724,1, +0xa726,1, +0xa728,1, +0xa72a,1, +0xa72c,1, +0xa72e,1, +0xa732,1, +0xa734,1, +0xa736,1, +0xa738,1, +0xa73a,1, +0xa73c,1, +0xa73e,1, +0xa740,1, +0xa742,1, +0xa744,1, +0xa746,1, +0xa748,1, +0xa74a,1, +0xa74c,1, +0xa74e,1, +0xa750,1, +0xa752,1, +0xa754,1, +0xa756,1, +0xa758,1, +0xa75a,1, +0xa75c,1, +0xa75e,1, +0xa760,1, +0xa762,1, +0xa764,1, +0xa766,1, +0xa768,1, +0xa76a,1, +0xa76c,1, +0xa76e,1, +0xa779,1, +0xa77b,1, +0xa77d,-35332, +0xa77e,1, +0xa780,1, +0xa782,1, +0xa784,1, +0xa786,1, +0xa78b,1, +0xa78d,-42280, +0xa790,1, +0xa792,1, +0xa796,1, +0xa798,1, +0xa79a,1, +0xa79c,1, +0xa79e,1, +0xa7a0,1, +0xa7a2,1, +0xa7a4,1, +0xa7a6,1, +0xa7a8,1, +0xa7aa,-42308, +0xa7ab,-42319, +0xa7ac,-42315, +0xa7ad,-42305, +0xa7ae,-42308, +0xa7b0,-42258, +0xa7b1,-42282, +0xa7b2,-42261, +0xa7b3,928, +0xa7b4,1, +0xa7b6,1, +0xa7b8,1, +0xa7ba,1, +0xa7bc,1, +0xa7be,1, +0xa7c2,1, +0xa7c4,-48, +0xa7c5,-42307, +0xa7c6,-35384, +0xa7c7,1, +0xa7c9,1, +0xa7f5,1, +}; + +static const Rune ucd_toupper2[] = { +0x61,0x7a,-32, +0xe0,0xf6,-32, +0xf8,0xfe,-32, +0x23f,0x240,10815, +0x256,0x257,-205, +0x28a,0x28b,-217, +0x37b,0x37d,130, +0x3ad,0x3af,-37, +0x3b1,0x3c1,-32, +0x3c3,0x3cb,-32, +0x3cd,0x3ce,-63, +0x430,0x44f,-32, +0x450,0x45f,-80, +0x561,0x586,-48, +0x10d0,0x10fa,3008, +0x10fd,0x10ff,3008, +0x13f8,0x13fd,-8, +0x1c83,0x1c84,-6242, +0x1f00,0x1f07,8, +0x1f10,0x1f15,8, +0x1f20,0x1f27,8, +0x1f30,0x1f37,8, +0x1f40,0x1f45,8, +0x1f60,0x1f67,8, +0x1f70,0x1f71,74, +0x1f72,0x1f75,86, +0x1f76,0x1f77,100, +0x1f78,0x1f79,128, +0x1f7a,0x1f7b,112, +0x1f7c,0x1f7d,126, +0x1f80,0x1f87,8, +0x1f90,0x1f97,8, +0x1fa0,0x1fa7,8, +0x1fb0,0x1fb1,8, +0x1fd0,0x1fd1,8, +0x1fe0,0x1fe1,8, +0x2170,0x217f,-16, +0x24d0,0x24e9,-26, +0x2c30,0x2c5e,-48, +0x2d00,0x2d25,-7264, +0xab70,0xabbf,-38864, +0xff41,0xff5a,-32, +0x10428,0x1044f,-40, +0x104d8,0x104fb,-40, +0x10cc0,0x10cf2,-64, +0x118c0,0x118df,-32, +0x16e60,0x16e7f,-32, +0x1e922,0x1e943,-34, +}; + +static const Rune ucd_toupper1[] = { +0xb5,743, +0xff,121, +0x101,-1, +0x103,-1, +0x105,-1, +0x107,-1, +0x109,-1, +0x10b,-1, +0x10d,-1, +0x10f,-1, +0x111,-1, +0x113,-1, +0x115,-1, +0x117,-1, +0x119,-1, +0x11b,-1, +0x11d,-1, +0x11f,-1, +0x121,-1, +0x123,-1, +0x125,-1, +0x127,-1, +0x129,-1, +0x12b,-1, +0x12d,-1, +0x12f,-1, +0x131,-232, +0x133,-1, +0x135,-1, +0x137,-1, +0x13a,-1, +0x13c,-1, +0x13e,-1, +0x140,-1, +0x142,-1, +0x144,-1, +0x146,-1, +0x148,-1, +0x14b,-1, +0x14d,-1, +0x14f,-1, +0x151,-1, +0x153,-1, +0x155,-1, +0x157,-1, +0x159,-1, +0x15b,-1, +0x15d,-1, +0x15f,-1, +0x161,-1, +0x163,-1, +0x165,-1, +0x167,-1, +0x169,-1, +0x16b,-1, +0x16d,-1, +0x16f,-1, +0x171,-1, +0x173,-1, +0x175,-1, +0x177,-1, +0x17a,-1, +0x17c,-1, +0x17e,-1, +0x17f,-300, +0x180,195, +0x183,-1, +0x185,-1, +0x188,-1, +0x18c,-1, +0x192,-1, +0x195,97, +0x199,-1, +0x19a,163, +0x19e,130, +0x1a1,-1, +0x1a3,-1, +0x1a5,-1, +0x1a8,-1, +0x1ad,-1, +0x1b0,-1, +0x1b4,-1, +0x1b6,-1, +0x1b9,-1, +0x1bd,-1, +0x1bf,56, +0x1c5,-1, +0x1c6,-2, +0x1c8,-1, +0x1c9,-2, +0x1cb,-1, +0x1cc,-2, +0x1ce,-1, +0x1d0,-1, +0x1d2,-1, +0x1d4,-1, +0x1d6,-1, +0x1d8,-1, +0x1da,-1, +0x1dc,-1, +0x1dd,-79, +0x1df,-1, +0x1e1,-1, +0x1e3,-1, +0x1e5,-1, +0x1e7,-1, +0x1e9,-1, +0x1eb,-1, +0x1ed,-1, +0x1ef,-1, +0x1f2,-1, +0x1f3,-2, +0x1f5,-1, +0x1f9,-1, +0x1fb,-1, +0x1fd,-1, +0x1ff,-1, +0x201,-1, +0x203,-1, +0x205,-1, +0x207,-1, +0x209,-1, +0x20b,-1, +0x20d,-1, +0x20f,-1, +0x211,-1, +0x213,-1, +0x215,-1, +0x217,-1, +0x219,-1, +0x21b,-1, +0x21d,-1, +0x21f,-1, +0x223,-1, +0x225,-1, +0x227,-1, +0x229,-1, +0x22b,-1, +0x22d,-1, +0x22f,-1, +0x231,-1, +0x233,-1, +0x23c,-1, +0x242,-1, +0x247,-1, +0x249,-1, +0x24b,-1, +0x24d,-1, +0x24f,-1, +0x250,10783, +0x251,10780, +0x252,10782, +0x253,-210, +0x254,-206, +0x259,-202, +0x25b,-203, +0x25c,42319, +0x260,-205, +0x261,42315, +0x263,-207, +0x265,42280, +0x266,42308, +0x268,-209, +0x269,-211, +0x26a,42308, +0x26b,10743, +0x26c,42305, +0x26f,-211, +0x271,10749, +0x272,-213, +0x275,-214, +0x27d,10727, +0x280,-218, +0x282,42307, +0x283,-218, +0x287,42282, +0x288,-218, +0x289,-69, +0x28c,-71, +0x292,-219, +0x29d,42261, +0x29e,42258, +0x345,84, +0x371,-1, +0x373,-1, +0x377,-1, +0x3ac,-38, +0x3c2,-31, +0x3cc,-64, +0x3d0,-62, +0x3d1,-57, +0x3d5,-47, +0x3d6,-54, +0x3d7,-8, +0x3d9,-1, +0x3db,-1, +0x3dd,-1, +0x3df,-1, +0x3e1,-1, +0x3e3,-1, +0x3e5,-1, +0x3e7,-1, +0x3e9,-1, +0x3eb,-1, +0x3ed,-1, +0x3ef,-1, +0x3f0,-86, +0x3f1,-80, +0x3f2,7, +0x3f3,-116, +0x3f5,-96, +0x3f8,-1, +0x3fb,-1, +0x461,-1, +0x463,-1, +0x465,-1, +0x467,-1, +0x469,-1, +0x46b,-1, +0x46d,-1, +0x46f,-1, +0x471,-1, +0x473,-1, +0x475,-1, +0x477,-1, +0x479,-1, +0x47b,-1, +0x47d,-1, +0x47f,-1, +0x481,-1, +0x48b,-1, +0x48d,-1, +0x48f,-1, +0x491,-1, +0x493,-1, +0x495,-1, +0x497,-1, +0x499,-1, +0x49b,-1, +0x49d,-1, +0x49f,-1, +0x4a1,-1, +0x4a3,-1, +0x4a5,-1, +0x4a7,-1, +0x4a9,-1, +0x4ab,-1, +0x4