Changelog¶

Version 0.7.0¶

Overview¶

This release adds JSON intermediate-representation generation, configurable boolean tokens, and a runtime parser API. It also redesigns the @Shared pattern, improves comment configuration, modernizes generated-source formatting, and publishes the documentation site.

New Features¶

• IR generator module - Added yajco-ir-generator-module, which exports annotated language models as JSON intermediate representation for the editor and IDE plugin generator currently in development (#51).
• The downstream generator produces TextMate-style highlighting rules, Tree-sitter grammars, and an LSP server.
• Added the package-level @Language annotation for language metadata such as name, version, and file extensions.
• Configurable boolean tokens - Added @BooleanValue to define custom tokens for true and false, including multiple tokens for each value (#59).
• Default boolean tokens remain true and false.
• Parser factory - Added ParserFactory.getParser(mainNodeType) for discovering and instantiating generated parsers without importing their generated classes directly (#63).
• Generated parsers are annotated with @ParserFor to identify their main node type.

Improvements¶

• Comment configuration - @Skip now has explicit lineComment and blockComment properties, with validation for block-comment delimiters (#58).
• Source formatting - Replaced google-java-format with the JavaParser pretty-printer, removing the reliance on JDK-internal APIs and supporting newer JDKs more reliably (#61).
• Annotation processor - Refactored language-model construction, concept registration, type resolution, and pattern mapping into dedicated components, with expanded unit-test coverage (#52).
• Documentation - Migrated documentation from the wiki to the GitHub Pages site and added a documentation publishing workflow.
• Dependencies - Updated jackson-databind in the IR generator to 2.22.1 (#62).

Breaking Changes¶

• @Shared pattern redesign - @Shared is now applied directly to the shared element rather than the collection that uses it (#64).
• The annotation no longer has a value property; configure only its separator on the shared element.
• The specialized shared-list type and its parser support were removed.
• Runtime API relocation - Parser, ParseException, lexer classes, and ReferenceResolver moved to the new yajco-runtime artifact.
• Parser and ParseException moved to the runtime module as part of the parser-factory API (#63).
• Update dependencies and imports to use yajco-runtime and the yajco.parser package.
• Comment syntax migration - Configure comments through @Parser(skips = ...); comment syntax is no longer specified in @Language (#58).
• Migrate deprecated @Skip.start and @Skip.end properties to lineComment and blockComment.

Changes from Version 0.5.9 to 0.6.0¶

Overview¶

This release includes 149 non-merge commits and represents a significant evolution of YAJCo, focusing on language pattern expansion, Java modernization, improved multi-parser capabilities, and better type system support. It also adds support for ANTLR4 parser generator.

Major Changes¶

New Language Patterns¶

• Flag Pattern - Initial implementation for Beaver and Antlr4 (d727693)
• Allows boolean flags in language syntax

• Represented by the @Flag annotation

• Mixed Repetition Pattern - Initial implementation (4c1ebb0)

• Supports complex repetition patterns in grammar rules

• Unordered Group Pattern - Represented by @UnorderedParameters annotation

• Allows parameters to appear in any order in the syntax

• Quoted String Pattern - Represented by @StringToken annotation

• Handles string literals with proper quoting

• Shared Pattern - Represented by @Shared annotation

• Enables sharing of language elements across different contexts

• Whitespace and Comment Patterns - Implemented as parameters to existing @Skip annotation (39c52ad)

• Simplified definition of whitespace and comment handling

Other New Features¶

• Automatic IDENTIFIER Token - Now automatically added and used for properties marked with @Identifier and @References (bb8ab1c)

• Reduces boilerplate in language definitions

• @Keyword - Added as an alias for @Before, can be placed on parameters (experimental) (4e02f01, d612759)

• More intuitive annotation name for keyword tokens

• @UniqueValues - Ensures uniqueness constraints on language elements

Type System Enhancements¶

• Optional Type Support - Added comprehensive support for java.util.Optional<T> throughout the codebase
• Visitor and Printer properly handle optional types
• No need to return Optional<T> from getter methods (method overloading resolves it)

Java & Build System Updates¶

• Upgraded to Java 11 - Annotation processor and annotations updated to support Java 11 (fe8272e, 973fda4)
• Code Formatter Change - Replaced Jalopy with google-java-format (c914dcf)

