anil.recoil.org
1/* encoding/utf8.c -- VERSION 1.0
2 *
3 * Guerrilla line editing library against the idea that a line editing lib
4 * needs to be 20,000 lines of C code.
5 *
6 * You can find the latest source code at:
7 *
8 * http://github.com/antirez/linenoise
9 *
10 * Does a number of crazy assumptions that happen to be true in 99.9999% of
11 * the 2010 UNIX computers around.
12 *
13 * ------------------------------------------------------------------------
14 *
15 * Copyright (c) 2010-2014, Salvatore Sanfilippo <antirez at gmail dot com>
16 * Copyright (c) 2010-2013, Pieter Noordhuis <pcnoordhuis at gmail dot com>
17 *
18 * All rights reserved.
19 *
20 * Redistribution and use in source and binary forms, with or without
21 * modification, are permitted provided that the following conditions are
22 * met:
23 *
24 * * Redistributions of source code must retain the above copyright
25 * notice, this list of conditions and the following disclaimer.
26 *
27 * * Redistributions in binary form must reproduce the above copyright
28 * notice, this list of conditions and the following disclaimer in the
29 * documentation and/or other materials provided with the distribution.
30 *
31 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
32 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
33 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
34 * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
35 * HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
36 * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
37 * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
38 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
39 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
40 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
41 * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
42 */
43
44#include <unistd.h>
45#include <stdio.h>
46
47#define UNUSED(x) (void)(x)
48
49/* ============================ UTF8 utilities ============================== */
50
51static unsigned long wideCharTable[][2] = {
52 { 0x1100, 0x115F }, { 0x231A, 0x231B }, { 0x2329, 0x232A }, { 0x23E9, 0x23EC },
53 { 0x23F0, 0x23F0 }, { 0x23F3, 0x23F3 }, { 0x25FD, 0x25FE }, { 0x2614, 0x2615 },
54 { 0x2630, 0x2637 }, { 0x2648, 0x2653 }, { 0x267F, 0x267F }, { 0x268A, 0x268F },
55 { 0x2693, 0x2693 }, { 0x26A1, 0x26A1 }, { 0x26AA, 0x26AB }, { 0x26BD, 0x26BE },
56 { 0x26C4, 0x26C5 }, { 0x26CE, 0x26CE }, { 0x26D4, 0x26D4 }, { 0x26EA, 0x26EA },
57 { 0x26F2, 0x26F3 }, { 0x26F5, 0x26F5 }, { 0x26FA, 0x26FA }, { 0x26FD, 0x26FD },
58 { 0x2705, 0x2705 }, { 0x270A, 0x270B }, { 0x2728, 0x2728 }, { 0x274C, 0x274C },
59 { 0x274E, 0x274E }, { 0x2753, 0x2755 }, { 0x2757, 0x2757 }, { 0x2795, 0x2797 },
60 { 0x27B0, 0x27B0 }, { 0x27BF, 0x27BF }, { 0x2B1B, 0x2B1C }, { 0x2B50, 0x2B50 },
61 { 0x2B55, 0x2B55 }, { 0x2E80, 0x2E99 }, { 0x2E9B, 0x2EF3 }, { 0x2F00, 0x2FD5 },
62 { 0x2FF0, 0x303E }, { 0x3041, 0x3096 }, { 0x3099, 0x30FF }, { 0x3105, 0x312F },
63 { 0x3131, 0x318E }, { 0x3190, 0x31E5 }, { 0x31EF, 0x321E }, { 0x3220, 0x3247 },
64 { 0x3250, 0xA48C }, { 0xA490, 0xA4C6 }, { 0xA960, 0xA97C }, { 0xAC00, 0xD7A3 },
65 { 0xF900, 0xFAFF }, { 0xFE10, 0xFE19 }, { 0xFE30, 0xFE52 }, { 0xFE54, 0xFE66 },
66 { 0xFE68, 0xFE6B }, { 0xFF01, 0xFF60 }, { 0xFFE0, 0xFFE6 }, { 0x16FE0, 0x16FE4 },