ad,-1, +0x4af,-1, +0x4b1,-1, +0x4b3,-1, +0x4b5,-1, +0x4b7,-1, +0x4b9,-1, +0x4bb,-1, +0x4bd,-1, +0x4bf,-1, +0x4c2,-1, +0x4c4,-1, +0x4c6,-1, +0x4c8,-1, +0x4ca,-1, +0x4cc,-1, +0x4ce,-1, +0x4cf,-15, +0x4d1,-1, +0x4d3,-1, +0x4d5,-1, +0x4d7,-1, +0x4d9,-1, +0x4db,-1, +0x4dd,-1, +0x4df,-1, +0x4e1,-1, +0x4e3,-1, +0x4e5,-1, +0x4e7,-1, +0x4e9,-1, +0x4eb,-1, +0x4ed,-1, +0x4ef,-1, +0x4f1,-1, +0x4f3,-1, +0x4f5,-1, +0x4f7,-1, +0x4f9,-1, +0x4fb,-1, +0x4fd,-1, +0x4ff,-1, +0x501,-1, +0x503,-1, +0x505,-1, +0x507,-1, +0x509,-1, +0x50b,-1, +0x50d,-1, +0x50f,-1, +0x511,-1, +0x513,-1, +0x515,-1, +0x517,-1, +0x519,-1, +0x51b,-1, +0x51d,-1, +0x51f,-1, +0x521,-1, +0x523,-1, +0x525,-1, +0x527,-1, +0x529,-1, +0x52b,-1, +0x52d,-1, +0x52f,-1, +0x1c80,-6254, +0x1c81,-6253, +0x1c82,-6244, +0x1c85,-6243, +0x1c86,-6236, +0x1c87,-6181, +0x1c88,35266, +0x1d79,35332, +0x1d7d,3814, +0x1d8e,35384, +0x1e01,-1, +0x1e03,-1, +0x1e05,-1, +0x1e07,-1, +0x1e09,-1, +0x1e0b,-1, +0x1e0d,-1, +0x1e0f,-1, +0x1e11,-1, +0x1e13,-1, +0x1e15,-1, +0x1e17,-1, +0x1e19,-1, +0x1e1b,-1, +0x1e1d,-1, +0x1e1f,-1, +0x1e21,-1, +0x1e23,-1, +0x1e25,-1, +0x1e27,-1, +0x1e29,-1, +0x1e2b,-1, +0x1e2d,-1, +0x1e2f,-1, +0x1e31,-1, +0x1e33,-1, +0x1e35,-1, +0x1e37,-1, +0x1e39,-1, +0x1e3b,-1, +0x1e3d,-1, +0x1e3f,-1, +0x1e41,-1, +0x1e43,-1, +0x1e45,-1, +0x1e47,-1, +0x1e49,-1, +0x1e4b,-1, +0x1e4d,-1, +0x1e4f,-1, +0x1e51,-1, +0x1e53,-1, +0x1e55,-1, +0x1e57,-1, +0x1e59,-1, +0x1e5b,-1, +0x1e5d,-1, +0x1e5f,-1, +0x1e61,-1, +0x1e63,-1, +0x1e65,-1, +0x1e67,-1, +0x1e69,-1, +0x1e6b,-1, +0x1e6d,-1, +0x1e6f,-1, +0x1e71,-1, +0x1e73,-1, +0x1e75,-1, +0x1e77,-1, +0x1e79,-1, +0x1e7b,-1, +0x1e7d,-1, +0x1e7f,-1, +0x1e81,-1, +0x1e83,-1, +0x1e85,-1, +0x1e87,-1, +0x1e89,-1, +0x1e8b,-1, +0x1e8d,-1, +0x1e8f,-1, +0x1e91,-1, +0x1e93,-1, +0x1e95,-1, +0x1e9b,-59, +0x1ea1,-1, +0x1ea3,-1, +0x1ea5,-1, +0x1ea7,-1, +0x1ea9,-1, +0x1eab,-1, +0x1ead,-1, +0x1eaf,-1, +0x1eb1,-1, +0x1eb3,-1, +0x1eb5,-1, +0x1eb7,-1, +0x1eb9,-1, +0x1ebb,-1, +0x1ebd,-1, +0x1ebf,-1, +0x1ec1,-1, +0x1ec3,-1, +0x1ec5,-1, +0x1ec7,-1, +0x1ec9,-1, +0x1ecb,-1, +0x1ecd,-1, +0x1ecf,-1, +0x1ed1,-1, +0x1ed3,-1, +0x1ed5,-1, +0x1ed7,-1, +0x1ed9,-1, +0x1edb,-1, +0x1edd,-1, +0x1edf,-1, +0x1ee1,-1, +0x1ee3,-1, +0x1ee5,-1, +0x1ee7,-1, +0x1ee9,-1, +0x1eeb,-1, +0x1eed,-1, +0x1eef,-1, +0x1ef1,-1, +0x1ef3,-1, +0x1ef5,-1, +0x1ef7,-1, +0x1ef9,-1, +0x1efb,-1, +0x1efd,-1, +0x1eff,-1, +0x1f51,8, +0x1f53,8, +0x1f55,8, +0x1f57,8, +0x1fb3,9, +0x1fbe,-7205, +0x1fc3,9, +0x1fe5,7, +0x1ff3,9, +0x214e,-28, +0x2184,-1, +0x2c61,-1, +0x2c65,-10795, +0x2c66,-10792, +0x2c68,-1, +0x2c6a,-1, +0x2c6c,-1, +0x2c73,-1, +0x2c76,-1, +0x2c81,-1, +0x2c83,-1, +0x2c85,-1, +0x2c87,-1, +0x2c89,-1, +0x2c8b,-1, +0x2c8d,-1, +0x2c8f,-1, +0x2c91,-1, +0x2c93,-1, +0x2c95,-1, +0x2c97,-1, +0x2c99,-1, +0x2c9b,-1, +0x2c9d,-1, +0x2c9f,-1, +0x2ca1,-1, +0x2ca3,-1, +0x2ca5,-1, +0x2ca7,-1, +0x2ca9,-1, +0x2cab,-1, +0x2cad,-1, +0x2caf,-1, +0x2cb1,-1, +0x2cb3,-1, +0x2cb5,-1, +0x2cb7,-1, +0x2cb9,-1, +0x2cbb,-1, +0x2cbd,-1, +0x2cbf,-1, +0x2cc1,-1, +0x2cc3,-1, +0x2cc5,-1, +0x2cc7,-1, +0x2cc9,-1, +0x2ccb,-1, +0x2ccd,-1, +0x2ccf,-1, +0x2cd1,-1, +0x2cd3,-1, +0x2cd5,-1, +0x2cd7,-1, +0x2cd9,-1, +0x2cdb,-1, +0x2cdd,-1, +0x2cdf,-1, +0x2ce1,-1, +0x2ce3,-1, +0x2cec,-1, +0x2cee,-1, +0x2cf3,-1, +0x2d27,-7264, +0x2d2d,-7264, +0xa641,-1, +0xa643,-1, +0xa645,-1, +0xa647,-1, +0xa649,-1, +0xa64b,-1, +0xa64d,-1, +0xa64f,-1, +0xa651,-1, +0xa653,-1, +0xa655,-1, +0xa657,-1, +0xa659,-1, +0xa65b,-1, +0xa65d,-1, +0xa65f,-1, +0xa661,-1, +0xa663,-1, +0xa665,-1, +0xa667,-1, +0xa669,-1, +0xa66b,-1, +0xa66d,-1, +0xa681,-1, +0xa683,-1, +0xa685,-1, +0xa687,-1, +0xa689,-1, +0xa68b,-1, +0xa68d,-1, +0xa68f,-1, +0xa691,-1, +0xa693,-1, +0xa695,-1, +0xa697,-1, +0xa699,-1, +0xa69b,-1, +0xa723,-1, +0xa725,-1, +0xa727,-1, +0xa729,-1, +0xa72b,-1, +0xa72d,-1, +0xa72f,-1, +0xa733,-1, +0xa735,-1, +0xa737,-1, +0xa739,-1, +0xa73b,-1, +0xa73d,-1, +0xa73f,-1, +0xa741,-1, +0xa743,-1, +0xa745,-1, +0xa747,-1, +0xa749,-1, +0xa74b,-1, +0xa74d,-1, +0xa74f,-1, +0xa751,-1, +0xa753,-1, +0xa755,-1, +0xa757,-1, +0xa759,-1, +0xa75b,-1, +0xa75d,-1, +0xa75f,-1, +0xa761,-1, +0xa763,-1, +0xa765,-1, +0xa767,-1, +0xa769,-1, +0xa76b,-1, +0xa76d,-1, +0xa76f,-1, +0xa77a,-1, +0xa77c,-1, +0xa77f,-1, +0xa781,-1, +0xa783,-1, +0xa785,-1, +0xa787,-1, +0xa78c,-1, +0xa791,-1, +0xa793,-1, +0xa794,48, +0xa797,-1, +0xa799,-1, +0xa79b,-1, +0xa79d,-1, +0xa79f,-1, +0xa7a1,-1, +0xa7a3,-1, +0xa7a5,-1, +0xa7a7,-1, +0xa7a9,-1, +0xa7b5,-1, +0xa7b7,-1, +0xa7b9,-1, +0xa7bb,-1, +0xa7bd,-1, +0xa7bf,-1, +0xa7c3,-1, +0xa7c8,-1, +0xa7ca,-1, +0xa7f6,-1, +0xab53,-928, +}; diff --git a/utftype.c b/utftype.c deleted file mode 100644 index 543b296..0000000 --- a/utftype.c +++ /dev/null @@ -1,1143 +0,0 @@ -/* - * The authors of this software are Rob Pike and Ken Thompson. - * Copyright (c) 2002 by Lucent