1 |
Wow, that's some kind of thread you started... :) I'll respond in |
2 |
general to a bunch of stuff on this list by topic. |
3 |
|
4 |
|
5 |
COUNCIL MEETING |
6 |
|
7 |
On Sat, Nov 17, 2012 at 10:29 PM, Greg KH <gregkh@g.o> wrote: |
8 |
> |
9 |
> So, that's a nice summary, but, what is the end result here? |
10 |
> |
11 |
|
12 |
Speaking as somebody who was there, but not for the council, the |
13 |
summary was the end result OF THE COUNCIL MEETING. |
14 |
|
15 |
The council was asked to set a deadline for everybody with a separate |
16 |
/usr to adopt one of the proposed mitigation solutions, like using a |
17 |
script, initramfs, or whatever. That is ALL that it was asked to |
18 |
decide on, and that was all it did decide on. The whole business |
19 |
about some devs wanting to fork udev came out about a day in advance, |
20 |
and speaking personally it only had a little influence on my vote. |
21 |
|
22 |
The reason I agree with chainsaw's proposal to defer the decision one |
23 |
month was that there seemed to be enough blockers on this that nothing |
24 |
was going to happen for almost another month anyway (best-case), and |
25 |
getting people to move to initramfs or mdev or |
26 |
[nu/eu]dev[-ng]/whatever wasn't actually going to be holding anything |
27 |
up for a while. I'd also have been willing to approve a plan to set a |
28 |
target for something like 90 days after all the necessary tools (like |
29 |
genkernel) were stable and news was sent out. Based on my questions |
30 |
for williamh I did not get the sense that delaying a month was |
31 |
actually hindering the udev project (the established udev). They were |
32 |
encouraged to continue working on their blockers, preparing news |
33 |
items, and so on - everything but having a deadline/go-ahead to break |
34 |
systems that didn't follow the news. |
35 |
|
36 |
So, a bunch of ideas were floating around in the meeting, and I |
37 |
embraced the wait a month option since that seemed to have the most |
38 |
support of any of the options out there. If williamh had identified |
39 |
some actual impact of delay on the udev team I'd have probably pushed |
40 |
for setting the deadline now, but just putting it far enough out there |
41 |
(90 days from genkernel/etc being ready) that all the various teams |
42 |
would have a shot at it. If the udev team gets their news items all |
43 |
worked out and perhaps even sent out (sans deadline) and all the |
44 |
blockers cleared before the next meeting I'd be supportive of setting |
45 |
the deadline around 60 days, but that would be just moral support |
46 |
since I'm not on the council. |
47 |
|
48 |
|
49 |
OFFICIAL UDEV PROJECT |
50 |
|
51 |
I have nothing to do with the new udev project, but I did pass the |
52 |
staff quiz with much help from calchan. :) |
53 |
|
54 |
Read the GLEPs - any Gentoo developer can start a project at any time. |
55 |
That's how things work around here. If I wanted to start a linux |
56 |
kernel fork as an official Gentoo project I could do so tomorrow. |
57 |
|
58 |
That doesn't mean that the new udev will become the default udev, any |
59 |
more than Gentoo hardened will ever become the default experience for |
60 |
new Gentoo users. Gentoo is about choice, and if we have devs |
61 |
interested in maintaining something new then we'll offer that choice |
62 |
to our users for as long as somebody takes care of it. |
63 |
|
64 |
If anybody wants to change the defaults/etc, I'd expect that to get a |
65 |
lot of discussion, and almost certainly a council vote. |
66 |
|
67 |
|
68 |
COPYRIGHT |
69 |
|
70 |
I think this issue is best dealt with on the side - it has no bearing |
71 |
on any of the really contentious points here. |
72 |
|
73 |
I note that the owners of the copyright on udev have announced to the |
74 |
world that (emphasis mine): |
75 |
You may modify your copy or copies of the Library or ANY PORTION OF |
76 |
IT, thus forming a work based on the Library, and copy and distribute |
77 |
such modifications or work under the terms of Section 1 above, |
78 |
provided that you also meet all of these conditions... |
79 |
|
80 |
None of those conditions included keeping the copyright line intact. |
81 |
|
82 |
Anybody can therefore alter the copyright line as they wish, as they |
83 |
have been given explicit permission to do so. They need only comply |
84 |
with the other terms in the LGPL to do so (the most important being |
85 |
licensing it under the LGPL and making the source available. |
86 |
|
87 |
In fact, (L)GPL v3 has an optional attribution clause, and the fact |
88 |
that they made this explicit is because some projects might not want |
89 |
to give out this authorization. |
90 |
|
91 |
So, if you want an official ruling from the trustees we would need to |
92 |
meet/vote on it and perhaps discuss with counsel, but my thinking is |
93 |
that anybody distributing work under the (L)GPL has waived their right |
94 |
to be named on the copyright line of any copies distributed by others, |
95 |
and as far as I can tell I have found nothing to the contrary from any |
96 |
authoritative source. The only way I think you could argue that |
97 |
removing copyright notices for a (L)GPL work is illegal is if you |
98 |
argue that an author doesn't have the legal power to license that |
99 |
right to another. However, I'd still think that promissory estoppel |
100 |
would probably interfere with any kind of recourse - you can't give |
101 |
somebody permission to do something, and then sue them for actually |
102 |
doing it. So, legal or not anybody with standing to sue over this has |
103 |
likely given up their rights to do so. |
104 |
|
105 |
Again, that's my two cents and not a license for anybody to do |
106 |
anything. This topic did come up recently with regard to accepting |
107 |
some other kind of outside work into Gentoo, and as I recall there was |
108 |
some debate over whether the copyright notices could be changed. I'd |
109 |
have to dig up the details - I think the issue might have been mooted |
110 |
before any kind of formal decision was reached... |
111 |
|
112 |
|
113 |
IS THE NEW UDEV A GOOD/BAD/UGLY IDEA |
114 |
|
115 |
Seems like this is the main point of this whole thread, and I don't |
116 |
find it terribly useful to harp on. If people want to start a udev |
117 |
fork more power to them. I'm supportive of that, just as I'm |
118 |
supportive of having systemd in the tree and unit files for as many |
119 |
packages as possible. As projects mature I'd be all for offering them |
120 |
as options in the handbook. Gentoo is about choice. |
121 |
|
122 |
Ditto for a /usr move or whatever else. I think we should have a |
123 |
reasonable default behavior, and as others have pointed out we could |
124 |
use profiles to control a bunch of these behaviors globally (like |
125 |
library install location, and so on). Again, offer the user the |
126 |
choice, and generally be conservative with the defaults. |
127 |
|
128 |
Will all these projects go the distance? Hard to say. I agree that |
129 |
projects inspired by "hate" or whatever often fizzle out, but several |
130 |
high profile forks have stuck around - usually because of conflict |
131 |
over the major goals of the project. The next few years should be |
132 |
interesting, as the amount of vertical integration seems to be |
133 |
creeping up, and if a major Gnome release is systemd-only or whatever |
134 |
that could really bring things to a head. This almost seems like an |
135 |
androidification of the traditional linux distro - where choice still |
136 |
exists but the ability to swap out layers starts to go away unless you |
137 |
stick with a lighter desktop environment. Should make things fun for |
138 |
the toolkit developers when a system might or might not have |
139 |
dbus/systemd/udev/X11/wayland/linux/bsd and who knows what else. |
140 |
|
141 |
|
142 |
Rich |