Skip to content

Commit ece66dd

Browse files
BDispBDisp
committed
After all, there is a pattern to calculate wider runes.
1 parent 8481cc1 commit ece66dd

File tree

3 files changed

+135
-82
lines changed

3 files changed

+135
-82
lines changed

‎NStack/unicode/Rune.ColumnWidth.cs

Lines changed: 127 additions & 74 deletions
Original file line numberDiff line numberDiff line change
@@ -8,69 +8,89 @@ namespace System
88
{
99
public partial struct Rune
1010
{
11-
static uint [,] combining = new uint [,] {
12-
{ 0x0300, 0x036F }, { 0x0483, 0x0486 }, { 0x0488, 0x0489 },
13-
{ 0x0591, 0x05BD }, { 0x05BF, 0x05BF }, { 0x05C1, 0x05C2 },
14-
{ 0x05C4, 0x05C5 }, { 0x05C7, 0x05C7 }, { 0x0600, 0x0603 },
15-
{ 0x0610, 0x0615 }, { 0x064B, 0x065E }, { 0x0670, 0x0670 },
16-
{ 0x06D6, 0x06E4 }, { 0x06E7, 0x06E8 }, { 0x06EA, 0x06ED },
17-
{ 0x070F, 0x070F }, { 0x0711, 0x0711 }, { 0x0730, 0x074A },
18-
{ 0x07A6, 0x07B0 }, { 0x07EB, 0x07F3 }, { 0x0901, 0x0902 },
19-
{ 0x093C, 0x093C }, { 0x0941, 0x0948 }, { 0x094D, 0x094D },
20-
{ 0x0951, 0x0954 }, { 0x0962, 0x0963 }, { 0x0981, 0x0981 },
21-
{ 0x09BC, 0x09BC }, { 0x09C1, 0x09C4 }, { 0x09CD, 0x09CD },
22-
{ 0x09E2, 0x09E3 }, { 0x0A01, 0x0A02 }, { 0x0A3C, 0x0A3C },
23-
{ 0x0A41, 0x0A42 }, { 0x0A47, 0x0A48 }, { 0x0A4B, 0x0A4D },
24-
{ 0x0A70, 0x0A71 }, { 0x0A81, 0x0A82 }, { 0x0ABC, 0x0ABC },
25-
{ 0x0AC1, 0x0AC5 }, { 0x0AC7, 0x0AC8 }, { 0x0ACD, 0x0ACD },
26-
{ 0x0AE2, 0x0AE3 }, { 0x0B01, 0x0B01 }, { 0x0B3C, 0x0B3C },
27-
{ 0x0B3F, 0x0B3F }, { 0x0B41, 0x0B43 }, { 0x0B4D, 0x0B4D },
28-
{ 0x0B56, 0x0B56 }, { 0x0B82, 0x0B82 }, { 0x0BC0, 0x0BC0 },
29-
{ 0x0BCD, 0x0BCD }, { 0x0C3E, 0x0C40 }, { 0x0C46, 0x0C48 },
30-
{ 0x0C4A, 0x0C4D }, { 0x0C55, 0x0C56 }, { 0x0CBC, 0x0CBC },
31-
{ 0x0CBF, 0x0CBF }, { 0x0CC6, 0x0CC6 }, { 0x0CCC, 0x0CCD },
32-
{ 0x0CE2, 0x0CE3 }, { 0x0D41, 0x0D43 }, { 0x0D4D, 0x0D4D },
33-
{ 0x0DCA, 0x0DCA }, { 0x0DD2, 0x0DD4 }, { 0x0DD6, 0x0DD6 },
34-
{ 0x0E31, 0x0E31 }, { 0x0E34, 0x0E3A }, { 0x0E47, 0x0E4E },
35-
{ 0x0EB1, 0x0EB1 }, { 0x0EB4, 0x0EB9 }, { 0x0EBB, 0x0EBC },
36-
{ 0x0EC8, 0x0ECD }, { 0x0F18, 0x0F19 }, { 0x0F35, 0x0F35 },
37-
{ 0x0F37, 0x0F37 }, { 0x0F39, 0x0F39 }, { 0x0F71, 0x0F7E },
38-
{ 0x0F80, 0x0F84 }, { 0x0F86, 0x0F87 }, { 0x0F90, 0x0F97 },
39-
{ 0x0F99, 0x0FBC }, { 0x0FC6, 0x0FC6 }, { 0x102D, 0x1030 },
40-
{ 0x1032, 0x1032 }, { 0x1036, 0x1037 }, { 0x1039, 0x1039 },
41-