Technologies. - * Permission to use, copy, modify, and distribute this software for any - * purpose without fee is hereby granted, provided that this entire notice - * is included in all copies of any software which is or includes a copy - * or modification of this software and in all copies of the supporting - * documentation for such software. - * THIS SOFTWARE IS BEING PROVIDED "AS IS", WITHOUT ANY EXPRESS OR IMPLIED - * WARRANTY. IN PARTICULAR, NEITHER THE AUTHORS NOR LUCENT TECHNOLOGIES MAKE - * ANY REPRESENTATION OR WARRANTY OF ANY KIND CONCERNING THE MERCHANTABILITY - * OF THIS SOFTWARE OR ITS FITNESS FOR ANY PARTICULAR PURPOSE. - */ -#include "utf.h" - -#define bsearch jsU_bsearch -#define nelem(a) (int)(sizeof (a) / sizeof (a)[0]) - -/* - * alpha ranges - - * only covers ranges not in lower||upper - */ -static const Rune __alpha2[] = -{ - 0x00d8, 0x00f6, /* Ø - ö */ - 0x00f8, 0x01f5, /* ø - ǵ */ - 0x0250, 0x02a8, /* ɐ - ʨ */ - 0x038e, 0x03a1, /* Ύ - Ρ */ - 0x03a3, 0x03ce, /* Σ - ώ */ - 0x03d0, 0x03d6, /* ϐ - ϖ */ - 0x03e2, 0x03f3, /* Ϣ - ϳ */ - 0x0490, 0x04c4, /* Ґ - ӄ */ - 0x0561, 0x0587, /* ա - և */ - 0x05d0, 0x05ea, /* א - ת */ - 0x05f0, 0x05f2, /* װ - ײ */ - 0x0621, 0x063a, /* ء - غ */ - 0x0640, 0x064a, /* ـ - ي */ - 0x0671, 0x06b7, /* ٱ - ڷ */ - 0x06ba, 0x06be, /* ں - ھ */ - 0x06c0, 0x06ce, /* ۀ - ێ */ - 0x06d0, 0x06d3, /* ې - ۓ */ - 0x0905, 0x0939, /* अ - ह */ - 0x0958, 0x0961, /* क़ - ॡ */ - 0x0985, 0x098c, /* অ - ঌ */ - 0x098f, 0x0990, /* এ - ঐ */ - 0x0993, 0x09a8, /* ও - ন */ - 0x09aa, 0x09b0, /* প - র */ - 0x09b6, 0x09b9, /* শ - হ */ - 0x09dc, 0x09dd, /* ড় - ঢ় */ - 0x09df, 0x09e1, /* য় - ৡ */ - 0x09f0, 0x09f1, /* ৰ - ৱ */ - 0x0a05, 0x0a0a, /* ਅ - ਊ */ - 0x0a0f, 0x0a10, /* ਏ - ਐ */ - 0x0a13, 0x0a28, /* ਓ - ਨ */ - 0x0a2a, 0x0a30, /* ਪ - ਰ */ - 0x0a32, 0x0a33, /* ਲ - ਲ਼ */ - 0x0a35, 0x0a36, /* ਵ - ਸ਼ */ - 0x0a38, 0x0a39, /* ਸ - ਹ */ - 0x0a59, 0x0a5c, /* ਖ਼ - ੜ */ - 0x0a85, 0x0a8b, /* અ - ઋ */ - 0x0a8f, 0x0a91, /* એ - ઑ */ - 0x0a93, 0x0aa8, /* ઓ - ન */ - 0x0aaa, 0x0ab0, /* પ - ર */ - 0x0ab2, 0x0ab3, /* લ - ળ */ - 0x0ab5, 0x0ab9, /* વ - હ */ - 0x0b05, 0x0b0c, /* ଅ - ଌ */ - 0x0b0f, 0x0b10, /* ଏ - ଐ */ - 0x0b13, 0x0b28, /* ଓ - ନ */ - 0x0b2a, 0x0b30, /* ପ - ର */ - 0x0b32, 0x0b33, /* ଲ - ଳ */ - 0x0b36, 0x0b39, /* ଶ - ହ */ - 0x0b5c, 0x0b5d, /* ଡ଼ - ଢ଼ */ - 0x0b5f, 0x0b61, /* ୟ - ୡ */ - 0x0b85, 0x0b8a, /* அ - ஊ */ - 0x0b8e, 0x0b90, /* எ - ஐ */ - 0x0b92, 0x0b95, /* ஒ - க */ - 0x0b99, 0x0b9a, /* ங - ச */ - 0x0b9e, 0x0b9f, /* ஞ - ட */ - 0x0ba3, 0x0ba4, /* ண - த */ - 0x0ba8, 0x0baa, /* ந - ப */ - 0x0bae, 0x0bb5, /* ம - வ */ - 0x0bb7, 0x0bb9, /* ஷ - ஹ */ - 0x0c05, 0x0c0c, /* అ - ఌ */ - 0x0c0e, 0x0c10, /* ఎ - ఐ */ - 0x0c12, 0x0c28, /* ఒ - న */ - 0x0c2a, 0x0c33, /* ప - ళ */ - 0x0c35, 0x0c39, /* వ - హ */ - 0x0c60, 0x0c61, /* ౠ - ౡ */ - 0x0c85, 0x0c8c, /* ಅ - ಌ */ - 0x0c8e, 0x0c90, /* ಎ - ಐ */ - 0x0c92, 0x0ca8, /* ಒ - ನ */ - 0x0caa, 0x0cb3, /* ಪ - ಳ */ - 0x0cb5, 0x0cb9, /* ವ - ಹ */ - 0x0ce0, 0x0ce1, /* ೠ - ೡ */ - 0x0d05, 0x0d0c, /* അ - ഌ */ - 0x0d0e, 0x0d10, /* എ - ഐ */ - 0x0d12, 0x0d28, /* ഒ - ന */ - 0x0d2a, 0x0d39, /* പ - ഹ */ - 0x0d60, 0x0d61, /* ൠ - ൡ */ - 0x0e01, 0x0e30, /* ก - ะ */ - 0x0e32, 0x0e33, /* า - ำ */ - 0x0e40, 0x0e46, /* เ - ๆ */ - 0x0e5a, 0x0e5b, /* ๚ - ๛ */ - 0x0e81, 0x0e82, /* ກ - ຂ */ - 0x0e87, 0x0e88, /* ງ - ຈ */ - 0x0e94, 0x0e97, /* ດ - ທ */ - 0x0e99, 0x0e9f, /* ນ - ຟ */ - 0x0ea1, 0x0ea3, /* ມ - ຣ */ - 0x0eaa, 0x0eab, /* ສ - ຫ */ - 0x0ead, 0x0eae, /* ອ - ຮ */ - 0x0eb2, 0x0eb3, /* າ - ຳ */ - 0x0ec0, 0x0ec4, /* ເ - ໄ */ - 0x0edc, 0x0edd, /* ໜ - ໝ */ - 0x0f18, 0x0f19, /* ༘ - ༙ */ - 0x0f40, 0x0f47, /* ཀ - ཇ */ - 0x0f49, 0x0f69, /* ཉ - ཀྵ */ - 0x10d0, 0x10f6, /* ა - ჶ */ - 0x1100, 0x1159, /* ᄀ - ᅙ */ - 0x115f, 0x11a2, /* ᅟ - ᆢ */ - 0x11a8, 0x11f9, /* ᆨ - ᇹ */ - 0x1e00, 0x1e9b, /* Ḁ - ẛ */ - 0x1f50, 0x1f57, /* ὐ - ὗ */ - 0x1f80, 0x1fb4, /* ᾀ - ᾴ */ - 0x1fb6, 0x1fbc, /* ᾶ - ᾼ */ - 0x1fc2, 0x1fc4, /* ῂ - ῄ */ - 0x1fc6, 0x1fcc, /* ῆ - ῌ */ - 0x1fd0, 0x1fd3, /* ῐ - ΐ */ - 0x1fd6, 0x1fdb, /* ῖ - Ί */ - 0x1fe0, 0x1fec, /* ῠ - Ῥ */ - 0x1ff2, 0x1ff4, /* ῲ - ῴ */ - 0x1ff6, 0x1ffc, /* ῶ - ῼ */ - 0x210a, 0x2113, /* ℊ - ℓ */ - 0x2115, 0x211d, /* ℕ - ℝ */ - 0x2120, 0x2122, /* ℠ - ™ */ - 0x212a, 0x2131, /* K - ℱ */ - 0x2133, 0x2138, /* ℳ - ℸ */ - 0x3041, 0x3094, /* ぁ - ゔ */ - 0x30a1, 0x30fa, /* ァ - ヺ */ - 0x3105, 0x312c, /* ㄅ - ㄬ */ - 0x3131, 0x318e, /* ㄱ - ㆎ */ - 0x3192, 0x319f, /* ㆒ - ㆟ */ - 0x3260, 0x327b, /* ㉠ - ㉻ */ - 0x328a, 0x32b0, /* ㊊ - ㊰ */ - 0x32d0, 0x32fe, /* ㋐ - ㋾ */ - 0x3300, 0x3357, /* ㌀ - ㍗ */ - 0x3371, 0x3376, /* ㍱ - ㍶ */ - 0x337b, 0x3394, /* ㍻ - ㎔ */ - 0x3399, 0x339e, /* ㎙ - ㎞ */ - 0x33a9, 0x33ad, /* ㎩ - ㎭ */ - 0x33b0, 0x33c1, /* ㎰ - ㏁ */ - 0x33c3, 0x33c5, /* ㏃ - ㏅ */ - 0x33c7, 0x33d7, /* ㏇ - ㏗ */ - 0x33d9, 0x33dd, /* ㏙ - ㏝ */ - 0x4e00, 0x9fff, /* 一 - 鿿 */ - 0xac00, 0xd7a3, /* 