1 |
On 17/9/2005 11:34:56, Brian Harring (ferringb@g.o) wrote: |
2 |
> On Sat, Sep 17, 2005 at 11:28:03AM +0200, Kevin F. Quinn wrote: |
3 |
> > The 30-day could be calculated from the $Header: of ebuilds that have |
4 |
> > no UNSTABLE, or where it's empty. |
5 |
> |
6 |
> Doesn't work for N arches keywording, or ebuild dev doing minor |
7 |
> syntax touch ups. |
8 |
|
9 |
Good point. The minor touch-up issue could be resolved by setting |
10 |
the string to the date the last issue was cleared instead of deleting |
11 |
it: |
12 |
|
13 |
UNSTABLE="2005/10/04" |
14 |
|
15 |
but to handle N arches needs a different approach (the 'maint' keyword |
16 |
idea also falls down here). |
17 |
|
18 |
|
19 |
My favourite idea so far is mike's '?arch' on the understanding that |
20 |
we have: |
21 |
|
22 |
package.mask - 'alpha' |
23 |
Not suitable for mainstream testing |
24 |
|
25 |
?arch - 'beta' |
26 |
Works on maintainers systems, worth testing |
27 |
Maintainer may not have tried it on arch. |
28 |
|
29 |
~arch - 'release candidate' |
30 |
Maintainer & arch team happy that it's a good candidate for arch |
31 |
30-day maturity phase, arch testing in progress |
32 |
|
33 |
arch - 'released' |
34 |
Arch team happy it's stable |
35 |
|
36 |
|
37 |
In particular it's worth noting that marking ?arch is not restricted |
38 |
the way marking ~arch is. Over time I expect the x86 arch team to |
39 |
impose more rigour on the use of ~x86, so that it behaves similarly |
40 |
to the other arches. |
41 |
|
42 |
In general, it would make sense for people to have arch or ~arch in |
43 |
make.conf, and use package.keywords to grab stuff from ?arch in a |
44 |
controlled fashion. |
45 |
|
46 |
-- |
47 |
gentoo-dev@g.o mailing list |