A slur spanning multiple notes under one sung syllable (melisma) must interact correctly with the lyric/extender line, and Gradus's schema cannot currently represent this at all

When several notes are sung to one syllable, the notes are joined by a slur AND the lyric syllable is followed by an extender line (or the slur substitutes for it, per house style) spanning the same note range — the two marks must be authored consistently or the notation misleads a singer.

Gradus's JSX `MelodyNote` type has NO lyric field at all (absent from the interface, lines 50-66); the public API's `NoteObject` object form DOES have `lyric?: string` (types.ts:84), but shorthand-string notes ('C5/q') have no way to attach a lyric at all, and the public API has no slur field whatsoever (see slur-representation-missing-in-public-api). Net result: Gradus currently cannot represent a melisma correctly in EITHER surface — the JSX surface has slurs but no lyrics; the public-API object form has lyrics but no slurs, and its shorthand form has neither slur nor lyric.

Gould's vocal-music chapter (lyrics/melisma) specifies the slur-over-melisma + extender-line convention explicitly; Read's vocal notation chapter and Ross concur the two marks must be kept in sync.

This rule can be decided from pitches, durations and voices — no layout required. A validator can check it before anything is drawn.