Le mercredi 10 mai 2006 à 16:05 +0200, Marco d'Itri a écrit : > I tried to be nice.
Sorry, but no. Playing bug ping-pong is not nice. > If you disagree with the resolution of this bug then > feel free to discuss this on debian-devel@ or bring it to the CTTE It has already been brought up on debian-devel several times, and you have always refused to implement the correct solution of having several udev versions installable in parallel. > It's the best possible solution given the constraints. If you disagree > then feel free to send a patch or at least an explanation of the changes > you propose. But please do your homeworks first: there is a long list > of people who tought that they could do something better and nobody ever > submitted a working proposal after being challanged. Maybe because not everyone has the time to discuss endlessly with an uncooperative person when there are so many other issues to fix. > > At the very least, the upgrade should fail at *config* time, not in the > > preinst when hundreds of packages have already been unpacked and are > > left unconfigured. > This would not work. If the system is not upgradable then the upgrade > must be stopped before the old udev installation is destroyed, and this > happens between preinst and postinst when the old files are deleted and > the old daemon is stopped. If the system isn't upgradeable, the upgrade should be stopped before it starts at all. You shouldn't think of your package alone, but in a whole distribution with hundreds of other packages installed. > > Anyway, don't close bugs that aren't fixed. Upgrades from sarge to etch > > fail, and this is a bug. > Upgrades do not fail and this is not a bug. Upgrades do fail. Do you want to see a log maybe? -- .''`. Josselin Mouette /\./\ : :' : [EMAIL PROTECTED] `. `' [EMAIL PROTECTED] `- Debian GNU/Linux -- The power of freedom
