Reverse a String in JavaScript: Preserve the Intended Text Units

The task

Reverse a decomposed accented letter followed by an emoji, keeping the accent attached to its letter. Verify Intl.Segmenter behavior on the actual Node22 runtime.

Runnable Node 22 example

const assert = require('node:assert/strict');

const text = 'a\u0301🙂';
const codeUnits = text.split('').reverse().join('');
const codePoints = [...text].reverse().join('');
const clusters = [...new Intl.Segmenter('en', {granularity: 'grapheme'}).segment(text)]
  .map(part => part.segment);
const reversed = clusters.reverse().join('');
assert.notEqual(codeUnits, '🙂a\u0301');
assert.equal(codePoints, '🙂\u0301a');
assert.equal(reversed, '🙂a\u0301');
console.log(reversed);

Expected output and explanation

The code-unit reversal fails the desired output and can separate the emoji’s surrogate pair. The code-point reversal preserves the emoji but moves the combining accent ahead of its letter, producing a different cluster arrangement.

The segmenter groups the original accented letter before reversing the segments, so the target output retains that relationship. This is a tested result for the explicit fixture and runtime, not a claim that reversing text preserves every language’s display or reading semantics.

Pitfall to check

Reversal can be inappropriate for bidirectional writing or complex presentation requirements even when clusters remain intact. Define the product’s text operation first. Do not call a code-unit algorithm Unicode-safe merely because it works for a short ASCII sample.