| 1 |
On Thursday 29 October 2009 21:17:23 Paul Hartman wrote: |
| 2 |
> On Thu, Oct 29, 2009 at 11:53 AM, Neil Bothwick <neil@××××××××××.uk> wrote: |
| 3 |
> > On Thu, 29 Oct 2009 11:45:54 -0500, Paul Hartman wrote: |
| 4 |
> >> >> Are there any tricks to remove all of kde4, or do |
| 5 |
> >> >> I have to do it manually? |
| 6 |
> >> > |
| 7 |
> >> > The set and the meta-package pull in basically the same packages. |
| 8 |
> >> > Unmerge the set, emerge kde-meta then do a depclean to catch any |
| 9 |
> >> > stragglers. |
| 10 |
> >> |
| 11 |
> >> I would just delete the set from the world_sets file, then emerge |
| 12 |
> >> kde-meta, then depclean... much easier than unmerging everything and |
| 13 |
> >> recompiling it. |
| 14 |
> > |
| 15 |
> > Sorry, that's what I meant. I was thinking that unmerging a set was like |
| 16 |
> > unmerging a meta-package, that only the set itself would go. |
| 17 |
> |
| 18 |
> Ah, okay, we meant the same thing then. :) AFAK emerge -C @set will |
| 19 |
> remove everything in @set but there's no portage command to remove the |
| 20 |
> set definition without unmerging what's in it. I could be wrong on |
| 21 |
> that, though. |
| 22 |
|
| 23 |
You could be wrong, but you are not :-) |
| 24 |
|
| 25 |
Your understanding of emerge -C @<set> is correct. |
| 26 |
|
| 27 |
Portage treats the contents of a set as strictly depending on the set itself. |
| 28 |
So there is no way to remove a set definition and leave it's child packages |
| 29 |
intact - they would then be orphans. |
| 30 |
|
| 31 |
Unless of course the child packages are themselves in world or depend on |
| 32 |
something in world. But that's a side-effect of something else altogether and |
| 33 |
not relevant to set behaviour. |
| 34 |
|
| 35 |
-- |
| 36 |
alan dot mckinnon at gmail dot com |
Archives