Re: Inclusion of kernel version in kernel package names: A followup
>>"Dale" == Dale Scheetz <firstname.lastname@example.org> writes:
Dale> On Tue, 14 May 1996, Ian Jackson wrote:
>> I don't think that using dpkg's dependency mechanisms is appropriate
>> here, because people are always compiling their own kernels.
>> Perhaps the kernel image prerm should check that the kernel being
>> removed isn't the default one in lilo.conf, and/or that it isn't the
>> only one.
Dale> This is not sufficient. Some of us use loadlin to boot linux and don't
Dale> even have a lilo.conf.
The discussion is in regards to removing image packages. All
I intend to do is to try to determine if I can see any other images
on the disk, or else warn the user that they might be rendering their
system un bootable. So, if you use loadlin, then the prerm script
will ask you if you want to continue. I think this aggravation is
worth the safety net the naive user will get; they still have a
chance to abort.
Dale> I contend that neither the image nor the source packages have
Dale> any business mucking around in /lib/modules.
Huh? The source package doesn't touch /lib, and the image
package just installs the precompiled modules. why should it not do
Dale> It currently
Dale> removes all modules dependancy files which requires running
Dale> depmod -a to fix.
Are you sure the new kernel-image packages do that? I think
that you are talking about the old case when each new image replaced
the older version. The new scheme explicitly *fixes* this problem.
Dale> I personally keep 3 to 4 kernels available on
Dale> my dos partition so that if I have trouble with one, I can move
Dale> back until I find one that works. I can imagine getting into a
Dale> state where the only modules with dependancy information are for
Dale> a kernel that only exists on the root file system. At the very
Dale> least the install should not mess with the currently running
Dale> kernel setup.
The pre/postrm scripts will not remove anything, really, they
just give the user a chance to abort, and remind them to run lilo/get
a bootdisk later. As to the rest, the remove action removes the
files the package came with, and the user has asked for their
I really fail to see your objections are, could you please
People love high ideals, but they got to be about 33-percent
plausible. -- Will Rogers %%
Manoj Srivastava Systems Research Programmer, Project Pilgrim,
Phone: (413) 545-3918 A143B Lederle Graduate Research Center,
Fax: (413) 545-1249 University of Massachusetts, Amherst, MA 01003