In `src/banjo/midi_writer.py:generate()`, the loop builds a 4-tuple `(spec, parsed, chord, voiced)` into `resolved_chords`, but `chord` is never read after construction (the metadata loop reads only `spec`, `parsed`, `voiced`; the events loop also doesn't touch it). Pre-dates the voice-leading work but became visible during that review.
Suggested: drop `chord` from the tuple — change to `(spec, parsed, voiced)` and update the two unpacking sites. ~5-line cleanup.
Cosmetic; doesn't affect behavior.
In `src/banjo/midi_writer.py:generate()`, the loop builds a 4-tuple `(spec, parsed, chord, voiced)` into `resolved_chords`, but `chord` is never read after construction (the metadata loop reads only `spec`, `parsed`, `voiced`; the events loop also doesn't touch it). Pre-dates the voice-leading work but became visible during that review.
Suggested: drop `chord` from the tuple — change to `(spec, parsed, voiced)` and update the two unpacking sites. ~5-line cleanup.
Cosmetic; doesn't affect behavior.