> 2012/1/6 Guillaume Rousse <[email protected]>: >> Le 06/01/2012 16:13, Wolfgang Bornath a écrit : >> >>> Ah, I see your reasoning, of course, if the packager forgot to name >>> the requires then urpmi declares them as orphans. But then, to be >>> safe, you have to forget about auto-orphans altogether because you can >>> not be sure that all packagers did their homework. >> >> Then you have to forget about using packages because you're not sure >> packagers did their work correctly. > > I'd argue like that as well if we were in court. But it's not the > same: if a packager misses something and the installation does not > work, so what? I can use another package or distribution. But if he > causes urpmi to regard a needed package as orphan and lets me remove > it the system can break, now that is a problem. > >> So far, still no one proved than 'orphan' status was wrong regarding >> urpmi >> definition of what is an orphan package, rather than regarding their own >> personal expectation. > > Yes, because the user does not care about any such definitions when he > reads on the console or in rpmdrake "These packages are orphans now, > you can safely remove them". I'd suggest to change this sentence ASAP > into "If you are sure that it will not break anything you can remove > them now". This would be a better advice for the user than "you can > safely remove". >
True that the user does not and should not care about definitions of an orphan, but also, the user should not be put in a situation where he/she will have to go hunting for what could or could not break anything. -- finid
