Control: tags -1 + pending Hi Stuart and Adrian,
Thank you for your reports! On Sat, Sep 5, 2020 at 10:03 AM Adrian Bunk <b...@debian.org> wrote: > > On Mon, Jul 27, 2020 at 09:09:45PM +1000, Stuart Prescott wrote: > >... > > guile segfaults as soon as it is started (with or without docs on stdin). > > The backtrace is as follows in case the problem jumps out. > > > > /build/lilypond-jv5Dao/lilypond-2.20.0/guile18/libguile/.libs$ > > LD_LIBRARY_PATH=. gdb ./guile > > Reading symbols from ./guile... > > (gdb) r > > Starting program: > > /build/lilypond-jv5Dao/lilypond-2.20.0/guile18/libguile/.libs/guile > > > > Program received signal SIGSEGV, Segmentation fault. > > scm_mark_locations (x=0x7fffffffe948, n=18446744073709551562) at > > gc-mark.c:435 > > 435 SCM obj = * (SCM *) &x[m]; > >... > > lilypond started to FTBFS when libgc 8 entered unstable. It's probably related to a change in gcc-10 too. I was able to compile guile-1.8 with gcc-9 to avoid FTBFS. It turns out the same bug got triggered with gcc-4 -O3 too, for the C code used for stack growth direction detection. According to Paul Eggert who fixed the bug in Autoconf, they had to outsmart gcc's optimization. ;-) Here is the changelog entry for the pending fix at https://salsa.debian.org/debian/lilypond/-/commit/f8b3bfe : Patch Guile-1.8 to the latest branch_release-1-8 branch found in Guile git repository on Savannah, see patches 1801 to 1825. In particular, 1812-Fix-stack-growth-direction-check.patch, originally written for Autoconf by Paul Eggert for Autoconf and pulled into Guile-2.0 in 2011, fixed a build error with -O3 back then, and fixes FTBFS with gcc-10 nowadays, due to a bug in find_stack_direction() code in configure.in. See http://lists.gnu.org/archive/html/bug-autoconf/2011-06/msg00035.html for the whole story. Thanks to David Kastrup for pointing me in that direction, even though I only subconsciously understood his email at the time: * https://lists.gnu.org/archive/html/lilypond-devel/2020-06/msg00045.html Also to Jonas Hahnfeld for explaining the issue in a way that I could understand: * https://lists.gnu.org/archive/html/lilypond-devel/2020-06/msg00047.html My summary post on lilypond-devel: * https://lists.gnu.org/archive/html/lilypond-devel/2020-09/msg00041.html Cheers, Anthony
