Support single-quoted attribute values.
Fix a mistake that broke highlighting for mismatched tag names.
Mark tags, attributes, and comments as isolating for bidirectional text.
The parser is now more precise about what is parsed as an entity.
Fix a bug that caused the end of a CDATA block, comment, or processing instruction to not be recognized when the character starting the end marker occured right before it.
Make the package work with new TS resolution styles.
Fix a bug where single-character comments, processing instructions, and CDATA blocks were not parsed correctly.
Fix a bug that would cause the tokenizer to miss ]]> and --> tokens when they were preceded by ] or -.
First stable version.
Move to 0.16 serialized parser format.
The parser now includes syntax highlighting information in its node types.
Fix parsing of CDATA sections.
The module's name changed from lezer-xml to @lezer/xml.
Upgrade to the 0.15.0 lezer interfaces.
Fix an issue that cause improper tokenizing during some types of error recovery.
Optimize tokenizing with a context tracker.
Make comments, processing instructions, and cdata consist of multiple tokens to avoid freezing the parser on huge inputs with unfinished elements of those types.
Fix versions of lezer packages depended on.
Adjust to changed serialized parser format.
Fix lezer depencency versions
Follow change in serialized parser format.
Make sure mismatched close tags name tokens have a different name than matching ones.
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
Follow 0.8.0 release of the library.
Use the lezer 0.7.0 parser format.
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.