Fix handling of empty lists and malformed PDF dictionary values

[pbottine]

This fixes issue #12 where certain malformed PDFs would cause "List index out of range" errors during parsing.

Changes to PDFList.load():

• Handle empty lists by returning zero-length wrapper at offset 0
• Filter items to only those with offset information before calculating bounds
• Log warnings when items lack position data instead of using incorrect defaults
• Change from @staticmethod to @classmethod for better conventions
• Add comprehensive docstring explaining edge case behavior

Changes to parse_object() dictionary handling:

• Fix logic ordering: check isinstance(value, list) before checking emptiness This prevents skipping falsy but valid values like 0, False, or empty strings
• Gracefully skip empty lists in dictionaries (log debug message)
• Catch ValueError from PDFList.load() for truly malformed lists
• Log warnings instead of raising exceptions for unexpected values
• Continue parsing to extract maximum data from malformed PDFs
• Update dictionary to keep it self-consistent after wrapping lists

Testing:

• Added comprehensive unit tests in tests/test_pdf.py covering:
  • Empty lists
  • Lists with/without offset information
  • Mixed offset scenarios
  • Empty lists in dictionaries
  • Malformed list values
  • Unexpected dictionary values
  • Preservation of falsy but valid values (0, False)
• All 8 new tests pass
• All 23 existing unit tests continue to pass
• PDF parsing verified with testdata/javascript.pdf

This builds on the approach from PR #3426 by @mrscottyrose with corrections to logic ordering, offset calculation, and comprehensive test coverage.

Fixes #12

cc: @smoelius

🤖 Generated with Claude Code