1 |
On 02/17/2016 01:47 PM, Rich Freeman wrote: |
2 |
> On Wed, Feb 17, 2016 at 11:04 AM, Richard Yao <ryao@g.o> wrote: |
3 |
>> |
4 |
>> This is something that I think many of us who had systems broken by |
5 |
>> sys-fs/udev multiple times before sys-fs/eudev was an option thought was |
6 |
>> obvious. |
7 |
> |
8 |
> About the only "system-breaking" change I'm aware of in udev over the |
9 |
> years was the change in default network interface names. That was |
10 |
> preceded by news on how to avoid the change, or how to adapt to the |
11 |
> change. |
12 |
> |
13 |
> There certainly wasn't some change introduced without warning that |
14 |
> just broke systems in random ways. |
15 |
> |
16 |
>> If a complete list of the breakages that lead to the creation of |
17 |
>> sys-fs/eudev were produced, I imagine that the list would have at least |
18 |
>> 3 to 5 items from the ~18 months before sys-fs/eudev with half of them |
19 |
>> were probably self inflicted by sys-fs/udev maintenance. |
20 |
> |
21 |
> Anytime upstream changes something it is up to package maintainers to |
22 |
> evaluate the change and adapt to it as needed, especially for critical |
23 |
> packages like udev. For the most part I think this is happening. |
24 |
> Whether it is the udev maintainers doing the QA or the eudev |
25 |
> maintainers doing the QA, somebody has to do the QA (and in Gentoo it |
26 |
> sounds like we do it twice, which is fine). |
27 |
> |
28 |
>> I recall one incident involving whether udev should be in /sbin or |
29 |
>> /usr/sbin being resolved after 6 months of debate between then future |
30 |
>> eudev founders and sys-fs/udev maintainers only because the systemd |
31 |
>> developers told the sys-fs/udev maintainers it should be in /sbin like |
32 |
>> others had told them. |
33 |
> |
34 |
> So, this sounds like a disagreement between the future eudev founders |
35 |
> and the udev maintainers in Gentoo. It really has nothing to do with |
36 |
> udev itself. |
37 |
|
38 |
That is just one thing that I remembered off the top of my head and |
39 |
quite frankly, that situation was the most absurd system breakage |
40 |
incident that ever occurred involving udev. The arguments by the |
41 |
sys-fs/udev maintainers at the time were that upstream wanted it that |
42 |
way when even the systemd developers did not agree with them. The matter |
43 |
was only resolved after one of the sys-fs/udev maintainers decided to |
44 |
ask the systemd developers what they actually thought after 6 months of |
45 |
claiming their way was the upstream way and were told that they thought |
46 |
the packaging was wrong. Everyone else had believed the sys-fs/udev |
47 |
maintainers' statements that upstream actually thought such things, and |
48 |
consequently, were adamant that the systemd maintainers had no idea what |
49 |
they were doing. |
50 |
|
51 |
There were multiple breakages caused by sys-fs/udev maintainance |
52 |
following systemd assimilating udev based on the principle that upstream |
53 |
wanted it that way. The outsourcing of responsibility for thought on |
54 |
such things were one of the things that contributed to eudev's creation. |
55 |
Another were absolute refusal by Kay Sievers at the time to fix |
56 |
regressions when given patches. It was not "rewrite the patch this way". |
57 |
It was along the lines of "we are not fixing that because we don't care |
58 |
about people affected by this and fixing that would add 40 lines to the |
59 |
codebase". |
60 |
|
61 |
> And that is OK - we're allowed to have the same package maintained |
62 |
> under two different names by two sets of maintainers. Obviously it |
63 |
> isn't ideal, and it would be better if everybody could agree. |
64 |
> |
65 |
>> Another broke support for older kernels for no apparent benefit (and |
66 |
>> this sort of regression naturally enters sys-fs/udev): |
67 |
> |
68 |
> That isn't really the same as "breaking Gentoo." And as was pointed |
69 |
> out they did accept patches to provide support back. |
70 |
|
71 |
That is a new development. |
72 |
|
73 |
> I think it is a bit unfair to characterize the udev maintainers as |
74 |
> breaking things left and right. They apparently differ with you in |
75 |
> how they prefer to set their defaults, and what their dependencies |
76 |
> are. They apparently also accept patches when you provide them for |
77 |
> older kernels, which is what upstreams are supposed to do. |
78 |
|
79 |
There were multiple incidents where either the sys-fs/udev maintainers |
80 |
or the systemd developers refused to listen to reason. Systems broke |
81 |
because of it and there were no warnings. |
82 |
|
83 |
> It really seems like the main reasons for eudev's existence right now |
84 |
> are (based on this thread): |
85 |
> 1. The eudev maintainers don't trust the udev maintainer's QA and |
86 |
> want to do their own layer of QA before introducing changes to the |
87 |
> tree. |
88 |
> 2. The eudev maintainers prefer a different default network interface |
89 |
> naming scheme (my understanding is that eudev can be configured to |
90 |
> behave as udev does by default, and vice-versa - for example, on the |
91 |
> box I'm typing this on my packets are going out over eth0 just fine on |
92 |
> systemd). |
93 |
> 3. The eudev tarballs are smaller lacking the systemd components, and |
94 |
> the udev build system doesn't have to build the systemd components (I |
95 |
> don't think the same is true of udev but I could be wrong). |
96 |
> |
97 |
> I'm not saying that eudev should go away, or that these concerns are |
98 |
> completely inappropriate. If somebody wanted to fork their own kernel |
99 |
> stable series and carefully curate patches they could choose to do so |
100 |
> and package it in Gentoo, and give it different default configuration |
101 |
> options, and so on. |
102 |
|
103 |
That is a false equivalence because: |
104 |
|
105 |
1. We have plenty of sys-kernel/*-sources packages, but Gentoo has no |
106 |
true default because the user must pick a package. |
107 |
2. Both mainline and the Gentoo kernel team do a decent job of not |
108 |
breaking things and fix them promptely when things do break. There are |
109 |
no drawn out arguments or debates about why things should be broken in |
110 |
lieu of fixing things when them being broken has been recognized. |