1 |
Yuri Vasilevski <yvasilev@g.o> posted |
2 |
20080827093427.36572652@×××××××××××××××××××.mx, excerpted below, on Wed, |
3 |
27 Aug 2008 09:34:27 -0500: |
4 |
|
5 |
> As Another example, the user might statically link bits of the exact |
6 |
> same library against a GPL-2 (not a GPL-2 or latter) program, just |
7 |
> because he is misinformed by portage that the program is GPL-2 and then |
8 |
> he gets into a legal problem. |
9 |
|
10 |
The original question didn't specify where the license change occurred, |
11 |
upstream, or whether upstream stayed the same, and the Gentoo change was |
12 |
simply correcting an earlier mistake. |
13 |
|
14 |
If the change occurred upstream, then the version of the code released as |
15 |
GPLv2 remains released at that, regardless of upstream changes; once |
16 |
released as GPLv2, that can't be revoked (tho GPLv3 or another license |
17 |
could be added), and there's no problem in this case because we are |
18 |
talking that same upstream version according to the question. This is |
19 |
how I read the original question, an upstream change. |
20 |
|
21 |
However, if we're correcting a Gentoo mistake on code that was never |
22 |
licensed as we said it was, then a bump should be mandatory, as will be |
23 |
removal of the previously licensed revision ASAP, because that revision |
24 |
was shipped under the wrong license, and the faster that's corrected and |
25 |
we're no longer violating the law, the better. |
26 |
|
27 |
In this latter case, I'd argue that the offending revision must be |
28 |
removed immediately, even if the new revision can't be stabilized fast |
29 |
enough and a downgrade or even missing dependency is forced. We had no |
30 |
right to be shipping the code licensed as it was, and if we end up |
31 |
breaking stuff by removing what wasn't ours to ship in the first place, |
32 |
so be it. |
33 |
|
34 |
However, if as stated in the original question, the only difference in |
35 |
the build is the license, then as someone else suggested, I believe a |
36 |
straight to stable policy should apply. If it does, then there won't be |
37 |
any breakage since the new revision will be stabilized in in place of and |
38 |
at removal of the old. |
39 |
|
40 |
-- |
41 |
Duncan - List replies preferred. No HTML msgs. |
42 |
"Every nonfree program has a lord, a master -- |
43 |
and if you use the program, he is your master." Richard Stallman |