Extended search drops any query term that contains a line break

Open Beginner friendly
#862 0 comments 0 reactions 0 assignees View on GitHub

@goranai is already working on this.

Since Nov 18, 2026.

  • #863 by @goranai — open

Assessment

Difficulty
2/5
Estimated time
1-3 hours
Newbie friendliness
78/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
javascript
Domain
search

Research direction

Start with the 16 detection regexes in src/search/extended/matchers.ts and compare their behavior with the line-break examples in the issue. Add a regression test for a quoted term containing a line break and for the AND query, then check that the quoted term finds “The Lock\nArtist” and the AND query returns no results.

Written by the indexing model from the issue text.

Description

Is there an existing issue for this?

  • I have searched the existing issues

Description of the bug

With useExtendedSearch: true, a query term that contains a line break (\n, \r, U+2028 or U+2029) is thrown away. It is not searched at all.

This is the same bug as #696. That issue and its PR #697 were closed by the stale bot in 2023 and the bug is still there on main (edf2fb6, 7.6.0-beta.0).

The cause is in src/search/extended/matchers.ts. All 16 detection regexes use (.*) without the s flag. In JavaScript . does not match a line break. So a term with a line break matches none of the regexes, not even the catch-all fuzzy one /^(.*)$/. parseQuery() then pushes no matcher for it, and the term is gone.

This gives wrong results in two ways:

  • a query that is only that term finds nothing, even when an item is equal to it
  • in an AND query the term is skipped, so items that do not contain it are returned

A tab inside a term already works, because . matches a tab. Only line breaks are lost.

The Fuse.js version where this bug is happening.
Other (please specify in description): 7.6.0-beta.0 (current main). #696 reported it on 6.6.2.

Is this a regression?
No.

Steps To Reproduce

import Fuse from 'fuse.js'

const books = [
  { title: "Old Man's War" },
  { title: 'The Lock\nArtist' },
  { title: 'Artist for Life' }
]

const search = (q, ext) =>
  new Fuse(books, { keys: ['title'], useExtendedSearch: ext, threshold: 0.3 })
    .search(q)
    .map((r) => r.item.title)

console.log(search('\'"Lock\nArtist"', true)) // []
console.log(search("'Lock\nArtist", true)) // []
console.log(search('Lock\nArtist', true)) // []
console.log(search('Lock\nArtist', false)) // [ 'The Lock\nArtist' ]
console.log(search('Lock\nArtist war', true)) // [ "Old Man's War" ]

The last query asks for items that match Lock\nArtist AND war. It returns "Old Man's War", because the first term was dropped.

Expected behavior

A term with a line break should be searched like any other term, as basic search does. '"Lock\nArtist" and 'Lock\nArtist should find "The Lock\nArtist". Lock\nArtist war should find nothing here.

A small fix is to add the s flag to the 16 regexes, as #696 suggested. The build target is ES2020, so the flag is supported. Then a line break is kept inside the term, the same way a tab is today. I can open a PR with that and a test.

One open question for you: the docs say white space acts as AND. If you would rather have an unquoted line break split terms like a space does, that is a change in tokenize() instead. Quoted terms like '"Lock\nArtist" would still need the s flag.

Validated with Claude Opus 5.5.

Dominant language
JavaScript
Stars
20.5k
Forks
828
PR merge metrics
No merged PRs in 30d

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from krisk/Fuse

All issues in krisk/Fuse

Similar issues

More JavaScript issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.