1 |
On 09/12/2017 02:47, Alexey Eschenko wrote: |
2 |
> Except that fact that I didn't unmasked it. |
3 |
|
4 |
You must have had a tree checkout slightly newer than mine. I just |
5 |
synced here and see that the mask has now been removed. |
6 |
|
7 |
It's quite unusual to unmask an alpha version, maybe raise it on b.g.o ? |
8 |
|
9 |
For the rest, that's just how blockers go unfortunately. There is no |
10 |
easy way for the maintainer to communicate to you at emerge time *why* |
11 |
the blocker is there, you just see the effect that it *is* there. |
12 |
|
13 |
It's proper to block package B if new version of package A provides the |
14 |
same features and they collide. But portage is stuck with nowhere to go |
15 |
if you happen to have package B in world. |
16 |
|
17 |
>> # fgrep -rni blueman /etc/portage |
18 |
>> /etc/portage/package.use/blueman:1:#net-wireless/blueman |
19 |
> But I understand other possible reasons. |
20 |
> |
21 |
> On 12/08/2017 07:37 PM, Alan McKinnon wrote: |
22 |
>> On 08/12/2017 15:22, Alexey Eschenko wrote: |
23 |
>>> It can be the issue. But older version (2.0.4) which is currently |
24 |
>>> installed works fine and has no conflicts. |
25 |
>>> |
26 |
>>> It's quite strange. |
27 |
>>> |
28 |
>>> |
29 |
>>> On 12/08/2017 03:39 PM, Vadim A. Misbakh-Soloviov wrote: |
30 |
>>>>> Is it really necessary to block one package when another installed? |
31 |
>>>> Most of the time, the reason to make packages to block each other is |
32 |
>>>> collisions (if they they contain files (like binaries or libraries) |
33 |
>>>> with same |
34 |
>>>> install paths). |
35 |
>>>> |
36 |
>>>> Although, I can't guarantee that it was the case here. |
37 |
>>>> |
38 |
>>>> I've noticed that Gnome Team makes some decisions, that doesn't looks |
39 |
>>>> logical |
40 |
>>>> for a few times already. |
41 |
>>>> |
42 |
>> |
43 |
>> It's not at all strange; it's quite ordinary actually. |
44 |
>> |
45 |
>> Keeping in mind that I do not use these packages, or gnome, look at the |
46 |
>> available blueman packages: |
47 |
>> |
48 |
>> # eix net-wireless/blueman |
49 |
>> * net-wireless/blueman |
50 |
>> Available versions: (~)2.0.3 (~)2.0.4 [M](~)2.1_alpha1 **9999 |
51 |
>> {appindicator network nls policykit pulseaudio thunar |
52 |
>> PYTHON_SINGLE_TARGET="python2_7 python3_4 python3_5 python3_6" |
53 |
>> PYTHON_TARGETS="python2_7 python3_4 python3_5 python3_6"} |
54 |
>> |
55 |
>> 2.1 is still in an alpha state, and it is p.masked: |
56 |
>> |
57 |
>> /var/portage/profiles/package.mask: |
58 |
>> # Michał Górny <mgorny@g.o> (26 Jan 2017) |
59 |
>> # Pre-release, masked for testing. Major changes since 2.0.4, |
60 |
>> # including dropped support for BlueZ 4. |
61 |
>> |
62 |
>> It is not unreasonable to conclude that blueman-2.1 intends to add |
63 |
>> features that conflict with gnome-bluetooth and they can't co-exist. As |
64 |
>> Vadim said, file collisions are often the underlying cause. |
65 |
>> |
66 |
>> You unmasked an alpha package, clearly tagged as "for testing". Nothing |
67 |
>> add abut the result you got at all. |
68 |
>> |
69 |
>> |
70 |
>> |
71 |
> |
72 |
|
73 |
|
74 |
-- |
75 |
Alan McKinnon |
76 |
alan.mckinnon@×××××.com |