On Tue, 2006-11-07 at 13:59 -0500, Willie Walker wrote: > Hi Dieter: > > Thanks for the response!
np > > Doing _only_ this leads to confusion as you are breaking simple commands > > like ps or killall. That's why the prctl stuff is necessary. Correcting myself: this is harmless. I've been confusing python's sys.argv[0] and the systems argv stored in /proc/id/cmdline. Sorry about that :S > Do you mean that ps and killall will end up showing the modified argv[0] > name? I just did a test on Ubuntu Edgy and I don't see this behavior. > In addition, killall continues to do the right thing. You're correct on the original name being used by ps/killall even after python's sys.argv[0] has been changed. Strangely at-poke and friends do pick up this change as you've described. I'm curious how it does it... > > This whole thing is nothing more than a non-portable hack (3). It's up > > to you guys if you're willing to use it. > > With the argv[0] hack, however, I think we cover the platforms we care > most about (Solaris and Linux). Correct. It would have to say Patrick is correct in his mail when he said sometimes a hack is all you've got. > It might be sufficient for our > purposes, unless we can figure out how to use "-m" and that just makes > our problem go away. There's a recipe called "Executing modules inside packages with '-m'" here: http://aspn.activestate.com/ASPN/Cookbook/Python/Recipe/307772 that works with python < 2.5. From 2.5 it (or something similar) is built in. > Thanks much for the insight and feedback. In this case I would recommend playing with sys.argv[0] after all... > Will HTH, sorry for the confusion, Dieter
signature.asc
Description: This is a digitally signed message part
_______________________________________________ pygtk mailing list [email protected] http://www.daa.com.au/mailman/listinfo/pygtk Read the PyGTK FAQ: http://www.async.com.br/faq/pygtk/
