| 1 |
On 15/08/2014 09:49, Сергей wrote: |
| 2 |
> sergey-pc sergey # emerge --depclean -a |
| 3 |
> |
| 4 |
> Calculating dependencies... done! |
| 5 |
>>>> No packages selected for removal by depclean |
| 6 |
>>>> To see reverse dependencies, use --verbose |
| 7 |
> Packages installed: 896 |
| 8 |
> Packages in world: 92 |
| 9 |
> Packages in system: 44 |
| 10 |
> Required packages: 896 |
| 11 |
> Number removed: 0 |
| 12 |
> sergey-pc sergey # emerge --depclean -a |
| 13 |
> |
| 14 |
> Calculating dependencies... done! |
| 15 |
>>>> Calculating removal order... |
| 16 |
> |
| 17 |
>>>> These are the packages that would be unmerged: |
| 18 |
> |
| 19 |
> |
| 20 |
> !!! 'app-editors/nano' (virtual/editor) is part of your system profile. |
| 21 |
> !!! Unmerging it may be damaging to your system. |
| 22 |
> |
| 23 |
> |
| 24 |
> app-editors/nano |
| 25 |
> selected: 2.3.2 |
| 26 |
> protected: none |
| 27 |
> omitted: none |
| 28 |
> |
| 29 |
> All selected packages: app-editors/nano-2.3.2 |
| 30 |
> |
| 31 |
>>>> 'Selected' packages are slated for removal. |
| 32 |
>>>> 'Protected' and 'omitted' packages will not be removed. |
| 33 |
> |
| 34 |
> Would you like to unmerge these packages? [Yes/No] |
| 35 |
> |
| 36 |
> NOTHING had changed between running these two commands. So, sometimes |
| 37 |
> depclean wants to remove nano, sometimes it doesn't. |
| 38 |
> If I look at emerge -pv --depclean, that's what I see: |
| 39 |
> If it removes nano: |
| 40 |
> sys-apps/ed-1.6 pulled in by: |
| 41 |
> app-text/texlive-core-2012-r1 requires sys-apps/ed |
| 42 |
> virtual/editor-0 requires sys-apps/ed |
| 43 |
> If it keeps nano: |
| 44 |
> app-editors/nano-2.3.2 pulled in by: |
| 45 |
> virtual/editor-0 requires app-editors/nano |
| 46 |
> |
| 47 |
> I have removed "ed" because I don't use it, but I think, depclean |
| 48 |
> should behave more exact. For example, say "Two different packages |
| 49 |
> provide *bla-bla* which is needed by installed virtual/*bla*"... |
| 50 |
> |
| 51 |
> Should I file a bug about this? |
| 52 |
|
| 53 |
|
| 54 |
This could be part of the whole undeterminate behaviour of recent |
| 55 |
portage that shows up with --backtrack being too small? |
| 56 |
|
| 57 |
I agree that -depclean must be exact and no matter what code path is |
| 58 |
selected the results must always be identical on every run. So filing a |
| 59 |
bug seems warranted |
| 60 |
|
| 61 |
|
| 62 |
-- |
| 63 |
Alan McKinnon |
| 64 |
alan.mckinnon@×××××.com |
Archives