On May 30, 2011, at 12:20 AM, k-ohara5...@oco.net wrote:
> This fails the "sniff test" because it removes some careful-looking code
> no explanation as to how the job of the old code is being done.
>
> What happens if the line on the next page is does not, in fact, have the
> same vertical reference point?
>
> \relative c''' { \override Staff . Glissando #'breakable = ##t
> { s3 a2 \glissando \break s1 \break s1 \break s1 \pageBreak \break s2
> c,2\trill\fermata^\ff }
> }
>
> I suppose a misaligned glissando is better than no glissando, but the
> old code made my the voice-change lines point to the right places:
> \paper {ragged-right = ##t }
> \new PianoStaff <<
> \new Staff = "one" {
> \showStaffSwitch
> c'1 \break
> \change Staff = "two"
> b2 a
> }
> \new Staff = "two" {
> \clef bass
> g'1^\trill^\f s1
> }
>
The problem doesn't come from the page break but from alignment with common Ys.
Line 242ish of line-spanner.cc is:
do
span_points[d][Y_AXIS] += common_y[d]->relative_coordinate (my_common_y,
Y_AXIS);
while (flip (&d) != LEFT);
The value added on will be different depending on what's on the staves. Try:
\relative c''' { \override Staff . Glissando #'breakable = ##t
s2 a2 \glissando \break s1 \break s1 \break s1 \break s2
c,2
}
versus
\relative c''' { \override Staff . Glissando #'breakable = ##t
s2 a2 \glissando \break s1 \break s1 \break s1 \break s2
c,2 \trill\fermata^\ff
}
I'm flying to Ireland today & will attempt to crunch this during the flight.
Cheers,
MS
_______________________________________________
lilypond-devel mailing list
lilypond-devel@gnu.org
https://lists.gnu.org/mailman/listinfo/lilypond-devel
