| 1 |
On Sun, 31 May 2015 12:59:32 +0200, Marc Joliet wrote: |
| 2 |
|
| 3 |
> > |
| 4 |
> > Yes, equery is still getting it wrong, gvim only depends on gettext |
| 5 |
> > when built with USE=nls. |
| 6 |
> |
| 7 |
> I don't get this. equery is showing *exactly* the same information as |
| 8 |
> your grep. It's not wrong, it's just that it only shows that gvim |
| 9 |
> *might* depend on gettext, namely when the nls USE flag is set, |
| 10 |
|
| 11 |
Yes the OP listed a long list of packages that he thought did depend on |
| 12 |
gettext, even though they do not with hi current settings. |
| 13 |
|
| 14 |
> while |
| 15 |
> "emerge --depclean -pv" will give you a definitive yes/no answer |
| 16 |
> (although automagic dependencies might still render its answer |
| 17 |
> incorrect). However, emerge won't tell you *why*, so strictly speaking |
| 18 |
> you need both for a full answer. |
| 19 |
|
| 20 |
depclean tells you which package depends on the package you want to |
| 21 |
remove, examination of the ebuild is usually the most reliable way of |
| 22 |
determining whether that need is USE controlled. |
| 23 |
|
| 24 |
The main disadvantage of depclean IMO is that it shows ony one package |
| 25 |
that depends on the package you want to remove, not all the relevant parts |
| 26 |
of the dependency chain. If the package you want to reove is not |
| 27 |
system-critical, the easiest option is often to quickpkg it, remove it |
| 28 |
and then look at the output from |
| 29 |
|
| 30 |
emerge --tree --update --deep @world |
| 31 |
|
| 32 |
|
| 33 |
-- |
| 34 |
Neil Bothwick |
| 35 |
|
| 36 |
Intel: where Quality is job number 0.9998782345! |
Archives