1 |
Hi, |
2 |
|
3 |
One thing that I find missing from this discussion is the actual meaning |
4 |
of the keywords. Currently, when a package is marked stable, it means |
5 |
that not only it compiled on the platform, but that it was also tested |
6 |
by a real human being and that it (most of the time) it has been (we |
7 |
hope) tested by many others without problem. The original GLEP22 |
8 |
proposal changes this, since the various "variables" are not truly |
9 |
orthogonal, we end up with many many untested situations. |
10 |
|
11 |
If we want to support a very large number of hardware/software |
12 |
combinations we have to either decide that we want to keep the same |
13 |
level of QA and then test on every possibility and mark each one or |
14 |
decide to change the level of QA and go for a debianization (if it |
15 |
compiles, its ok) or an even looser system. For all its deficiencies, |
16 |
rac's proposal seems like the only sane way to do this. And it would |
17 |
allow for for much more diversified QA information, so we could record |
18 |
different information from the results of a tinderbox to the number of |
19 |
users using it. But this is very heavy to do. As a short term solution, |
20 |
we could instead go with "multi-part" keywords (stuff like |
21 |
x86:linux:gnu:glibc or i386-pc-linux-gnu (why not use standardised that |
22 |
if we do that?)). Those keywords can always be later used to fill a |
23 |
database. |
24 |
|
25 |
This said, how much effort are we ready to put into satisfying what will |
26 |
probably be a very very small number of users. I have the feeling that |
27 |
almost all non-linux uses are purely academic (proving it can be done). |
28 |
There is very little demand or use for gentoo outside the Linux world |
29 |
(maybe except for macosx and that's just one more keyword), and no |
30 |
proper solution has been proposed. So my conclusion on this is that for |
31 |
the time being, we should stay with the current keywording system and |
32 |
maybe add two or three more. |
33 |
|
34 |
Olivier |
35 |
tester@g.o |
36 |
|
37 |
On Wed, 2004-03-17 at 00:12, Grant Goodyear wrote: |
38 |
> Okay, here's a consolidation of the various comments about GLEP 22. |
39 |
> |
40 |
> The first comment, by Method, pointed out the most significant flaw in |
41 |
> my proposal: the ARCH, LIBC, KERNEL, and USERLAND variables are not |
42 |
> orthogonal (to use rac's term), so if ARCH lists x86 and ppc while |
43 |
> KERNEL lists linux and obsd, how to we address the fact that obsd/ppc is |
44 |
> not, in fact, a valid combination? |
45 |
> |
46 |
> I'm going to ignore a large thread of messages that discuss how |
47 |
> sensitive alternative arch's are to requiring package patches to compile |
48 |
> and/or function properly, other than to note that with the large number |
49 |
> of possible ARCH/LIBC/KERNEL/USERLAND permutations even just enumerating |
50 |
> the patches might end up being a tad hairy. There was also a thread |
51 |
> suggesting that packages should not, in general, be marked stable unless |
52 |
> it is supported on all archs (or whatever). |
53 |
> |
54 |
> Drake Wyrm noted that the proposed system allowed people to test new |
55 |
> permutations that might just work. That sort of ability is something we |
56 |
> should probably strive to maintain in whatever our ultimate solution |
57 |
> should turn out to be. |
58 |
> |
59 |
> Rac has suggested a simple, if rather draconian, solution, which is to |
60 |
> just let the explosion of keywords to occur, but get them out of the |
61 |
> ebuilds. His suggestion is to provide a database (maintained external |
62 |
> to the ebuilds, perhaps even external to the users' systems) that uses a |
63 |
> five-dimensional (arch/os/libc/kernel/ebuild-version) key to determine |
64 |
> whether or not an ebuild should be built. |
65 |
> |
66 |
> I would prefer a somewhat more elegant approach than rac's brute-force |
67 |
> database, if I could only think of one. Any other thoughts out there? |
68 |
-- |
69 |
Olivier CrĂȘte |
70 |
tester@g.o |
71 |
Gentoo Developer |