severity 572003 wishlist
thanks

Hi,

On Sun, Feb 28 2010, M G Berberich wrote:

> generation of initrd is now handled by scripts in /etc/kernel. whoch
> have to be set up on the host, teh kernel ist to be installed. This is
> totaly broken.

> Needing an initrd or not, is a property of the kernel, not of the
> system the kernel get's installed. So it should be configured on a
> per-kernel basis and the configuration should be packaged with the
> kernel.

        There is a point here. It is indeed possible to set up a kernel
 such that it definitely needs a initrd (by compiling all file
 systems as modules), so this should be examined further.


> At the moment it is configured on a per-host basis und “packaged” with
> the host. 
>
> The effect is:
> * if the kernel needs an initrd and the install-host is not set up for
>   initrd-generation the initrd is missing and has to be generated per
>   hand.

        Right. I just set up all target hosts with the initrd scripts.

> * if the kernel does not need an initrd and the system is set up for
>   initrd-generation the initrd is generated and update-grub adds an
>   initrd-line into menu.conf.

        This is not the case. If the kernel was not set up with an
 initrd in mind, no inird is created.  My desktop never gets an initrd
 line. 


> make-kpkg should generate packages that can be installed on a system
> without needing manual chekup and interaction. In short they should
> behave like the Debian-kernel-packages.

        Debian official kernel packages behave like kernel-package used
 to (they used ancient kernel-package scripts, I believe). I think that
 the solution of using /etc/kernel.d, while not optimal, is better than
 the mess that was being created with tryng to keep up with the fast
 changing bootup actions needed on diferent architectres; the new system
 is more flexible, though it does need  a minimal setup.

        manoj
-- 
Price's Advice: It's all a game -- play it to have fun.
Manoj Srivastava <sriva...@acm.org> <http://www.golden-gryphon.com/>  
4096R/C5779A1C E37E 5EC5 2A01 DA25 AD20  05B6 CF48 9438 C577 9A1C



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org

Reply via email to