1 |
On Tue, Feb 9, 2016 at 1:25 PM, Alon Bar-Lev <alonbl@g.o> wrote: |
2 |
> |
3 |
> I also mask all systemd files that somehow find their way into my |
4 |
> system thanks to your past decisions as systemd supporter, instead of |
5 |
> installing these only when systemd USE flag is set and adding openrc |
6 |
> USE flag for the systemd users folks. |
7 |
> |
8 |
|
9 |
FWIW those decisions largely predate me, or at least my position on |
10 |
the Council. They're still decisions I support for a few reasons, |
11 |
which I'll outline because I think they add value: |
12 |
|
13 |
1. Most of this concerns text files, such as systemd units or init.d |
14 |
scripts. Their installation doesn't have much practical impact on the |
15 |
operation of a system which doesn't use them, and having them around |
16 |
makes it easier for users to switch implementations either way without |
17 |
reinstalling half of their system. These are maskable in any case |
18 |
fairly trivially. |
19 |
|
20 |
2. I've also seen concerns when upstream renames paths to systemd-* |
21 |
for things that used to be stored someplace else. While I know the |
22 |
name tends to concern people, I think that it is beneficial to stick |
23 |
to upstream paths because it doesn't add value to depart from them. |
24 |
One of the things I like about Gentoo is that we follow upstream |
25 |
closely, which means that if you use openrc you have init.d scripts |
26 |
written for openrc and not LSB, and if you use systemd you have |
27 |
systemd units and not wrappers around LSB. |
28 |
|
29 |
I think these are both compelling reasons to maintain the current |
30 |
policy, even if many don't like the direction upstream is going in the |
31 |
case of systemd. As is the case with eudev if upstream ticks off too |
32 |
many people it can be forked. |
33 |
|
34 |
> It may in future that udev will completely relay on systemd and we |
35 |
> left with eudev as the only choice for non systemd-eco system, it is |
36 |
> better to start building this eco-system now, make sure we have a |
37 |
> working solution at any given point in time. |
38 |
|
39 |
I think you'll find that the eudev maintainers have largely created |
40 |
that eco-system already, though others can judge how well they did. |
41 |
This is really just about changing a default, and will not on its own |
42 |
make the eudev experience any better or worse, or more supported or |
43 |
less supported. Largely that is going to be the result of decisions |
44 |
made by many upstream communities, and whether those who want to have |
45 |
strong eudev support in Gentoo help deal with any issues that arise. |
46 |
|
47 |
-- |
48 |
Rich |