Re: zenmap as root error

["João Medeiros" <ignotus21 () gmail com>]

On Tue, Apr 1, 2008 at 11:39 PM, Guilherme Polo <ggpolo () gmail com> wrote:
> 2008/4/1, João Medeiros <ignotus21 () gmail com>:
>  > On Tue, Apr 1, 2008 at 11:21 PM, Guilherme Polo <ggpolo () gmail com> wrote:
>  >  > 2008/4/1, João Medeiros <ignotus21 () gmail com>:
>  >  >
>  >  > > On Tue, Apr 1, 2008 at 10:52 PM, Guilherme Polo <ggpolo () gmail com> wrote:
>  >  >  >  >  But, for PyGtk 2.8 and earlier I would suggest doing this in a
>  >  >  >  >  different manner (and much simpler and less error-prone):
>  >  >  >  >
>  >  >  >  >  try:
>  >  >  >  >     import gtk
>  >  >  >  >  except ImportError, e:
>  >  >  >  >     print e
>  >  >  >  >
>  >  >  >  >  For PyGtk 2.10 and newer, ImportError is actually a warning so the
>  >  >  >  >  code changes a bit:
>  >  >  >  >
>  >  >  >  >  import warnings
>  >  >  >  >  warnings.filterwarnings('error', module='gtk')
>  >  >  >  >  try:
>  >  >  >  >     import gtk
>  >  >  >  >  except Warning, w:
>  >  >  >  >     print w
>  >  >  >  >  warnings.resetwarnings()
>  >  >  >  >
>  >  >  >
>  >  >  >
>  >  >  > I think that just display the an eventually warning is not enough. We
>  >  >  >  have to leave segfault.
>  >  >
>  >  >  If this Warning happens, it implies the gtk engine couldn't start and
>  >  >  the app wouldn't run anyway (or could try and segfault). Also, you
>  >  >  could just call exit after your print that message, it was just a
>  >  >  correct template on how to proceed when importing gtk fails.
>  >
>  >
>  > Hum, sorry if I'm wrong, but if I just call exit in a warning
>  >  exception the programing will exit with any warning. Maybe some
>  >  warning does not cause segfault.
>  >
>
>  You could then alternatively catch ImportWarning, but this one I
>  haven't tested. The problem with your "solution" was already pointed
>  by Fyodor, setting a wrong DISPLAY will bypass your solution and the
>  segfault will still happen.
>
>
>  >
>  >  >
>  >  >  > By this way we have to check if the waring is
>  >  >  >  associated to display too.
>  >  >
>  >  >  Why ? Just checking for DISPLAY, like you proposed, doesn't solve this.
>  >
>  >
>  > Because of what I say above I think we have to check display
>  >  condition. Checking the if DISPLAY var is set and the system is
>  >  *nix-like the problem is solved. I have just tested in my linux
>  >  machine few minutes ago.
>  >
>
>  Gtk already checks for dispaly.

I'm just trying to say to you that you have to check if the warning is
associated to "Can't open display". Because this is not a import
problem. I think this solve the problem. Good lucky.

Att, João Medeiros.

_______________________________________________
Sent through the nmap-dev mailing list
http://cgi.insecure.org/mailman/listinfo/nmap-dev
Archived at http://SecLists.Org