1 |
On Sat, Sep 09, 2006 at 02:26:03PM +0000, Peter wrote: |
2 |
> On Sat, 09 Sep 2006 12:41:14 +0100, Mick wrote: |
3 |
> |
4 |
> > Hi All, |
5 |
> > |
6 |
> > Just ran depclean and the following candicates are shown: |
7 |
> > ======================================= |
8 |
> > Calculating dependencies... done! |
9 |
|
10 |
[snip] |
11 |
|
12 |
> With the exception of tcsh, probably no. Run equery d {package} and see if |
13 |
> anything depends on them. The ones you show are pretty standard. You might |
14 |
> be able to dump the framebuffer, but check. |
15 |
|
16 |
If depclean has listed the packages, I'm fairly sure that means |
17 |
portage couldn't find anything in "system" or "world" that depends on |
18 |
it (or anything that depends on something that depends on it, |
19 |
etc.). So querying for dependencies is, by definition, going to return |
20 |
nothing. |
21 |
|
22 |
That doesn't necessarily mean the packages can safely be removed (I |
23 |
once borked my system badly by making that assumption). Look at the |
24 |
package descriptions or google them to try to find out what exactly |
25 |
they do before you decide it's safe to remove them. Usually, even if |
26 |
the package *is* required, all that happens is some program will no |
27 |
longer run and you'll need to re-emerge the package. You can then add |
28 |
it to world so that it's not picked up by depclean in the future. In |
29 |
the worst case, you find you've remove something essential and the |
30 |
system no longer even boots. (That's what happened to me, with libcap |
31 |
if I remember right.) |
32 |
|
33 |
That was with older versions of portage. I don't know if newer ones |
34 |
are safer. |
35 |
|
36 |
HTH, |
37 |
|
38 |
Toby |
39 |
-- |
40 |
PhD Student |
41 |
Quantum Information Theory group |
42 |
Max Planck Institute for Quantum Optics |
43 |
Garching, Germany |
44 |
|
45 |
email: toby@××××××××.org |
46 |
web: www.dr-qubit.org |
47 |
-- |
48 |
gentoo-user@g.o mailing list |