Re: Kernel Recompile
Hi,
>>"Raul" == Raul Miller <rdm@test.legislate.com> writes:
Raul> Manoj Srivastava <srivasta@datasync.com> wrote:
>> Well, if they install a Debian packaed kernel source package, it is
>> documented in there. It is documented in the kernel-package package.
>> If you wish, I shall even add it to the kernel-image package, about
>> what one should do to upgrade the kernel the debian way.
Raul> I think it should be mentioned each of these places and in modutils
Raul> as well.
It has been now so documented in the kernel-* packages. Not
that I think this shall make much of a difference, because I think
few people actually read the documentation.
Raul> I don't care if it's standard or not. However, if other debian packages
Raul> depend on this package in some fashion (and near as I can tell, every
Raul> debian package which deals with modules does), they should document its
Raul> existence (probably by suggesting it).
If they actually depend on it, I guess so. I really think,
though, apart from packages that really depend on kernel-mage-VERSION
explicitly, nothing really requires the kernel be compiled with
kernel-package. They just require the kernel be compiled and
installed *correctly*. Like, that the kernel actually can find the
modules and packages can find the System.map.
The problem really seems to happen when people can't get the
kernel compiled and installed right. In that case, it is
really easier to point them to kernel-package than repeat the whole
thing over in excruciating detail (which, in the past, people still
did not manage to get right).
Raul> I don't care if people actually use the code in the package to build
Raul> kernel images or not. However, I feel that what the code does should be
Raul> documented, so that those who feel the urge to do it manually can be
Raul> referred to the documentation. Unfortunately, with the current state of
Raul> affairs, such folk have a valid objection.
What valid objection? What, pray, does the kernel-package do
which is special (apart from not making very many mistakes any more)?
What should be documented? The developers corner describes pretty
well already what it takes to create a Debian package. The kernel
documentation (apart from the /usr/include symlinks) describes how to
compile a kernel. /usr/doc/kernel-* (various kernel-* [ackages now)
document how to use make-kpkg.
I am *tired* of spoon feeding people, and then have them come
around and complain that the spoon feeding was not in their
face enough.
Raul> I don't think this needs to turn into a policy issue, but then other
Raul> issues which I'd consider less worthy (emacs oriented stuff) have had
Raul> more thorough documentation, and are better represented across packages.
I find this to be untrue, and mildly insulting. Maybe I am
reading this wrong. The README files in the kernel image and
kernel-package documentation document this quite well enough.
And really, one should, if only one had a clue, be able to
compile a kernel with out all this fandango. the Documentation
directory in the kernel sources are all that people seem to require
every where else in Linux. kernel-package merely makes this more
convenient; but by no means is it mandatory. If you want to make it
mandatory, raise its priority to standard.
manoj
violently in agreement with his sig
--
No matter how much you do, you'll never do enough.
Manoj Srivastava <srivasta@acm.org> <http://www.datasync.com/%7Esrivasta/>
Key C7261095 fingerprint = CB D9 F4 12 68 07 E4 05 CC 2D 27 12 1D F5 E8 6E
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: