1 |
I just had another thought... would it be possible to impliment |
2 |
genkernel as an eclass? or perhaps split off related functions into a |
3 |
"library" and have both an eclass and a command line implimentation? |
4 |
|
5 |
I know I'm thinking big, but it is something that could really help. |
6 |
|
7 |
Nathaniel |
8 |
|
9 |
On Wed, 2004-04-28 at 22:25 -0400, Nathaniel McCallum wrote: |
10 |
> Just a quick perspective from the Gentoo-Installer project... NOT having |
11 |
> a full binary installation for us it not an option. We ARE providing |
12 |
> users with the ability to use the installer to do a stage one install. |
13 |
> HOWEVER, we should also be able to have a system up and running in under |
14 |
> an hour. This brings me to a thought I've had for a while: with |
15 |
> genkernel working so well, we should be able to provide binary kernel |
16 |
> packages. Let me put it this way... |
17 |
> |
18 |
> In the ebuild for kernels there should be a flag (unfortunately "build" |
19 |
> is already taken) that once downloading and installing kernel source |
20 |
> code it builds a binary kernel with genkernel. Then, if you make a |
21 |
> binary package of the kernel ebuild it could install not only the source |
22 |
> but a pre-built kernel as well. We could then provide a few default |
23 |
> binary kernels with the GRP release. Eventually, when gentoo tools get |
24 |
> a little better (we may actually work on this as a subset of Gentoo |
25 |
> Installer, we could even auto configure boot info (grub, lilo, etc...). |
26 |
> If this is the case, we could get a full system install into under half |
27 |
> an hour, which would be of GREAT benefit to both desktop users and |
28 |
> enterprise users. |
29 |
> |
30 |
> Just my $.02. |
31 |
> |
32 |
> Nathaniel |
33 |
> |
34 |
> |
35 |
> -- |
36 |
> gentoo-releng@g.o mailing list |
37 |
> |
38 |
> |
39 |
|
40 |
|
41 |
-- |
42 |
gentoo-releng@g.o mailing list |