67 { 0x16FF0, 0x16FF1 }, { 0x17000, 0x187F7 }, { 0x18800, 0x18CD5 }, { 0x18CFF, 0x18D08 },
68 { 0x1AFF0, 0x1AFF3 }, { 0x1AFF5, 0x1AFFB }, { 0x1AFFD, 0x1AFFE }, { 0x1B000, 0x1B122 },
69 { 0x1B132, 0x1B132 }, { 0x1B150, 0x1B152 }, { 0x1B155, 0x1B155 }, { 0x1B164, 0x1B167 },
70 { 0x1B170, 0x1B2FB }, { 0x1D300, 0x1D356 }, { 0x1D360, 0x1D376 }, { 0x1F004, 0x1F004 },
71 { 0x1F0CF, 0x1F0CF }, { 0x1F18E, 0x1F18E }, { 0x1F191, 0x1F19A }, { 0x1F200, 0x1F202 },
72 { 0x1F210, 0x1F23B }, { 0x1F240, 0x1F248 }, { 0x1F250, 0x1F251 }, { 0x1F260, 0x1F265 },
73 { 0x1F300, 0x1F320 }, { 0x1F32D, 0x1F335 }, { 0x1F337, 0x1F37C }, { 0x1F37E, 0x1F393 },
74 { 0x1F3A0, 0x1F3CA }, { 0x1F3CF, 0x1F3D3 }, { 0x1F3E0, 0x1F3F0 }, { 0x1F3F4, 0x1F3F4 },
75 { 0x1F3F8, 0x1F43E }, { 0x1F440, 0x1F440 }, { 0x1F442, 0x1F4FC }, { 0x1F4FF, 0x1F53D },
76 { 0x1F54B, 0x1F54E }, { 0x1F550, 0x1F567 }, { 0x1F57A, 0x1F57A }, { 0x1F595, 0x1F596 },
77 { 0x1F5A4, 0x1F5A4 }, { 0x1F5FB, 0x1F64F }, { 0x1F680, 0x1F6C5 }, { 0x1F6CC, 0x1F6CC },
78 { 0x1F6D0, 0x1F6D2 }, { 0x1F6D5, 0x1F6D7 }, { 0x1F6DC, 0x1F6DF }, { 0x1F6EB, 0x1F6EC },
79 { 0x1F6F4, 0x1F6FC }, { 0x1F7E0, 0x1F7EB }, { 0x1F7F0, 0x1F7F0 }, { 0x1F90C, 0x1F93A },
80 { 0x1F93C, 0x1F945 }, { 0x1F947, 0x1F9FF }, { 0x1FA70, 0x1FA7C }, { 0x1FA80, 0x1FA89 },
81 { 0x1FA8F, 0x1FAC6 }, { 0x1FACE, 0x1FADC }, { 0x1FADF, 0x1FAE9 }, { 0x1FAF0, 0x1FAF8 },
82 { 0x20000, 0x2FFFD }, { 0x30000, 0x3FFFD },
83};
84
85static size_t wideCharTableSize = sizeof(wideCharTable) / sizeof(wideCharTable[0]);
86
87static unsigned long combiningCharTable[] = {
88 0x0300, 0x0301, 0x0302, 0x0303, 0x0304, 0x0305, 0x0306, 0x0307,
89 0x0308, 0x0309, 0x030A, 0x030B, 0x030C, 0x030D, 0x030E, 0x030F,
90 0x0310, 0x0311, 0x0312, 0x0313, 0x0314, 0x0315, 0x0316, 0x0317,
91 0x0318, 0x0319, 0x031A, 0x031B, 0x031C, 0x031D, 0x031E, 0x031F,
92 0x0320, 0x0321, 0x0322, 0x0323, 0x0324, 0x0325, 0x0326, 0x0327,
93 0x0328, 0x0329, 0x032A, 0x032B, 0x032C, 0x032D, 0x032E, 0x032F,
94 0x0330, 0x0331, 0x0332, 0x0333, 0x0334, 0x0335, 0x0336, 0x0337,
95 0x0338, 0x0339, 0x033A, 0x033B, 0x033C, 0x033D, 0x033E, 0x033F,
96 0x0340, 0x0341, 0x0342, 0x0343, 0x0344, 0x0345, 0x0346, 0x0347,
97 0x0348, 0x0349, 0x034A, 0x034B, 0x034C, 0x034D, 0x034E, 0x034F,
98 0x0350, 0x0351, 0x0352, 0x0353, 0x0354, 0x0355, 0x0356, 0x0357,
99 0x0358, 0x0359, 0x035A, 0x035B, 0x035C, 0x035D, 0x035E, 0x035F,
100 0x0360, 0x0361, 0x0362, 0x0363, 0x0364, 0x0365, 0x0366, 0x0367,
101 0x0368, 0x0369, 0x036A, 0x036B, 0x036C, 0x036D, 0x036E, 0x036F,
102 0x0483, 0x0484, 0x0485, 0x0486, 0x0487, 0x0591, 0x0592, 0x0593,
103 0x0594, 0x0595, 0x0596, 0x0597, 0x0598, 0x0599, 0x059A, 0x059B,
104 0x059C, 0x059D, 0x059E, 0x059F, 0x05A0, 0x05A1, 0x05A2, 0x05A3,
105 0x05A4, 0x05A5, 0x05A6, 0x05A7, 0x05A8, 0x05A9, 0x05AA, 0x05AB,
106 0x05AC, 0x05AD, 0x05AE, 0x05AF, 0x05B0, 0x05B1, 0x05B2, 0x05B3,
107 0x05B4, 0x05B5, 0x05B6, 0x05B7, 0x05B8, 0x05B9, 0x05BA, 0x05BB,
108 0x05BC, 0x05BD, 0x05BF, 0x05C1, 0x05C2, 0x05C4, 0x05C5, 0x05C7,
109 0x0610, 0x0611, 0x0612, 0x0613, 0x0614, 0x0615, 0x0616, 0x0617,
110 0x0618, 0x0619, 0x061A, 0x064B, 0x064C, 0x064D, 0x064E, 0x064F,
111 0x0650, 0x0651, 0x0652, 0x0653, 0x0654, 0x0655, 0x0656, 0x0657,
112 0x0658, 0x0659, 0x065A, 0x065B, 0x065C, 0x065D, 0x065E, 0x065F,
113 0x0670, 0x06D6, 0x06D7, 0x06D8, 0x06D9, 0x06DA, 0x06DB, 0x06DC,
114 0x06DF, 0x06E0, 0x06E1, 0x06E2, 0x06E3, 0x06E4, 0x06E7, 0x06E8,