가 - 힣 */ - 0xf900, 0xfb06, /* 豈 - st */ - 0xfb13, 0xfb17, /* ﬓ - ﬗ */ - 0xfb1f, 0xfb28, /* ײַ - ﬨ */ - 0xfb2a, 0xfb36, /* שׁ - זּ */ - 0xfb38, 0xfb3c, /* טּ - לּ */ - 0xfb40, 0xfb41, /* נּ - סּ */ - 0xfb43, 0xfb44, /* ףּ - פּ */ - 0xfb46, 0xfbb1, /* צּ - ﮱ */ - 0xfbd3, 0xfd3d, /* ﯓ - ﴽ */ - 0xfd50, 0xfd8f, /* ﵐ - ﶏ */ - 0xfd92, 0xfdc7, /* ﶒ - ﷇ */ - 0xfdf0, 0xfdf9, /* ﷰ - ﷹ */ - 0xfe70, 0xfe72, /* ﹰ - ﹲ */ - 0xfe76, 0xfefc, /* ﹶ - ﻼ */ - 0xff66, 0xff6f, /* ヲ - ッ */ - 0xff71, 0xff9d, /* ア - ン */ - 0xffa0, 0xffbe, /* ᅠ - ᄒ */ - 0xffc2, 0xffc7, /* ᅡ - ᅦ */ - 0xffca, 0xffcf, /* ᅧ - ᅬ */ - 0xffd2, 0xffd7, /* ᅭ - ᅲ */ - 0xffda, 0xffdc, /* ᅳ - ᅵ */ -}; - -/* - * alpha singlets - - * only covers ranges not in lower||upper - */ -static const Rune __alpha1[] = -{ - 0x00aa, /* ª */ - 0x00b5, /* µ */ - 0x00ba, /* º */ - 0x03da, /* Ϛ */ - 0x03dc, /* Ϝ */ - 0x03de, /* Ϟ */ - 0x03e0, /* Ϡ */ - 0x06d5, /* ە */ - 0x09b2, /* ল */ - 0x0a5e, /* ਫ਼ */ - 0x0a8d, /* ઍ */ - 0x0ae0, /* ૠ */ - 0x0b9c, /* ஜ */ - 0x0cde, /* ೞ */ - 0x0e4f, /* ๏ */ - 0x0e84, /* ຄ */ - 0x0e8a, /* ຊ */ - 0x0e8d, /* ຍ */ - 0x0ea5, /* ລ */ - 0x0ea7, /* ວ */ - 0x0eb0, /* ະ */ - 0x0ebd, /* ຽ */ - 0x1fbe, /* ι */ - 0x207f, /* ⁿ */ - 0x20a8, /* ₨ */ - 0x2102, /* ℂ */ - 0x2107, /* ℇ */ - 0x2124, /* ℤ */ - 0x2126, /* Ω */ - 0x2128, /* ℨ */ - 0xfb3e, /* מּ */ - 0xfe74, /* ﹴ */ -}; - -/* - * space ranges - */ -static const Rune __space2[] = -{ - 0x0009, 0x000a, /* tab and newline */ - 0x0020, 0x0020, /* space */ - 0x00a0, 0x00a0, /*   */ - 0x2000, 0x200b, /*   - ​ */ - 0x2028, 0x2029, /* 
 - 
 */ - 0x3000, 0x3000, /*   */ - 0xfeff, 0xfeff, /*  */ -}; - -/* - * lower case ranges - * 3rd col is conversion excess 500 - */ -static const Rune __toupper2[] = -{ - 0x0061, 0x007a, 468, /* a-z A-Z */ - 0x00e0, 0x00f6, 468, /* à-ö À-Ö */ - 0x00f8, 0x00fe, 468, /* ø-þ Ø-Þ */ - 0x0256, 0x0257, 295, /* ɖ-ɗ Ɖ-Ɗ */ - 0x0258, 0x0259, 298, /* ɘ-ə Ǝ-Ə */ - 0x028a, 0x028b, 283, /* ʊ-ʋ Ʊ-Ʋ */ - 0x03ad, 0x03af, 463, /* έ-ί Έ-Ί */ - 0x03b1, 0x03c1, 468, /* α-ρ Α-Ρ */ - 0x03c3, 0x03cb, 468, /* σ-ϋ Σ-Ϋ */ - 0x03cd, 0x03ce, 437, /* ύ-ώ Ύ-Ώ */ - 0x0430, 0x044f, 468, /* а-я А-Я */ - 0x0451, 0x045c, 420, /* ё-ќ Ё-Ќ */ - 0x045e, 0x045f, 420, /* ў-џ Ў-Џ */ - 0x0561, 0x0586, 452, /* ա-ֆ Ա-Ֆ */ - 0x1f00, 0x1f07, 508, /* ἀ-ἇ Ἀ-Ἇ */ - 0x1f10, 0x1f15, 508, /* ἐ-ἕ Ἐ-Ἕ */ - 0x1f20, 0x1f27, 508, /* ἠ-ἧ Ἠ-Ἧ */ - 0x1f30, 0x1f37, 508, /* ἰ-ἷ Ἰ-Ἷ */ - 0x1f40, 0x1f45, 508, /* ὀ-ὅ Ὀ-Ὅ */ - 0x1f60, 0x1f67, 508, /* ὠ-ὧ Ὠ-Ὧ */ - 0x1f70, 0x1f71, 574, /* ὰ-ά Ὰ-Ά */ - 0x1f72, 0x1f75, 586, /* ὲ-ή Ὲ-Ή */ - 0x1f76, 0x1f77, 600, /* ὶ-ί Ὶ-Ί */ - 0x1f78, 0x1f79, 628, /* ὸ-ό Ὸ-Ό */ - 0x1f7a, 0x1f7b, 612, /* ὺ-ύ Ὺ-Ύ */ - 0x1f7c, 0x1f7d, 626, /* ὼ-ώ Ὼ-Ώ */ - 0x1f80, 0x1f87, 508, /* ᾀ-ᾇ ᾈ-ᾏ */ - 0x1f90, 0x1f97, 508, /* ᾐ-ᾗ ᾘ-ᾟ */ - 0x1fa0, 0x1fa7, 508, /* ᾠ-ᾧ ᾨ-ᾯ */ - 0x1fb0, 0x1fb1, 508, /* ᾰ-ᾱ Ᾰ-Ᾱ */ - 0x1fd0, 0x1fd1, 508, /* ῐ-ῑ Ῐ-Ῑ */ - 0x1fe0, 0x1fe1, 508, /* ῠ-ῡ Ῠ-Ῡ */ - 0x2170, 0x217f, 484, /* ⅰ-ⅿ Ⅰ-Ⅿ */ - 0x24d0, 0x24e9, 474, /* ⓐ-ⓩ Ⓐ-Ⓩ */ - 0xff41, 0xff5a, 468, /* a-z A-Z */ -}; - -/* - * lower case singlets - * 2nd col is conversion excess 500 - */ -static const Rune __toupper1[] = -{ - 0x00ff, 621, /* ÿ Ÿ */ - 0x0101, 499, /* ā Ā */ - 0x0103, 499, /* ă Ă */ - 0x0105, 499, /* ą Ą */ - 0x0107, 499, /* ć Ć */ - 0x0109, 499, /* ĉ Ĉ */ - 0x010b, 499, /* ċ Ċ */ - 0x010d, 499, /* č Č */ - 0x010f, 499, /* ď Ď */ - 0x0111, 499, /* đ Đ */ - 0x0113, 499, /* ē Ē */ - 0x0115, 499, /* ĕ Ĕ */ - 0x0117, 499, /* ė Ė */ - 0x0119, 499, /* ę Ę */ - 0x011b, 499, /* ě Ě */ - 0x011d, 499, /* ĝ Ĝ */ - 0x011f, 499, /* ğ Ğ */ - 0x0121, 499, /* ġ Ġ */ - 0x0123, 499, /* ģ Ģ */ - 0x0125, 499, /* ĥ Ĥ */ - 0x0127, 499, /* ħ Ħ */ - 0x0129, 499, /* ĩ Ĩ */ - 0x012b, 499, /* ī Ī */ - 0x012d, 499, /* ĭ Ĭ */ - 0x012f, 499, /* į Į */ - 0x0131, 268, /* ı I */ - 0x0133, 499, /* ij IJ */ - 0x0135, 499, /* ĵ Ĵ */ - 0x0137, 499, /* ķ Ķ */ - 0x013a, 499, /* ĺ Ĺ */ - 0x013c, 499, /* ļ Ļ */ - 0x013e, 499, /* ľ Ľ */ - 0x0140, 499, /* ŀ Ŀ */ - 0x0142, 499, /* ł Ł */ - 0x0144, 499, /* ń Ń */ - 0x0146, 499, /* ņ Ņ */ - 0x0148, 499, /* ň Ň */ - 0x014b, 499, /* ŋ Ŋ */ - 0x014d, 499, /* ō Ō */ - 0x014f, 499, /* ŏ Ŏ */ - 0x0151, 499, /* ő Ő */ - 0x0153, 499, /* œ Œ */ - 0x0155, 499, /* ŕ Ŕ */ - 0x0157, 499, /* ŗ Ŗ */ - 0x0159, 499, /* ř Ř */ - 0x015b, 499, /* ś Ś */ - 0x015d, 499, /* ŝ Ŝ */ - 0x015f, 499, /* ş Ş */ - 0x0161, 499, /* š Š */ - 0x0163, 499, /* ţ Ţ */ - 0x0165, 499, /* ť Ť */ - 0x0167, 499, /* ŧ Ŧ */ - 0x0169, 499, /* ũ Ũ */ - 0x016b, 499, /* ū Ū */ - 0x016d, 499, /* ŭ Ŭ */ - 0x016f, 499, /* ů Ů */ - 0x0171, 499, /* ű Ű */ - 0x0173, 499, /* ų Ų */ - 0x0175, 499, /* ŵ Ŵ */ - 0x0177, 499, /* ŷ Ŷ */ - 0x017a, 499, /* ź Ź */ - 0x017c, 499, /* ż Ż */ - 0x017e, 499, /* ž Ž */ - 0x017f, 200, /* ſ S */ - 0x0183, 499, /* ƃ Ƃ */ - 0x0185, 499, /* ƅ Ƅ */ - 0x0188, 499, /* ƈ Ƈ */ - 0x018c, 499, /* ƌ Ƌ */ - 0x0192, 499, /* ƒ Ƒ */ - 0x0199, 499, /* ƙ Ƙ */ - 0x01a1, 499, /* ơ Ơ */ - 0x01a3, 499, /* ƣ Ƣ */ - 0x01a5, 499, /* ƥ Ƥ */ - 0x01a8, 499, /* ƨ Ƨ */ - 0x01ad, 499, /* ƭ Ƭ */ - 0x01b0, 499, /* ư Ư */ - 0x01b4, 499, /* ƴ Ƴ */ - 0x01b6, 499, /* ƶ Ƶ */ - 0x01b9, 499, /* ƹ Ƹ */ - 0x01bd, 499, /* ƽ Ƽ */ - 0x01c5, 499, /* Dž DŽ */ - 0x01c6, 498, /* dž DŽ */ - 0x01c8, 499, /* Lj LJ */ - 0x01c9, 498, /* lj LJ */ - 0x01cb, 499, /* Nj NJ */ - 0x01cc, 498, /* nj NJ */ - 0x01ce, 499, /* ǎ Ǎ */ - 0x01d0, 499, /* ǐ Ǐ */ - 0x01d2, 499, /* ǒ Ǒ */ - 0x01d4, 499, /* ǔ Ǔ */ - 0x01d6, 499, /* ǖ Ǖ */ - 0x01d8, 499, /* ǘ Ǘ */ - 0x01da, 499, /* ǚ Ǚ */ - 0x01dc, 499, /* ǜ Ǜ */ - 0x01df, 499, /* ǟ Ǟ */ - 0x01e1, 499, /* ǡ Ǡ */ - 0x01e3, 499, /* ǣ Ǣ */ - 0x01e5, 499, /* ǥ Ǥ */ - 0x01e7, 499, /* ǧ Ǧ */ - 0x01e9, 499, /* ǩ Ǩ */ - 0x01eb, 499, /* ǫ Ǫ */ - 0x01ed, 499, /* ǭ Ǭ */ - 0x01ef, 499, /* ǯ Ǯ */ - 0x01f2, 499, /* Dz DZ */ - 0x01f3, 498, /* dz DZ */ - 0x01f5, 499, /* ǵ Ǵ */ - 0x01fb, 499, /* ǻ Ǻ */ - 0x01fd, 499, /* ǽ Ǽ */ - 0x01ff, 499, /* ǿ Ǿ */ - 0x0201, 499, /* ȁ Ȁ */ - 0x0203, 499, /* ȃ Ȃ */ - 0x0205, 499, /* ȅ Ȅ */ - 0x0207, 499, /* ȇ Ȇ */ - 0x0209, 499, /* ȉ Ȉ */ - 0x020b, 499, /* ȋ Ȋ */ - 0x020d, 499, /* ȍ Ȍ */ - 0x020f, 499, /* ȏ Ȏ */ - 0x0211, 499, /* ȑ Ȑ */ - 0x0213, 499, /* ȓ Ȓ */ - 0x0215, 499, /* ȕ Ȕ */ - 0x0217, 499, /* ȗ Ȗ */ - 0x0253, 290, /* ɓ Ɓ */ - 0x0254, 294, /* ɔ Ɔ */ - 0x025b, 297, /* ɛ Ɛ */ - 0x0260, 295, /* ɠ Ɠ */ - 0x0263, 293, /* ɣ Ɣ */ - 0x0268, 291, /* ɨ Ɨ */ - 0x0269, 289, /* ɩ Ɩ */ - 0x026f, 289, /* ɯ Ɯ */ - 0x0272, 287, /* ɲ Ɲ */ - 0x0283, 282, /* ʃ Ʃ */ - 0x0288, 282, /* ʈ Ʈ */ - 0x0292, 281, /* ʒ Ʒ */ - 0x03ac, 462, /* ά Ά */ - 0x03cc, 436, /* ό Ό */ - 0x03d0, 438, /* ϐ Β */ - 0x03d1, 443, /* ϑ Θ */ - 0x03d5, 453, /* ϕ Φ */ - 0x03d6, 446, /* ϖ Π */ - 0x03e3, 499, /* ϣ Ϣ */ - 0x03e5, 499, /* ϥ Ϥ */ - 0x03e7, 499, /* ϧ Ϧ */ - 0x03e9, 499, /* ϩ Ϩ */ - 0x03eb, 499, /* ϫ Ϫ */ - 0x03ed, 499, /* ϭ Ϭ */ - 0x03ef, 499, /* ϯ Ϯ */ - 0x03f0, 414, /* ϰ Κ */ - 0x03f1, 420, /* ϱ Ρ */ - 0x0461, 499, /* ѡ Ѡ */ - 0x0463, 499, /* ѣ Ѣ */ - 0x0465, 499, /* ѥ Ѥ */ - 0x0467, 499, /* ѧ Ѧ */ - 0x0469, 499, /* ѩ Ѩ */ - 0x046b, 499, /* ѫ Ѫ */ - 0x046d, 499, /* ѭ Ѭ */ - 0x046f, 499, /* ѯ Ѯ */ - 0x0471, 499, /* ѱ Ѱ */ - 0x0473, 499, /* ѳ Ѳ */ - 0x0475, 499, /* ѵ Ѵ */ - 0x0477, 499, /* ѷ Ѷ */ - 0x0479, 499, /* ѹ Ѹ */ - 0x047b, 499, /* ѻ Ѻ */ - 0x047d, 499, /* ѽ Ѽ */ - 0x047f, 499, /* ѿ Ѿ */ - 0x0481, 499, /* ҁ Ҁ */ - 0x0491, 499, /* ґ Ґ */ - 0x0493, 499, /* ғ Ғ */ - 0x0495, 499, /* ҕ Ҕ */ - 0x0497, 499, /* җ Җ */ - 0x0499, 499, /* ҙ Ҙ */ - 0x049b, 499, /* қ Қ */ - 0x049d, 499, /* ҝ Ҝ */ - 0x049f, 499, /* ҟ Ҟ */ - 0x04a1, 499, /* ҡ Ҡ */ - 0x04a3, 499, /* ң Ң */ - 0x04a5, 499, /* ҥ Ҥ */ - 0x04a7, 499, /* ҧ Ҧ */ - 0x04a9, 499, /* ҩ Ҩ */ - 0x04ab, 499, /* ҫ Ҫ */ - 0x04ad, 499, /* ҭ Ҭ */ - 0x04af, 499, /* ү Ү */ - 0x04b1, 499, /* ұ Ұ */ - 0x04b3, 499, /* ҳ Ҳ */ - 0x04b5, 499, /* ҵ Ҵ */ - 0x04b7, 499, /* ҷ Ҷ */ - 0x04b9, 499, /* ҹ Ҹ */ - 0x04bb, 499, /* һ Һ */ - 0x04bd, 499, /* ҽ Ҽ */ - 0x04bf, 499, /* ҿ Ҿ */ - 0x04c2, 499, /* ӂ Ӂ */ - 0x04c4, 499, /* ӄ Ӄ */ - 0x04c8, 499, /* ӈ Ӈ */ - 0x04cc, 499, /* ӌ Ӌ */ - 0x04d1, 499, /* ӑ Ӑ */ - 0x04d3, 499, /* ӓ Ӓ */ - 0x04d5, 499, /* ӕ Ӕ */ - 0x04d7, 499, /* ӗ Ӗ */ - 0x04d9, 499, /* ә Ә */ - 0x04db, 499, /* ӛ Ӛ */ - 0x04dd, 499, /* ӝ Ӝ */ - 0x04df, 499, /* ӟ Ӟ */ - 0x04e1, 499, /* ӡ Ӡ */ - 0x04e3, 499, /* ӣ Ӣ */ - 0x04e5, 499, /* ӥ Ӥ */ - 0x04e7, 499, /* ӧ Ӧ */ - 0x04e9, 499, /* ө Ө */ - 0x04eb, 499, /* ӫ Ӫ */ - 0x04ef, 499, /* ӯ Ӯ */ - 0x04f1, 499, /* ӱ Ӱ */ - 0x04f3, 499, /* ӳ Ӳ */ - 0x04f5, 499, /* ӵ Ӵ */ - 0x04f9, 499, /* ӹ Ӹ */ - 0x1e01, 499, /* ḁ Ḁ */ - 0x1e03, 499, /* ḃ Ḃ */ - 0x1e05, 499, /* ḅ Ḅ */ - 0x1e07, 499, /* ḇ Ḇ */ - 0x1e09, 499, /* ḉ Ḉ */ - 0x1e0b, 499, /* ḋ Ḋ */ - 0x1e0d, 499, /* ḍ Ḍ */ - 0x1e0f, 499, /* ḏ Ḏ */ - 0x1e11, 499, /* ḑ Ḑ */ - 0x1e13, 499, /* ḓ Ḓ */ - 0x1e15, 499, /* ḕ Ḕ */ - 0x1e17, 499, /* ḗ Ḗ */ - 0x1e19, 499, /* ḙ Ḙ */ - 0x1e1b, 499, /* ḛ Ḛ */ - 0x1e1d, 499, /* ḝ Ḝ */ - 0x1e1f, 499, /* ḟ Ḟ */ - 0x1e21, 499, /* ḡ Ḡ */ - 0x1e23, 499, /* ḣ Ḣ */ - 0x1e25, 499, /* ḥ Ḥ */ - 0x1e27, 499, /* ḧ Ḧ */ - 0x1e29, 499, /* ḩ Ḩ */ - 0x1e2b, 499, /* ḫ Ḫ */ - 0x1e2d, 499, /* ḭ Ḭ */ - 0x1e2f, 499, /* ḯ Ḯ */ - 0x1e31, 499, /* ḱ Ḱ */ - 0x1e33, 499, /* ḳ Ḳ */ - 0x1e35, 499, /* ḵ Ḵ */ - 0x1e37, 499, /* ḷ Ḷ */ - 0x1e39, 499, /* ḹ Ḹ */ - 0x1e3b, 499, /* ḻ Ḻ */ - 0x1e3d, 499, /* ḽ Ḽ */ - 0x1e3f, 499, /* ḿ Ḿ */ - 0x1e41, 499, /* ṁ Ṁ */ - 0x1e43, 499, /* ṃ Ṃ */ - 0x1e45, 499, /* ṅ Ṅ */ - 0x1e47, 499, /* ṇ Ṇ */ - 0x1e49, 499, /* ṉ Ṉ */ - 0x1e4b, 499, /* ṋ Ṋ */ - 0x1e4d, 499, /* ṍ Ṍ */ - 0x1e4f, 499, /* ṏ Ṏ */ - 0x1e51, 499, /* ṑ Ṑ */ - 0x1e53, 499, /* ṓ Ṓ */ - 0x1e55, 499, /* ṕ Ṕ */ - 0x1e57, 499, /* ṗ Ṗ */ - 0x1e59, 