| 1 |
On Wednesday 30 November 2011 11:34:05 Zac Medico wrote: |
| 2 |
> On 11/30/2011 08:09 AM, Mike Frysinger wrote: |
| 3 |
> > On Wednesday 30 November 2011 01:23:59 Zac Medico wrote: |
| 4 |
> >> If it wasn't for implicit system dependencies, the system set and its |
| 5 |
> >> dependencies wouldn't need this kind of special treatment. |
| 6 |
> > |
| 7 |
> > if it wasn't for implicit system dependencies, we'd have significant |
| 8 |
> > bloat in the tree |
| 9 |
> |
| 10 |
> I'm not suggesting that we do away with implicit dependencies entirely. |
| 11 |
> However, I think it's good to minimize them, as removing packages like |
| 12 |
> sys-libs/zlib from the system set tends to do. |
| 13 |
|
| 14 |
i'm good with this as well which is why i've been culling the simpler stuff |
| 15 |
|
| 16 |
> > and circular deps out the wazoo |
| 17 |
> |
| 18 |
> Ignoring circular dependencies doesn't make them go away. Ignoring |
| 19 |
> dependencies can lead to build failures that could have been avoided if |
| 20 |
> they were expressed in a way that the dependency resolver could properly |
| 21 |
> account for them. |
| 22 |
|
| 23 |
that doesn't address the previously mentioned issue. there's absolutely no |
| 24 |
reason why the majority of packages in the tree should have to list |
| 25 |
"coreutils" or "sed" or "grep" in their DEPEND, nor is there any reason why |
| 26 |
the system should have been lacking them in the first place. |
| 27 |
-mike |
Archives