Bug#923527: bliss: example segfaults in gmp

Bernhard Übelacker bernhardu at mailbox.org
Mon Mar 25 22:52:52 GMT 2019


Hello David Bremner,

>> Maybe something got lost in the input file,
>> so you could attach it to the email?
> ok

Thanks for the file, but unfortunately that makes no
difference - works fine in my test VM.


>> Is the backtrace you copied all of the output or
>> did you remove the frame in function main?
> Here is the full backtrace
> Program received signal SIGSEGV, Segmentation fault.
> 0x00007fd9b3fea0f9 in __gmpz_set_si () from /usr/lib/x86_64-linux-gnu/libgmp.so.10
> (gdb) bt
> #0  0x00007fd9b3fea0f9 in __gmpz_set_si () from /usr/lib/x86_64-linux-gnu/libgmp.so.10
> #1  0x00007fd9b453ac10 in bliss::AbstractGraph::search(bool, bliss::Stats&) () from /usr/lib/x86_64-linux-gnu/libbliss.so.2
> #2  0x00007fd9b453da3c in bliss::AbstractGraph::find_automorphisms(bliss::Stats&, void (*)(void*, unsigned int, unsigned int const*), void*) () from /usr/lib/x86_64-linux-gnu/libbliss.so.2
> #3  0x000055d66f737cd0 in main (argc=2, argv=0x7ffc262cb408) at bliss.cc:297

I guess from frame 3 having debug information, you either installed
the dbgsym package or rebuilt the bliss binary package locally?

As the address offset ...cd0 does not appear in the current
debian binary package's main function, I guess you did a
rebuild with debug information enabled?

Just to be sure, what does a 'which bliss' return?

Can you reproduce the crash also with the official
debian package bliss?

If yes, possibly you can install these dbgsym packages [1]:
    bliss-dbgsym libbliss2-dbgsym libgmp10-dbgsym

And forward the output of these gdb commands,
when the crash happened:
    info reg
    info share
    bt full

Kind regards,
Bernhard

[1] https://wiki.debian.org/HowToGetABacktrace#Installing_the_debugging_symbols



More information about the debian-science-maintainers mailing list