[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: kernel-package Help



On Sun, Aug 22, 2004 at 11:43:10PM -0500, Manoj Srivastava wrote:
> On Mon, 23 Aug 2004 13:03:56 +0900, Horms  <horms@debian.org> said: 
> 
> > This is to get around the problem where these packages are rejected
> > as their version, 2.4.26-6 is less than the version currently in
> > debian provided by kernel-image-1-i386-2.4.27, that is version
> > 2.4.27-X. This will allow 2.4.26-6 to be uploaded.
> 
> 	I guess I haven't been following along too closely.  The idea
>  behind kernel image naming used to be that version 2.4.26 was a
>  different package from version 2.4.27, and thus was independent;
>  uploading 2.4.27 would in no way impact 2.4.26 uploads.  Has this
>  been somehow changed?

Ok, let me resume, and propose a solution here.

The real problem here is that the kernel packages contain both the real
packages (kernel-image-2.4.26-arch or something), and also meta packages
pointing to the latest kernel-image, or let's say the one we decide to be the
default one.

Since 2.4.26 and 2.4.27 where both uploaded with different versions of this
same package, there was a problem since the new one replaced the new one.

This caused problems in two cases : i386, where the new upload of 2.4.26-5 or
whatever was older than the version present in the archive which was 2.4.27-1
i think. And in the sparc case, where the 2.4.26 packages had version number
39, while the 2.4.27 packages just uploaded had version 2.4.27-1.

The solution to this is to create a kernel-image-meta or whatever source
package, which provides the kernel-image-<arch> and the kernel-image-<arch>-2.[46]
packages. This would be fully independent from the real kernel-image packages,
and hold the default kernels per arch for both 2.4 and 2.6 and general.

It seems thought, and that is where you come in, that from horms experience,
removing those meta-packages from the packages makes make-kpkg fail to build
the package, which is what horms was trying to tell you. Could you look into
this with him, and maybe upload a new kernel-package which fix this, if
necessary, or help out horms to fix this problem ?

Friendly,

Sven Luther



Reply to: