On Fri, Jun 18, 2004 at 10:28:50PM +0200, Jens Schmalzing wrote: > Hi, > > Sven Luther writes: > > > > The existing mkvmlinuz can do this if you put the line > > > /boot/vmlinuz-$release in /etc/mkvmlinuz/output . > > > Well, it doesn't cost much by setting it by hand, > > Yes. > > > unless we want to set /etc/mkvmlinuz/output to it by default. > > There are packages that contain /boot/vmlinuz-$release, the risk of > clobbering that by accident with a default setting is too high for my > taste.
Well, only 2.4.25 packages, which is why we should make a test for non 2.4 kernels or something such. 2.4 kernels don't have the the necessary mkvmlinuz magic to make it work anyway. > > > mkvmlinuz works for both non-initrd and 2.4 kernels. > > > > Yeah, but what would happen if i tried to install a > > kernel-image-2.4.25-powerpc-chrp package ? There would be no vmlinux > > for it, and thus the problem. > > What exactly is the problem here? Installing 2.4 and 2.6 side by side? Well, these contain a vmlinuz, but no vmlinux, so mkvmlinuz, as called by the post inst would fail, and also the problem you mentioned above. > > Ok. I would still like to have kernel-image... depend on mkvmlinuz > > though. > > On many systems it's perfectly reasonable to install a kernel-image > package without mkvmlinuz. So a dependency is too strong, a > suggestion should be sufficient IMHO. And how please do you want to make sure someone with a pegasos machine, not knowing about a mkvmlinuz, will be able to boot its machine when he thought he had installed the kernel-image package ? That would be a RC bug, making the system unusable and all. > > > Sounds good. If Policy doesn't allow 3), the postinst could still > > > check the file and present a warning. > > > > Why should it not ? > > Policy is not clear to me there, I need time to understand it. > > > You are asking the user if he would like to have it added, a bit > > like passwd or whatever asking about changing the groups and > > such. > > Actually, changing of /etc/passwd by base-passwd is explicitly allowed > in the Policy and that's one of the things that puzzles me. Why do > they bother to allow that one case? > > > If this is not allowed by Policy, then the policy is buggy. > > In the past, when I thought there was a bug in Policy it eventually > turned out to be in my brain. Always. > > > I hope to have this in the alioth svn repo soon though, and we can > > then start moving the powerpc package there also. > > Can you please give me a very brief tutorial for checking things in > and out? Sure, but once the repository is there. The difficult thing is the first checkout, after you just do update and ci kind of stuff. Friendly, Sven Luther
