1 |
On Sat, 2007-03-10 at 01:48 +0000, Steve Long wrote: |
2 |
> Thomas Rösner wrote: |
3 |
> >> Once we have USE-based dependencies across the board, then yes. Until |
4 |
> >> that time, we should really be building both client and server for *all* |
5 |
> >> packages. |
6 |
> > |
7 |
> > I can understand that rationale for the client part, but which packages |
8 |
> > would depend on the server part of e.g. MySQL if they could? |
9 |
> > And building the server part to get the small client lib is a larger |
10 |
> > PITA than building the client lib to get the server, no? |
11 |
> > |
12 |
> > In other words: this is a sound argument against the client use flag, |
13 |
> > but I don't think it's quite as convincing regarding the server flag, |
14 |
> > which is more important IMHO. |
15 |
> > |
16 |
> Yeah on reflection I think just a server flag would be better. There's a |
17 |
> topic on the forum about it: |
18 |
> http://forums.gentoo.org/viewtopic-t-525893.html |
19 |
> |
20 |
> After all, people who want a server know they do. The default would thus be |
21 |
> simply to build the client end. This is obviously why there are seven |
22 |
> ebuilds including cvs doing this already with no complaints. |
23 |
|
24 |
I wouldn't have too much of a problem with only a "server" USE flag, but |
25 |
I'd still prefer things were held off until USE-based dependencies. If |
26 |
we did decide on a USE=server flag, then it should also likely have |
27 |
USE=dedicated merged with it so there's only one flag for enabling |
28 |
server builds. |
29 |
|
30 |
-- |
31 |
Chris Gianelloni |
32 |
Release Engineering Strategic Lead |
33 |
Alpha/AMD64/x86 Architecture Teams |
34 |
Games Developer/Council Member/Foundation Trustee |
35 |
Gentoo Foundation |