Ok, everybody will love me for my late thinking again. But here we go: we have two cases of footnotes: one that will attach itself to whatever happens at a given point of time, being an independent event. Those footnotes, having a natural duration of 0, come (unless they are in simultaneous music) _before_ the things they attach themselves to, possibly restricted by Grob type so that they don't grab too much.
Then you can use footnotes as _articulations_. Then they come behind what they refer to, separated by - ^ _. I think that's less intuitive than necessary. I propose we _always_ let them try to attach themselves as postevents/articulations, by defining the footnote command with define-event-function. Then the - before the footnote call is optional, like it is with things like \p. I think that makes more sense to explain. Yes, it is a nuisance to rewrite the documentation _again_, but everything else is really not all that helpful. Strictly speaking, this would also require a convert-ly rule, but I can't think of a good one. http://codereview.appspot.com/6137050/ _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
