Release Notes v0.4.1

Major Features

Regex Support in Search

Fixes #11: OR search (regex pipe operator) not working in search_code_advanced

Added comprehensive regex support to the search_code_advanced tool with the new regex parameter:

# OR searches now work!
search_code_advanced("ERROR|WARN", regex=True)      # Returns OR matches

# Other regex patterns
search_code_advanced("function|class", regex=True)  # OR search
search_code_advanced("^import", regex=True)         # Start of line
search_code_advanced("Error$", regex=True)          # End of line
search_code_advanced("[A-Z]+", regex=True)          # Character classes

Key Benefits:

• OR searches: Use | to search for multiple patterns simultaneously
• Pattern matching: Full regex support with anchors, character classes, etc.
• Safety: Built-in ReDoS attack protection
• Backward compatible: Default behavior unchanged (literal search)

Bug Fixes

Fixed ag (Silver Searcher) File Pattern Handling

Resolved issue where glob patterns like *.py were incorrectly passed to ag's -G parameter, causing incorrect file matching behavior.

Improvements:

• *.py now correctly converts to \.py$ for ag's regex engine
• test_*.js now correctly converts to ^test_.*\.js$
• src/*.css now correctly converts to ^src/.*\.css$
• All search tools (ugrep, ripgrep, ag, grep) now handle file patterns consistently

Technical Changes

New Features

• regex parameter: Explicit control over regex vs literal search modes
• Pattern safety validation: is_safe_regex_pattern() function with ReDoS protection
• Automatic glob-to-regex conversion: Smart pattern conversion for ag search strategy
• Enhanced file pattern support: Consistent behavior across all search backends

Security Enhancements

• Protection against ReDoS (Regular Expression Denial of Service) attacks
• Validates regex patterns before execution
• Rejects potentially dangerous patterns like (.+)+, (.*)*
• Only allows safe regex metacharacters: |, (, ), [, ], ^, $

Code Improvements

• Renamed create_safe_fuzzy_pattern() to create_word_boundary_pattern() for clarity
• Improved parameter documentation across all search strategies
• Enhanced error handling and validation

Usage Examples

Basic Regex Search

# Default behavior (literal search) - unchanged
search_code_advanced("ERROR|WARN")  # Searches for literal "ERROR|WARN"

# New regex support
search_code_advanced("ERROR|WARN", regex=True)  # OR search

File Pattern Examples

# All these now work consistently across all search tools
search_code_advanced("def main", file_pattern="*.py")       # Python files
search_code_advanced("function", file_pattern="test_*.js")  # Test JS files
search_code_advanced("import", file_pattern="src/*.ts")     # TypeScript in src/

Advanced Regex Patterns

search_code_advanced("class [A-Z][a-zA-Z0-9]*", regex=True)  # Class names
search_code_advanced("(TODO|FIXME|HACK)", regex=True)        # Code comments
search_code_advanced("^\\s*def\\s+", regex=True)            # Function definitions

Security Notes

• All regex patterns are validated for safety before execution
• Dangerous patterns that could cause ReDoS attacks are automatically rejected
• Pattern validation ensures server stability and performance
• Safe regex features are clearly documented

Backward Compatibility

This release maintains 100% backward compatibility:

• Default search behavior remains unchanged (literal string search)
• Existing code will continue to work without any modifications
• New regex parameter is optional with False as default
• File pattern handling improved transparently without API changes

Migration Guide

To use regex features:

# Old way (still works)
search_code_advanced("simple text")

# New regex capabilities
search_code_advanced("pattern1|pattern2", regex=True)  # OR search
search_code_advanced("^start", regex=True)             # Line start
search_code_advanced("end$", regex=True)               # Line end

No migration needed for file patterns:

File pattern handling is automatically improved - existing code works better without changes.

Testing

This release has been thoroughly tested with:

• Basic search functionality validation
• OR pattern matching verification across all search backends
• File pattern conversion for ag, ripgrep, ugrep, and grep
• Safety checks for dangerous regex patterns
• Backward compatibility confirmation with existing codebases
• Performance testing with various pattern complexities

Related Issues

• Fixed #11: OR search (regex pipe operator) not working in search_code_advanced

Version Updates

• Bumped version to 0.4.1 reflecting the major regex feature addition
• Updated version strings in pyproject.toml and src/code_index_mcp/__init__.py

---

Full Changelog: v0.3.1...v0.4.1