1 |
On 6/18/21 8:29 PM, Zac Medico wrote: |
2 |
> On 6/18/21 8:13 PM, Zac Medico wrote: |
3 |
>> On 6/18/21 6:01 PM, Zac Medico wrote: |
4 |
>>> If emerge --depclean fails to resolve any dependencies, then it will |
5 |
>>> now suggest emerge @unsatisfied-deps as the simplest possible |
6 |
>>> solution, and will also suggest to unmerge @unavailable where |
7 |
>>> appropriate at the end: |
8 |
>>> |
9 |
>>> $ emerge --depclean |
10 |
>>> |
11 |
>>> Calculating dependencies... done! |
12 |
>>> * Dependencies could not be completely resolved due to |
13 |
>>> * the following required packages not being installed: |
14 |
>>> * |
15 |
>>> * virtual/cdrtools pulled in by: |
16 |
>>> * app-cdr/cdrdao-1.2.4 |
17 |
>>> * |
18 |
>>> * Have you forgotten to resolve unsatisfied dependencies prior to |
19 |
>>> * depclean? The simplest possible command for this purpose is as |
20 |
>>> * follows: |
21 |
>>> * |
22 |
>>> * emerge @unsatisfied-deps |
23 |
>> |
24 |
>> It turns out that @unsatisfied-deps is often unsuitable here because it |
25 |
>> pulls in a bunch of @installed packages, when you really want to use |
26 |
>> @world as the source of truth. |
27 |
> |
28 |
> The underlying reason is the same as the reason that we've never used |
29 |
> @installed updates as a substitute for @world updates. It just doesn't |
30 |
> work, because @installed is polluted in comparison to @world. |
31 |
> |
32 |
|
33 |
My plan is it introduce an @unsatisfied-world set which is equivalent to |
34 |
@unsatisfied-deps but filters out any non @world packages. |
35 |
-- |
36 |
Thanks, |
37 |
Zac |