Simon Albrecht <[email protected]> writes: > On 13.02.2017 18:35, Michael Rivers wrote: >> I guess it must be incredibly difficult to solve or it would >> have been solved by now. > > Very recently, David Kastrup wrote about having made like half a dozen > attempts that didn’t work out. And he’s definitely the most capable > programmer and the one with the most understanding of LilyPond’s > intestines that we have.
Oh, that doesn't mean that it's hard. It's just not obvious. I am pretty confident that transferring the "grace fixup" approach from sequential music to simultaneous music would likely do the trick in some manner. There are just several problems: a) grace fixup is precalculated in sequential music which is a mistake for several reasons and does not transfer well to simultaneous music. b) the consequences of stuff going wrong with grace fixups in simultaneous music are more dire than in sequential music. c) several callbacks for music start/length behave somewhat quirky in connection with grace music and one would need to straighten this out to avoid case b) kicking in with a vengeance. That's why I said that I don't consider this a good student project: it's hit or miss. Miss does not cause any progress, hit will likely be not all that spectacular with regard to the actual coding work getting done. It's not likely "incredibly difficult". Just elusive. I am pretty sure that whatever solution might eventually be done will be of the "so why was this so hard?" variety. -- David Kastrup _______________________________________________ lilypond-user mailing list [email protected] https://lists.gnu.org/mailman/listinfo/lilypond-user
