better-strip-comments

Strip line and/or block comments from a string. Blazing fast, and works with JavaScript, Sass, CSS, Less.js, and a number of other languages.

This is a fork of jonschlinkert/strip-comments maintained at imsyedabdullah/better-strip-comments. The public API is 100% compatible with the original — if your code works against [email protected], it will work against this fork unchanged.

• What's new in this fork
• Install
• What does this do?
• Usage
• API
• About

What's new in this fork

String-aware parsing (bug fix)

The original parser used a regex with [^\1] inside a character class to match string bodies. Backreferences don't work inside [...] in JavaScript — \1 there is interpreted as the octal character \x01, not as "the opening quote." In practice this meant that strings containing comment markers would sometimes break the parser, causing it to treat /* inside a string as the start of a real block comment and eat everything after it until EOF.

Real-world example that broke the original:

const strip = require('better-strip-comments');

// PHP file containing a glob pattern inside a single-quoted string
const input = "$candidates = glob($plugin_root . '/*.php') ?: [];\nforeach ($candidates as $c) { echo $c; }";

// Original strip-comments 2.0.1: truncates everything after the '/*
// This fork:                      returns the input unchanged
console.log(strip(input));

The fix rewrites string tokenization to use three separate, correct regexes — one each for single quotes, double quotes, and template literals — with proper handling of escape sequences and newline bounding. URLs, glob patterns, and any other comment-like characters inside strings are now preserved correctly.

strip('const url = "https://example.com/path";');
// => 'const url = "https://example.com/path";'   ✓ preserved

strip("const glob = '/path/**/*.js'; // trailing");
// => "const glob = '/path/**/*.js'; "             ✓ glob preserved, comment stripped

strip("const s = '/* not a comment */'; /* real */ const x = 1;");
// => "const s = '/* not a comment */';  const x = 1;"

PHP trailing-comment-at-EOF (bug fix)

The original PHP line-comment regex required one of ?> or \n to follow the comment. A trailing # or // comment on the last line of a file (no trailing newline) was silently left in the output. Fixed by also accepting end-of-string as a terminator.

strip('<?php\n$x = 1; # tail', { language: 'php' });
// Original: '<?php\n$x = 1; # tail'   ✗ comment kept
// Fork:     '<?php\n$x = 1; '         ✓ comment stripped

Newly supported languages

In addition to the languages the original supports, this fork adds:

| Language | Identifier(s) | Notes | |---|---|---| | Rust | rust | C-style // and /* */, including /// and //! doc comments | | Go | go | C-style | | Kotlin | kotlin, kt | C-style | | Dart | dart | C-style including /// doc comments | | Scala | scala | C-style | | JSON with comments | jsonc, json5 | C-style | | JSX / TSX | jsx, tsx | Same as JS/TS; note that JSX comments are {/* ... */} | | Vue / Svelte | vue, svelte | Defaults to JS-style for <script> blocks | | TOML | toml | # line comments | | INI | ini | Both # and ; line comments | | YAML | yaml, yml | # line comments | | Shell / Bash | shell, bash, sh, zsh | # line comments | | PowerShell | powershell, ps1 | # line comments and <# ... #> block comments | | Windows Batch | batch, bat, cmd | REM (case-insensitive) and :: line comments | | R | r | # line comments | | Elixir | elixir | # line comments | | SCSS | scss | Alias for JS-style (previously only sass was aliased) | | TypeScript | typescript | Canonical spelling (the original typscript typo-alias is still accepted) |

All of these correctly preserve comment-like characters inside strings. Example:

strip('name: "#not-a-comment"\n# real comment\nage: 30', { language: 'yaml' });
// => 'name: "#not-a-comment"\n\nage: 30'

strip('echo "#not a comment" # trailing', { language: 'bash' });
// => 'echo "#not a comment" '

strip('set REMOTE=1\nREM actual comment', { language: 'batch' });
// => 'set REMOTE=1\n'   // REMOTE is not mistaken for a REM comment

Test coverage

The fork ships 91 new tests across three files (test/regression.js, test/new-languages.js, test/strings.js) plus fixture pairs for the new languages. The full test suite — original 60 upstream tests plus 91 new ones — runs with npm test.

