Re: [user] segfault w/ error 4

["U.Mutlu" <for-gmane at mutluit dot com>]

U.Mutlu wrote, On 09/11/2012 10:38 AM:
> Jonathan Wakely wrote, On 09/11/2012 10:29 AM:
> > On Sep 11, 2012 9:22 AM, "U.Mutlu" <for-gmane@mutluit.com> wrote:
> > > Daniel KrÃgler wrote, On 09/11/2012 09:07 AM:
> > >
> > > > 2012/9/11 U.Mutlu <for-gmane@mutluit.com>:
> > > > > U.Mutlu wrote, On 09/11/2012 07:52 AM:
> > > > > >
> > > > > There was in a function a string comparisions whereby a const std::string
> > > > > was compared against a const ptr to char ("if (s == psz) ...").
> > > > > Not sure if that could be the reason (IMHO yes), as it indeed looked ugly &
> > > > > dangerous
> > > >
> > > >
> > > > There should be nothing wrong with this, *unless* the pointer to char
> > > > was a null
> > > > pointer. basic_string has overloaded operator== with const char* as a
> > > > single argument as
> > > > well.
> > > >
> > > > > I now have replaced the char ptr by an ordinary std::string.
> > > > > Will wait & observe what happens... :-)
> > > >
> > > >
> > > > I would expect no difference, if the std::string is initialized with
> > > > the same char pointer.
> > >
> > >
> > > Right you are, the crashes continue... :-(
> >
> > Have you tried using a debugger?
>
> It's a daemon program, and I have only basic experience with gdb,
> and not sure how to debug such a daemon program in gdb.

Last year or so I had a similar hard to find bug and at that time
I had to tell the system to generate a coredump file, and later I had analysed
it in gdb. But since I haven't done it for a long time now, the steps escaped me,
so I have to research how it was done... :-)

> > Can you run the program under valgrind?
>
> will test this as next later in the evening