1 |
On 06/06/2014 16:34, Gevisz wrote: |
2 |
> After today's emerge-webrsync, I have found out that usual |
3 |
> # emerge --update --deep --with-bdeps=y --newuse --ask world |
4 |
> does not work, trying but being unable to emerge systemd. |
5 |
> |
6 |
> As I have found out, the reason for it was that upower |
7 |
> suddenly decided that it needs systemd but, for some |
8 |
> (lucky :) reason, it could not be emerged as my system |
9 |
> uses OpenRC. |
10 |
> |
11 |
> The latest news said the following: |
12 |
> |
13 |
> "UPower discontinued hibernate and suspend support in favor of systemd. |
14 |
> Because of this, we have created a compability package at |
15 |
> sys-power/upower-pm-utils which will give you the old UPower with |
16 |
> sys-power/pm-utils support back. |
17 |
> Some desktops have integrated the sys-power/pm-utils support directly |
18 |
> to their code, like Xfce, and as a result, they work also with the new |
19 |
> UPower as expected. |
20 |
> |
21 |
> All non-systemd users are recommended to choose between: |
22 |
> |
23 |
> # emerge --oneshot --noreplace 'sys-power/upower-pm-utils' |
24 |
> |
25 |
> or |
26 |
> |
27 |
> # emerge --oneshot --noreplace '>=sys-power/upower-0.99.0' |
28 |
> |
29 |
> However, all systemd users are recommended to stay with |
30 |
> sys-power/upower." |
31 |
> |
32 |
> However, that news did *not* say that without |
33 |
> # emerge --oneshot --noreplace 'sys-power/upower-pm-utils' |
34 |
> it is impossible to update the system even if you use xfce. |
35 |
|
36 |
|
37 |
The news item is a compromise as the problem you ran into is really a |
38 |
very minor one; it's the presence of systemd and two blocking upower |
39 |
packages that cause confusion. Extra points for the magic flame |
40 |
war-inducing trigger-word "systemd" - see the enormous threads here and |
41 |
on -dev for proof. |
42 |
|
43 |
So the news item clarifies that you are not being forced to use systemd |
44 |
and explains the alternatives in a concise manner. |
45 |
|
46 |
Once you have that done, anything that comes next is simple routine |
47 |
portage blockers which you are expected to know how to deal with, and |
48 |
are not at all worthy of mention in a news item. |
49 |
|
50 |
|
51 |
> Only after executing the last command, which installed |
52 |
> upower-pm-utils and unistalled upower, the |
53 |
> # emerge --update --deep --with-bdeps=y --newuse --ask world |
54 |
> worked as desired. |
55 |
> |
56 |
> upower was not in my world file, so I think it's a some kind of a bug. |
57 |
|
58 |
|
59 |
It's not a bug and upower is not supposed to be in world. It's a simple |
60 |
dep library that portage will pull in if you have any other packages |
61 |
that need it. Keep it and upower-pm-utils out of world so that Samuli's |
62 |
next planned phase of this change in UPower will go smoothly. |
63 |
|
64 |
The reason for all the confusion is due to how portage works internally. |
65 |
Briefly, it sees you have a choice between upower and upower-pm-utils |
66 |
and usually ends up picking upower. |
67 |
|
68 |
It's similar to virtuals where portage doesn't know what you want so |
69 |
just picks the first in the list. If you want the second, then you have |
70 |
to emerge it yourself first, stopping portage from deciding. |
71 |
|
72 |
As I said above, this simple package addition produces the magic |
73 |
flamewar trigger word "systemd" whcih always gets half the audience very |
74 |
very upset indeed. |
75 |
|
76 |
|
77 |
|
78 |
|
79 |
-- |
80 |
Alan McKinnon |
81 |
alan.mckinnon@×××××.com |