1 |
On 24/07/2003 at 10:28:57(+0200), Sven Vermeulen used 1.6K just to say: |
2 |
> On Wed, Jul 23, 2003 at 12:16:03PM -0400, Matt Rickard wrote: |
3 |
> > I'm wondering what's the best way to handle these? First option is to |
4 |
> > put them in vanilla-sources, for example, |
5 |
> > vanilla-sources-2.4.22_pre7.ebuild. If we ~ mask them, people using the |
6 |
> > unstable profile will get these prerelease kernels if merging |
7 |
> > vanilla-sources. I'm not sure this is desired behavior, however, it IS |
8 |
> > the unstable profile :) |
9 |
> |
10 |
> Its more like the testing. ~arch is when a developer deems his package ready |
11 |
> for inclusion. |
12 |
> |
13 |
> > Another option is to create a separate category for these, something |
14 |
> > like pre-vanilla-sources. This has the benefit that people who want |
15 |
> > vanilla will get only vanilla, and not a prerelease -- regardless of |
16 |
> > which profile they are using. |
17 |
> |
18 |
> I don't like this idea (even though I'm in the minority here :). First of |
19 |
> all, when the _pre's are finished and the stable kernel is released, people |
20 |
> which are using pre-vanilla-sources will be stuck at the latest _pre (or |
21 |
> _testing). |
22 |
> |
23 |
> And when they emerge the vanilla-sources, they'll miss the next _pre cycle. |
24 |
> |
25 |
> Personally, I would go for package.mask |
26 |
|
27 |
Wow... I was getting scared seeing how this thread developed, but in the end, |
28 |
after reading this post, I realized that the world has not gone *completely* |
29 |
mad, after all. |
30 |
|
31 |
I fully support the opinions stated above, and I simply cannot comprehend what |
32 |
the big deal with ~arch masking vanilla-sources is. Even *considering* the |
33 |
option of a separate package is ridiculous. As long as developers are careful |
34 |
enough to not remove the ~arch mask from any _pre kernel, I am perfectly fine, |
35 |
and I believe there wouldn't be anyone who isn't fine. What are we trying to do |
36 |
-- make sure people who insist on running the *unstable* profile actually don't |
37 |
get the "unstable" sources. There is also the package.mask if you feel that is |
38 |
not enough, but this was also mentioned already. And the decision that was |
39 |
taken is ... weird. What is the idea in having unstable (i.e. ~masked) |
40 |
packages in the first place? Are you going to keep only stable versions in |
41 |
vanilla-sources? What's the point? Why not move all _pre, _alpha ane _beta |
42 |
versions of packages in separate directories? As it was already mentioned, _pre |
43 |
kernels are more stable than many other packages. The first one I can think of |
44 |
is gentoo-sources, that insisted on corrupting my filesystem every now and |
45 |
then, so I couldn't upgrade my glibc, without upgrading to vanilla-sources |
46 |
first (some files were having funny contents during compilation but it was hard |
47 |
to reproduce), and I am running vanilla-sources ever since. |
48 |
|
49 |
Sorry for the tone, but I feel frustrated. |
50 |
|
51 |
-- |
52 |
() Georgi Georgiev () A committee is a life form with six or more () |
53 |
() chutz@×××.net () legs and no brain. -- Lazarus Long, "Time () |
54 |
() +81(90)6266-1163 () Enough For Love" () |