Control: severity -1 grave Hi,
[Jaromír: please send mail to the bug, not the list] On 21/01/17 14:18, Christian Grigis wrote: > Package: gwc > Version: 0.21.19~dfsg0-6 > Severity: important > > Running 'gnome_wave_cleaner' from the testing package version > (0.21.19~dfsg0-6) crashes immediately at startup: > > $ gnome_wave_cleaner > Current stack limit: 8388608 bytes > Segmentation fault This bug is clearly RC. Jaromír, you did test this before uploading it right? [...] > The gdb backtrace gives: > > (gdb) run > Starting program: /home/glove/tmp/gwc-testing/gwc-0.21.19~dfsg0/gwc > [Thread debugging using libthread_db enabled] > Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1". > Current stack limit: 8388608 bytes > > Program received signal SIGSEGV, Segmentation fault. > 0x00007ffff6e047f9 in g_type_is_a () from > /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0 > (gdb) bt > #0 0x00007ffff6e047f9 in g_type_is_a () from > /usr/lib/x86_64-linux-gnu/libgobject-2.0.so.0 > #1 0x00007ffff7519084 in gtk_type_new () from > /usr/lib/x86_64-linux-gnu/libgtk-x11-2.0.so.0 > #2 0x000055555557223c in led_bar_new (segments=20, orientation=0) at > gtkledbar.c:82 The problem is here. led_bar_get_type returns an unsigned int, but gtk_type_new expects a "GtkType" which is an integer with the same size as a pointer. This code is going to need porting to work on 64-bit arches. > I experimented with optimization levels but even -O0 gives the crash. > > However, it looks like it is related to the compiler, because forcing to > compile with gcc-5 instead of gcc-6 produces a binary that starts up > normally. This is probably related to PIE being enabled. I suspect that without PIE (and therefore ASLR), by chance all the pointers were < 4GB so the bug never occurred. James
signature.asc
Description: OpenPGP digital signature