Fix/emoji support (#6187)

Co-authored-by: elasticdotventures <brianh@elastic.ventures>
This commit is contained in:
Jacob Richman 2025-08-13 17:33:01 -07:00 committed by GitHub
parent 6d01ba65a2
commit 342820cf5e
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 75 additions and 11 deletions

View File

@ -5,6 +5,7 @@
*/ */
import { describe, it, expect, beforeEach } from 'vitest'; import { describe, it, expect, beforeEach } from 'vitest';
import stripAnsi from 'strip-ansi';
import { renderHook, act } from '@testing-library/react'; import { renderHook, act } from '@testing-library/react';
import { import {
useTextBuffer, useTextBuffer,
@ -1278,6 +1279,45 @@ Contrary to popular belief, Lorem Ipsum is not simply random text. It has roots
); );
expect(getBufferState(result).text).toBe('Pasted Text'); expect(getBufferState(result).text).toBe('Pasted Text');
}); });
it('should not strip popular emojis', () => {
const { result } = renderHook(() =>
useTextBuffer({ viewport, isValidPath: () => false }),
);
const emojis = '🐍🐳🦀🦄';
act(() =>
result.current.handleInput({
name: '',
ctrl: false,
meta: false,
shift: false,
paste: false,
sequence: emojis,
}),
);
expect(getBufferState(result).text).toBe(emojis);
});
});
describe('stripAnsi', () => {
it('should correctly strip ANSI escape codes', () => {
const textWithAnsi = '\x1B[31mHello\x1B[0m World';
expect(stripAnsi(textWithAnsi)).toBe('Hello World');
});
it('should handle multiple ANSI codes', () => {
const textWithMultipleAnsi = '\x1B[1m\x1B[34mBold Blue\x1B[0m Text';
expect(stripAnsi(textWithMultipleAnsi)).toBe('Bold Blue Text');
});
it('should not modify text without ANSI codes', () => {
const plainText = 'Plain text';
expect(stripAnsi(plainText)).toBe('Plain text');
});
it('should handle empty string', () => {
expect(stripAnsi('')).toBe('');
});
}); });
}); });

View File

@ -5,6 +5,7 @@
*/ */
import stripAnsi from 'strip-ansi'; import stripAnsi from 'strip-ansi';
import { stripVTControlCharacters } from 'util';
import { spawnSync } from 'child_process'; import { spawnSync } from 'child_process';
import fs from 'fs'; import fs from 'fs';
import os from 'os'; import os from 'os';
@ -496,21 +497,44 @@ export const replaceRangeInternal = (
/** /**
* Strip characters that can break terminal rendering. * Strip characters that can break terminal rendering.
* *
* Strip ANSI escape codes and control characters except for line breaks. * Uses Node.js built-in stripVTControlCharacters to handle VT sequences,
* Control characters such as delete break terminal UI rendering. * then filters remaining control characters that can disrupt display.
*
* Characters stripped:
* - ANSI escape sequences (via strip-ansi)
* - VT control sequences (via Node.js util.stripVTControlCharacters)
* - C0 control chars (0x00-0x1F) except CR/LF which are handled elsewhere
* - C1 control chars (0x80-0x9F) that can cause display issues
*
* Characters preserved:
* - All printable Unicode including emojis
* - DEL (0x7F) - handled functionally by applyOperations, not a display issue
* - CR/LF (0x0D/0x0A) - needed for line breaks
*/ */
function stripUnsafeCharacters(str: string): string { function stripUnsafeCharacters(str: string): string {
const stripped = stripAnsi(str); const strippedAnsi = stripAnsi(str);
return toCodePoints(stripped) const strippedVT = stripVTControlCharacters(strippedAnsi);
return toCodePoints(strippedVT)
.filter((char) => { .filter((char) => {
if (char.length > 1) return false;
const code = char.codePointAt(0); const code = char.codePointAt(0);
if (code === undefined) { if (code === undefined) return false;
return false;
} // Preserve CR/LF for line handling
const isUnsafe = if (code === 0x0a || code === 0x0d) return true;
code === 127 || (code <= 31 && code !== 13 && code !== 10);
return !isUnsafe; // Remove C0 control chars (except CR/LF) that can break display
// Examples: BELL(0x07) makes noise, BS(0x08) moves cursor, VT(0x0B), FF(0x0C)
if (code >= 0x00 && code <= 0x1f) return false;
// Remove C1 control chars (0x80-0x9F) - legacy 8-bit control codes
if (code >= 0x80 && code <= 0x9f) return false;
// Preserve DEL (0x7F) - it's handled functionally by applyOperations as backspace
// and doesn't cause rendering issues when displayed
// Preserve all other characters including Unicode/emojis
return true;
}) })
.join(''); .join('');
} }