Given that the error is "Illegal instruction", and reproducibly happens
on x86-bm-01 and not the other machines we've tried, could it be
something assuming CPU features more recent than the amd64 baseline?
If so, it's not obvious where: scilab does contain some C/C++ code (as
well as Java), but there aren't any asm() or __builtin_ia32* calls in
it, or any -march options in the build log. Maybe it's in a dependency,
not scilab itself?
There is an open "illegal instruction crash in Xcos" bug upstream, which
may or may not be related:
Alexis Murzeau wrote:
will such bug-fix upstream release be accepted in
buster, now that we are in full freeze ?
Generally not: Debian prefers not to fix minor bugs during freeze
because this might accidentally create major ones.