1 |
On Wed, Feb 04, 2004 at 12:45:46PM +0100, Matt Steven wrote: |
2 |
> Maybe I'm putting the carriage before the horse, but I don't see how having a |
3 |
> separate tree would be as efficient as having a 'x86-server' keyword for |
4 |
> example added to all server-stable packages and security updates. |
5 |
|
6 |
The problem would be people deleting things when they shouldn't. If we |
7 |
are to provide an enterprise system it needs to be absolutely solid. I'd |
8 |
worry about new developers "cleaning" the stable ebuilds. While this |
9 |
shouldn't happen and needs to be fixed in the main tree anyway I really |
10 |
think that its a good plan to have people specifically "trained" to look |
11 |
after the enterprise tree to be the only ones able to add or remove |
12 |
ebuilds from it. |
13 |
|
14 |
> In any case it's obvious there's a demand for a more stable portage option, |
15 |
> however it's implemented. I think the keyword approach might be the simplest |
16 |
> because it doesn't take away any functionality, and minimizes bloat. |
17 |
|
18 |
Well whether it minimizes bloat depends on which tree you plan to use. |
19 |
We're talking about adding ebuilds (because remember these ebuilds will |
20 |
be there for a year, long after newer and greater versions are in |
21 |
portage) to the normal gentoo tree which increases bloat for people who |
22 |
don't plan to use that tree. |
23 |
|
24 |
Also, people who do plan to use only the stable stuff will have an |
25 |
absolute shedload of ebuilds they need to store/parse etc even though |
26 |
they will never use them. |
27 |
|
28 |
So I'm not sure bloat is a valid reason either way. |
29 |
|
30 |
-- |
31 |
rob holland - [ tigger@g.o ] |
32 |
irc://irc.freenode.net/#gentoo as tigger^ |
33 |
http://dev.gentoo.org/~tigger/tigger@××××××××××.asc |