v0.1.1

Patch release with correctness and reliability fixes for the compiled matcher and its FFI adapter.

Fixed

• Corrected charset union and intersection when either operand is complemented.
• Constrained charset arithmetic to the Unicode codepoint domain and removed boundary-overflow cases.
• Corrected logical emptiness checks for complemented full-range charsets.
• Fixed \uXXXX and surrogate-pair decoding in the optimized wire reader.
• Rejected malformed control characters and trailing wire input.
• Removed the per-call response-buffer leak in scanFast.
• Made differential fuzz scripts fail reliably when compilation or execution fails.

Validation

• 377 interpreted matcher checks and 18 grammar-error cases.
• 51 compiled-core checks, wire tests, and 78 differential cases.
• 10,000 randomized differential cases with zero disagreements.
• 168 Arturo corpus files with zero disagreements and 287 matching definitions.
• Strict documentation build, package validation, shell lint, and both Nim refc and orc memory managers.

v0.1.0

First release on pkgr.art.

A PARSE-class grammar dialect for Arturo: rules are ordinary Arturo blocks that compose, nest and recurse, and they match strings and blocks, so Arturo can pattern-match its own source. Quantifiers, ordered choice, captures with rollback, charsets, recursive rules, block matching with type and literal-word terminals, lookahead, collect/keep, host escapes, cut, opt-in memoization, and farthest-failure error reporting through scanError.

One file, no dependencies, Arturo 0.10.0 or later.