1 |
On Thu, Dec 11, 2014 at 10:58:12AM +0000, Peter Humphrey wrote: |
2 |
> Hello list, |
3 |
> |
4 |
> Today when I ran my usual upgrade I ran into a problem. Portage told me it |
5 |
> had skipped net-libs/libkgapi:4 because of unsatisfied dependencies. All it |
6 |
> could tell me was that >=kde-base/kdepimlibs-4.14:4 required it. |
7 |
> |
8 |
> So I embarked on one of those fruitless cycles of adding packages to |
9 |
> package.keywords, one at a time as they caused portage to stop, until after |
10 |
> about 8 additions I hit the inevitable show-stopper. The mixing of versions |
11 |
> of parts of KDE resulted in both 4.14.3 and 4.12.5 of kdelibs being required |
12 |
> at the same time. |
13 |
> |
14 |
> It looks as though someone has allowed something from ~amd64 out of the wild |
15 |
> and into the peace of amd64. Does anyone have a clue to what it is? I |
16 |
> thought it might be chromium, since libkgapi is for handling Google |
17 |
> calendars and contacts, but eix tells me I already have the latest stable |
18 |
> version. |
19 |
> |
20 |
|
21 |
Hello Peter, |
22 |
|
23 |
if I'd be in your place I'd try to run emerge @world -uvDNta (the most |
24 |
important one here is the -t for showing the dependency tree, you could also |
25 |
play with --backtrack=30 to resolve possible issues with the dependency |
26 |
calculation), after you removed the keyword stuff you just added. |
27 |
From your fist statement (required by >=kde-base/kdepimlibs-4.14:4) I guess |
28 |
that you keyworded something that depends on kdepimlibs or kdepimlibs itself in |
29 |
the past. |
30 |
Another option would be to try equery d kdepimlibs : it should show everything |
31 |
that depends on it (and you can crosscheck that with your keywords). |
32 |
As a final straw (or first option - depending on you preferences) you could |
33 |
resync the tree: maybe you've gotten an incomplete version/revbump that already |
34 |
is resolved in later versions of the tree. |
35 |
|
36 |
|
37 |
WKR |
38 |
Hinnerk |