1 |
On Thu, Aug 29, 2013 at 11:22 AM, Jack Morgan <jmorgan@g.o> wrote: |
2 |
> As I mentioned in the -dev ML, I don't think this is the right approach |
3 |
> to your concern. There should be a clear definition of what is expected |
4 |
> from an arch that is offically supported by Gentoo Linux. By offically |
5 |
> supportd I mean ARCH/stable keyworded. If an arch fails to meet those |
6 |
> requirements, then "demote" it to ~arch only status. This should be a |
7 |
> GLEP. Otherwise, you are asking others to base their decision on someones |
8 |
> perception. |
9 |
|
10 |
That seems like a reasonable question for the council to address. |
11 |
|
12 |
Personally my sense is that if a package has a STABLEREQ pending for |
13 |
60 days maintainers should be able to drop the stable version of the |
14 |
package at their discretion. This really should be a max though. If |
15 |
this is happening with any regularity then the arch should be dropped |
16 |
from stable keywords, unless some other arrangement is made (like |
17 |
@system-only stable - an arrangement with its own pros and cons). |
18 |
|
19 |
It probably makes sense for those calls to be made by the Council |
20 |
unless the arch project lead agrees. |
21 |
|
22 |
> If keywording an ARCH is a real concern, then Gentoo Linux should have |
23 |
> a long hard look as what it wants to support as a developer community. I |
24 |
> want to challange the council to take this as an opportunity to define |
25 |
> this. If developer resources are limited, then Gentoo Linux can't |
26 |
> support everything it has in the portage tree. |
27 |
> |
28 |
|
29 |
Developer resources would be limited even if we had a paid staff of |
30 |
20k developers. |
31 |
|
32 |
Unmaintained packages do get treecleaned. However, the issue here is |
33 |
with arch maintenance, not package maintenance. It isn't the |
34 |
responsibility of package maintainers to do arch testing, though many |
35 |
choose to do so with the agreement of the appropriate arch projects. |
36 |
|
37 |
Rich |