115 0x06EA, 0x06EB, 0x06EC, 0x06ED, 0x0711, 0x0730, 0x0731, 0x0732,
116 0x0733, 0x0734, 0x0735, 0x0736, 0x0737, 0x0738, 0x0739, 0x073A,
117 0x073B, 0x073C, 0x073D, 0x073E, 0x073F, 0x0740, 0x0741, 0x0742,
118 0x0743, 0x0744, 0x0745, 0x0746, 0x0747, 0x0748, 0x0749, 0x074A,
119 0x07A6, 0x07A7, 0x07A8, 0x07A9, 0x07AA, 0x07AB, 0x07AC, 0x07AD,
120 0x07AE, 0x07AF, 0x07B0, 0x07EB, 0x07EC, 0x07ED, 0x07EE, 0x07EF,
121 0x07F0, 0x07F1, 0x07F2, 0x07F3, 0x07FD, 0x0816, 0x0817, 0x0818,
122 0x0819, 0x081B, 0x081C, 0x081D, 0x081E, 0x081F, 0x0820, 0x0821,
123 0x0822, 0x0823, 0x0825, 0x0826, 0x0827, 0x0829, 0x082A, 0x082B,
124 0x082C, 0x082D, 0x0859, 0x085A, 0x085B, 0x0897, 0x0898, 0x0899,
125 0x089A, 0x089B, 0x089C, 0x089D, 0x089E, 0x089F, 0x08CA, 0x08CB,
126 0x08CC, 0x08CD, 0x08CE, 0x08CF, 0x08D0, 0x08D1, 0x08D2, 0x08D3,
127 0x08D4, 0x08D5, 0x08D6, 0x08D7, 0x08D8, 0x08D9, 0x08DA, 0x08DB,
128 0x08DC, 0x08DD, 0x08DE, 0x08DF, 0x08E0, 0x08E1, 0x08E3, 0x08E4,
129 0x08E5, 0x08E6, 0x08E7, 0x08E8, 0x08E9, 0x08EA, 0x08EB, 0x08EC,
130 0x08ED, 0x08EE, 0x08EF, 0x08F0, 0x08F1, 0x08F2, 0x08F3, 0x08F4,
131 0x08F5, 0x08F6, 0x08F7, 0x08F8, 0x08F9, 0x08FA, 0x08FB, 0x08FC,
132 0x08FD, 0x08FE, 0x08FF, 0x0900, 0x0901, 0x0902, 0x093A, 0x093C,
133 0x0941, 0x0942, 0x0943, 0x0944, 0x0945, 0x0946, 0x0947, 0x0948,
134 0x094D, 0x0951, 0x0952, 0x0953, 0x0954, 0x0955, 0x0956, 0x0957,
135 0x0962, 0x0963, 0x0981, 0x09BC, 0x09C1, 0x09C2, 0x09C3, 0x09C4,
136 0x09CD, 0x09E2, 0x09E3, 0x09FE, 0x0A01, 0x0A02, 0x0A3C, 0x0A41,
137 0x0A42, 0x0A47, 0x0A48, 0x0A4B, 0x0A4C, 0x0A4D, 0x0A51, 0x0A70,
138 0x0A71, 0x0A75, 0x0A81, 0x0A82, 0x0ABC, 0x0AC1, 0x0AC2, 0x0AC3,
139 0x0AC4, 0x0AC5, 0x0AC7, 0x0AC8, 0x0ACD, 0x0AE2, 0x0AE3, 0x0AFA,
140 0x0AFB, 0x0AFC, 0x0AFD, 0x0AFE, 0x0AFF, 0x0B01, 0x0B3C, 0x0B3F,
141 0x0B41, 0x0B42, 0x0B43, 0x0B44, 0x0B4D, 0x0B55, 0x0B56, 0x0B62,
142 0x0B63, 0x0B82, 0x0BC0, 0x0BCD, 0x0C00, 0x0C04, 0x0C3C, 0x0C3E,
143 0x0C3F, 0x0C40, 0x0C46, 0x0C47, 0x0C48, 0x0C4A, 0x0C4B, 0x0C4C,
144 0x0C4D, 0x0C55, 0x0C56, 0x0C62, 0x0C63, 0x0C81, 0x0CBC, 0x0CBF,
145 0x0CC6, 0x0CCC, 0x0CCD, 0x0CE2, 0x0CE3, 0x0D00, 0x0D01, 0x0D3B,
146 0x0D3C, 0x0D41, 0x0D42, 0x0D43, 0x0D44, 0x0D4D, 0x0D62, 0x0D63,
147 0x0D81, 0x0DCA, 0x0DD2, 0x0DD3, 0x0DD4, 0x0DD6, 0x0E31, 0x0E34,
148 0x0E35, 0x0E36, 0x0E37, 0x0E38, 0x0E39, 0x0E3A, 0x0E47, 0x0E48,
149 0x0E49, 0x0E4A, 0x0E4B, 0x0E4C, 0x0E4D, 0x0E4E, 0x0EB1, 0x0EB4,
150 0x0EB5, 0x0EB6, 0x0EB7, 0x0EB8, 0x0EB9, 0x0EBA, 0x0EBB, 0x0EBC,
151 0x0EC8, 0x0EC9, 0x0ECA, 0x0ECB, 0x0ECC, 0x0ECD, 0x0ECE, 0x0F18,
152 0x0F19, 0x0F35, 0x0F37, 0x0F39, 0x0F71, 0x0F72, 0x0F73, 0x0F74,
153 0x0F75, 0x0F76, 0x0F77, 0x0F78, 0x0F79, 0x0F7A, 0x0F7B, 0x0F7C,
154 0x0F7D, 0x0F7E, 0x0F80, 0x0F81, 0x0F82, 0x0F83, 0x0F84, 0x0F86,
155 0x0F87, 0x0F8D, 0x0F8E, 0x0F8F, 0x0F90, 0x0F91, 0x0F92, 0x0F93,
156 0x0F94, 0x0F95, 0x0F96, 0x0F97, 0x0F99, 0x0F9A, 0x0F9B, 0x0F9C,
157 0x0F9D, 0x0F9E, 0x0F9F, 0x0FA0, 0x0FA1, 0x0FA2, 0x0FA3, 0x0FA4,
158 0x0FA5, 0x0FA6, 0x0FA7, 0x0FA8, 0x0FA9, 0x0FAA, 0x0FAB, 0x0FAC,
159 0x0FAD, 0x0FAE, 0x0FAF, 0x0FB0, 0x0FB1, 0x0FB2, 0x0FB3, 0x0FB4,
160 0x0FB5, 0x0FB6, 0x0FB7, 0x0FB8, 0x0FB9, 0x0FBA, 0x0FBB, 0x0FBC,
161 0x0FC6, 0x102D, 0x102E, 0x102F, 0x1030, 0x1032, 0x1033, 0x1034,
162 0x1035, 0x1036, 0x1037, 0x1039, 0x103A, 0x103D, 0x103E, 0x1058,
163 0x1059, 0x105E, 0x105F, 0x1060, 0x1071, 0x1072, 0x1073, 0x1074,
