1 |
On Tue, 17 Apr 2007 15:30:14 +0400, Andrey Gerasimenko wrote: |
2 |
|
3 |
> >> I do not see how it is hard to create a minimal installation CD image |
4 |
> >> every time new hardware support is added into the kernel, |
5 |
> > |
6 |
> > Then do it. Open source gives you the opportunity to make things |
7 |
> > happen yourself instead of whining that others won't do it. The build |
8 |
> > tools are in portage, so there's nothing stopping anyone from |
9 |
> > producing an updated minimal install CD, as has already been posted |
10 |
> > to the forums. |
11 |
> > |
12 |
> |
13 |
> Sorry, where do you see whining? |
14 |
|
15 |
Sorry, that came across as rather harsh. It was intended as a general |
16 |
comment, not a criticism of you. |
17 |
|
18 |
> Even if I say that if whatever posted |
19 |
> to forums is good then it should go to the official Gentoo site, this |
20 |
> would not be whining. |
21 |
|
22 |
No, but it would not be practical either, because an official release |
23 |
needs a lot more testing. |
24 |
|
25 |
> Once again, there should be some problem with my English. It is |
26 |
> official Gentoo release policy to have minimal, live, and platform |
27 |
> releases in sync. Posting a new image to forums is not that tightly |
28 |
> related to policies. |
29 |
|
30 |
No it's not, and I never suggested it was. As an Open Source project, |
31 |
ANYONE can build a new, unofficial image that supports brand new |
32 |
hardware. They don't need to wait for the full releng cycle of testing on |
33 |
all packages. |
34 |
|
35 |
> This very thread, as explained in my post, is just |
36 |
> one reason to change the policy. |
37 |
|
38 |
Then you should file a bug suggesting this. |
39 |
|
40 |
> >> or new gcc |
41 |
> >> version goes stable, or new portage version goes stable. |
42 |
> > |
43 |
> > These are irrelevant. As long as the CD boots, recognises your core |
44 |
> > hardware - which really comes down to disk controllers and network |
45 |
> > interfaces - and installs a working system, the rest can be updated |
46 |
> > post-install. |
47 |
> > |
48 |
> |
49 |
> No, GCC and portage are relevant. The fact that the installation |
50 |
> process succeeds does not help much when a new user, just after |
51 |
> downloading the latest and greatest, has to recompile something as |
52 |
> basic and huge as GCC or just interrupt the install getting the scary |
53 |
> message "you better do nothing until you upgrade Portage". |
54 |
|
55 |
Whatever is included, something big will have a new version by the time |
56 |
the full install has been comprehensively tested on all supported |
57 |
platforms and put on the mirrors for a week. A Gentoo install is supposed |
58 |
to give you a working system that is a starting point, not an end in |
59 |
itself. |
60 |
|
61 |
The only time a new install disc is really necessary is when the old one |
62 |
doesn't support your hardware. |
63 |
|
64 |
> > A major GCC update is the exception to this rule, but that is |
65 |
> > precisely the sort of thing that needs extensive testing on a range |
66 |
> > of platforms rather than a rushed release. |
67 |
|
68 |
> Just in case you already deleted my post, I recommend new minimal CD |
69 |
> release each time a new GCC version, major or not, goes stable. |
70 |
|
71 |
I still maintain that minor GCC upgrades are not an issue, kernel |
72 |
upgrades are far more relevant as that is where most hardware support |
73 |
takes place. Why do you consider a GCC upgrade such a big deal? After a |
74 |
Stage 3 install, you are likely to want to do an emerge -e world anyway, |
75 |
to apply your customisations, so GCC will probably be recompiled anyway. |
76 |
As long as the latest stable version is not incompatible with the CD, |
77 |
what's the big deal? |
78 |
|
79 |
> What extra testing does a stable version need? |
80 |
|
81 |
To ensure that everything works as a cohesive whole. |
82 |
|
83 |
|
84 |
-- |
85 |
Neil Bothwick |
86 |
|
87 |
If at first you don't succeed, skydiving is not for you. |