bible_io_references

Added

• Conventional package:bible_io_references/bible_io_references.dart entrypoint
• Non-throwing tryParse and typed ParseResult APIs
• Typed parse error classifications while preserving legacy string codes
• Value equality, checked copying, comparison, and JSON round-tripping
• Localized long/short reference formatting and supported-language introspection
• Configurable ReferenceParser aliases, language priority, and ambiguity rejection
• Detected-language, selected-book, custom-alias, and alternate-match metadata
• Immutable Passage values and PassageParser support for whole books, chapters, chapter ranges, verse lists, semicolon-separated sequences, and configurable single-chapter shorthand
• Localized passage formatting plus OSIS and USFM serialization for every passage variant
• Reference-oriented Unicode normalization for compatibility punctuation, fullwidth forms, Arabic-Indic digits, combining/vocalization marks, RTL controls, Unicode spacing, and CJK book/chapter adjacency
• UTF-16 source-span mapping through ReferenceInputNormalization
• Parser-driven ReferenceExtractor matches with exact original offsets, one-pass replacement, generic linkification, and escaped Markdown links
• Complete OSIS and USFM book/reference identifiers with reverse parsing
• Published CLI executable with rich-passage language/text/JSON/OSIS/USFM output and a package example
• UTF-8 CLI batch processing from stdin or files with JSON Lines and exit codes
• Linux and Windows CI for formatting, analysis, and the full test suite

Changed

• Reference validation remains edition-neutral: chapter and verse numbers use broad sanity limits, while canon membership and edition-specific numbering are intentionally not enforced

Fixed

• Cross-book ranges are now recognized by Reference.parse
• Empty input has a dedicated empty_reference diagnostic
• Horizontal-bar range separators are supported
• Pathological chapter and verse numbers are rejected by broad sanity limits
• Previously undiscovered property, integration, concurrency, and benchmark tests now run by default
• Ambiguous two-character Arabic aliases no longer silently select different books
• README license badge and wording now match the repository's AGPL-3.0 license

🎉 Initial Release

Core Features:

• Parse single verse references (VerseRef)
• Parse verse ranges (VerseRangeRef)
• Support for flexible formatting (colons, dots, various dash types)
• Multi-language support with intelligent auto-detection
• Comprehensive error handling with detailed diagnostics

API Design:

• Immutable, thread-safe data structures
• Sealed class hierarchy for type safety
• Static parsing methods on classes
• Backward-compatible legacy functions

Languages Supported:

• English (complete)
• Spanish (complete)
• French (complete)
• German (complete)
• Portuguese (complete)
• Russian (complete)
• Korean (complete)
• Chinese (complete)
• Hebrew (complete)
• Hindi (complete)
• Indonesian (complete)
• Tagalog (complete)

Quality Assurance:

• 2400+ comprehensive tests
• Property-based testing
• Performance benchmarks
• Language data quality validation
• Concurrent operation testing

Performance:

• Sub-millisecond parsing performance
• Zero external dependencies
• Memory efficient implementation
• Thread-safe operations