CHANGELOG.md

Changelog

0.2.0

Features

• Tolerant parsing: parse(input, { tolerant: true }) collects all errors and returns a best-effort CST instead of throwing on the first syntax error. Useful for editor integrations, as-you-type validation, and multi-error diagnostics.
• Tolerant tokenization: tokenize(input, { tolerant: true }) returns a TokenizeResult with the token stream and any lexer errors.
• toCleanTree(): narrows a tolerant ParseResult to a strict FilterNode when error-free, or returns null. Safe to pass directly to transform().
• hasErrorNodes(): walks a tolerant CST and returns true if any ErrorNode is present.
• ErrorNode CST type: new node type representing recovered syntax errors, with the original error, skipped tokens, and expectedAt span for editor diagnostics.

New exports

toCleanTree, hasErrorNodes, ErrorNode, ParseResult, TokenizeResult, MaybeError

0.1.3

Minor test improvements.

0.1.2

Enable npm provenance attestation on publish.

0.1.1

Package metadata and CI fixes.

0.1.0

Initial release.

Features

• Full AIP-160 grammar: comparisons (=, !=, <, <=, >, >=), AND/OR/NOT, field traversal (.), has operator (:), functions, wildcards, parentheses
• filter(): one-shot parse + evaluate in a single call
• compile(): parse once, evaluate many times with CompiledFilter.evaluate()
• evaluateAsync(): async variant supporting custom functions that return promises
• Serialization: CompiledFilter.toJSON() / CompiledFilter.fromJSON() for persistence to databases, caches, or network transfer
• Structured errors: typed error classes (InputLengthError, DepthLimitError, UnexpectedTokenError, etc.) with source spans and contextual hints
• Full CST and AST: parse() produces a concrete syntax tree (for tooling), transform() produces an abstract syntax tree (for evaluation or custom backends)
• Security hardening: input length limits, recursion depth limits, prototype chain guards, error message truncation
• Zero dependencies