Faster decoding & escaping, with smaller bundles. Encoding output stays compatible with v8.

Improvements

• Decoding: decodeHTML 1.19×, decodeHTMLStrict 1.14×, decodeHTMLAttribute 1.18×, decodeXML 1.46×, default decode 1.44× as fast. #2248, #2348
• Escaping: escapeUTF8 1.49×, escapeAttribute 1.23×, escapeText 1.33× as fast. #2247, #2348
• Encoding mostly ASCII 4 KiB prose: encodeHTML 1.70×, encodeNonAsciiHTML 2.24×, encodeXML / escape 2.46× as fast. On the short, dense benchmark input, HTML encoding is unchanged; XML encoding and default encode take about 6% longer. #2246, #2348
• 19% smaller full bundle (18% gzipped), targeting ES2022.

Compared with v8.0.0; speedups are medians of nine runs on Node 26.8.1, Linux/x64. Short inputs use the repository benchmark fixtures; results vary with input.

Fixes

• decodeHTMLAttribute preserves interrupted references such as &notin<, previously incorrectly decoded to ¬in<. #2213
• decodeHTML preserves unterminated names such as &Xi$, previously incorrectly decoded to Â. #2331
• decodeXML and default decode preserve XML numeric code points: € now yields U+0080 instead of €. HTML decoding still yields €. By @adarshx01 in #2347.

Deprecations

• Use String.prototype.codePointAt instead of getCodePoint from entities/escape. #2268

Full Changelog: v8.0.0...v8.1.0

What's Changed

• BREAKING CHANGE: entities is now ESM-only #2183
  • BREAKING: Update engines to node >= 20.19.0
    • Update tsconfig: target es2022
  • BREAKING: Remove deprecated functions & polyfills
• Use shared eslint config #2157

Full Changelog: v7.0.1...v8.0.0

What's Changed

• Addresses an issue that broke streaming for htmlparser2 by @fb55 in #2117
• Exclude test files from published package by @fb55 in 2651348

Full Changelog: v7.0.0...v7.0.1

No changes in public API; a lot of internals changed though, for faster encoding & decoding and reduced memory usage.

Improvements

• Improve decode trie size by @fb55 in #1944

Reduced length of the encoded trie by 21% for HTML (from 15,242 words (uint16) to 12075), by:

• Adding a semicolon flag on value nodes, which means semicolons don't have to be encoded anymore.
• Compacting dict keys: two keys are now stored in each word, instead of one
• Compacting runs of 3 or more characters: successive characters are now also stored as two characters per word

Also reduces the size of the encoded table by 32% (after the previous change) by encoding it as base64.

• Improve encode trie file size by @fb55 in #1948

Encodes the encode trie as a string, with a simple parser turning it into the previous structure on load. This leads to a 44% decrease in file size (from 24,034 to 13,392 bytes).

• Speed up encodes using bitsets by @fb55 in #1952

Uses bitmaps to replace regex when looking for ASCII code points that should be replaced with entities. Leads to a ~1.6x speed up for encodeHTML.

Miscellaneous

• Fix broken repository URL in package.json by @ExplodingCabbage in https://github.com/fb55/entities/pull/1919pull/1931
• Use Biome by @fb55 in #1953

New Contributors

• @ExplodingCabbage made their first contribution in #1919

Full Changelog: v6.0.1...v7.0.0

What's Changed

• fix: add types for old node entry points by @43081j in #1861

Full Changelog: v6.0.0...v6.0.1

Features

• Improved bundler tree shaking by adding pure annotations by @Gravitonic in #1501
• BREAKING: Simplified deep import paths, to make entities work in old NodeJS versions by @fb55 in #1531
  • If you are using deep imports, you will have to update them. Instead of importing entities/dist/decode.js, import entities/decode. The same applies to escape.
• Automated publishes by @fb55 in #1532
  • entities is now available on both npm and JSR
• Enabled isolated declarations by @fb55 in #1522

Full Changelog: v5.0.0...v6.0.0

What's Changed

• Improved code style (based on eslint-plugin-unicorn) #1496
• Upgraded toolchain (to tshy, vitest, and tsx) #1497

Breaking Changes

• ⚠️ BREAKING: The lib directory was renamed to dist in #1497. Deep imports will have to be updated.

Full Changelog: v4.5.0...v5.0.0

New Features

entities now features an EntityDecoder class that can be used to decode entities across multiple chunks. This is needed to parse streams, and will be the foundation for entity decoding in htmlparser2 and parse5.

For users of this library, the most welcome addition will be a new attribute decoding mode, which will ignore some entities that were previously parsed (eg. `foo=bar&amp=boo' will be left unchanged).

entities' decode functions all use the new decoder under the hood. There is a new decodeHTMLAttribute function for attribute parsing, and some rare edge-cases where entities diverged from the HTML spec have been eliminated.

Relevant PRs

• feat(decode): Add EntityDecoder class by @fb55 in #1136
• fix(decode): Handle chunks for numeric entities by @fb55 in #1146

Small Changes

• Typo fixes by @styfle #1032, @Jiralite #1060, and @sandeepkambham08 #1129
• Harden GitHub Actions by @step-security-bot in #928

Full Changelog: v4.4.0...v4.5.0

• Smaller encode/decode maps by @sapphi-red in #909
  • The encode and decode maps are now stored more space-efficiently, saving 44% in minified bundles, and cutting the size of the installed module by 52%.

Please note that some JavaScript minifiers, such as terser, will now produce non-ASCII output. If you use entities in the browser, please make sure you are set up to ship UTF-8 to clients, or set eg. terser's ascii_only option to false.

Full Changelog: v4.3.1...v4.4.0

• Fixed an off-by-one error that caused some entities to be decoded to different values (#858) 3b0d48b

v4.3.0...v4.3.1