• Improved formatting of generated Java files

• Dependency Updates:

• Updated Velocity to version 2.2

• Updated Maven dependencies

• Security Fixes:

• Fixed xstream compatibility with latest Java
• Bumped junit from 4.12 to 4.13.1

New Parser Generator Backend: ANTLR4¶

• ANTLR4 Support Added - Complete implementation of ANTLR4 parser generator backend (8391ced and subsequent commits)
• Previously YAJCo supported Beaver and JavaCC backends
• ANTLR4 grammar generation with semantic actions
• Custom regex-matching lexer for unified lexical analysis
• Support for references, arrays, lists, sets, enums, and factory methods
• Improved error reporting for lexer and parser errors
• Note: Not all language patterns are supported yet (work in progress)

Experimental Xtext Support¶

• Xtext Integration - Experimental support for Xtext framework (e086621)
• Added integration of YAJCo with Xtext
• Xtext module moved into yajco-modules (2eb6d4d)
• Removed dependency of annotation processor on Xtext module (c9edc84)

Multi-Parser Support¶

• Multiple Parsers - Now allows multiple parsers in the same project (a5b9bcc)
• Can use different parser backends (Beaver, JavaCC, ANTLR4) simultaneously
• Common Parser Interface - All generated parsers now implement a common interface (085d91d)
• Added common ParseException (20d1925)
• Better interoperability between different parser implementations
• Service Provider Interface (SPI) - Generated parsers registered as service providers in each CompilerGenerator

License Change¶

• Changed from GNU LGPL v3 to MIT License (f4bc11a)
• More permissive licensing for broader adoption

Bug Fixes & Improvements¶

Parser & Grammar Generation¶

• Fixed @Identifier recognition in base classes (7af4908)
• Fixed @FactoryMethod causing exceptions in ReferenceResolver (49e234e)
• Fixed unordered parameters processing with @After tokens (7950b7e)
• Fixed missing separator in sequences with finite @Range.maxOccurs (5fd0279)
• More reliable generation of actions for UnorderedParameters grammar rules (e203576)

Code Generation¶

• Fixed Printer optional unboxing issues (829bf51, 0997b24)
• Improved indentation in generated Visitor (6b5634b)
• Fixed compile errors in generated visitor by unboxing Optional values (97acc60)
• Optimized imports in Utilities.java (6b76997)

Reference Resolution¶

• Fixed ReferenceResolver hash issues (bded7b3)
• Changed HashMap to IdentityHashMap to handle objects with custom hashCode()

• Use IdentityHashSet in generated Visitor to ensure all objects are visited

• Resolve identifier fields from parent classes recursively (ff68f45)

ANTLR4 Module¶

• Removed SecurityManager hack for running Antlr4 (2284628)
• Unified lexical analysis across parser generators
• Multiple bug fixes and improvements for ANTLR4 grammar generation

Project Infrastructure¶

• CI/CD Migration - Migrated from Travis CI to GitHub Actions (ea984bb)
• Build Scripts - Created Maven workflow for GitHub Actions
• Project Organization:
• Reorganized Xtext module into yajco-modules
• Simplified pom.xml files across modules
• Updated project metadata and URLs
• Cleaned up examples and removed unnecessary documentation files (45a02e8)

Refactoring & Code Quality¶

• Symbol Management:
• Symbol is now Cloneable (483bdf4)

• Changed withVarName() to return a clone (5b23f26)

• Type Safety:

• Return patterns cast to requested type (0997b24)

• Improved generic type handling (964eb9a)

• Code Cleanup:

• Moved types not used in language model to yajco-grammar-module (896804e)
• Consistent indentation using spaces throughout project (b2cf7b8)
• Removed backwards-compatibility code and unnecessary abstractions

Breaking Changes¶

• Minimum Java version is now 11
• License changed from LGPL v3 to MIT
• Some internal APIs refactored for better type safety

Contributors¶

Special thanks to all contributors who made this release possible, including pull requests from:

• Miroslav Remák - Complete ANTLR4 backend implementation, visitor enhancements, and numerous bug fixes
• Michal Fecenko - Multiple feature additions including new annotations and patterns
• Ján Halama - Integration with Xtext
• Adrián Szegedi - Bug fixes and refactoring improvements