Re: Crash in 2.25.4

Le mercredi 17 mai 2023 à 20:02 +0000, Gwyn Ciesla via bug-lilypond a écrit :
> Attaching reproducer.
> 
> From Fedora bug:
> 
> When I attempt to build a score using a PianoStaff or GrandStaff (which contains more than one sub-staff), lilypond crashes.
> 
> Reproducible: Always
> 
> Steps to Reproduce:
> 1. Attempt to build a minimal reproducer (which I will attach) with lilypond-2.25.4
> Actual Results:  
> Slightly varying modes of crashing.
> 
> #1) Without dumping core:
> 
> --- 8< ---
> nils@makake:~/music> lilypond minimal.ly 
> 
> GNU LilyPond 2.25.4 (running Guile 2.2)
> Processing `minimal.ly'
> Parsing...
> Interpreting music...
> Preprocessing graphical objects...
> Finding the ideal number of pages...
> Fitting music on 1 page...
> Drawing systems...
> /usr/share/lilypond/2.25.4/ly/init.ly:64:2: error: Guile signaled an error for the expression beginning here
> #
>  (let ((book-handler (if (defined? 'default-toplevel-book-handler)
> In procedure scm_hash_fn_create_handle_x: Error while printing exception.
> --- >8 ---
> 
> #2) With dumping core:
> 
> --- 8< ---
> nils@makake:~/music> lilypond minimal.ly 
> 
> GNU LilyPond 2.25.4 (running Guile 2.2)
> Processing `minimal.ly'
> Parsing...
> Interpreting music...
> Preprocessing graphical objects...
> Finding the ideal number of pages...
> Fitting music on 1 page...
> Drawing systems...
> /usr/share/lilypond/2.25.4/ly/init.ly:64:2: error: Guile signaled an error for the expression beginning here
> #
>  (let ((book-handler (if (defined? 'default-toplevel-book-handler)
> In procedure scm_hash_fn_create_handle_x: Wrong type argument in position 1 (expecting hash-table): Segmentation fault (core dumped)
> --- >8 ---
> 
> Expected Results:  
> It builds a PDF from the score.
> 
> - I get the same behavior on Fedora 38 and Rawhide.
> - Using the lilypond-2.25.3 Fedora package _or_ the 2.25.4 binary from upstream doesn’t exhibit the issue.


Thank you for forwarding the report. I can reproduce it on my Fedora system. Apparently we're crashing while looking up the font for the system start brace, but I don't understand why yet. It's quite
strange that this doesn't reproduce with the official binaries or local builds. Judging from the backtraces, there might be a memory corruption somewhere, but I have to find out.

-----BEGIN PGP SIGNATURE-----

iHUEABYKAB0WIQSZ7TKxnKGyBvBjzBmj8PYLiTOX/gUCZGU7pwAKCRCj8PYLiTOX
/sSBAQDa2KULyL/Hy2x+wVH66n0+5xqmhXfCVi8QHN2Hj+OOUgEA45PJN7H4pjYY
3goYFFEDqPZw+2qp1Bqu/fAL1mI1WQE=
=aj1l
-----END PGP SIGNATURE-----