v7.0.0

[email protected] changes a lot of internals, resulting in an 20% overall performance improvement in AndreasMadsen's htmlparser-benchmark.

Breaking changes:

• Fixed how start & end index positions are calculated (#910) 5ab080e
  • Some indices, especially end indices, will now have changed. Most importantly, end indices will now always be greater or equal than start indices (whoops!).

Features:

• Added an isVoidElement method to the parser (#785) 00ce57a

Refactors:

• Use a trie to decode HTML & XML entities in the tokenizer (#863) 9a47a55
  • Leads to large speed-ups when dealing with entities.
• Iterate over char codes in the tokenizer (#894) f5aed75
  • Improved tokenizer performance by ~40%.
• Use Map for openImpliesClose in the parser (#911) 39a8109
• Moved logic of FeedHandler to a function (#912) 3a672ff

v6.1.0

Features:

• Export tokenizer callback interface from main module (#751) ab0b3fc f59473a

Fixes:

• Allow XML tags to start with any character (#778) 0b94ab5

Upgrades:

• Bump domhandler from 4.0.0 to 4.1.0 e64e8e5
• Bump domelementtype from 2.1.0 to 2.2.0 8bc1719
• Bump domutils from 2.4.4 to 2.5.2 8b91d97 cf77476 7c233de

v6.0.1...v6.1.0

v6.0.1

• Fix parsing special closing tags (#746) 214ab08
  • Thanks to @BenoitZugmeyer for the report (#745)!

v6.0.0...v6.0.1

v6.0.0

Breaking:

• Bump domhandler, domutils 4dd4233 0d278fd
  • The new version of domhandler now comes with an actual root element for the document. This might break tests in a few cases. See the domhandler release notes for more details.
• Make some private properties actually private 1c71e60

Features:

• Add a parseDocument method 4653f23
  • This returns the root node of the document, instead of an array of the first nodes. You likely want to use this instead of the now deprecated getDOM method.
• Improve docs df7ea98 1ce1d3b 0437d9c

Minor:

• FeedHandler: Slightly restructure code b6b4382

v5.0.1...v6.0.0

v5.0.1

• Fix: Parse entities in <title> tags (#614, #615 by @billneff79) 3295a8b
• Fix: Remove @types/node as a peer dependency 1ace384

Also pulls in a new version of the entities module, which features more compact entity maps.

v5.0.0

Breaking changes:

• Default the decodeEntities option to true 8ac01e0
• Removes underscores in front of many private properties & methods. 6e296d2
• Removes EVENTS, WritableStream and CollectingHandler exports from module import. The latter two are still part of the module, but now have to be imported explicitly. 6e296d2
• The parser no longer extends EventEmitter f30f13c
• HTML <title> tag content is now processed as text (#483 by @billneff79) 0189e56

Features:

• Add media content parsing to FeedHandler (#560 by @gcandal) a85e4e0
• Expose the quotes that were used in the onattribute event 3c86256
• Add "sideEffects: false" to package.json (#474 by @ericjeney) d90dd64
• Explain stream usage in README (#446 by @mnmkng) 4c0fba8

Bug Fixes:

• Properly back out of numeric entities, decode entities in attributes (fixes #276) eaf2872
• Fix broken parsing after self-closing special tags (#515 by @warriordog) 4ec596f
• Fix parse bug when tag name is not ASCII alpha (#497 by @Zuckjet) bc010de

Diverse:

• Improve Coverage (#540 by @brettz9) 6d8a2ff
• Check missing elem with getOneElement (#543 by @brettz9) 1cf297e
• Add test for #125 40d9556

Thanks to everyone that contributed to this release!

Commit Range:
v4.1.0...v5.0.0

v4.1.0

• Don't fail when parsing <__proto__> (fixes #387)
• Add types field to package.json
• Update dependencies

v4.0.0

• Port to TypeScript, Jest
• Remove the Stream and ProxyHandler exports
• Order some conditionals in Tokenizer by their likelihood to be hit
• Fix implicit closing of certain tags — @voithos
• Fix: options.Tokenizer modified outer scope — @thorn0

3.3.0: [tests] write only single characters for testing chunked data

failed previously (only for FeedHandler tests), fixed now due to
DomHandler upgrade (which removed the `ignoreWhitespace` option)