Install

Install with npm (requires Node.js >=10):

$ npm install --save better-strip-comments

Or install this fork directly from GitHub:

$ npm install --save imsyedabdullah/better-strip-comments

What does this do?

Takes a string and returns a new string with comments removed. Works with line comments and/or block comments. Optionally removes the first comment only or ignores protected comments.

Works with:

• ada
• apl
• applescript
• bash / sh / zsh / shell
• batch / bat / cmd
• c
• csharp
• css
• dart
• elixir
• go
• hashbang
• haskell
• html
• ini
• java
• javascript
• jsonc / json5
• jsx / tsx
• kotlin / kt
• less
• lua
• matlab
• ocaml
• pascal
• perl
• php
• powershell / ps1
• python
• r
• ruby
• rust
• sass / scss
• scala
• shebang
• sql
• swift
• toml
• typescript / typscript / ts
• vue / svelte
• xml
• yaml / yml

Usage

By default all comments are stripped.

const strip = require('better-strip-comments');
const str = strip('const foo = "bar";// this is a comment\n /* me too */');
console.log(str);
// => 'const foo = "bar";\n'

To strip comments for a specific language, pass the language option:

strip('x = 1  # comment', { language: 'python' });
// => 'x = 1  '

strip('$x = 1 # trailing\n<#\n  doc\n#>\n$y = 2', { language: 'powershell' });
// => '$x = 1 \n\n$y = 2'

For more use-cases see the tests.

API

strip

Strip all code comments from the given input, including protected comments that start with !, unless disabled by setting options.keepProtected to true.

Params

• input {String}: string from which to strip comments

• options {Object}: optional options

  • line {Boolean}: if false strip only block comments, default true
  • block {Boolean}: if false strip only line comments, default true
  • language {String}: language name (default: 'javascript')
  • keepProtected {Boolean}: Keep ignored comments (e.g. /*! and //!)
  • preserveNewlines {Boolean}: Preserve newlines after comments are stripped

• returns {String}: modified input

Example

const str = strip('const foo = "bar";// this is a comment\n /* me too */');
console.log(str);
// => 'const foo = "bar";'

.block

Strip only block comments.

Params

• input {String}: string from which to strip comments
• options {Object}: pass opts.keepProtected: true to keep ignored comments (e.g. /*!)
• returns {String}: modified string

Example

const strip = require('better-strip-comments');
const str = strip.block('const foo = "bar";// this is a comment\n /* me too */');
console.log(str);
// => 'const foo = "bar";// this is a comment'

.line

Strip only line comments.

Params

• input {String}: string from which to strip comments
• options {Object}: pass opts.keepProtected: true to keep ignored comments (e.g. //!)
• returns {String}: modified string

Example

const str = strip.line('const foo = "bar";// this is a comment\n /* me too */');
console.log(str);
// => 'const foo = "bar";\n/* me too */'

.first

Strip the first comment from the given input. Or, if opts.keepProtected is true, the first non-protected comment will be stripped.

Params

• input {String}
• options {Object}: pass opts.keepProtected: true to keep comments with !
• returns {String}

Example

const output = strip.first(input, { keepProtected: true });
console.log(output);
// => '//! first comment\nfoo; '

.parse

Parses a string and returns a basic CST (Concrete Syntax Tree).

Params

• input {String}: string to parse
• options {Object}: parse options
• returns {Object}: CST

About

Install dependencies and run tests with:

$ npm install && npm test

This is a fork; contributions here should be fork-specific (fixes, new languages, regression tests). For issues that also affect the upstream package, please consider opening them against the original repo as well.

Pull requests and stars are welcome. For bugs and feature requests, please open an issue.

Credit

The original strip-comments was written by Jon Schlinkert — see jonschlinkert/strip-comments. The parser architecture (CST-based parse/compile pipeline, per-language regex tables) and the public API are entirely his work. This fork's contribution is limited to bug fixes in the tokenizer and the addition of new language entries, with the original design preserved.

License

Copyright (c) 2014-present, Jon Schlinkert. Copyright (c) 2026, Syed Abdullah (fork modifications).

Released under the MIT License.