164 0x1082, 0x1085, 0x1086, 0x108D, 0x109D, 0x135D, 0x135E, 0x135F,
165 0x1712, 0x1713, 0x1714, 0x1732, 0x1733, 0x1752, 0x1753, 0x1772,
166 0x1773, 0x17B4, 0x17B5, 0x17B7, 0x17B8, 0x17B9, 0x17BA, 0x17BB,
167 0x17BC, 0x17BD, 0x17C6, 0x17C9, 0x17CA, 0x17CB, 0x17CC, 0x17CD,
168 0x17CE, 0x17CF, 0x17D0, 0x17D1, 0x17D2, 0x17D3, 0x17DD, 0x180B,
169 0x180C, 0x180D, 0x180F, 0x1885, 0x1886, 0x18A9, 0x1920, 0x1921,
170 0x1922, 0x1927, 0x1928, 0x1932, 0x1939, 0x193A, 0x193B, 0x1A17,
171 0x1A18, 0x1A1B, 0x1A56, 0x1A58, 0x1A59, 0x1A5A, 0x1A5B, 0x1A5C,
172 0x1A5D, 0x1A5E, 0x1A60, 0x1A62, 0x1A65, 0x1A66, 0x1A67, 0x1A68,
173 0x1A69, 0x1A6A, 0x1A6B, 0x1A6C, 0x1A73, 0x1A74, 0x1A75, 0x1A76,
174 0x1A77, 0x1A78, 0x1A79, 0x1A7A, 0x1A7B, 0x1A7C, 0x1A7F, 0x1AB0,
175 0x1AB1, 0x1AB2, 0x1AB3, 0x1AB4, 0x1AB5, 0x1AB6, 0x1AB7, 0x1AB8,
176 0x1AB9, 0x1ABA, 0x1ABB, 0x1ABC, 0x1ABD, 0x1ABF, 0x1AC0, 0x1AC1,
177 0x1AC2, 0x1AC3, 0x1AC4, 0x1AC5, 0x1AC6, 0x1AC7, 0x1AC8, 0x1AC9,
178 0x1ACA, 0x1ACB, 0x1ACC, 0x1ACD, 0x1ACE, 0x1B00, 0x1B01, 0x1B02,
179 0x1B03, 0x1B34, 0x1B36, 0x1B37, 0x1B38, 0x1B39, 0x1B3A, 0x1B3C,
180 0x1B42, 0x1B6B, 0x1B6C, 0x1B6D, 0x1B6E, 0x1B6F, 0x1B70, 0x1B71,
181 0x1B72, 0x1B73, 0x1B80, 0x1B81, 0x1BA2, 0x1BA3, 0x1BA4, 0x1BA5,
182 0x1BA8, 0x1BA9, 0x1BAB, 0x1BAC, 0x1BAD, 0x1BE6, 0x1BE8, 0x1BE9,
183 0x1BED, 0x1BEF, 0x1BF0, 0x1BF1, 0x1C2C, 0x1C2D, 0x1C2E, 0x1C2F,
184 0x1C30, 0x1C31, 0x1C32, 0x1C33, 0x1C36, 0x1C37, 0x1CD0, 0x1CD1,
185 0x1CD2, 0x1CD4, 0x1CD5, 0x1CD6, 0x1CD7, 0x1CD8, 0x1CD9, 0x1CDA,
186 0x1CDB, 0x1CDC, 0x1CDD, 0x1CDE, 0x1CDF, 0x1CE0, 0x1CE2, 0x1CE3,
187 0x1CE4, 0x1CE5, 0x1CE6, 0x1CE7, 0x1CE8, 0x1CED, 0x1CF4, 0x1CF8,
188 0x1CF9, 0x1DC0, 0x1DC1, 0x1DC2, 0x1DC3, 0x1DC4, 0x1DC5, 0x1DC6,
189 0x1DC7, 0x1DC8, 0x1DC9, 0x1DCA, 0x1DCB, 0x1DCC, 0x1DCD, 0x1DCE,
190 0x1DCF, 0x1DD0, 0x1DD1, 0x1DD2, 0x1DD3, 0x1DD4, 0x1DD5, 0x1DD6,
191 0x1DD7, 0x1DD8, 0x1DD9, 0x1DDA, 0x1DDB, 0x1DDC, 0x1DDD, 0x1DDE,
192 0x1DDF, 0x1DE0, 0x1DE1, 0x1DE2, 0x1DE3, 0x1DE4, 0x1DE5, 0x1DE6,
193 0x1DE7, 0x1DE8, 0x1DE9, 0x1DEA, 0x1DEB, 0x1DEC, 0x1DED, 0x1DEE,
194 0x1DEF, 0x1DF0, 0x1DF1, 0x1DF2, 0x1DF3, 0x1DF4, 0x1DF5, 0x1DF6,
195 0x1DF7, 0x1DF8, 0x1DF9, 0x1DFA, 0x1DFB, 0x1DFC, 0x1DFD, 0x1DFE,
196 0x1DFF, 0x20D0, 0x20D1, 0x20D2, 0x20D3, 0x20D4, 0x20D5, 0x20D6,
197 0x20D7, 0x20D8, 0x20D9, 0x20DA, 0x20DB, 0x20DC, 0x20E1, 0x20E5,
198 0x20E6, 0x20E7, 0x20E8, 0x20E9, 0x20EA, 0x20EB, 0x20EC, 0x20ED,
199 0x20EE, 0x20EF, 0x20F0, 0x2CEF, 0x2CF0, 0x2CF1, 0x2D7F, 0x2DE0,
200 0x2DE1, 0x2DE2, 0x2DE3, 0x2DE4, 0x2DE5, 0x2DE6, 0x2DE7, 0x2DE8,
201 0x2DE9, 0x2DEA, 0x2DEB, 0x2DEC, 0x2DED, 0x2DEE, 0x2DEF, 0x2DF0,
202 0x2DF1, 0x2DF2, 0x2DF3, 0x2DF4, 0x2DF5, 0x2DF6, 0x2DF7, 0x2DF8,
203 0x2DF9, 0x2DFA, 0x2DFB, 0x2DFC, 0x2DFD, 0x2DFE, 0x2DFF, 0x302A,
204 0x302B, 0x302C, 0x302D, 0x3099, 0x309A, 0xA66F, 0xA674, 0xA675,
205 0xA676, 0xA677, 0xA678, 0xA679, 0xA67A, 0xA67B, 0xA67C, 0xA67D,
206 0xA69E, 0xA69F, 0xA6F0, 0xA6F1, 0xA802, 0xA806, 0xA80B, 0xA825,
207 0xA826, 0xA82C, 0xA8C4, 0xA8C5, 0xA8E0, 0xA8E1, 0xA8E2, 0xA8E3,
208 0xA8E4, 0xA8E5, 0xA8E6, 0xA8E7, 0xA8E8, 0xA8E9, 0xA8EA, 0xA8EB,
209 0xA8EC, 0xA8ED, 0xA8EE, 0xA8EF, 0xA8F0, 0xA8F1, 0xA8FF, 0xA926,
210 0xA927, 0xA928, 0xA929, 0xA92A, 0xA92B, 0xA92C, 0xA92D, 0xA947,
211 0xA948, 0xA949, 0xA94A, 0xA94B, 0xA94C, 0xA94D, 0xA94E, 0xA94F,
212 0xA950, 0xA951, 0xA980, 0xA981, 0xA982, 0xA9B3, 0xA9B6, 0xA9B7,
