1 |
On Sat, 24 Feb 2007 13:51:09 +0900 Jason Stubbs <jstubbs@g.o> |
2 |
wrote: |
3 |
| So with your DEPEND="|| ( tetex ptex )" case, you're saying that it |
4 |
| is valid because the choice of tetex or ptex doesn't affect the |
5 |
| resultant binaries? Extrapolating that, specifying link dependencies |
6 |
| within an || construct is flat out wrong? |
7 |
|
8 |
Yep. |
9 |
|
10 |
| If so, it's way out of my domain so I can't really comment other than |
11 |
| you haven't given a reason for this requirement. |
12 |
|
13 |
The reason is that that's all Portage and the current VDB and |
14 |
binary formats permit. |
15 |
|
16 |
| Having said that, I'll accept it if we're strictly talking in the |
17 |
| EAPI-0 domain as there is no way for a package manager to guarantee a |
18 |
| safe --depclean implmentation given that raw *DEPENDs are stored in |
19 |
| the current installed package database. |
20 |
|
21 |
Right. Even as a >EAPI-0 requirement, the way it is now (slots |
22 |
excluded) makes sense because it's the easiest way of guaranteeing |
23 |
consistency of binary packages. Users should never (excluding |
24 |
blockers, which are a different issue) be forced to uninstall a package |
25 |
just so that a binary builds in a particular way. |
26 |
|
27 |
-- |
28 |
Ciaran McCreesh |
29 |
Mail : ciaranm at ciaranm.org |
30 |
Web : http://ciaranm.org/ |
31 |
Paludis, the secure package manager : http://paludis.pioto.org/ |