{ 0x1058, 0x1059 }, { 0x1160, 0x11FF }, { 0x135F, 0x135F },
42-
{ 0x1712, 0x1714 }, { 0x1732, 0x1734 }, { 0x1752, 0x1753 },
43-
{ 0x1772, 0x1773 }, { 0x17B4, 0x17B5 }, { 0x17B7, 0x17BD },
44-
{ 0x17C6, 0x17C6 }, { 0x17C9, 0x17D3 }, { 0x17DD, 0x17DD },
45-
{ 0x180B, 0x180D }, { 0x18A9, 0x18A9 }, { 0x1920, 0x1922 },
46-
{ 0x1927, 0x1928 }, { 0x1932, 0x1932 }, { 0x1939, 0x193B },
47-
{ 0x1A17, 0x1A18 }, { 0x1B00, 0x1B03 }, { 0x1B34, 0x1B34 },
48-
{ 0x1B36, 0x1B3A }, { 0x1B3C, 0x1B3C }, { 0x1B42, 0x1B42 },
49-
{ 0x1B6B, 0x1B73 }, { 0x1DC0, 0x1DCA }, { 0x1DFE, 0x1DFF },
50-
{ 0x200B, 0x200F }, { 0x202A, 0x202E }, { 0x2060, 0x2063 },
51-
{ 0x206A, 0x206F }, { 0x20D0, 0x20EF }, { 0x302A, 0x302F },
52-
{ 0x3099, 0x309A }, { 0xA806, 0xA806 }, { 0xA80B, 0xA80B },
53-
{ 0xA825, 0xA826 }, { 0xFB1E, 0xFB1E }, { 0xFE00, 0xFE0F },
54-
{ 0xFE20, 0xFE23 }, { 0xFEFF, 0xFEFF }, { 0xFFF9, 0xFFFB },
55-
{ 0x10A01, 0x10A03 }, { 0x10A05, 0x10A06 }, { 0x10A0C, 0x10A0F },
56-
{ 0x10A38, 0x10A3A }, { 0x10A3F, 0x10A3F }, { 0x1D167, 0x1D169 },
57-
{ 0x1D173, 0x1D182 }, { 0x1D185, 0x1D18B }, { 0x1D1AA, 0x1D1AD },
58-
{ 0x1D242, 0x1D244 }, { 0xE0001, 0xE0001 }, { 0xE0020, 0xE007F },
59-
{ 0xE0100, 0xE01EF }
11+
static uint[,] combining = new uint[,] {
12+
{ 0x0300, 0x036F, 1 }, { 0x0483, 0x0486, 1 }, { 0x0488, 0x0489, 1 },
13+
{ 0x0591, 0x05BD, 1 }, { 0x05BF, 0x05BF, 1 }, { 0x05C1, 0x05C2, 1 },
14+
{ 0x05C4, 0x05C5, 1 }, { 0x05C7, 0x05C7, 1 }, { 0x0600, 0x0603, 1 },
15+
{ 0x0610, 0x0615, 1 }, { 0x064B, 0x065E, 1 }, { 0x0670, 0x0670, 1 },
16+
{ 0x06D6, 0x06E4, 1 }, { 0x06E7, 0x06E8, 1 }, { 0x06EA, 0x06ED, 1 },
17+
{ 0x070F, 0x070F, 1 }, { 0x0711, 0x0711, 1 }, { 0x0730, 0x074A, 1 },
18+
{ 0x07A6, 0x07B0, 1 }, { 0x07EB, 0x07F3, 1 }, { 0x0901, 0x0902, 1 },
19+
{ 0x093C, 0x093C, 1 }, { 0x0941, 0x0948, 1 }, { 0x094D, 0x094D, 1 },
20+
{ 0x0951, 0x0954, 1 }, { 0x0962, 0x0963, 1 }, { 0x0981, 0x0981, 1 },
21+
{ 0x09BC, 0x09BC, 1 }, { 0x09C1, 0x09C4, 1 }, { 0x09CD, 0x09CD, 1 },
22+
{ 0x09E2, 0x09E3, 1 }, { 0x0A01, 0x0A02, 1 }, { 0x0A3C, 0x0A3C, 1 },
23+
{ 0x0A41, 0x0A42, 1 }, { 0x0A47, 0x0A48, 1 }, { 0x0A4B, 0x0A4D, 1 },
24+
{ 0x0A70, 0x0A71, 1 }, { 0x0A81, 0x0A82, 1 }, { 0x0ABC, 0x0ABC, 1 },
25+
{ 0x0AC1, 0x0AC5, 1 }, { 0x0AC7, 0x0AC8, 1 }, { 0x0ACD, 0x0ACD, 1 },
26+
{ 0x0AE2, 0x0AE3, 1 }, { 0x0B01, 0x0B01, 1 }, { 0x0B3C, 0x0B3C, 1 },
27+
