Menu
▾
▴
#1800 Segfault compiling LSR 346
Originally created by: *anonymous
Originally created by: ColinPKC...@gmail.com
Originally owned by: k-ohara5...@oco.net
User Tim Reeves (Tim.Reeves@tokamerica.com) reports:
I tried to compile the "Creating music with Scheme (music box)" snippet (
http://lsr.dsi.unimi.it/LSR/Item?id=346 ) and when I do, Lilypond crashes
with the following output:
> > %lilypond %args "C:\Documents and Settings\timr\Desktop\Bach Prelude in
C major.ly"
Processing `C:/Documents and Settings/timr/Desktop/Bach Prelude in C
major.ly'
Parsing...
Interpreting music... [8][16][24][32]
Processing time: 4 seconds
So it seems to get through compiling except it fails to generate the ps or
pdf.
Any idea why?
I should add: I didn't know what version to tell Lilypond the snippet was,
so I assumed the previous stable version, 2.12.3. Is that correct?
(LP 2.15.1 on Windows XP SP3)
**************
On my machine (Ubuntu 11.04 x86-64), it compiled w/no errors under 2.12 and segfaults for both 2.14.2 D/L today and 2.15.3, although James L reports success with 2.14.1
Discussion
Originally posted by: k-ohara5...@oco.net
Tinyfying the example, the essentials are
cross-staff automatic beams, and
implicitly-created Staff that is abandoned when we change away.
<http://codereview.appspot.com/4564041/diff/5001/lily/auto-beam-engraver.cc>
<<
\new Voice {
c8 c c
\change Staff = "down"
c
}
\context Staff = "down" s2
>>
Originally posted by: k-ohara5...@oco.net
The Beam now remembers which Staff it started in, so that it can be engraved by that Staff. In the troublesome example, the starting Staff (implicitly created) was abandoned and thinks it can disappear before the beam is finished.
The automatic staff changer solves a similar problem, keeping two Staves alive while a voice hops from one to the other. We can adopt this solution to fix this bug
<http://codereview.appspot.com/4830064>. Not a polished patch yet, but a start, so we don't have to revert the improvement that caused this bug.
Owner: k-ohara5...@oco.net
Status: Started
Originally posted by: k-ohara5...@oco.net
The patch at <http://codereview.appspot.com/4830064> is done.
No regression tests change
(except that the beams in beam-skip.ly got slightly shorter for me, and stayed shorter when I went back to master).
Labels: Patch-new
Originally posted by: pkx1...@gmail.com
Passes make and reg tests
Labels: -Patch-new Patch-review
Originally posted by: k-ohara5...@oco.net
Fixed with the pair of commits 7aa537da5 9ff88d127
Labels: -Patch-review fixed_2_15_9
Status: Fixed
Originally posted by: brownian.box@gmail.com
2.15.9 gives three-staff-system for that snippet. Last three measures are "strange". Is this ok? Should it become another issue?
Originally posted by: k-ohara5...@oco.net
This bug was the segmentation fault. Since you were able to get output, then the bug is fixed.
The three-staff system, and resulting strange staff-change in the final three measures, are due to some misplaced { } in th15 LSR snippet. I've been intending
to put the attached correction on the LSR.
Status: Verified
Originally posted by: n.putt...@gmail.com
> I've been intending to put the attached correction on the LSR.
Thanks, I've updated the snippet.
Originally posted by: k-ohara5...@oco.net
Issue 2562 has been merged into this issue.