v1.1.1 - Critical Bug Fixes & Security Enhancements (2025-02-02)

Fixed

• Critical: Pattern Matching Word Boundary Detection

  • Fixed false positive pattern matching causing incorrect element removal
  • Elements like <section class="section-heading"> were incorrectly treated as ads (contained "ad")
  • Implemented proper word boundary detection with separators: -, _, space, tab
  • Text extraction from affected pages increased by 1,273x (87 → 111,010 characters)

• Test Output Formatting

  • Fixed fmt.Printf misuse that caused format errors with % characters
  • Prevented %!f(MISSING), %!a(MISSING) errors in test output

• Cache Double-Check Locking Race Condition

  • Fixed potential race condition in cache Get method
  • Properly re-checks entry after acquiring write lock

• HTML Entity Parsing Logic

  • Simplified numeric entity validation (removed redundant parsing)
  • Eliminated unnecessary validation loops and goto statements

• URI Security Validation

  • Reordered checks to block dangerous protocols first (javascript:, vbscript:, file:)
  • Fixed potential bypass through leading/trailing whitespace
  • Corrected data URL character validation (was rejecting valid UTF-8)

Changed

• Code Quality
  • Simplified re-exported types and constants (25 → 14 lines)
  • Removed unused re-exports: Tokenizer, ParseOption, ParseWithOptions, etc.
  • Cleaned up redundant comments throughout codebase
  • Maintained 100% backward compatibility

Security

• Enhanced protocol validation order for safer URL handling
• Fixed data URL validation to properly handle base64-encoded content
• Corrected cache concurrency issues for thread-safe operation

Migration Notes

• Zero Breaking Changes - All existing API calls work without modification
• Tests: All existing tests pass successfully