Re: Inclusion of kernel version in kernel package names: A followup
On Wed, 15 May 1996, Guy Maor wrote:
> On Wed, 15 May 1996, Dale Scheetz wrote:
> > On Wed, 15 May 1996, Guy Maor wrote:
> > > I think it's also a mistake, Dale, to install your own kernels with
> > > your own lilo.conf and still expect kernel images that you've installed
> > > to work.
> > There must be a typo here. My point was that I don't use lilo at all! I
> > use loadlin. As a result looking a lilo.conf doesn't tell you what kernel
> > I run.
> This only strengthens my argument. You're just one of the many people
> who've slipped through the cracks of the default scripts.
If the argument you are strengthening is that it is a mistake to build
your own kernel without using lilo.conf, I disagree.
My expectation was not that the newly installed kernel image should work
for me, as installed (I know where to copy the new vmlinuz to, if I want
to use the new kernel), it was that installing the kernel image should not
"break" the current installed kernel. Image-1.3.64 did that by deleting
the modules.dep file (and complained that it couldn't delete all the
modules) for the current kernel image. I understand the new kernel-image
package doesn't do this any more (I'll get a chance to test that this
evening, I hope)
I get a choice about this by upgrading with dpkg. Even relatively
experienced users of dselect tend to just "let" dselect do what it wants
to and allow a full installation. Outside of any conflicts or depends this
means that image and source WILL be installed. My points were always that
this installation shouldn't break the current, or previously current
kernel functionality the way image-1.3.64 did.
aka Dale Scheetz Phone: 1 (904) 877-0257
Flexible Software Fax: NONE
Black Creek Critters e-mail: firstname.lastname@example.org
------------ If you don't see what you want, just ask --------------