Summary

hello_structure.pdf is still a severe gap and remains ignored.

Evidence

Run:
cargo test -p pdfplumber --test cross_validation -- --include-ignored --nocapture

Current result:

• hello_structure.pdf: chars 38.9%, words 44.4%

Ignored reason in test file already points to tagged PDF + TrueType handling gap.

Scope

• Investigate tagged PDF extraction path for this fixture
• Close TrueType/encoding mapping gap causing low char/word recovery

Acceptance Criteria

• Fixture reaches >=95% chars and >=95% words
• Convert from cross_validate_ignored! to asserting cross_validate!