[At 18.11.2005 16:35, Steve Langasek kindly sent the following quotation.] >> (void)kill( -getprocgroup(), SIGHUP ); > >> in the cleanup() method. It's probably more than most of the parent >> processes do out there, but at least it reveals such megabugs like >> that of vacation. > > No, I think this is a bogus assumption on the part of maildrop, not a > "megabug" in vacation. I don't see any reason why maildrop should > be either setting a process group, or killing the group, under such > circumstances. > > I think it's still a bug in vacation to not check the exit value of > the commmand it spawns, but I believe the larger bug is maildrop's.
Well, I stand for a different philosophy. Unless some process is designed and set up to run standalone (daemon), it should never be left on its own by a parent. There should never be any zombies. Zombies mean lame programmers. So when there is a cop (maildrop here), who can avoid the state of zombies slowly polluting the machine, which is easily reachable especially in mass mail delivery sites, I count it's ok. What would be a reason of running any child-child-child zombies after the maildrop exit? None, I think. If the process chain is alive and long running, then it's IMO not optimal, but tolerable. Therefore that must be a major vacation bug and its authors should go read Stevens (again). I would let maildrop doing a good work protecting the machine, but of course I have no word here except this jaw-jaw. :-) Have a nice $DAY_PHASE. -- \//\/\ (Sometimes credited as 1494 F8DD 6379 4CD7 E7E3 1FC9 D750 4243 1F05 9424.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