{ 0x0B3F, 0x0B3F, 1 }, { 0x0B41, 0x0B43, 1 }, { 0x0B4D, 0x0B4D, 1 },
28+
{ 0x0B56, 0x0B56, 1 }, { 0x0B82, 0x0B82, 1 }, { 0x0BC0, 0x0BC0, 1 },
29+
{ 0x0BCD, 0x0BCD, 1 }, { 0x0C3E, 0x0C40, 1 }, { 0x0C46, 0x0C48, 1 },
30+
{ 0x0C4A, 0x0C4D, 1 }, { 0x0C55, 0x0C56, 1 }, { 0x0CBC, 0x0CBC, 1 },
31+
{ 0x0CBF, 0x0CBF, 1 }, { 0x0CC6, 0x0CC6, 1 }, { 0x0CCC, 0x0CCD, 1 },
32+
{ 0x0CE2, 0x0CE3, 1 }, { 0x0D41, 0x0D43, 1 }, { 0x0D4D, 0x0D4D, 1 },
33+
{ 0x0DCA, 0x0DCA, 1 }, { 0x0DD2, 0x0DD4, 1 }, { 0x0DD6, 0x0DD6, 1 },
34+
{ 0x0E31, 0x0E31, 1 }, { 0x0E34, 0x0E3A, 1 }, { 0x0E47, 0x0E4E, 1 },
35+
{ 0x0EB1, 0x0EB1, 1 }, { 0x0EB4, 0x0EB9, 1 }, { 0x0EBB, 0x0EBC, 1 },
36+
{ 0x0EC8, 0x0ECD, 1 }, { 0x0F18, 0x0F19, 1 }, { 0x0F35, 0x0F35, 1 },
37+
{ 0x0F37, 0x0F37, 1 }, { 0x0F39, 0x0F39, 1 }, { 0x0F71, 0x0F7E, 1 },
38+
{ 0x0F80, 0x0F84, 1 }, { 0x0F86, 0x0F87, 1 }, { 0x0F90, 0x0F97, 1 },
39+
{ 0x0F99, 0x0FBC, 1 }, { 0x0FC6, 0x0FC6, 1 }, { 0x102D, 0x1030, 1 },
40+
{ 0x1032, 0x1032, 1 }, { 0x1036, 0x1037, 1 }, { 0x1039, 0x1039, 1 },
41+
{ 0x1058, 0x1059, 1 }, { 0x1160, 0x11FF, 1 }, { 0x135F, 0x135F, 1 },
42+
{ 0x1712, 0x1714, 1 }, { 0x1732, 0x1734, 1 }, { 0x1752, 0x1753, 1 },
43+
{ 0x1772, 0x1773, 1 }, { 0x17B4, 0x17B5, 1 }, { 0x17B7, 0x17BD, 1 },
44+
{ 0x17C6, 0x17C6, 1 }, { 0x17C9, 0x17D3, 1 }, { 0x17DD, 0x17DD, 1 },
45+
{ 0x180B, 0x180D, 1 }, { 0x18A9, 0x18A9, 1 }, { 0x1920, 0x1922, 1 },
46+
{ 0x1927, 0x1928, 1 }, { 0x1932, 0x1932, 1 }, { 0x1939, 0x193B, 1 },
47+
{ 0x1A17, 0x1A18, 1 }, { 0x1B00, 0x1B03, 1 }, { 0x1B34, 0x1B34, 1 },
48+
{ 0x1B36, 0x1B3A, 1 }, { 0x1B3C, 0x1B3C, 1 }, { 0x1B42, 0x1B42, 1 },
49+
{ 0x1B6B, 0x1B73, 1 }, { 0x1DC0, 0x1DCA, 1 }, { 0x1DFE, 0x1DFF, 1 },
50+
{ 0x200B, 0x200F, 1 }, { 0x202A, 0x202E, 1 }, { 0x2060, 0x2063, 1 },
51+
{ 0x206A, 0x206F, 1 }, { 0x20D0, 0x20EF, 1 }, { 0x302A, 0x302F, 2 },
52+
{ 0x3099, 0x309A, 2 }, { 0xA806, 0xA806, 1 }, { 0xA80B, 0xA80B, 1 },
53+
{ 0xA825, 0xA826, 1 }, { 0xFB1E, 0xFB1E, 1 }, { 0xFE00, 0xFE0F, 1 },
54+
{ 0xFE20, 0xFE23, 1 }, { 0xFEFF, 0xFEFF, 1 }, { 0xFFF9, 0xFFFB, 1 },
55+
{ 0x10A01, 0x10A03, 1 }, { 0x10A05, 0x10A06, 1 }, { 0x10A0C, 0x10A0F, 1 },
56+
{ 0x10A38, 0x10A3A, 1 }, { 0x10A3F, 0x10A3F, 1 }, { 0x1D167, 0x1D169, 2 },
57+
{ 0x1D173, 0x1D182, 2 }, { 0x1D185, 0x1D18B, 2 }, { 0x1D1AA, 0x1D1AD, 2 },
58+
{ 0x1D242, 0x1D244, 2 }, { 0xE0001, 0xE0001, 1 }, { 0xE0020, 0xE007F, 1 },
59+
{ 0xE0100, 0xE01EF, 1 }
6060
};
6161

