Fix an issue that broke attribute reading when defining a nested language on a tag that isn't a style, textearea, or script tag.
When using the "selfClosing" dialect, fix parse errors for <br/> (tags that implicitly self-close) and <script/> (special tags).
Add bidiIsolate node props for tags, comments, and attributes.
Mark tags, attributes, and comments as isolating for bidirectional text.
Fix a bug that caused the parser to not properly recognize empty HTML comments. Don't emit nested parses for empty elements in configureNesting
Fix an issue where configureNesting could emit empty overlay ranges for empty elements.
Improve parsing of multiple unfinished opening tags by disallowing less-than characters in attribute names.
Make the package work with new TS resolution styles.
Fix a bug that caused mixed parsing of non-style/script elements to attach the nested parse tree to the incorrect node.
Make sure build output is strictly ES6 again.
Fix a crash caused by nexted parsing of attributes returning invalid ranges for attributes missing their closing quote.
Always allow self-closing tags in <svg> and <math> tags.
configureNesting now supports targeting tags other than style, script, and textarea.
configureNesting now takes a second argument that can be used to specify that the value of some attributes should be parsed with an external parser.
Align allowed characters in unquoted attribute values with the spec. Replace another unicode escape with a character
Allow more characters in attribute names, following HTML spec.
The new "selfClosing" dialect allows /> syntax to be used to create self-closing tags.
Continue parsing when an invalid entity reference appears in an attribute value.
First stable version.
Fix a bug where comment end tokens preceded by dashes were sometimes not recognized.
Move to 0.16 serialized parser format.
The parser now includes syntax highlighting information in its node types.
Make sure the tree for unfinished self-closing tags shows them as self-closing.
The module's name changed from lezer-html to @lezer/html.
Upgrade to the 0.15.0 lezer interfaces.
Add a rule for invalid & syntax to avoid error-recovery around stray ampersands.
Fix a problem where like attributes inappropriately included trailing whitespace.
Strip quotes from attribute values passed to attrs predicates in configureNesting.
Optimize the tokenizer by using a context tracker.
Make comments consist of multiple tokens, so that huge comments don't freeze the parser.
Fix versions of lezer packages depended on.
The nested parser configuration utility is now called configureNesting, and returns an object to pass to Parser.configure's nested option instead of a new parser.
The parser can now be given a "noMatch" dialect to not mark mismatched tags.
Adjust to changed serialized parser format.
The parser now more effectively matches close and open tags, even in the presence of mismatched tags.
Fix lezer depencency versions
Follow change in serialized parser format.
Upgrade to 0.10 parser serialization
Upgrade to 0.9 parser serialization
Tag start/end tokens now have NodeProp.openedBy/closedBy props.
Regenerate parser with a fix in lezer-generator so that the top node prop is properly assigned.
Make the package load as an ES module on node
Provide an ES module file.
Adds support for single-quoted attribute values.
Don't treat /> tag ends as self-closing, just ignore them instead.
Follow 0.8.0 release of the library.
Use the lezer 0.7.0 parser format.
Allow whitespace between the < and / in a close tag.
Fix top prop missing from build output.
Move from lang to top prop on document node.
Adjust to 0.4.0 parse table format.
First numbered release.