Bug#613662: unmet dependencies race condition
Le Thursday 17 February 2011 11:30:40 jidanni@jidanni.org, vous avez écrit :
> Why can't they simply ensure the lower layers in this particular
> dependency pyramid are sent out before or at the same time as the other
> parts?
Because it's not important in unstable (as it gets solved over time)!
You are using unstable, that is a suite _designed_ to suffer those issues
_all_the_time_ (and will continue to do so).
If you want a system that doesn't have this type of problems; use testing!
> Anyway, there upon full upgrade we get asked:
> (…)
> One slip of the finger and we have removed our operating system, "never
> to boot again", "it was time for grandpa to retire anyway".
>
> Yes of course it will probably ask us "are we sure?".
… This happens because you try a "full upgrade", which is the _WRONG_ thing to
do when using unstable. I mentioned you already that using "safe-upgrade" or
"upgrade" doesn't trigger that type of problems. So you should really try using
the correct tools for the suite you are using; expectations you can have on
stable are not the same than the ones you can have on unstable; both imply
different use cases, different tools and different temporary issues.
Now there is plenty of documentation out there that will explain this better
than me; please let this non-bugreport die.
--
Didier Raboud, proud Debian Developer.
CH-1020 Renens
odyx@debian.org
Reply to: