1 |
Mike Frysinger posted <200509161817.01181.vapier@g.o>, excerpted |
2 |
below, on Fri, 16 Sep 2005 18:17:01 -0400: |
3 |
|
4 |
>> We still have KEYWORDS="-*". Sure, I know many do not like it, and if |
5 |
>> something was decided in regards to it, I missed it, but it is generally |
6 |
>> seen as 'less severe' than a package.mask'd mask, and its local to the |
7 |
>> package, so should not get stale. |
8 |
> |
9 |
> that would address the 'arch teams marking ahead of maintainer' issue, but in |
10 |
> general, i think we need a testing mask of some sort separate from |
11 |
> package.mask where we can put things like modular X, new KDE betas, new GNOME |
12 |
> betas, e17 packages, etc... |
13 |
|
14 |
Exactly. I'm a full ~arch user, and regularly load packages such as gcc4 |
15 |
(including snapshots once in awhile, plus the accompanying glibc and |
16 |
binutils) and kde and xorg snapshots, as well as pre versions of |
17 |
baselayout and portage, at times. |
18 |
|
19 |
I'd DEFINITELY be more convenient if I could easily separate the security |
20 |
and confirmed system munching stuff in package.mask out from these testing |
21 |
packages. ? arch would be great, here, as it would mean I could simply |
22 |
package.keyword as necessary, rather than (1) package.unmask, then |
23 |
sometimes (2) copying to overlay, and (3) adding ~arch and redigesting so |
24 |
portage will work with it. However, a testing.mask and parallel |
25 |
testing.unmask in /etc/portage, would work fine as well, provided when |
26 |
they were used, ~archs were carried over, to prevent having to overlay the |
27 |
package simply to add the appropriate keyword. (Being amd64 and having |
28 |
some packages do amd64 conditionals, I don't like adding ~x86 or -* to |
29 |
package.keywords, so overlay it has to be.) |
30 |
|
31 |
The point has been made that snapshots/pres/rcs and the like should never |
32 |
make it to ~arch, because they are never reasonable candidates for |
33 |
arch-stable. Point taken. However, that's certainly far easier for |
34 |
testers such as myself, than having to move a hundred kde-split-pkgs to |
35 |
overlay and keyword them, to be able to test the latest kde snapshot, then |
36 |
do the exact /same/ thing a week or two later for the /next/ one. For |
37 |
devs and users alike, an upstream package testing area (to parallel ~arch |
38 |
which is effectively ebuild testing and stable candidate area) that was |
39 |
separate and distinct from known actively harmful package.mask, would be |
40 |
/very/ useful, giving both advanced users and devs a way to know with no |
41 |
doubt what was considered ready for testing of the upstream-package, as |
42 |
deployed on Gentoo. |
43 |
|
44 |
-- |
45 |
Duncan - List replies preferred. No HTML msgs. |
46 |
"Every nonfree program has a lord, a master -- |
47 |
and if you use the program, he is your master." Richard Stallman in |
48 |
http://www.linuxdevcenter.com/pub/a/linux/2004/12/22/rms_interview.html |
49 |
|
50 |
|
51 |
-- |
52 |
gentoo-dev@g.o mailing list |