62-
static int bisearch (uint rune, uint [,] table, int max)
62+
static uint[,] combiningWideChars = new uint[,] {
63+
/* Hangul Jamo init. consonants - 0x1100, 0x11ff */
64+
/* Miscellaneous Technical - 0x2300, 0x23ff */
65+
/* Hangul Syllables - 0x11a8, 0x11c2 */
66+
/* CJK Compatibility Ideographs - f900, fad9 */
67+
/* Vertical forms - fe10, fe19 */
68+
/* CJK Compatibility Forms - fe30, fe4f */
69+
/* Fullwidth Forms - ff01, ffee */
70+
/* Alphabetic Presentation Forms - 0xFB00, 0xFb4f */
71+
/* Chess Symbols - 0x1FA00, 0x1FA0f */
72+
73+
{ 0x1100, 0x115f, 2 }, { 0x2329, 0x232a, 2 }, { 0x2e80, 0x303e, 2 },
74+
{ 0x3041, 0x3096, 2 }, { 0x3099, 0x30ff, 2 }, { 0x3105, 0x312e, 2 },
75+
{ 0x3131, 0x318e, 2 }, { 0x3190, 0x3247, 2 }, { 0x3250, 0x4dbf, 2 },
76+
{ 0x4e00, 0xa4c6, 2 }, { 0xa960, 0xa97c, 2 }, { 0xac00 ,0xd7a3, 2 },
77+
{ 0xf900, 0xfaff, 2 }, { 0xfe10, 0xfe1f, 2 }, { 0xfe30 ,0xfe6b, 2 },
78+
{ 0xff01, 0xff60, 2 }, { 0xffe0, 0xffe6, 2 }
79+
};
80+
81+
static int bisearch(uint rune, uint[,] table, int max)
6382
{
6483
int min = 0;
6584
int mid;
6685

