Release v0.30.17 - "Tactical Tiger"

Release Date: December 03, 2025

Features

• Add Rust zero_angle for long-range zeroing (1000+ yards)
• Replace tests/generated/ with new AI-generated tests
• Add SAAMI spec sheet export for wildcat cartridges (MBA-98) MBA-98
• Add dead code detection using pgvector semantic search
• Add duplicate code detection to semantic search client
• Add multi-project code indexer with Rust support
• Improve exclusion pattern matching with fnmatch
• Add CLI args to scanner for local testing
• Integrate pgvector semantic search into code scanner
• Add incremental checkpoint saving for test generation
• Add rejection tracking to prevent regenerating failed tests
• Add multi-stage quality gate for AI test generation
• Integrate semantic search for finding similar tests
• Add HTTP search server for test orchestrator integration
• Add standalone indexer script for MCP code search
• Add test generator system with git stash workflow and server sync docs
• Add daily/weekly scanner modules with JIRA sync and email reporting
• Add AI code scanner agent with vLLM integration
• Replace Amazon PA-API with Canopy API for product search
• Add recoil calculation API endpoint

Bug Fixes

• Skip tests using legacy Python solver with known zero_angle issues
• Add execute permissions to code scanner cron scripts
• Guard against division by zero in form_factor.py
• Guard against division by zero in 4DOF similarity calculation
• Add division-by-zero guards in bullet_analyzer.py (MBA-229) MBA-229
• Raise exception when zero_angle fails instead of returning 0.0
• Use NaN instead of -999 for trajectory failures in zero_angle
• Create individual JIRA tickets instead of summary tickets
• Handle string type for avg_complexity in semantic search context formatting
• Add conditional imports to all mcp-server modules for direct execution
• Use conditional imports for multi_indexer to support direct execution
• Improve JSON output reliability for code scanner
• Scanner path handling for native vs Docker execution
• Handle absolute paths in scanner tools read_file
• Write test files to generated/ dir during PR creation
• Correct R² assertion in drag synthesis test
• Add disk cleanup step to GitHub Actions workflow
• Re-enable test_bc_segments_interpolation test
• Restore proper Poetry config for full dependency resolution
• Add functions-framework to poetry.lock, bump to v0.30.16
• Add gunicorn dependency for Cloud Run deployment
• Cache test file contents before git stash to prevent loss
• Update .gitignore to allow test files in tests/ directory
• Make run_indexer.py self-contained without relative imports
• Repair context_builder.py syntax and improve file I/O handling in prompts
• Skip test_bc_segments_interpolation due to pre-existing Rust backend issue
• Resolve torch and google-cloud-secret-manager dependency issues
• Implement rifle weight-based recoil multiplier
• Refactor duplicate code and remove duplicate import
• Update version references to 0.30.11
• Move Canopy API key from hardcoded value to environment variable
• Register Amazon affiliate products API endpoint (MBA-164) MBA-164

Documentation

• Fix coordinate system documentation to reflect standard ballistics convention

Chores

• Bump version to 0.30.17
• Replace vLLM references with llama.cpp across code-scanner
• Update submodule, release notes, and llama.cpp migration
• Bump version to 0.30.14 and update ballistics-engine to v0.13.7
• Switch to Ollama with gpt-oss:120b for test generation
• Remove '15-20x performance' marketing text from health endpoint and docs
• Bump version to 0.30.13
• Update ballistics_rust to use ballistics-engine v0.13.3
• Bump version to 0.30.12
• Bump version to 0.30.11
• Bump version to 0.30.7 for MBA-164 recoil calculator deployment MBA-164

Other

• Add generated ML tests and update Ollama documentation (#7)
• Add weekly download report Cloud Function
• hotfix: Remove duplicate products_bp registration causing startup crash (MBA-164)

← All Releases