213 0xA9B8, 0xA9B9, 0xA9BC, 0xA9BD, 0xA9E5, 0xAA29, 0xAA2A, 0xAA2B,
214 0xAA2C, 0xAA2D, 0xAA2E, 0xAA31, 0xAA32, 0xAA35, 0xAA36, 0xAA43,
215 0xAA4C, 0xAA7C, 0xAAB0, 0xAAB2, 0xAAB3, 0xAAB4, 0xAAB7, 0xAAB8,
216 0xAABE, 0xAABF, 0xAAC1, 0xAAEC, 0xAAED, 0xAAF6, 0xABE5, 0xABE8,
217 0xABED, 0xFB1E, 0xFE00, 0xFE01, 0xFE02, 0xFE03, 0xFE04, 0xFE05,
218 0xFE06, 0xFE07, 0xFE08, 0xFE09, 0xFE0A, 0xFE0B, 0xFE0C, 0xFE0D,
219 0xFE0E, 0xFE0F, 0xFE20, 0xFE21, 0xFE22, 0xFE23, 0xFE24, 0xFE25,
220 0xFE26, 0xFE27, 0xFE28, 0xFE29, 0xFE2A, 0xFE2B, 0xFE2C, 0xFE2D,
221 0xFE2E, 0xFE2F, 0x101FD, 0x102E0, 0x10376, 0x10377, 0x10378, 0x10379,
222 0x1037A, 0x10A01, 0x10A02, 0x10A03, 0x10A05, 0x10A06, 0x10A0C, 0x10A0D,
223 0x10A0E, 0x10A0F, 0x10A38, 0x10A39, 0x10A3A, 0x10A3F, 0x10AE5, 0x10AE6,
224 0x10D24, 0x10D25, 0x10D26, 0x10D27, 0x10D69, 0x10D6A, 0x10D6B, 0x10D6C,
225 0x10D6D, 0x10EAB, 0x10EAC, 0x10EFC, 0x10EFD, 0x10EFE, 0x10EFF, 0x10F46,
226 0x10F47, 0x10F48, 0x10F49, 0x10F4A, 0x10F4B, 0x10F4C, 0x10F4D, 0x10F4E,
227 0x10F4F, 0x10F50, 0x10F82, 0x10F83, 0x10F84, 0x10F85, 0x11001, 0x11038,
228 0x11039, 0x1103A, 0x1103B, 0x1103C, 0x1103D, 0x1103E, 0x1103F, 0x11040,
229 0x11041, 0x11042, 0x11043, 0x11044, 0x11045, 0x11046, 0x11070, 0x11073,
230 0x11074, 0x1107F, 0x11080, 0x11081, 0x110B3, 0x110B4, 0x110B5, 0x110B6,
231 0x110B9, 0x110BA, 0x110C2, 0x11100, 0x11101, 0x11102, 0x11127, 0x11128,
232 0x11129, 0x1112A, 0x1112B, 0x1112D, 0x1112E, 0x1112F, 0x11130, 0x11131,
233 0x11132, 0x11133, 0x11134, 0x11173, 0x11180, 0x11181, 0x111B6, 0x111B7,
234 0x111B8, 0x111B9, 0x111BA, 0x111BB, 0x111BC, 0x111BD, 0x111BE, 0x111C9,
235 0x111CA, 0x111CB, 0x111CC, 0x111CF, 0x1122F, 0x11230, 0x11231, 0x11234,
236 0x11236, 0x11237, 0x1123E, 0x11241, 0x112DF, 0x112E3, 0x112E4, 0x112E5,
237 0x112E6, 0x112E7, 0x112E8, 0x112E9, 0x112EA, 0x11300, 0x11301, 0x1133B,
238 0x1133C, 0x11340, 0x11366, 0x11367, 0x11368, 0x11369, 0x1136A, 0x1136B,
239 0x1136C, 0x11370, 0x11371, 0x11372, 0x11373, 0x11374, 0x113BB, 0x113BC,
240 0x113BD, 0x113BE, 0x113BF, 0x113C0, 0x113CE, 0x113D0, 0x113D2, 0x113E1,
241 0x113E2, 0x11438, 0x11439, 0x1143A, 0x1143B, 0x1143C, 0x1143D, 0x1143E,
242 0x1143F, 0x11442, 0x11443, 0x11444, 0x11446, 0x1145E, 0x114B3, 0x114B4,
243 0x114B5, 0x114B6, 0x114B7, 0x114B8, 0x114BA, 0x114BF, 0x114C0, 0x114C2,
244 0x114C3, 0x115B2, 0x115B3, 0x115B4, 0x115B5, 0x115BC, 0x115BD, 0x115BF,
245 0x115C0, 0x115DC, 0x115DD, 0x11633, 0x11634, 0x11635, 0x11636, 0x11637,
246 0x11638, 0x11639, 0x1163A, 0x1163D, 0x1163F, 0x11640, 0x116AB, 0x116AD,
247 0x116B0, 0x116B1, 0x116B2, 0x116B3, 0x116B4, 0x116B5, 0x116B7, 0x1171D,
248 0x1171F, 0x11722, 0x11723, 0x11724, 0x11725, 0x11727, 0x11728, 0x11729,
249 0x1172A, 0x1172B, 0x1182F, 0x11830, 0x11831, 0x11832, 0x11833, 0x11834,
250 0x11835, 0x11836, 0x11837, 0x11839, 0x1183A, 0x1193B, 0x1193C, 0x1193E,
251 0x11943, 0x119D4, 0x119D5, 0x119D6, 0x119D7, 0x119DA, 0x119DB, 0x119E0,
252 0x11A01, 0x11A02, 0x11A03, 0x11A04, 0x11A05, 0x11A06, 0x11A07, 0x11A08,
253 0x11A09, 0x11A0A, 0x11A33, 0x11A34, 0x11A35, 0x11A36, 0x11A37, 0x11A38,
254 0x11A3B, 0x11A3C, 0x11A3D, 0x11A3E, 0x11A47, 0x11A51, 0x11A52, 0x11A53,
255 0x11A54, 0x11A55, 0x11A56, 0x11A59, 0x11A5A, 0x11A5B, 0x11A8A, 0x11A8B,
256 0x11A8C, 0x11A8D, 0x11A8E, 0x11A8F, 0x11A90, 0x11A91, 0x11A92, 0x11A93,
