segfault with CVS version of lilypond

[Werner LEMBERG]

[CVS 2006-01-27 06:52]

Sigh.  After a few months I'm trying lilypond again -- very nice tie
output, BTW! -- and I immediately hit a segfault :-)  I know that
`GrandStaff' is not suitable for cross staff beaming, but...

Below is the smallest possible input file I was able to create,
together with the first few levels of a backtrace.


    Werner


======================================================================


\version "2.7.28"

rechts = \context Staff = "rechts" \relative c' {
  \clef treble
  f1 |
  f1 |
}


links = \context Staff = "links" \relative c {
  \clef bass
  f2 f4( f4 |
  f4) f8.
\change Staff = "rechts"
                         g16
\change Staff = "links"
                              r2 |
}


\score {
  << \context GrandStaff
     << \rechts
        \links
     >>
  >>
}

% EOF


======================================================================


#0  std::String_handle::copy (this=Variable "this" is not available.
) at string-handle.icc:45
#1  0x081f172d in std::String::operator+= (this=0x1bd, s=Cannot access memory 
at address 0xb
)
    at string-handle.icc:131
#2  0x0816f8d1 in Slur_score_state::get_best_curve (this=0xbfffa8c0)
    at parray.hh:62
#3  0x0817270f in Slur::calc_control_points (smob=0x40a644c8)
    at slur-scoring.cc:387
#4  0x40067313 in scm_dapply (proc=0x40696df8, arg1=Variable "arg1" is not 
available.
) at eval.c:3576
#5  0x400629e3 in scm_apply (proc=0x40696df8, arg1=Variable "arg1" is not 
available.
) at eval.c:3453
#6  0x40067a89 in scm_call_1 (proc=0x40696df8, arg1=0x40a644c8) at eval.c:3315
#7  0x080bde68 in Grob::try_callback (this=0x85db410, sym=0x40544610,
    proc=0x40696df8) at grob.hh:20
#8  0x080bdf93 in Grob::internal_get_property (this=0x85db410, sym=0x40544610)
    at grob-property.cc:104