On Thu, Jul 16, 2015, at 04:01, Paul Wise wrote: > You can see the compiler commands used by clicking the Result column > for the architectures on this page. > > https://buildd.debian.org/status/logs.php?pkg=nano
It uses -fstack-protector-strong, which my compiler doesn't yet have. But if that was kicking in, it would have reported a "stack smashing detected" and have aborted instead of segfaulted. I cannot reproduce it with -fstack-protector nor -fstack-protector-all. Maybe it has to do with being a 64-bit system? Can you test on a 32-bit machine? If the line number given in your original backtrace is accurate, then it is segfaulting on a do_gotolinecolumn(). Strange. Does the segfault also occur if you start both nanos with nano +1 foo ? > Interestingly if I run nano under valgrind it doesn't segfault but > there are definitely some coding problems; several uninitialised > values, memory leaks and reading into unaddressable bytes. I have zero experience with valgrind or any other code-checking tools. Two days ago I ran valgrind on src/nano, but am puzzled by what all this means. I did find and plug a little memleak (totally unrelated to this crash), but will need a lot of time to learn to interpret valgrind's (and other tools') results. Benno -- http://www.fastmail.com - A fast, anti-spam email service. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
