1 |
Joel Konkle-Parker posted <cavtfr$2a7$1@×××××××××.org>, excerpted below, |
2 |
on Fri, 18 Jun 2004 19:22:31 -0400: |
3 |
|
4 |
> My understanding of the portage system is that when an ebuild is made |
5 |
> and is determined to be stable enough, it is put into ~arch for testing. |
6 |
> Once it's been there for a while, and there are no major problems, it |
7 |
> gets manually bumped to arch. |
8 |
[] |
9 |
> It seems like [Debian's] system is more efficient, since it doesn't rely |
10 |
> on someone manually deciding that it's time to push $package to stable. |
11 |
> |
12 |
> Is there a reason for this, or has this been considered? |
13 |
|
14 |
The recommended time at ~ is a month, on Gentoo, but it's left up to the |
15 |
individual developer/herd responsible for the ebuild. Additionally, the |
16 |
various arch teams can mark it stable for their arch. |
17 |
|
18 |
There is, however, a bot mailer in place, I am told, that mails reminders |
19 |
to those responsible, if an app remains in ~ for to long (a month or two, |
20 |
don't recall). In addition, users can file bugs on an app saying it works |
21 |
and asking that it be moved to stable for their arch, if there are no |
22 |
serious bugs. |
23 |
|
24 |
This would seem rather safer than a set period, since for example x86 gets |
25 |
far more testing and therefore would logically advance out of ~ far faster |
26 |
than sparc or amd64, for instance. (I'm on the latter, here, as just a |
27 |
user, and a fairly new one at that, but I do follow this list.) |
28 |
|
29 |
-- |
30 |
Duncan - List replies preferred. No HTML msgs. |
31 |
"They that can give up essential liberty to obtain a little |
32 |
temporary safety, deserve neither liberty nor safety." -- |
33 |
Benjamin Franklin |
34 |
|
35 |
|
36 |
|
37 |
-- |
38 |
gentoo-dev@g.o mailing list |