xterm
Version:
Full xterm terminal, in your browser
172 lines (163 loc) • 6.32 kB
text/typescript
/**
* Copyright (c) 2016 The xterm.js authors. All rights reserved.
* @license MIT
*/
import { fill } from 'common/TypedArrayUtils';
export const wcwidth = (function(opts: {nul: number, control: number}): (ucs: number) => number {
// extracted from https://www.cl.cam.ac.uk/%7Emgk25/ucs/wcwidth.c
// combining characters
const COMBINING_BMP = [
[], [0x0483, 0x0486], [0x0488, 0x0489],
[], [0x05BF, 0x05BF], [0x05C1, 0x05C2],
[], [0x05C7, 0x05C7], [0x0600, 0x0603],
[], [0x064B, 0x065E], [0x0670, 0x0670],
[], [0x06E7, 0x06E8], [0x06EA, 0x06ED],
[], [0x0711, 0x0711], [0x0730, 0x074A],
[], [0x07EB, 0x07F3], [0x0901, 0x0902],
[], [0x0941, 0x0948], [0x094D, 0x094D],
[], [0x0962, 0x0963], [0x0981, 0x0981],
[], [0x09C1, 0x09C4], [0x09CD, 0x09CD],
[], [0x0A01, 0x0A02], [0x0A3C, 0x0A3C],
[], [0x0A47, 0x0A48], [0x0A4B, 0x0A4D],
[], [0x0A81, 0x0A82], [0x0ABC, 0x0ABC],
[], [0x0AC7, 0x0AC8], [0x0ACD, 0x0ACD],
[], [0x0B01, 0x0B01], [0x0B3C, 0x0B3C],
[], [0x0B41, 0x0B43], [0x0B4D, 0x0B4D],
[], [0x0B82, 0x0B82], [0x0BC0, 0x0BC0],
[], [0x0C3E, 0x0C40], [0x0C46, 0x0C48],
[], [0x0C55, 0x0C56], [0x0CBC, 0x0CBC],
[], [0x0CC6, 0x0CC6], [0x0CCC, 0x0CCD],
[], [0x0D41, 0x0D43], [0x0D4D, 0x0D4D],
[], [0x0DD2, 0x0DD4], [0x0DD6, 0x0DD6],
[], [0x0E34, 0x0E3A], [0x0E47, 0x0E4E],
[], [0x0EB4, 0x0EB9], [0x0EBB, 0x0EBC],
[], [0x0F18, 0x0F19], [0x0F35, 0x0F35],
[], [0x0F39, 0x0F39], [0x0F71, 0x0F7E],
[], [0x0F86, 0x0F87], [0x0F90, 0x0F97],
[], [0x0FC6, 0x0FC6], [0x102D, 0x1030],
[], [0x1036, 0x1037], [0x1039, 0x1039],
[], [0x1160, 0x11FF], [0x135F, 0x135F],
[], [0x1732, 0x1734], [0x1752, 0x1753],
[], [0x17B4, 0x17B5], [0x17B7, 0x17BD],
[], [0x17C9, 0x17D3], [0x17DD, 0x17DD],
[], [0x18A9, 0x18A9], [0x1920, 0x1922],
[], [0x1932, 0x1932], [0x1939, 0x193B],
[], [0x1B00, 0x1B03], [0x1B34, 0x1B34],
[], [0x1B3C, 0x1B3C], [0x1B42, 0x1B42],
[], [0x1DC0, 0x1DCA], [0x1DFE, 0x1DFF],
[], [0x202A, 0x202E], [0x2060, 0x2063],
[], [0x20D0, 0x20EF], [0x302A, 0x302F],
[], [0xA806, 0xA806], [0xA80B, 0xA80B],
[], [0xFB1E, 0xFB1E], [0xFE00, 0xFE0F],
[], [0xFEFF, 0xFEFF], [0xFFF9, 0xFFFB]
];
const COMBINING_HIGH = [
[], [0x10A05, 0x10A06], [0x10A0C, 0x10A0F],
[], [0x10A3F, 0x10A3F], [0x1D167, 0x1D169],
[], [0x1D185, 0x1D18B], [0x1D1AA, 0x1D1AD],
[], [0xE0001, 0xE0001], [0xE0020, 0xE007F],
[]
];
// binary search
function bisearch(ucs: number, data: number[][]): boolean {
let min = 0;
let max = data.length - 1;
let mid;
if (ucs < data[0][0] || ucs > data[max][1]) {
return false;
}
while (max >= min) {
mid = (min + max) >> 1;
if (ucs > data[mid][1]) {
min = mid + 1;
} else if (ucs < data[mid][0]) {
max = mid - 1;
} else {
return true;
}
}
return false;
}
function wcwidthHigh(ucs: number): 0 | 1 | 2 {
if (bisearch(ucs, COMBINING_HIGH)) {
return 0;
}
if ((ucs >= 0x20000 && ucs <= 0x2fffd) || (ucs >= 0x30000 && ucs <= 0x3fffd)) {
return 2;
}
return 1;
}
const control = opts.control | 0;
// create lookup table for BMP plane
const table = new Uint8Array(65536);
fill(table, 1);
table[0] = opts.nul;
// control chars
fill(table, opts.control, 1, 32);
fill(table, opts.control, 0x7f, 0xa0);
// apply wide char rules first
// wide chars
fill(table, 2, 0x1100, 0x1160);
table[0x2329] = 2;
table[0x232a] = 2;
fill(table, 2, 0x2e80, 0xa4d0);
table[0x303f] = 1; // wrongly in last line
fill(table, 2, 0xac00, 0xd7a4);
fill(table, 2, 0xf900, 0xfb00);
fill(table, 2, 0xfe10, 0xfe1a);
fill(table, 2, 0xfe30, 0xfe70);
fill(table, 2, 0xff00, 0xff61);
fill(table, 2, 0xffe0, 0xffe7);
// apply combining last to ensure we overwrite
// wrongly wide set chars:
// the original algo evals combining first and falls
// through to wide check so we simply do here the opposite
// combining 0
for (let r = 0; r < COMBINING_BMP.length; ++r) {
fill(table, 0, COMBINING_BMP[r][0], COMBINING_BMP[r][1] + 1);
}
return function (num: number): number {
if (num < 32) {
return control | 0;
}
if (num < 127) {
return 1;
}
if (num < 65536) {
return table[num];
}
// do a full search for high codepoints
return wcwidthHigh(num);
};
})({nul: 0, control: 0}); // configurable options
/**
* Get the terminal cell width for a string.
*/
export function getStringCellWidth(s: string): number {
let result = 0;
const length = s.length;
for (let i = 0; i < length; ++i) {
let code = s.charCodeAt(i);
// surrogate pair first
if (0xD800 <= code && code <= 0xDBFF) {
if (++i >= length) {
// this should not happen with strings retrieved from
// Buffer.translateToString as it converts from UTF-32
// and therefore always should contain the second part
// for any other string we still have to handle it somehow:
// simply treat the lonely surrogate first as a single char (UCS-2 behavior)
return result + wcwidth(code);
}
const second = s.charCodeAt(i);
// convert surrogate pair to high codepoint only for valid second part (UTF-16)
// otherwise treat them independently (UCS-2 behavior)
if (0xDC00 <= second && second <= 0xDFFF) {
code = (code - 0xD800) * 0x400 + second - 0xDC00 + 0x10000;
} else {
result += wcwidth(second);
}
}
result += wcwidth(code);
}
return result;
}