499, /* ṙ Ṙ */ - 0x1e5b, 499, /* ṛ Ṛ */ - 0x1e5d, 499, /* ṝ Ṝ */ - 0x1e5f, 499, /* ṟ Ṟ */ - 0x1e61, 499, /* ṡ Ṡ */ - 0x1e63, 499, /* ṣ Ṣ */ - 0x1e65, 499, /* ṥ Ṥ */ - 0x1e67, 499, /* ṧ Ṧ */ - 0x1e69, 499, /* ṩ Ṩ */ - 0x1e6b, 499, /* ṫ Ṫ */ - 0x1e6d, 499, /* ṭ Ṭ */ - 0x1e6f, 499, /* ṯ Ṯ */ - 0x1e71, 499, /* ṱ Ṱ */ - 0x1e73, 499, /* ṳ Ṳ */ - 0x1e75, 499, /* ṵ Ṵ */ - 0x1e77, 499, /* ṷ Ṷ */ - 0x1e79, 499, /* ṹ Ṹ */ - 0x1e7b, 499, /* ṻ Ṻ */ - 0x1e7d, 499, /* ṽ Ṽ */ - 0x1e7f, 499, /* ṿ Ṿ */ - 0x1e81, 499, /* ẁ Ẁ */ - 0x1e83, 499, /* ẃ Ẃ */ - 0x1e85, 499, /* ẅ Ẅ */ - 0x1e87, 499, /* ẇ Ẇ */ - 0x1e89, 499, /* ẉ Ẉ */ - 0x1e8b, 499, /* ẋ Ẋ */ - 0x1e8d, 499, /* ẍ Ẍ */ - 0x1e8f, 499, /* ẏ Ẏ */ - 0x1e91, 499, /* ẑ Ẑ */ - 0x1e93, 499, /* ẓ Ẓ */ - 0x1e95, 499, /* ẕ Ẕ */ - 0x1ea1, 499, /* ạ Ạ */ - 0x1ea3, 499, /* ả Ả */ - 0x1ea5, 499, /* ấ Ấ */ - 0x1ea7, 499, /* ầ Ầ */ - 0x1ea9, 499, /* ẩ Ẩ */ - 0x1eab, 499, /* ẫ Ẫ */ - 0x1ead, 499, /* ậ Ậ */ - 0x1eaf, 499, /* ắ Ắ */ - 0x1eb1, 499, /* ằ Ằ */ - 0x1eb3, 499, /* ẳ Ẳ */ - 0x1eb5, 499, /* ẵ Ẵ */ - 0x1eb7, 499, /* ặ Ặ */ - 0x1eb9, 499, /* ẹ Ẹ */ - 0x1ebb, 499, /* ẻ Ẻ */ - 0x1ebd, 499, /* ẽ Ẽ */ - 0x1ebf, 499, /* ế Ế */ - 0x1ec1, 499, /* ề Ề */ - 0x1ec3, 499, /* ể Ể */ - 0x1ec5, 499, /* ễ Ễ */ - 0x1ec7, 499, /* ệ Ệ */ - 0x1ec9, 499, /* ỉ Ỉ */ - 0x1ecb, 499, /* ị Ị */ - 0x1ecd, 499, /* ọ Ọ */ - 0x1ecf, 499, /* ỏ Ỏ */ - 0x1ed1, 499, /* ố Ố */ - 0x1ed3, 499, /* ồ Ồ */ - 0x1ed5, 499, /* ổ Ổ */ - 0x1ed7, 499, /* ỗ Ỗ */ - 0x1ed9, 499, /* ộ Ộ */ - 0x1edb, 499, /* ớ Ớ */ - 0x1edd, 499, /* ờ Ờ */ - 0x1edf, 499, /* ở Ở */ - 0x1ee1, 499, /* ỡ Ỡ */ - 0x1ee3, 499, /* ợ Ợ */ - 0x1ee5, 499, /* ụ Ụ */ - 0x1ee7, 499, /* ủ Ủ */ - 0x1ee9, 499, /* ứ Ứ */ - 0x1eeb, 499, /* ừ Ừ */ - 0x1eed, 499, /* ử Ử */ - 0x1eef, 499, /* ữ Ữ */ - 0x1ef1, 499, /* ự Ự */ - 0x1ef3, 499, /* ỳ Ỳ */ - 0x1ef5, 499, /* ỵ Ỵ */ - 0x1ef7, 499, /* ỷ Ỷ */ - 0x1ef9, 499, /* ỹ Ỹ */ - 0x1f51, 508, /* ὑ Ὑ */ - 0x1f53, 508, /* ὓ Ὓ */ - 0x1f55, 508, /* ὕ Ὕ */ - 0x1f57, 508, /* ὗ Ὗ */ - 0x1fb3, 509, /* ᾳ ᾼ */ - 0x1fc3, 509, /* ῃ ῌ */ - 0x1fe5, 507, /* ῥ Ῥ */ - 0x1ff3, 509, /* ῳ ῼ */ -}; - -/* - * upper case ranges - * 3rd col is conversion excess 500 - */ -static const Rune __tolower2[] = -{ - 0x0041, 0x005a, 532, /* A-Z a-z */ - 0x00c0, 0x00d6, 532, /* À-Ö à-ö */ - 0x00d8, 0x00de, 532, /* Ø-Þ ø-þ */ - 0x0189, 0x018a, 705, /* Ɖ-Ɗ ɖ-ɗ */ - 0x018e, 0x018f, 702, /* Ǝ-Ə ɘ-ə */ - 0x01b1, 0x01b2, 717, /* Ʊ-Ʋ ʊ-ʋ */ - 0x0388, 0x038a, 537, /* Έ-Ί έ-ί */ - 0x038e, 0x038f, 563, /* Ύ-Ώ ύ-ώ */ - 0x0391, 0x03a1, 532, /* Α-Ρ α-ρ */ - 0x03a3, 0x03ab, 532, /* Σ-Ϋ σ-ϋ */ - 0x0401, 0x040c, 580, /* Ё-Ќ ё-ќ */ - 0x040e, 0x040f, 580, /* Ў-Џ ў-џ */ - 0x0410, 0x042f, 532, /* А-Я а-я */ - 0x0531, 0x0556, 548, /* Ա-Ֆ ա-ֆ */ - 0x10a0, 0x10c5, 548, /* Ⴀ-Ⴥ ა-ჵ */ - 0x1f08, 0x1f0f, 492, /* Ἀ-Ἇ ἀ-ἇ */ - 0x1f18, 0x1f1d, 492, /* Ἐ-Ἕ ἐ-ἕ */ - 0x1f28, 0x1f2f, 492, /* Ἠ-Ἧ ἠ-ἧ */ - 0x1f38, 0x1f3f, 492, /* Ἰ-Ἷ ἰ-ἷ */ - 0x1f48, 0x1f4d, 492, /* Ὀ-Ὅ ὀ-ὅ */ - 0x1f68, 0x1f6f, 492, /* Ὠ-Ὧ ὠ-ὧ */ - 0x1f88, 0x1f8f, 492, /* ᾈ-ᾏ ᾀ-ᾇ */ - 0x1f98, 0x1f9f, 492, /* ᾘ-ᾟ ᾐ-ᾗ */ - 0x1fa8, 0x1faf, 492, /* ᾨ-ᾯ ᾠ-ᾧ */ - 0x1fb8, 0x1fb9, 492, /* Ᾰ-Ᾱ ᾰ-ᾱ */ - 0x1fba, 0x1fbb, 426, /* Ὰ-Ά ὰ-ά */ - 0x1fc8, 0x1fcb, 414, /* Ὲ-Ή ὲ-ή */ - 0x1fd8, 0x1fd9, 492, /* Ῐ-Ῑ ῐ-ῑ */ - 0x1fda, 0x1fdb, 400, /* Ὶ-Ί ὶ-ί */ - 0x1fe8, 0x1fe9, 492, /* Ῠ-Ῡ ῠ-ῡ */ - 0x1fea, 0x1feb, 388, /* Ὺ-Ύ ὺ-ύ */ - 0x1ff8, 0x1ff9, 372, /* Ὸ-Ό ὸ-ό */ - 0x1ffa, 0x1ffb, 374, /* Ὼ-Ώ ὼ-ώ */ - 0x2160, 0x216f, 516, /* Ⅰ-Ⅿ ⅰ-ⅿ */ - 0x24b6, 0x24cf, 526, /* Ⓐ-Ⓩ ⓐ-ⓩ */ - 0xff21, 0xff3a, 532, /* A-Z a-z */ -}; - -/* - * upper case singlets - * 2nd col is conversion excess 500 - */ -static const Rune __tolower1[] = -{ - 0x0100, 501, /* Ā ā */ - 0x0102, 501, /* Ă ă */ - 0x0104, 501, /* Ą ą */ - 0x0106, 501, /* Ć ć */ - 0x0108, 501, /* Ĉ ĉ */ - 0x010a, 501, /* Ċ ċ */ - 0x010c, 501, /* Č č */ - 0x010e, 501, /* Ď ď */ - 0x0110, 501, /* Đ đ */ - 0x0112, 501, /* Ē ē */ - 0x0114, 501, /* Ĕ ĕ */ - 0x0116, 501, /* Ė ė */ - 0x0118, 501, /* Ę ę */ - 0x011a, 501, /* Ě ě */ - 0x011c, 501, /* Ĝ ĝ */ - 0x011e, 501, /* Ğ ğ */ - 0x0120, 501, /* Ġ ġ */ - 0x0122, 501, /* Ģ ģ */ - 0x0124, 501, /* Ĥ ĥ */ - 0x0126, 501, /* Ħ ħ */ - 0x0128, 501, /* Ĩ ĩ */ - 0x012a, 501, /* Ī ī */ - 0x012c, 501, /* Ĭ ĭ */ - 0x012e, 501, /* Į į */ - 0x0130, 301, /* İ i */ - 0x0132, 501, /* IJ ij */ - 0x0134, 501, /* Ĵ ĵ */ - 0x0136, 501, /* Ķ ķ */ - 0x0139, 501, /* Ĺ ĺ */ - 0x013b, 501, /* Ļ ļ */ - 0x013d, 501, /* Ľ ľ */ - 0x013f, 501, /* Ŀ ŀ */ - 0x0141, 501, /* Ł ł */ - 0x0143, 501, /* Ń ń */ - 0x0145, 501, /* Ņ ņ */ - 0x0147, 501, /* Ň ň */ - 0x014a, 501, /* Ŋ ŋ */ - 0x014c, 501, /* Ō ō */ - 0x014e, 501, /* Ŏ ŏ */ - 0x0150, 501, /* Ő ő */ - 0x0152, 501, /* Œ œ */ - 0x0154, 501, /* Ŕ ŕ */ - 0x0156, 501, /* Ŗ ŗ */ - 0x0158, 501, /* Ř ř */ - 0x015a, 501, /* Ś ś */ - 0x015c, 