257 0x11A94, 0x11A95, 0x11A96, 0x11A98, 0x11A99, 0x11C30, 0x11C31, 0x11C32,
258 0x11C33, 0x11C34, 0x11C35, 0x11C36, 0x11C38, 0x11C39, 0x11C3A, 0x11C3B,
259 0x11C3C, 0x11C3D, 0x11C3F, 0x11C92, 0x11C93, 0x11C94, 0x11C95, 0x11C96,
260 0x11C97, 0x11C98, 0x11C99, 0x11C9A, 0x11C9B, 0x11C9C, 0x11C9D, 0x11C9E,
261 0x11C9F, 0x11CA0, 0x11CA1, 0x11CA2, 0x11CA3, 0x11CA4, 0x11CA5, 0x11CA6,
262 0x11CA7, 0x11CAA, 0x11CAB, 0x11CAC, 0x11CAD, 0x11CAE, 0x11CAF, 0x11CB0,
263 0x11CB2, 0x11CB3, 0x11CB5, 0x11CB6, 0x11D31, 0x11D32, 0x11D33, 0x11D34,
264 0x11D35, 0x11D36, 0x11D3A, 0x11D3C, 0x11D3D, 0x11D3F, 0x11D40, 0x11D41,
265 0x11D42, 0x11D43, 0x11D44, 0x11D45, 0x11D47, 0x11D90, 0x11D91, 0x11D95,
266 0x11D97, 0x11EF3, 0x11EF4, 0x11F00, 0x11F01, 0x11F36, 0x11F37, 0x11F38,
267 0x11F39, 0x11F3A, 0x11F40, 0x11F42, 0x11F5A, 0x13440, 0x13447, 0x13448,
268 0x13449, 0x1344A, 0x1344B, 0x1344C, 0x1344D, 0x1344E, 0x1344F, 0x13450,
269 0x13451, 0x13452, 0x13453, 0x13454, 0x13455, 0x1611E, 0x1611F, 0x16120,
270 0x16121, 0x16122, 0x16123, 0x16124, 0x16125, 0x16126, 0x16127, 0x16128,
271 0x16129, 0x1612D, 0x1612E, 0x1612F, 0x16AF0, 0x16AF1, 0x16AF2, 0x16AF3,
272 0x16AF4, 0x16B30, 0x16B31, 0x16B32, 0x16B33, 0x16B34, 0x16B35, 0x16B36,
273 0x16F4F, 0x16F8F, 0x16F90, 0x16F91, 0x16F92, 0x16FE4, 0x1BC9D, 0x1BC9E,
274 0x1CF00, 0x1CF01, 0x1CF02, 0x1CF03, 0x1CF04, 0x1CF05, 0x1CF06, 0x1CF07,
275 0x1CF08, 0x1CF09, 0x1CF0A, 0x1CF0B, 0x1CF0C, 0x1CF0D, 0x1CF0E, 0x1CF0F,
276 0x1CF10, 0x1CF11, 0x1CF12, 0x1CF13, 0x1CF14, 0x1CF15, 0x1CF16, 0x1CF17,
277 0x1CF18, 0x1CF19, 0x1CF1A, 0x1CF1B, 0x1CF1C, 0x1CF1D, 0x1CF1E, 0x1CF1F,
278 0x1CF20, 0x1CF21, 0x1CF22, 0x1CF23, 0x1CF24, 0x1CF25, 0x1CF26, 0x1CF27,
279 0x1CF28, 0x1CF29, 0x1CF2A, 0x1CF2B, 0x1CF2C, 0x1CF2D, 0x1CF30, 0x1CF31,
280 0x1CF32, 0x1CF33, 0x1CF34, 0x1CF35, 0x1CF36, 0x1CF37, 0x1CF38, 0x1CF39,
281 0x1CF3A, 0x1CF3B, 0x1CF3C, 0x1CF3D, 0x1CF3E, 0x1CF3F, 0x1CF40, 0x1CF41,
282 0x1CF42, 0x1CF43, 0x1CF44, 0x1CF45, 0x1CF46, 0x1D167, 0x1D168, 0x1D169,
283 0x1D17B, 0x1D17C, 0x1D17D, 0x1D17E, 0x1D17F, 0x1D180, 0x1D181, 0x1D182,
284 0x1D185, 0x1D186, 0x1D187, 0x1D188, 0x1D189, 0x1D18A, 0x1D18B, 0x1D1AA,
285 0x1D1AB, 0x1D1AC, 0x1D1AD, 0x1D242, 0x1D243, 0x1D244, 0x1DA00, 0x1DA01,
286 0x1DA02, 0x1DA03, 0x1DA04, 0x1DA05, 0x1DA06, 0x1DA07, 0x1DA08, 0x1DA09,
287 0x1DA0A, 0x1DA0B, 0x1DA0C, 0x1DA0D, 0x1DA0E, 0x1DA0F, 0x1DA10, 0x1DA11,
288 0x1DA12, 0x1DA13, 0x1DA14, 0x1DA15, 0x1DA16, 0x1DA17, 0x1DA18, 0x1DA19,
289 0x1DA1A, 0x1DA1B, 0x1DA1C, 0x1DA1D, 0x1DA1E, 0x1DA1F, 0x1DA20, 0x1DA21,
290 0x1DA22, 0x1DA23, 0x1DA24, 0x1DA25, 0x1DA26, 0x1DA27, 0x1DA28, 0x1DA29,
291 0x1DA2A, 0x1DA2B, 0x1DA2C, 0x1DA2D, 0x1DA2E, 0x1DA2F, 0x1DA30, 0x1DA31,
292 0x1DA32, 0x1DA33, 0x1DA34, 0x1DA35, 0x1DA36, 0x1DA3B, 0x1DA3C, 0x1DA3D,
293 0x1DA3E, 0x1DA3F, 0x1DA40, 0x1DA41, 0x1DA42, 0x1DA43, 0x1DA44, 0x1DA45,
294 0x1DA46, 0x1DA47, 0x1DA48, 0x1DA49, 0x1DA4A, 0x1DA4B, 0x1DA4C, 0x1DA4D,
295 0x1DA4E, 0x1DA4F, 0x1DA50, 0x1DA51, 0x1DA52, 0x1DA53, 0x1DA54, 0x1DA55,
296 0x1DA56, 0x1DA57, 0x1DA58, 0x1DA59, 0x1DA5A, 0x1DA5B, 0x1DA5C, 0x1DA5D,
297 0x1DA5E, 0x1DA5F, 0x1DA60, 0x1DA61, 0x1DA62, 0x1DA63, 0x1DA64, 0x1DA65,
298 0x1DA66, 0x1DA67, 0x1DA68, 0x1DA69, 0x1DA6A, 0x1DA6B, 0x1DA6C, 0x1DA75,
299 0x1DA84, 0x1DA9B, 0x1DA9C, 0x1DA9D, 0x1DA9E, 0x1DA9F, 0x1DAA1, 0x1DAA2,
300 0x1DAA3, 0x1DAA4, 0x1DAA5, 0x1DAA6, 0x1DAA7, 0x1DAA8, 0x1DAA9, 0x1DAAA,
