1 |
On Fri, 10 Feb 2006 01:11:00 +0100 "Diego 'Flameeyes' Pettenò" |
2 |
<flameeyes@g.o> wrote: |
3 |
| On Friday 10 February 2006 00:50, Ciaran McCreesh wrote: |
4 |
| > Provision should be made for future ports that require more than two |
5 |
| > keywords. There's no particular reason to artificially limit this to |
6 |
| > two at this stage. |
7 |
| |
8 |
| There's a reason: simplicity. More than two part keywords simply are |
9 |
| difficult to read, handle and are proving themselves pointless. |
10 |
| We just need to identify an operating system (or a variant of it) not |
11 |
| a whole complex of dependencies. |
12 |
|
13 |
Mmm. Do you consider the cost of slightly simplified wording in the |
14 |
GLEP now to be worth extra effort that may be required in the future by |
15 |
some especially whacked out port? I'd prefer to see at least a note in |
16 |
the GLEP stating that three part keywords are legal if they ever become |
17 |
necessary. |
18 |
|
19 |
| > Examples of how this lot is to be used in DEPEND= etc would be good |
20 |
| > for clarity. |
21 |
| |
22 |
| Hmm well they are already being used, anyway, the following comes |
23 |
| from rpm2targz: |
24 |
|
25 |
Yeah, *I* know how they work. I suspect that many people reading the |
26 |
GLEP, however, are going to go "Huh? Why do we need this lot at all?". |
27 |
|
28 |
| > You should clarify the env map behaviour when multiple matches all |
29 |
| > occur (first match, last match or merge?). |
30 |
| |
31 |
| Never thought about that actually, good point. I think last match |
32 |
| would be good. |
33 |
|
34 |
Then your example code is wrong :) |
35 |
|
36 |
-- |
37 |
Ciaran McCreesh : Gentoo Developer (Wearer of the shiny hat) |
38 |
Mail : ciaranm at gentoo.org |
39 |
Web : http://dev.gentoo.org/~ciaranm |