[subotac]

Summary

Fixes #2172.

Match Lodash when a separator starts exactly at the truncation boundary or at the beginning of the string.

import { truncate as lodashTruncate } from 'lodash';
import { truncate } from 'es-toolkit/compat';

const options = { length: 14, separator: /,? +/ };

lodashTruncate('hello world remainder', options); // 'hello world...'
truncate('hello world remainder', options);
// Before: 'hello...'; after: 'hello world...'

lodashTruncate(' abcdef', { length: 5, separator: /,? +/ }); // '...'
truncate(' abcdef', { length: 5, separator: /,? +/ });
// Before: ' a...'; after: '...'

The Lodash calls compile against @types/lodash with strict: true.
The same separator is used in the Stratus truncation filter.

Changes

• Preserve the retained prefix when a separator begins at the truncation boundary.
• Handle string and regular-expression separators at index zero.
• Advance past empty regular-expression matches, including Unicode matches, without changing the supplied expression's lastIndex.
• Add regression tests and a bundle size benchmark (benchmarks/bundle-size/truncate.spec.ts).

Bundle size

truncate from es-toolkit/compat, bundled and minified with esbuild as in benchmarks/bundle-size:

	main	This PR
Minified	813 B	1,005 B (+192 B)
Minified + gzip	497 B	585 B (+88 B)

The escapeRegExp import is gone, but the boundary check and the loop that finds the last separator (including the step past empty matches) add more code than it saved. lodash-es is 5,354 B.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
es-toolkit	Ready	Preview	Nov 2, 2026 1:36am UTC