- Add tolerantMode for generic Nim code with < > brackets
- Fix JS ?. and ?? operator tokenization
- Add Python 3.10 match/case keywords, underscore prefix support
- Fix f-string parsing operator precedence bug
- Fix duplicate foundClosing assignments in JS/PHP tokenizers
- Treat unknown flavors as info instead of error
- Add comprehensive test coverage for all fixes
Expand nimcheck with validators and tokenizers for C, C++, C#, Go, Rust,
Ruby, CSS, SQL, Markdown, Dockerfile, Makefile, Kotlin, Lua, Swift,
TypeScript, and XML. Register all flavors in the validator factory and
improve auto-detection scoring for the new languages.
Fix infinite tokenizer loops that caused OOM kills: closeBracket now
advances position, finishTokenizeStep guards stalled tokenization, and
Jinja/JS tokenizers no longer double-advance on brackets.
Fix block-balance false positives in Lua (for/do) and Ruby (postfix
unless), SQL trailing-comma detection across whitespace, and Makefile
tab literals in test fixtures.
Establish the Validatrix source code validation framework in Nim, including core detection, tokenization, and validation infrastructure for 10 languages (bash, HTML, JavaScript, Jinja, JSON, Nim, PHP, Python, TOML, YAML). The initial commit introduces the main validatrix.nim entry point with public API (validateSource, validateFile, inspectSource), a comprehensive test suite with per-language test files, build configuration via Makefile and .nimble, detailed LESSONS_LEARNED.md documenting 14 bug classes found during development, and a .gitignore excluding compiled binaries, logs, and build artifacts. The detector module provides extension-to-flavor mapping for 40+ file extensions and content-based language detection, while the debug module supports conditional compilation with -d:validatrixDebug for detailed tokenization logging.
