Re: [Nano-devel] Nano 1.3.8 and 1.3.8-cvs crashes on suse 9.3 x86_64

[David Lawrence Ramsey]

Sorry for the delay in getting back to you.  I've been a bit burned out
lately.

Simon Strandman wrote:

<snip>

> Okay. I installed valgrind 3.0 and ran it on cvs checkout of nano from
> earlier today. The problem is that I can't reproduce the bug when
> running under valgrind. I could search and do anything without it
> crashing it seemed. But without valgrind it still crashes. Here is the
> output anyway, I don't know if it's of any use. (It's a suse 9.3
> system running in a chroot, that's why /proc/version reports gentoo)

<snip>

Thanks for trying.  I think it is of some use after all.  Since
valgrind's output shows that it's redirecting the string functions, it
could mean that the problem isn't with nano per se.

Was the version tested under valgrind built with the same optimization
flags that the crashing version was?  I just ran into a similar problem
with an entirely different program: a default build that used "-O2"
would crash reliably, but a debug build that didn't use it wouldn't
crash at all.  If the same thing is happening here, maybe the optimized
string functions are being miscompiled when certain optimization flags
are used with gcc.  Does building current CVS with e.g. different levels
of "-O" change anything?