1 |
> > .. which may not be received too well. There is a perception that |
2 |
> > Developers *support* ~arch, which is a skewed outlook; it's there for |
3 |
> > testing, it is *not* meant to be used by 99.5% of end users. It is a |
4 |
> > means to an end, a way to track packages which *may* be stable, a QA |
5 |
> > process. |
6 |
> > |
7 |
> > ie: The following would/should be entirely acceptable: |
8 |
> > |
9 |
> > <User> I'm running ~arch of libfoo and it's breaking appwoo, help! |
10 |
> > Need this to work, really *REALLY* badly! |
11 |
> > |
12 |
> > <Dev> We're aware of those issues, but libfoo works fine for most |
13 |
> > of the other apps which require it. No ETA on the fix, |
14 |
> > tough sh*t for running ~arch on a critical box. |
15 |
> > |
16 |
> > <User> Arrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrgh! |
17 |
> > |
18 |
> > If you're interested in helping that QA process, most of the |
19 |
> > architecture teams now have an 'Arch Tester' (AT) setup you could help |
20 |
> > out with... |
21 |
> |
22 |
> Well ... I've been running ~x86 and ~amd64 for a long time and I can't |
23 |
> remember an instance where I needed to drop back to stable for the |
24 |
> things I regularly use, such as R, maxima, Ruby, Lyx, and I can't |
25 |
> remember a time when I needed to drop back to stable for a core |
26 |
> component like the kernel, gcc, perl, or python either. But -- that's |
27 |
> x86 and amd64 -- it might be much riskier on something less common, like |
28 |
> powerpc. |
29 |
|
30 |
I wasn't attempting to state "This does not work!"; merely expressing |
31 |
that ~arch isn't really a supported platform. Dropping back to stable |
32 |
isn't really a viable route, once your system is ~arch there's quite a |
33 |
lot to go <BOOM!> if you tried to globally undo that. Wanna try it? |
34 |
;) |
35 |
|
36 |
At the moment Gentoo Linux has a reputation as a "ricer" distribution, |
37 |
and a large proportion of users on ~arch does nothing to solve that... |
38 |
Speaking entirely frankly I'd love to see increased adoption in |
39 |
enterprise, there's a whole lot this distribution has to offer to |
40 |
server farms, for example. |
41 |
|
42 |
Look at it this way: by running ~arch whilst *not* a Developer or |
43 |
Arch Tester you're having a very limited impact, or possibly a |
44 |
negative one. Getting onto the 'track' of contributing to the project |
45 |
through the various 'Arch Tester' teams is a great way for a "Power |
46 |
User" to help out; should you feel you're more technically inclined, |
47 |
can write a useful language or three / hack ebuilds as naturally as |
48 |
breathing, I know we need Developers! Especially in understaffed |
49 |
areas like Release Engineering. :) |
50 |
|
51 |
I'd have liked to see two main things happen with Gentoo 2008.0: |
52 |
|
53 |
* Get rid of stage3 - all our install documentation works with |
54 |
just the stage3 right now, we don't "support" stage1/2 |
55 |
installs yet users are /always/ asking on IRC and MLs |
56 |
for help with a stage1 install because they think it's l33t. |
57 |
Remove it from mirrors, put it in /experimental, whatever; |
58 |
we need the stage1/2 somewhere for lotsa reasons, but lets |
59 |
make it less obvious to weed out those clueless ricers. |
60 |
|
61 |
(the next one is more of a Portage change) |
62 |
* Have some warning banners on ~arch and a toggle option for |
63 |
make.conf to disable them. There are *far* too many people |
64 |
on IRC suggesting newbies adopt ~arch, and they do so.. :( |
65 |
They've got no clue what it means, then they bitch/whine |
66 |
when they hit ABI issues or other problems and blame Gentoo. |
67 |
Don't document the toggle option in the Install Manual ;) |
68 |
|
69 |
Suggested value for disabling the big flashy warning banners :P |
70 |
MODIFYING_ACCEPT_KEYWORDS_MAY_BREAK_MY_BOX_AND_I_UNDERSTAND_THIS |
71 |
-- |
72 |
gentoo-releng@l.g.o mailing list |