v7.0.0

7.0.0

Highlights

• Major rewrite: parser now builds on css-tree for correctness and performance.
• Package is pure ESM and uses exports; no CommonJS output.
• Walker helpers are exposed via registerWalkers() and can enable walk* methods like walkWords.

Breaking changes

• ESM‑only: require('postcss-values-parser') is no longer supported. Use import.
• Node.js 20.19+ required (per engines).
• Walker helpers aren’t attached by default. If you relied on root.walkWords(...) in v6, call registerWalkers(Container) once to add the walk* methods.
• Internal AST is compatible but not identical (migrated to css-tree under the hood). Class names like Word, Numeric, Func, etc., remain.

Migration

• Replace CommonJS with ESM imports.

• Register walker helpers before using any walk* APIs:

  import { parse, Container, registerWalkers } from 'postcss-values-parser';
  registerWalkers(Container);

  const root = parse('var(--foo)');
  root.walkWords((word) => {
  if (word.isVariable) console.log(word.value);
  });

• Review docs for updated node shapes and stringifier behavior.

Commits since v6.0.2

• 58c3ac0 feat!: use css-tree, rewrite package for ESM (#149) — Andrew Powell
• 1d367aa docs: update docs for v7 — shellscape
• c8b1647 chore: add github workflows — shellscape
• b19c7eb chore: use nvmrc — shellscape

Merged pull requests

• #149 use css-tree, rewrite package for ESM — Andrew Powell

Full Changelog: v6.0.2...v7.0.0

v6.0.2

Bugfixes

• fix: support parsing */ expressions without surrounding spaces (#144)
• fix: Quoted – clone throw an error (#141) (#142 )

Updates

• docs: fix numeric parse result (#139)

v6.0.0

Breaking Changes

• feat!: update PostCSS, use NodeErrorOptions directly from PostCSS types (#133)

Bugfixes

• fix: Function with dot inside name (#134)
• fix: TypeScript types, add validation (#132)

v4.0.0

Updates

• chore!: upgrade is-url-superb to ^4.0.0 (#125)

Breaking Changes

Node v10+ now required

v3.2.0

Bugfixes

fix: comma separated funcs with no spaces. fixes #91
fix: wonky function names. fixes #92
fix: numeric parsing (#109)

Features

feat: add hwb, lab, and lch to functions matching isColor (#111) …

Updates

test: add test for #104
test: add tests for #106

v3.0.5

Bugfixes

• fix: add time value units (#87)

v3.0.4

Updates

• Dependency updates and npm audit update

v3.0.3

Bugfixes

• fix: quater milimeters not detected (#81)

Updates

• docs: fix comment type (#78)

v3.0.2

Bugfixes

• fix: toString should stringify node (#74)
• fix: manipulation of node structures, proper stringifying (#73)

v3.0.0

Major Version 3

This is a major rewrite from top to bottom and contains many breaking changes. The rewrite leverages PostCSS itself under the hood. That means it uses the same tokenizer and parser, with some modifications. The last major update to PostCSS provided consumers with an outstanding model to inherit and extend from the base classes to create derived parsers. That means we now benefit from the rock-solid tokenizing in PostCSS without having to reinvent the wheel, and without introducing new bugs.

This also allows for use of the base PostCSS Node classes, without redefining or duplicating their functionality. Tasks like walking nodes, or inspecting Node properties you'd expect in PostCSS are now identical.

Breaking Changes

Some Node types have changed. Most notably, all characters which can be considered "punctuation" have been moved into the Punctuation Node. Additionally, decorative property names on Nodes have changed in some instances. The full API documentation has more details.

Old Node Type	New Node Type
AtWord	AtWord
Colon	Punctuation
Comma	Punctuation
Comment	Comment
Func	Func
Numbr	Numeric
Operator	Operator
Paren	Punctuation
Str	Quoted
UnicodeRange	UnicodeRange
Word	Word