"m...@mikesolomon.org" <m...@mikesolomon.org> writes: > On 6 févr. 2013, at 00:17, David Kastrup <d...@gnu.org> wrote: > >> "m...@mikesolomon.org" <m...@mikesolomon.org> writes: >> >>> Find #3 was wrong. There should have been no suicides in this precise >>> case. LilyPond just does not know how to do spacing on empty >>> skylines. This makes sense, as it is difficult to estimate the >>> distance between something and nothing. So, I'll use my previous >>> solution of flattening the skyline to 0 with a nice comment. >> >> That sounds nonsensical. A skyline should be able to contain empty >> stretches or be completely empty; that should be "neutral", namely let >> any neighboring skyline take over. A skyline flattened to 0, however, >> will kill the neighboring skyline and replace it with its >> maximum/minimum. >> > > Good call. I think the problem is that the empty skyline is the > lowermost, so it impacts system-system spacing with the neighboring > system. Not sure how yet...
I suspect adding -infinity to the depth or something like that. Our representation of empty intervals does not combine perfectly with everything without being careful. -- David Kastrup _______________________________________________ bug-lilypond mailing list bug-lilypond@gnu.org https://lists.gnu.org/mailman/listinfo/bug-lilypond