67-
if (rune < table [0, 0] || rune > table [max, 1])
86+
if (rune < table[0, 0] || rune > table[max, 1])
6887
return 0;
69-
while (max >= min) {
88+
while (max >= min)
89+
{
7090
mid = (min + max) / 2;
71-
if (rune > table [mid, 1])
91+
if (rune > table[mid, 1])
7292
min = mid + 1;
73-
else if (rune < table [mid, 0])
93+
else if (rune < table[mid, 0])
7494
max = mid - 1;
7595
else
7696
return 1;
@@ -79,37 +99,70 @@ static int bisearch (uint rune, uint [,] table, int max)
7999
return 0;
80100
}
81101

102+
static bool bisearch(uint rune, uint[,] table, out int width)
103+
{
104+
width = 0;
105+
var length = table.GetLength(0);
106+
if (length == 0 || rune < table[0, 0] || rune > table[length - 1, 1])
107+
return false;
108+
109+
for (int i = 0; i < length; i++)
110+
{
111+
if (rune >= table[i, 0] && rune <= table[i, 1]) {
112+
width = (int)table[i, 2];
113+
return true;
114+
}
115+
}
116+
117+
return false;
118+
}
119+
120+
static uint gethexaformat(uint rune, int length)
121+
{
122+
var hex = rune.ToString($"x{length}");
123+
var hexstr = hex.Substring(hex.Length - length, length);
124+
return (uint)int.Parse(hexstr, System.Globalization.NumberStyles.HexNumber);
125+
}
126+
127+
/// <summary>
128+
/// Check if the rune is a non-spacing character.
129+
/// </summary>
130+
/// <param name="rune">The rune.<param>
131+
/// <returns>True if is a non-spacing character, false otherwise.</returns>
132+
public static bool IsNonSpacingChar(uint rune)
133+
{
134+
return bisearch(rune, combining, out _);
135+
}
136+
137+
/// <summary>
138+
/// Check if the rune is a wide character.
139+
/// </summary>
140+
/// <param name="rune">The rune.</param>
141+
/// <returns>True if is a wide character, false otherwise.</returns>
142+
public static bool IsWideChar(uint rune)
143+
{
144+
return bisearch(gethexaformat(rune, 4), combiningWideChars, out _);
145+
}
146+
82147
/// <summary>
83148
/// Number of column positions of a wide-character code. This is used to measure runes as displayed by text-based terminals.
84149
/// </summary>
85150
/// <returns>The width in columns, 0 if the argument is the null character, -1 if the value is not printable, otherwise the number of columns that the rune occupies.</returns>
86151
/// <param name="rune">The rune.</param>
87-
public static int ColumnWidth (Rune rune)
88-
{
152+
public static int ColumnWidth(Rune rune)
153+
{
89154
uint irune = (uint)rune;
90155
if (irune < 32 || (irune >= 0x7f && irune <= 0xa0))
91156
return -1;
92157
if (irune < 127)
93158
return 1;
94159
/* binary search in table of non-spacing characters */
95-
if (bisearch (irune, combining, combining.GetLength (0)-1) != 0)
96-
return 0;
160+
if (bisearch(irune, combining, out int width))
161+
// if (bisearch(irune, combining, combining.GetLength(0) - 1) != 0)
162+
// return 0;
163+
return width;
97164
/* if we arrive here, ucs is not a combining or C0/C1 control character */
98-
return 1 +
99-
((irune >= 0x1100 &&
100-
(irune <= 0x115f || /* Hangul Jamo init. consonants */
101-
irune == 0x2329 || irune == 0x232a || /* Miscellaneous Technical */
102-
(irune >= 0x2e80 && irune <= 0xa4cf &&
103-
irune != 0x303f) || /* CJK ... Yi */
104-
(irune >= 0xac00 && irune <= 0xd7a3) || /* Hangul Syllables */
105-
(irune >= 0xf900 && irune <= 0xfaff) || /* CJK Compatibility Ideographs */
106-
(irune >= 0xfe10 && irune <= 0xfe19) || /* Vertical forms */
107-
(irune >= 0xfe30 && irune <= 0xfe6f) || /* CJK Compatibility Forms */
108-
(irune >= 0xff00 && irune <= 0xff60) || /* Fullwidth Forms */
109-
(irune >= 0xffe0 && irune <= 0xffe6) || /* Alphabetic Presentation Forms*/
110-
(irune >= 0x1fa00 && irune <= 0x1facf) || /* Chess Symbols*/
111-
(irune >= 0x20000 && irune <= 0x2fffd) ||
112-
(irune >= 0x30000 && irune <= 0x3fffd))) ? 1 : 0);
113-
}
165+
return bisearch(gethexaformat(irune, 4), combiningWideChars, out width) ? width : 1;
166+
}
114167
}
115168
}