501, /* Ŝ ŝ */ - 0x015e, 501, /* Ş ş */ - 0x0160, 501, /* Š š */ - 0x0162, 501, /* Ţ ţ */ - 0x0164, 501, /* Ť ť */ - 0x0166, 501, /* Ŧ ŧ */ - 0x0168, 501, /* Ũ ũ */ - 0x016a, 501, /* Ū ū */ - 0x016c, 501, /* Ŭ ŭ */ - 0x016e, 501, /* Ů ů */ - 0x0170, 501, /* Ű ű */ - 0x0172, 501, /* Ų ų */ - 0x0174, 501, /* Ŵ ŵ */ - 0x0176, 501, /* Ŷ ŷ */ - 0x0178, 379, /* Ÿ ÿ */ - 0x0179, 501, /* Ź ź */ - 0x017b, 501, /* Ż ż */ - 0x017d, 501, /* Ž ž */ - 0x0181, 710, /* Ɓ ɓ */ - 0x0182, 501, /* Ƃ ƃ */ - 0x0184, 501, /* Ƅ ƅ */ - 0x0186, 706, /* Ɔ ɔ */ - 0x0187, 501, /* Ƈ ƈ */ - 0x018b, 501, /* Ƌ ƌ */ - 0x0190, 703, /* Ɛ ɛ */ - 0x0191, 501, /* Ƒ ƒ */ - 0x0193, 705, /* Ɠ ɠ */ - 0x0194, 707, /* Ɣ ɣ */ - 0x0196, 711, /* Ɩ ɩ */ - 0x0197, 709, /* Ɨ ɨ */ - 0x0198, 501, /* Ƙ ƙ */ - 0x019c, 711, /* Ɯ ɯ */ - 0x019d, 713, /* Ɲ ɲ */ - 0x01a0, 501, /* Ơ ơ */ - 0x01a2, 501, /* Ƣ ƣ */ - 0x01a4, 501, /* Ƥ ƥ */ - 0x01a7, 501, /* Ƨ ƨ */ - 0x01a9, 718, /* Ʃ ʃ */ - 0x01ac, 501, /* Ƭ ƭ */ - 0x01ae, 718, /* Ʈ ʈ */ - 0x01af, 501, /* Ư ư */ - 0x01b3, 501, /* Ƴ ƴ */ - 0x01b5, 501, /* Ƶ ƶ */ - 0x01b7, 719, /* Ʒ ʒ */ - 0x01b8, 501, /* Ƹ ƹ */ - 0x01bc, 501, /* Ƽ ƽ */ - 0x01c4, 502, /* DŽ dž */ - 0x01c5, 501, /* Dž dž */ - 0x01c7, 502, /* LJ lj */ - 0x01c8, 501, /* Lj lj */ - 0x01ca, 502, /* NJ nj */ - 0x01cb, 501, /* Nj nj */ - 0x01cd, 501, /* Ǎ ǎ */ - 0x01cf, 501, /* Ǐ ǐ */ - 0x01d1, 501, /* Ǒ ǒ */ - 0x01d3, 501, /* Ǔ ǔ */ - 0x01d5, 501, /* Ǖ ǖ */ - 0x01d7, 501, /* Ǘ ǘ */ - 0x01d9, 501, /* Ǚ ǚ */ - 0x01db, 501, /* Ǜ ǜ */ - 0x01de, 501, /* Ǟ ǟ */ - 0x01e0, 501, /* Ǡ ǡ */ - 0x01e2, 501, /* Ǣ ǣ */ - 0x01e4, 501, /* Ǥ ǥ */ - 0x01e6, 501, /* Ǧ ǧ */ - 0x01e8, 501, /* Ǩ ǩ */ - 0x01ea, 501, /* Ǫ ǫ */ - 0x01ec, 501, /* Ǭ ǭ */ - 0x01ee, 501, /* Ǯ ǯ */ - 0x01f1, 502, /* DZ dz */ - 0x01f2, 501, /* Dz dz */ - 0x01f4, 501, /* Ǵ ǵ */ - 0x01fa, 501, /* Ǻ ǻ */ - 0x01fc, 501, /* Ǽ ǽ */ - 0x01fe, 501, /* Ǿ ǿ */ - 0x0200, 501, /* Ȁ ȁ */ - 0x0202, 501, /* Ȃ ȃ */ - 0x0204, 501, /* Ȅ ȅ */ - 0x0206, 501, /* Ȇ ȇ */ - 0x0208, 501, /* Ȉ ȉ */ - 0x020a, 501, /* Ȋ ȋ */ - 0x020c, 501, /* Ȍ ȍ */ - 0x020e, 501, /* Ȏ ȏ */ - 0x0210, 501, /* Ȑ ȑ */ - 0x0212, 501, /* Ȓ ȓ */ - 0x0214, 501, /* Ȕ ȕ */ - 0x0216, 501, /* Ȗ ȗ */ - 0x0386, 538, /* Ά ά */ - 0x038c, 564, /* Ό ό */ - 0x03e2, 501, /* Ϣ ϣ */ - 0x03e4, 501, /* Ϥ ϥ */ - 0x03e6, 501, /* Ϧ ϧ */ - 0x03e8, 501, /* Ϩ ϩ */ - 0x03ea, 501, /* Ϫ ϫ */ - 0x03ec, 501, /* Ϭ ϭ */ - 0x03ee, 501, /* Ϯ ϯ */ - 0x0460, 501, /* Ѡ ѡ */ - 0x0462, 501, /* Ѣ ѣ */ - 0x0464, 501, /* Ѥ ѥ */ - 0x0466, 501, /* Ѧ ѧ */ - 0x0468, 501, /* Ѩ ѩ */ - 0x046a, 501, /* Ѫ ѫ */ - 0x046c, 501, /* Ѭ ѭ */ - 0x046e, 501, /* Ѯ ѯ */ - 0x0470, 501, /* Ѱ ѱ */ - 0x0472, 501, /* Ѳ ѳ */ - 0x0474, 501, /* Ѵ ѵ */ - 0x0476, 501, /* Ѷ ѷ */ - 0x0478, 501, /* Ѹ ѹ */ - 0x047a, 501, /* Ѻ ѻ */ - 0x047c, 501, /* Ѽ ѽ */ - 0x047e, 501, /* Ѿ ѿ */ - 0x0480, 501, /* Ҁ ҁ */ - 0x0490, 501, /* Ґ ґ */ - 0x0492, 501, /* Ғ ғ */ - 0x0494, 501, /* Ҕ ҕ */ - 0x0496, 501, /* Җ җ */ - 0x0498, 501, /* Ҙ ҙ */ - 0x049a, 501, /* Қ қ */ - 0x049c, 501, /* Ҝ ҝ */ - 0x049e, 501, /* Ҟ ҟ */ - 0x04a0, 501, /* Ҡ ҡ */ - 0x04a2, 501, /* Ң ң */ - 0x04a4, 501, /* Ҥ ҥ */ - 0x04a6, 501, /* Ҧ ҧ */ - 0x04a8, 501, /* Ҩ ҩ */ - 0x04aa, 501, /* Ҫ ҫ */ - 0x04ac, 501, /* Ҭ ҭ */ - 0x04ae, 501, /* Ү ү */ - 0x04b0, 501, /* Ұ ұ */ - 0x04b2, 501, /* Ҳ ҳ */ - 0x04b4, 501, /* Ҵ ҵ */ - 0x04b6, 501, /* Ҷ ҷ */ - 0x04b8, 501, /* Ҹ ҹ */ - 0x04ba, 501, /* Һ һ */ - 0x04bc, 501, /* Ҽ ҽ */ - 0x04be, 501, /* Ҿ ҿ */ - 0x04c1, 501, /* Ӂ ӂ */ - 0x04c3, 501, /* Ӄ ӄ */ - 0x04c7, 501, /* Ӈ ӈ */ - 0x04cb, 501, /* Ӌ ӌ */ - 0x04d0, 501, /* Ӑ ӑ */ - 0x04d2, 501, /* Ӓ ӓ */ - 0x04d4, 501, /* Ӕ ӕ */ - 0x04d6, 501, /* Ӗ ӗ */ - 0x04d8, 501, /* Ә ә */ - 0x04da, 501, /* Ӛ ӛ */ - 0x04dc, 501, /* Ӝ ӝ */ - 0x04de, 501, /* Ӟ ӟ */ - 0x04e0, 501, /* Ӡ ӡ */ - 0x04e2, 501, /* Ӣ ӣ */ - 0x04e4, 501, /* Ӥ ӥ */ - 0x04e6, 501, /* Ӧ ӧ */ - 0x04e8, 501, /* Ө ө */ - 0x04ea, 501, /* Ӫ ӫ */ - 0x04ee, 501, /* Ӯ ӯ */ - 0x04f0, 501, /* Ӱ ӱ */ - 0x04f2, 501, /* Ӳ ӳ */ - 0x04f4, 501, /* Ӵ ӵ */ - 0x04f8, 501, /* Ӹ ӹ */ - 0x1e00, 501, /* Ḁ ḁ */ - 0x1e02, 501, /* Ḃ ḃ */ - 0x1e04, 501, /* Ḅ ḅ */ - 0x1e06, 501, /* Ḇ ḇ */ - 0x1e08, 501, /* Ḉ ḉ */ - 0x1e0a, 501, /* Ḋ ḋ */ - 0x1e0c, 501, /* Ḍ ḍ */ - 0x1e0e, 501, /* Ḏ ḏ */ - 0x1e10, 501, /* Ḑ ḑ */ - 0x1e12, 501, /* Ḓ ḓ */ - 0x1e14, 501, /* Ḕ ḕ */ - 0x1e16, 501, /* Ḗ ḗ */ - 0x1e18, 501, /* Ḙ ḙ */ - 0x1e1a, 501, /* Ḛ ḛ */ - 0x1e1c, 501, /* Ḝ ḝ */ - 0x1e1e, 501, /* Ḟ ḟ */ - 0x1e20, 501, /* Ḡ ḡ */ - 0x1e22, 501, /* Ḣ ḣ */ - 0x1e24, 501, /* Ḥ ḥ */ - 0x1e26, 501, /* Ḧ ḧ */ - 0x1e28, 501, /* Ḩ ḩ */ - 0x1e2a, 501, /* Ḫ ḫ */ - 0x1e2c, 501, /* Ḭ ḭ */ - 0x1e2e, 501, /* Ḯ ḯ */ - 0x1e30, 501, /* Ḱ ḱ */ - 0x1e32, 501, /* Ḳ ḳ */ - 0x1e34, 501, /* Ḵ ḵ */ - 0x1e36, 501, /* Ḷ ḷ */ - 0x1e38, 501, /* Ḹ ḹ */ - 0x1e3a, 501, /* Ḻ ḻ */ - 0x1e3c, 501, /* Ḽ ḽ */ - 