1 |
On Sun, Jan 4, 2009 at 10:28 PM, David Leverton |
2 |
<levertond@××××××××××.com> wrote: |
3 |
> On Sunday 04 January 2009 16:48:38 Nirbheek Chauhan wrote: |
4 |
>> On the contrary, the reverse of what you say is true. A simple grep of |
5 |
>> the tree showed that: |
6 |
> |
7 |
> In how many of those ebuilds would the long form be |
8 |
> use1? ( cat/pkg[use2] ) |
9 |
> rather than |
10 |
> use1? ( cat/pkg[use2] ) !use1? ( cat/pkg ) |
11 |
> ? Obviously I didn't look through all the hits, but the former seems quite |
12 |
> common, and any possible shortcut would only save a few characters. |
13 |
> |
14 |
|
15 |
Well, to get that info, you'd have to do even more shell-foo :) |
16 |
I believe that KDE-related packages often have dependencies of the |
17 |
former sort; for example |
18 |
|
19 |
use imlib && built_with_use dev-libs/imlib X |
20 |
|
21 |
I've personally encountered use-deps of the type: |
22 |
|
23 |
use cairo && built_with_use category/cairo X |
24 |
use avahi && built_with_use net-dns/avahi gtk |
25 |
|
26 |
etc. |
27 |
|
28 |
Also, the same logic would apply to the existing pkg[use?] deps too. |
29 |
However, pkg[use1?use2] seem to be (at the very least) common in |
30 |
relation to pkg[use?] deps. |
31 |
|
32 |
-- |
33 |
~Nirbheek Chauhan |