1 |
On Tue, Nov 21, 2006 at 12:56:22AM +0100, Marius Mauch wrote: |
2 |
> Motivation |
3 |
> ========== |
4 |
<snip> |
5 |
> Furthermore, some software requires that a user interactively accept its |
6 |
> license for its author's to consider it legally binding. This is |
7 |
> currently implemented using ``eutils.eclass``. |
8 |
|
9 |
This seems to *imply* ACCEPT_LICENSE will replace this mechanism. Is |
10 |
that the case? If it is, perhaps it could explicitly say so here? |
11 |
|
12 |
<snip> |
13 |
> Backwards Compatibility |
14 |
> ======================= |
15 |
> |
16 |
> There should be no change to the user experience without the user |
17 |
> explicitly choosing to do so. This mandates that the |
18 |
> configuration variable be named ``ACCEPT_LICENSE`` as some users may |
19 |
> already have it set due to ebuilds using ``eutil.eclass``'s |
20 |
> implementation. It also mandates that the default ``ACCEPT_LICENSE`` be |
21 |
> set to ``@NON-MUST-HAVE-READ`` in the main gentoo repository as implementations |
22 |
> are not required to provide an internal default. |
23 |
|
24 |
At first I thought this paragraph was incorrect. But when I looked at |
25 |
the check_license code in eutils.eclass I noticed it also uses an |
26 |
ACCEPT_LICENSE var. I never realised it did this since I only |
27 |
encountered check_license interactively and it did not tell me about |
28 |
ACCEPT_LICENSE then. |
29 |
|
30 |
Will portage export ACCEPT_LICENSE with the groups, wildcards and |
31 |
package.license entries expanded? If I try to emerge something with a |
32 |
license not in NON-MUST-HAVE-READ, then accept the license by setting |
33 |
it to "*" or including a group including this license, then merge the |
34 |
ebuild, I would expect "check_license" to treat the license as |
35 |
"accepted". |
36 |
|
37 |
Since check_license was (I assume) originally added because it was |
38 |
required for certain (mostly games) ebuilds: is the possibility to |
39 |
accept the license by putting a wildcard or group in ACCEPT_LICENSE |
40 |
"compatible" with those licenses? If it is not this would need some |
41 |
more thought: it would be quite confusing if certain licenses did not |
42 |
follow the same "rules" for groups and wildcards as other licenses, or |
43 |
if portage followed different rules at resolve time than check_license |
44 |
in eutils does. |
45 |
|
46 |
-- |
47 |
Marien. |