‎NStackTests/NStackTests.csproj

Lines changed: 7 additions & 7 deletions
Original file line numberDiff line numberDiff line change
@@ -1,6 +1,6 @@
11
<?xml version="1.0" encoding="utf-8"?>
22
<Project DefaultTargets="Build" ToolsVersion="4.0" xmlns="http://schemas.microsoft.com/developer/msbuild/2003">
3-
<Import Project="..\packages\NUnit.3.12.0\build\NUnit.props" Condition="Exists('..\packages\NUnit.3.12.0\build\NUnit.props')" />
3+
<Import Project="..\packages\NUnit.3.13.2\build\NUnit.props" Condition="Exists('..\packages\NUnit.3.13.2\build\NUnit.props')" />
44
<PropertyGroup>
55
<Configuration Condition=" '$(Configuration)' == '' ">Debug</Configuration>
66
<Platform Condition=" '$(Platform)' == '' ">AnyCPU</Platform>
@@ -31,13 +31,13 @@
3131
<AllowUnsafeBlocks>true</AllowUnsafeBlocks>
3232
</PropertyGroup>
3333
<ItemGroup>
34-
<Reference Include="nunit.framework, Version=3.12.0.0, Culture=neutral, PublicKeyToken=2638cd05610744eb, processorArchitecture=MSIL">
35-
<HintPath>..\packages\NUnit.3.12.0\lib\net45\nunit.framework.dll</HintPath>
34+
<Reference Include="nunit.framework, Version=3.13.2.0, Culture=neutral, PublicKeyToken=2638cd05610744eb, processorArchitecture=MSIL">
35+
<HintPath>..\packages\NUnit.3.13.2\lib\net45\nunit.framework.dll</HintPath>
3636
</Reference>
3737
<Reference Include="System" />
38-
<Reference Include="System.ValueTuple, Version=4.0.3.0, Culture=neutral, PublicKeyToken=cc7b13ffcd2ddd51, processorArchitecture=MSIL">
39-
<HintPath>..\packages\System.ValueTuple.4.5.0\lib\net461\System.ValueTuple.dll</HintPath>
40-
</Reference>
38+
<Reference Include="System.ValueTuple, Version=4.0.3.0, Culture=neutral, PublicKeyToken=cc7b13ffcd2ddd51, processorArchitecture=MSIL">
39+
<HintPath>..\packages\System.ValueTuple.4.5.0\lib\net461\System.ValueTuple.dll</HintPath>
40+
</Reference>
4141
</ItemGroup>
4242
<ItemGroup>
4343
<Compile Include="ustringTest.cs" />
@@ -59,6 +59,6 @@
5959
<PropertyGroup>
6060
<ErrorText>This project references NuGet package(s) that are missing on this computer. Use NuGet Package Restore to download them. For more information, see http://go.microsoft.com/fwlink/?LinkID=322105. The missing file is {0}.</ErrorText>
6161
</PropertyGroup>
62-
<Error Condition="!Exists('..\packages\NUnit.3.12.0\build\NUnit.props')" Text="$([System.String]::Format('$(ErrorText)', '..\packages\NUnit.3.12.0\build\NUnit.props'))" />
62+
<Error Condition="!Exists('..\packages\NUnit.3.13.2\build\NUnit.props')" Text="$([System.String]::Format('$(ErrorText)', '..\packages\NUnit.3.13.2\build\NUnit.props'))" />
6363
</Target>
6464
</Project>

‎NStackTests/packages.config

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,5 +1,5 @@
11
<?xml version="1.0" encoding="utf-8"?>
22
<packages>
3-
<package id="NUnit" version="3.12.0" targetFramework="net462" />
3+
<package id="NUnit" version="3.13.2" targetFramework="net462" />
44
<package id="System.ValueTuple" version="4.5.0" targetFramework="net462" />
55
</packages>

0 commit comments

Comments
 (0)