301 0x1DAAB, 0x1DAAC, 0x1DAAD, 0x1DAAE, 0x1DAAF, 0x1E000, 0x1E001, 0x1E002,
302 0x1E003, 0x1E004, 0x1E005, 0x1E006, 0x1E008, 0x1E009, 0x1E00A, 0x1E00B,
303 0x1E00C, 0x1E00D, 0x1E00E, 0x1E00F, 0x1E010, 0x1E011, 0x1E012, 0x1E013,
304 0x1E014, 0x1E015, 0x1E016, 0x1E017, 0x1E018, 0x1E01B, 0x1E01C, 0x1E01D,
305 0x1E01E, 0x1E01F, 0x1E020, 0x1E021, 0x1E023, 0x1E024, 0x1E026, 0x1E027,
306 0x1E028, 0x1E029, 0x1E02A, 0x1E08F, 0x1E130, 0x1E131, 0x1E132, 0x1E133,
307 0x1E134, 0x1E135, 0x1E136, 0x1E2AE, 0x1E2EC, 0x1E2ED, 0x1E2EE, 0x1E2EF,
308 0x1E4EC, 0x1E4ED, 0x1E4EE, 0x1E4EF, 0x1E5EE, 0x1E5EF, 0x1E8D0, 0x1E8D1,
309 0x1E8D2, 0x1E8D3, 0x1E8D4, 0x1E8D5, 0x1E8D6, 0x1E944, 0x1E945, 0x1E946,
310 0x1E947, 0x1E948, 0x1E949, 0x1E94A, 0xE0100, 0xE0101, 0xE0102, 0xE0103,
311 0xE0104, 0xE0105, 0xE0106, 0xE0107, 0xE0108, 0xE0109, 0xE010A, 0xE010B,
312 0xE010C, 0xE010D, 0xE010E, 0xE010F, 0xE0110, 0xE0111, 0xE0112, 0xE0113,
313 0xE0114, 0xE0115, 0xE0116, 0xE0117, 0xE0118, 0xE0119, 0xE011A, 0xE011B,
314 0xE011C, 0xE011D, 0xE011E, 0xE011F, 0xE0120, 0xE0121, 0xE0122, 0xE0123,
315 0xE0124, 0xE0125, 0xE0126, 0xE0127, 0xE0128, 0xE0129, 0xE012A, 0xE012B,
316 0xE012C, 0xE012D, 0xE012E, 0xE012F, 0xE0130, 0xE0131, 0xE0132, 0xE0133,
317 0xE0134, 0xE0135, 0xE0136, 0xE0137, 0xE0138, 0xE0139, 0xE013A, 0xE013B,
318 0xE013C, 0xE013D, 0xE013E, 0xE013F, 0xE0140, 0xE0141, 0xE0142, 0xE0143,
319 0xE0144, 0xE0145, 0xE0146, 0xE0147, 0xE0148, 0xE0149, 0xE014A, 0xE014B,
320 0xE014C, 0xE014D, 0xE014E, 0xE014F, 0xE0150, 0xE0151, 0xE0152, 0xE0153,
321 0xE0154, 0xE0155, 0xE0156, 0xE0157, 0xE0158, 0xE0159, 0xE015A, 0xE015B,
322 0xE015C, 0xE015D, 0xE015E, 0xE015F, 0xE0160, 0xE0161, 0xE0162, 0xE0163,
323 0xE0164, 0xE0165, 0xE0166, 0xE0167, 0xE0168, 0xE0169, 0xE016A, 0xE016B,
324 0xE016C, 0xE016D, 0xE016E, 0xE016F, 0xE0170, 0xE0171, 0xE0172, 0xE0173,
325 0xE0174, 0xE0175, 0xE0176, 0xE0177, 0xE0178, 0xE0179, 0xE017A, 0xE017B,
326 0xE017C, 0xE017D, 0xE017E, 0xE017F, 0xE0180, 0xE0181, 0xE0182, 0xE0183,
327 0xE0184, 0xE0185, 0xE0186, 0xE0187, 0xE0188, 0xE0189, 0xE018A, 0xE018B,
328 0xE018C, 0xE018D, 0xE018E, 0xE018F, 0xE0190, 0xE0191, 0xE0192, 0xE0193,
329 0xE0194, 0xE0195, 0xE0196, 0xE0197, 0xE0198, 0xE0199, 0xE019A, 0xE019B,
330 0xE019C, 0xE019D, 0xE019E, 0xE019F, 0xE01A0, 0xE01A1, 0xE01A2, 0xE01A3,
331 0xE01A4, 0xE01A5, 0xE01A6, 0xE01A7, 0xE01A8, 0xE01A9, 0xE01AA, 0xE01AB,
332 0xE01AC, 0xE01AD, 0xE01AE, 0xE01AF, 0xE01B0, 0xE01B1, 0xE01B2, 0xE01B3,
333 0xE01B4, 0xE01B5, 0xE01B6, 0xE01B7, 0xE01B8, 0xE01B9, 0xE01BA, 0xE01BB,
334 0xE01BC, 0xE01BD, 0xE01BE, 0xE01BF, 0xE01C0, 0xE01C1, 0xE01C2, 0xE01C3,
335 0xE01C4, 0xE01C5, 0xE01C6, 0xE01C7, 0xE01C8, 0xE01C9, 0xE01CA, 0xE01CB,
336 0xE01CC, 0xE01CD, 0xE01CE, 0xE01CF, 0xE01D0, 0xE01D1, 0xE01D2, 0xE01D3,
337 0xE01D4, 0xE01D5, 0xE01D6, 0xE01D7, 0xE01D8, 0xE01D9, 0xE01DA, 0xE01DB,
338 0xE01DC, 0xE01DD, 0xE01DE, 0xE01DF, 0xE01E0, 0xE01E1, 0xE01E2, 0xE01E3,
339 0xE01E4, 0xE01E5, 0xE01E6, 0xE01E7, 0xE01E8, 0xE01E9, 0xE01EA, 0xE01EB,
340 0xE01EC, 0xE01ED, 0xE01EE, 0xE01EF,
341};
342
343static unsigned long combiningCharTableSize = sizeof(combiningCharTable) / sizeof(combiningCharTable[0]);
344
345/* Check if the code is a wide character
346 */
347static int isWideChar(unsigned long cp) {
348 size_t i;
349 for (i = 0; i < wideCharTableSize; i++)
350 if (wideCharTable[i][0] <= cp && cp <= wideCharTable[i][1]) return 1;