0x1e3e, 501, /* Ḿ ḿ */ - 0x1e40, 501, /* Ṁ ṁ */ - 0x1e42, 501, /* Ṃ ṃ */ - 0x1e44, 501, /* Ṅ ṅ */ - 0x1e46, 501, /* Ṇ ṇ */ - 0x1e48, 501, /* Ṉ ṉ */ - 0x1e4a, 501, /* Ṋ ṋ */ - 0x1e4c, 501, /* Ṍ ṍ */ - 0x1e4e, 501, /* Ṏ ṏ */ - 0x1e50, 501, /* Ṑ ṑ */ - 0x1e52, 501, /* Ṓ ṓ */ - 0x1e54, 501, /* Ṕ ṕ */ - 0x1e56, 501, /* Ṗ ṗ */ - 0x1e58, 501, /* Ṙ ṙ */ - 0x1e5a, 501, /* Ṛ ṛ */ - 0x1e5c, 501, /* Ṝ ṝ */ - 0x1e5e, 501, /* Ṟ ṟ */ - 0x1e60, 501, /* Ṡ ṡ */ - 0x1e62, 501, /* Ṣ ṣ */ - 0x1e64, 501, /* Ṥ ṥ */ - 0x1e66, 501, /* Ṧ ṧ */ - 0x1e68, 501, /* Ṩ ṩ */ - 0x1e6a, 501, /* Ṫ ṫ */ - 0x1e6c, 501, /* Ṭ ṭ */ - 0x1e6e, 501, /* Ṯ ṯ */ - 0x1e70, 501, /* Ṱ ṱ */ - 0x1e72, 501, /* Ṳ ṳ */ - 0x1e74, 501, /* Ṵ ṵ */ - 0x1e76, 501, /* Ṷ ṷ */ - 0x1e78, 501, /* Ṹ ṹ */ - 0x1e7a, 501, /* Ṻ ṻ */ - 0x1e7c, 501, /* Ṽ ṽ */ - 0x1e7e, 501, /* Ṿ ṿ */ - 0x1e80, 501, /* Ẁ ẁ */ - 0x1e82, 501, /* Ẃ ẃ */ - 0x1e84, 501, /* Ẅ ẅ */ - 0x1e86, 501, /* Ẇ ẇ */ - 0x1e88, 501, /* Ẉ ẉ */ - 0x1e8a, 501, /* Ẋ ẋ */ - 0x1e8c, 501, /* Ẍ ẍ */ - 0x1e8e, 501, /* Ẏ ẏ */ - 0x1e90, 501, /* Ẑ ẑ */ - 0x1e92, 501, /* Ẓ ẓ */ - 0x1e94, 501, /* Ẕ ẕ */ - 0x1ea0, 501, /* Ạ ạ */ - 0x1ea2, 501, /* Ả ả */ - 0x1ea4, 501, /* Ấ ấ */ - 0x1ea6, 501, /* Ầ ầ */ - 0x1ea8, 501, /* Ẩ ẩ */ - 0x1eaa, 501, /* Ẫ ẫ */ - 0x1eac, 501, /* Ậ ậ */ - 0x1eae, 501, /* Ắ ắ */ - 0x1eb0, 501, /* Ằ ằ */ - 0x1eb2, 501, /* Ẳ ẳ */ - 0x1eb4, 501, /* Ẵ ẵ */ - 0x1eb6, 501, /* Ặ ặ */ - 0x1eb8, 501, /* Ẹ ẹ */ - 0x1eba, 501, /* Ẻ ẻ */ - 0x1ebc, 501, /* Ẽ ẽ */ - 0x1ebe, 501, /* Ế ế */ - 0x1ec0, 501, /* Ề ề */ - 0x1ec2, 501, /* Ể ể */ - 0x1ec4, 501, /* Ễ ễ */ - 0x1ec6, 501, /* Ệ ệ */ - 0x1ec8, 501, /* Ỉ ỉ */ - 0x1eca, 501, /* Ị ị */ - 0x1ecc, 501, /* Ọ ọ */ - 0x1ece, 501, /* Ỏ ỏ */ - 0x1ed0, 501, /* Ố ố */ - 0x1ed2, 501, /* Ồ ồ */ - 0x1ed4, 501, /* Ổ ổ */ - 0x1ed6, 501, /* Ỗ ỗ */ - 0x1ed8, 501, /* Ộ ộ */ - 0x1eda, 501, /* Ớ ớ */ - 0x1edc, 501, /* Ờ ờ */ - 0x1ede, 501, /* Ở ở */ - 0x1ee0, 501, /* Ỡ ỡ */ - 0x1ee2, 501, /* Ợ ợ */ - 0x1ee4, 501, /* Ụ ụ */ - 0x1ee6, 501, /* Ủ ủ */ - 0x1ee8, 501, /* Ứ ứ */ - 0x1eea, 501, /* Ừ ừ */ - 0x1eec, 501, /* Ử ử */ - 0x1eee, 501, /* Ữ ữ */ - 0x1ef0, 501, /* Ự ự */ - 0x1ef2, 501, /* Ỳ ỳ */ - 0x1ef4, 501, /* Ỵ ỵ */ - 0x1ef6, 501, /* Ỷ ỷ */ - 0x1ef8, 501, /* Ỹ ỹ */ - 0x1f59, 492, /* Ὑ ὑ */ - 0x1f5b, 492, /* Ὓ ὓ */ - 0x1f5d, 492, /* Ὕ ὕ */ - 0x1f5f, 492, /* Ὗ ὗ */ - 0x1fbc, 491, /* ᾼ ᾳ */ - 0x1fcc, 491, /* ῌ ῃ */ - 0x1fec, 493, /* Ῥ ῥ */ - 0x1ffc, 491, /* ῼ ῳ */ -}; - -/* - * title characters are those between - * upper and lower case. ie DZ Dz dz - */ -static const Rune __totitle1[] = -{ - 0x01c4, 501, /* DŽ Dž */ - 0x01c6, 499, /* dž Dž */ - 0x01c7, 501, /* LJ Lj */ - 0x01c9, 499, /* lj Lj */ - 0x01ca, 501, /* NJ Nj */ - 0x01cc, 499, /* nj Nj */ - 0x01f1, 501, /* DZ Dz */ - 0x01f3, 499, /* dz Dz */ -}; - -static const Rune * -bsearch(Rune c, const Rune *t, int n, int ne) -{ - const Rune *p; - int m; - - while(n > 1) { - m = n/2; - p = t + m*ne; - if(c >= p[0]) { - t = p; - n = n-m; - } else - n = m; - } - if(n && c >= t[0]) - return t; - return 0; -} - -Rune -tolowerrune(Rune c) -{ - const Rune *p; - - p = bsearch(c, __tolower2, nelem(__tolower2)/3, 3); - if(p && c >= p[0] && c <= p[1]) - return c + p[2] - 500; - p = bsearch(c, __tolower1, nelem(__tolower1)/2, 2); - if(p && c == p[0]) - return c + p[1] - 500; - return c; -} - -Rune -toupperrune(Rune c) -{ - const Rune *p; - - p = bsearch(c, __toupper2, nelem(__toupper2)/3, 3); - if(p && c >= p[0] && c <= p[1]) - return c + p[2] - 500; - p = bsearch(c, __toupper1, nelem(__toupper1)/2, 2); - if(p && c == p[0]) - return c + p[1] - 500; - return c; -} - -Rune -totitlerune(Rune c) -{ - const Rune *p; - - p = bsearch(c, __totitle1, nelem(__totitle1)/2, 2); - if(p && c == p[0]) - return c + p[1] - 500; - return c; -} - -int -islowerrune(Rune c) -{ - const Rune *p; - - p = bsearch(c, __toupper2, nelem(__toupper2)/3, 3); - if(p && c >= p[0] && c <= p[1]) - return 1; - p = bsearch(c, __toupper1, nelem(__toupper1)/2, 2); - if(p && c == p[0]) - return 1; - return 0; -} - -int -isupperrune(Rune c) -{ - const Rune *p; - - p = bsearch(c, __tolower2, nelem(__tolower2)/3, 3); - if(p && c >= p[0] && c <= p[1]) - return 1; - p = bsearch(c, __tolower1, nelem(__tolower1)/2, 2); - if(p && c == p[0]) - return 1; - return 0; -} - -int -isalpharune(Rune c) -{ - const Rune *p; - - if(isupperrune(c) || islowerrune(c)) - return 1; - p = bsearch(c, __alpha2, nelem(__alpha2)/2, 2); - if(p && c >= p[0] && c <= p[1]) - return 1; - p = bsearch(c, __alpha1, nelem(__alpha1), 1); - if(p && c == p[0]) - return 1; - return 0; -} - -int -istitlerune(Rune c) -{ - return isupperrune(c) && islowerrune(c); -} - -int -isspacerune(Rune c) -{ - const Rune *p; - - p = bsearch(c, __space2, nelem(__space2)/2, 2); - if(p && c >= p[0] && c <= p[1]) - return 1; - return 0; -}