v0.1.8

Changes

Changed

• Java 25 Grammar Sync - Synced grammar improvements from jbct-cli
  • Added cut operators (^) after discriminating keywords for better error messages
  • Added keyword helper rules with word boundaries (ClassKW, InterfaceKW, EnumKW, etc.)
  • Added token boundaries to PrimType, Modifier, Literal, Primary rules
  • Updated RefType lookahead to handle Type.@Annotation Inner correctly
  • Added TypeExpr rule for Type.class and Type::new expressions
  • Updated operator rules with lookaheads to prevent compound operator conflicts
  • Added RecordDecl lookahead to distinguish from methods/fields named 'record'

Fixed

• Farthest Failure Tracking - Error positions now report at the furthest parsing position instead of 1:1 after backtracking
  • Added furthestPos/furthestFailure tracking to both AST and CST generated parsers
  • Replaces null checks with Option<T> in CST parser generator for consistency
  • Fixed infinite recursion in AST parser when whitespace rule contained * quantifier
  • Fixed "unexpected input" errors to also use furthest failure position

Added

• Error Position Tests - 3 new tests verifying farthest failure tracking in generated parsers
• Test count: 305 → 308

Release 0.1.7

Changes

Changed

• JBCT Compliance Refactoring - Replaced null usage with Option<T> throughout the codebase
  • ParseResultWithDiagnostics.node() returns Option<CstNode>
  • ParseMode uses Option for nullable fields
  • ParsingContext packrat cache uses Option
  • Diagnostic.code() returns Option<String>
  • Generated ParseResult and CstParseResult use Option for nullable fields

Added

• JBCT Maven Plugin - Added jbct-maven-plugin 0.4.1 for code formatting and linting
• Internal Type Tests - 24 new tests for ParseResult, ParsingContext, and generated parser diagnostics

Test count: 271 → 305

Maven

<dependency>
    <groupId>org.pragmatica-lite</groupId>
    <artifactId>peglib</artifactId>
    <version>0.1.7</version>
</dependency>

v0.1.6

Added

• AST Support in Generated Parsers

  • Generated CST parsers now include AstNode type and parseAst() method
  • Allows parsing directly to AST (without trivia) from generated parsers

• Packrat Toggle in Generated Parsers

  • Added setPackratEnabled(boolean) method to generated parsers
  • Allows disabling memoization at runtime to reduce memory usage for large inputs

• Unlimited Action Variable Support

  • Action code now supports unlimited $N positional variables (previously limited to $1-$20)
  • Uses regex-based substitution for flexibility

Fixed

• Grammar Validation

  • Implemented Grammar.validate() to detect undefined rule references
  • Recursively walks all expressions and reports first undefined reference with location

• Thread Safety in Whitespace Skipping

  • Moved skippingWhitespace flag from PegEngine (per-instance) to ParsingContext (per-parse)
  • Fixes potential race conditions when reusing parser instances across threads

• Packrat Cache Key Collision Risk

  • Changed cache key from hashCode() to unique sequential IDs
  • Eliminates theoretical collision bugs with different rule names having same hash

Changed

• Builder API Naming Standardized

  • PegParser.Builder methods renamed: withPackrat() → packrat(), withTrivia() → trivia(), withErrorRecovery() → recovery()
  • Removed duplicate ParserConfig.Builder (unused)

• Code Simplification

  • Consolidated 3 duplicate expression parsing switch statements into unified parseExpressionWithMode()
  • Extracted buildParseError() helper to eliminate duplicate error message construction
  • Removed unused SemanticValues.choice, SourceLocation.advanceColumn()/advanceLine() methods
  • ~120 lines of duplicate code eliminated

• Test count: 268 → 271

• Updated pragmatica-lite dependency: 0.8.4 → 0.9.0

---

📦 Maven Central
```xml

org.pragmatica-lite
peglib
0.1.6

```