#1703 Upward slur placed too high (in very specific situations)

[Anonymous]

Originally created by: *anonymous

Originally created by: brownian.box@gmail.com

Reported by Mark Polesky,
http://lists.gnu.org/archive/html/bug-lilypond/2011-06/msg00124.html

(2.15.2 does the same)

%------------------8<------------------------
I honestly tried to make this example file as tiny as possible,
but the combination of elements needed to trigger the bug seems
curiously specific.  For example, if I remove any one of these
elements, the problem disappears:

#(set-default-paper-size "letter")
#(set-global-staff-size 17)
left-margin = 0.75\in  --and--  right-margin = 0.75\in
\key bf \major
\break

Stranger still, if I change staff-size to 15, 16, 18, 19, or 20,
the problem goes away, and if I change both margins to 0.5\in or
1\in, the problem also goes away.

So, what's going on?  This is bizarre...

Thanks.
- Mark

\version "2.14.0"

#(set-default-paper-size "letter")
#(set-global-staff-size 17)
\paper {
  left-margin = 0.75\in
  right-margin = 0.75\in
}

\relative bes' {
  \key bes \major
  <<
    { bes='4 f8 f16( f-.) bes4 f8 f16( f-.) | }
    \\
    { d='8 d4 d16 r d8 d4 d16 r | }
    \\
    { d='8 d16 f s4 d8 d16 f s4 | }
  >>
  \break
  bes4 r2. |
}

[Google Importer]

Originally posted by: simon.al...@mail.de

(No comment was entered for this change.)

Labels: -Priority-Medium

[Simon Albrecht]

I found another instance of this bug, with a slightly smaller number of conditions:

\version "2.19.32"
{
  \voiceOne
  \clef bass
  b8( a d' bis)
}

Still, using another clef (from baritone to french) or omitting any of the notes or the accidental (omitting, not hiding) makes the problem disappear. A somewhat elusive bug, and apparently quite rarely observed.

[Simon Albrecht]

To drive the absurdity even further: I tested Mark’s initial example with 2.16.2, 2.18.2 and 2.19.32. Neither shows the bug.

[Simon Albrecht]

Mark’s initial example does trigger the bug with 2.19.83 (Ubuntu 16, 64-bit), while my smaller example does not…
A couple similar issues were found to be caused by rounding issues. Maybe this one too?