On Nov 1, 2011, at 12:52 AM, Keith OHara wrote: >> This fixes the SpanBar regression ID'd by Phil. > > Well, not exactly. The failing regression test said: > texidoc = "SpanBars participate in the horizontal collision system; > the accidentals should not collide with the bar lines." > > This patch implements issue 1955, keeping accidentals, fingerings, etc., > from folding above/below bar lines at all. Thus this patch moves the > accidental away from the extended bar line, whether there is a span bar > or not (e.g., whether in a PianoStaff or ChoirStaff). > > So if we take this patch, we should probably change the regression test, > as well, because the SpanBars will not actually be participating in the > horizontal collision system. > (e.g., if somebody does \remove "Bar_engraver", the accidental collides > with the span bar again.) >
Ah, I see - yes, you're right - will change in a new patch in a couple days. The problem with SpanBars having Y-extents is that they can never have pure Y extents, as it is not possible to figure out how high they'll be before line breaks. One solution would be to create a BarStub grob that works like SpanBarStub but in Staves. Give it all the same properties as the BarLine save the stencil. Then, have the extra-spacing-height callback apply to BarStub and not BarLine. But this seems a bit excessive - where are concrete cases that involve people removing the Bar_line_engraver and, in these cases, can they accomplish the same thing by setting #'transparent = ##t ? Cheers, MS _______________________________________________ lilypond-devel mailing list lilypond-devel@gnu.org https://lists.gnu.org/mailman/listinfo/lilypond-devel
