1 |
On Sun, 2004-05-02 at 20:06 -0700, Greg KH wrote: |
2 |
> On Sun, May 02, 2004 at 10:54:26PM -0400, Nathaniel McCallum wrote: |
3 |
> > On Sun, 2004-05-02 at 19:45 -0700, Greg KH wrote: |
4 |
> > > On Sun, May 02, 2004 at 10:04:35PM -0400, Nathaniel McCallum wrote: |
5 |
> > > > The new glep is here: http://www.gentoo.org/proj/en/glep/glep-0026.html |
6 |
> > > |
7 |
> > > Hm, did you consult the current kernel developers/maintainers/packagers |
8 |
> > > before creating this? At first glance I see a number of issues that |
9 |
> > > would make this completely impractical and impossible to implement. |
10 |
> > Yes. Actually, a lot of people have talked about it. |
11 |
> |
12 |
> Ok, did johnm sign off on this? As he is the current Gentoo kernel |
13 |
> maintainer I would expect you would have run this by him. |
14 |
|
15 |
I sent him an email. He hasn't replied yet. However, plasmaroo and I |
16 |
spoke extensively about this. |
17 |
|
18 |
> Do you want me to outline my objections to this proposal here, or is |
19 |
> there some other place to do it? |
20 |
|
21 |
Here is fine. |
22 |
|
23 |
> > > Also, have you taken a look at the current genkernel work being done for |
24 |
> > > 2.6? It probably resolves lots of the issues it sounds like you |
25 |
> > > currently have with the kernel packaging process. |
26 |
> > |
27 |
> > Yes. This GLEP is basically a reworking of genkernel into portage. |
28 |
> |
29 |
> Yes you have seen the new work? And no, you don't like it? Or yes, you |
30 |
> like it and think it can help out? Confused here... :) |
31 |
|
32 |
I do like the current process. It just needs to be integrated into |
33 |
portage. We need to be able to have easy building of binary kernel |
34 |
packages. We need a one step build for users, rather than two step |
35 |
portage first then genkernel. |
36 |
|
37 |
> > Please understand it is bad design to have two build systems. |
38 |
> |
39 |
> Ivory towers are over that-a-way. Real world is here :) Come on, the |
40 |
> build system for the kernel is just so much more complex than what the |
41 |
> current portage system can even begin to support, based on what you are |
42 |
> trying to propose. |
43 |
|
44 |
Not at all or in any sense is this true. People 5 years ago thought |
45 |
something like portage would never work. In the original version of |
46 |
GLIS (before genkernel even existed) I built a fully functinal automated |
47 |
building system. And, btw, both it an genkernel were written in bash. |
48 |
There is nothing stoping a re-working of genkernel's code (in BASH) into |
49 |
eclasses and ebuilds (ALSO in BASH). |
50 |
|
51 |
I'm not living in the ivory tower. Abstraction is simply good code. |
52 |
|
53 |
> |
54 |
> Also, please realize that the 2.6 kernel build process is infinitely |
55 |
> better than the 2.4 one, and based on it, a number of your listed |
56 |
> objections and proposals are pretty moot. |
57 |
|
58 |
Like binary packaging support? Being able to keep around just the |
59 |
kernel headers? A single API for all package building? Actually, none |
60 |
of the reasons for the GLEP are solved by the 2.6 kernel build process |
61 |
(although it is bettter). |
62 |
|
63 |
|
64 |
> > Two build systems means one of them has a bug (ie doesn't support |
65 |
> > needed types of building). |
66 |
> |
67 |
> Could you perhaps outline your current objections and problems with the |
68 |
> current build system, contrast it with the currently being worked on |
69 |
> changes to that build system, and maybe we can work from there? |
70 |
|
71 |
Please read the GLEP carefully. Everything is outlined in there. |
72 |
|
73 |
> |
74 |
> > > I do find the idea of packaging up a "binary" kernel a bit interesting, |
75 |
> > > but who is going to be doing that packaging? And who is going to |
76 |
> > > mediate the zillion different "but why don't you select this kernel |
77 |
> > > option instead" requests that will be caused by such a package? :) |
78 |
> > |
79 |
> > Users can do their own packageing, then can also supply their own binary |
80 |
> > kernels. |
81 |
> |
82 |
> Do we allow this with the current -bin packages already? I might be |
83 |
> mistaken, but I didn't think we did for some reason... |
84 |
|
85 |
-bin has nothing to do with kernels. Besides, -bin packages have to be |
86 |
made by hand. I'm talking about automation. |
87 |
|
88 |
> > This is very much desireable for enterprise users (think clusters |
89 |
> > especially) and for Gentoo Installer. We will not be accepting |
90 |
> > those kind of feature bugs as the user can always re-emerge the kernel |
91 |
> > with their own options. (its the same exact thing as "why didn't you |
92 |
> > build such and such GRP package with such and such flag", we just ignore |
93 |
> > those unless it is a major issue.) |
94 |
> |
95 |
> I see you don't realize the enormous headache and impossibility of |
96 |
> trying to use USE flags to enable or disable different kernel patches. |
97 |
We aren't planning to do any such thing. I really wish people would |
98 |
READ GLEPs before posting. |
99 |
|
100 |
|
101 |
> But hey, feel free to prove me wrong, I'm eagerly awaiting your |
102 |
> reference implementation. Any ETA on it? :) |
103 |
|
104 |
Yeah, as soon as I'm done with GentooInstaller. |
105 |
|
106 |
Nathaniel |
107 |
|
108 |
|
109 |
-- |
110 |
gentoo-dev@g.o mailing list |