351 return 0;
352}
353
354/* Check if the code is a combining character
355 */
356static int isCombiningChar(unsigned long cp) {
357 size_t i;
358 for (i = 0; i < combiningCharTableSize; i++)
359 if (combiningCharTable[i] == cp) return 1;
360 return 0;
361}
362
363/* Get length of previous UTF8 character
364 */
365static size_t prevUtf8CharLen(const char* buf, int pos) {
366 int end = pos--;
367 while (pos >= 0 && ((unsigned char)buf[pos] & 0xC0) == 0x80)
368 pos--;
369 return end - pos;
370}
371
372/* Convert UTF8 to Unicode code point
373 */
374static size_t utf8BytesToCodePoint(const char* buf, size_t len, int* cp) {
375 if (len) {
376 unsigned char byte = buf[0];
377 if ((byte & 0x80) == 0) {
378 *cp = byte;
379 return 1;
380 } else if ((byte & 0xE0) == 0xC0) {
381 if (len >= 2) {
382 *cp = (((unsigned long)(buf[0] & 0x1F)) << 6) |
383 ((unsigned long)(buf[1] & 0x3F));
384 return 2;
385 }
386 } else if ((byte & 0xF0) == 0xE0) {
387 if (len >= 3) {
388 *cp = (((unsigned long)(buf[0] & 0x0F)) << 12) |
389 (((unsigned long)(buf[1] & 0x3F)) << 6) |
390 ((unsigned long)(buf[2] & 0x3F));
391 return 3;
392 }
393 } else if ((byte & 0xF8) == 0xF0) {
394 if (len >= 4) {
395 *cp = (((unsigned long)(buf[0] & 0x07)) << 18) |
396 (((unsigned long)(buf[1] & 0x3F)) << 12) |
397 (((unsigned long)(buf[2] & 0x3F)) << 6) |
398 ((unsigned long)(buf[3] & 0x3F));
399 return 4;
400 }
401 }
402 }
403 return 0;
404}
405
406/* Get length of next grapheme
407 */
408size_t linenoiseUtf8NextCharLen(const char* buf, size_t buf_len, size_t pos, size_t *col_len) {
409 size_t beg = pos;
410 int cp;
411 size_t len = utf8BytesToCodePoint(buf + pos, buf_len - pos, &cp);
412 if (isCombiningChar(cp)) {
413 /* NOTREACHED */
414 return 0;
415 }
416 if (col_len != NULL) *col_len = isWideChar(cp) ? 2 : 1;
417 pos += len;
418 while (pos < buf_len) {
419 int cp;
420 len = utf8BytesToCodePoint(buf + pos, buf_len - pos, &cp);
421 if (!isCombiningChar(cp)) return pos - beg;
422 pos += len;
423 }
424 return pos - beg;
425}
426
427/* Get length of previous grapheme
428 */
429size_t linenoiseUtf8PrevCharLen(const char* buf, size_t buf_len, size_t pos, size_t *col_len) {
430 UNUSED(buf_len);
431 size_t end = pos;
432 while (pos > 0) {
433 size_t len = prevUtf8CharLen(buf, pos);
434 pos -= len;
435 int cp;
436 utf8BytesToCodePoint(buf + pos, len, &cp);
437 if (!isCombiningChar(cp)) {
438 if (col_len != NULL) *col_len = isWideChar(cp) ? 2 : 1;
439 return end - pos;
440 }
441 }
442 /* NOTREACHED */
443 return 0;
444}
445
446/* Read a Unicode from file.
447 */
448size_t linenoiseUtf8ReadCode(int fd, char* buf, size_t buf_len, int* cp) {
449 if (buf_len < 1) return -1;
450 size_t nread = read(fd,&buf[0],1);
451 if (nread <= 0) return nread;
452
453 unsigned char byte = buf[0];
454 if ((byte & 0x80) == 0) {
455 ;
456 } else if ((byte & 0xE0) == 0xC0) {
457 if (buf_len < 2) return -1;
458 nread = read(fd,&buf[1],1);
459 if (nread <= 0) return nread;
460 } else if ((byte & 0xF0) == 0xE0) {
461 if (buf_len < 3) return -1;
462 nread = read(fd,&buf[1],2);
463 if (nread <= 0) return nread;
464 } else if ((byte & 0xF8) == 0xF0) {
465 if (buf_len < 3) return -1;
466 nread = read(fd,&buf[1],3);
467 if (nread <= 0) return nread;
468 } else {
469 return -1;
470 }
471
472 return utf8BytesToCodePoint(buf, buf_len, cp);
473}