| 1 |
On Fri, 18 Apr 2008 19:49:55 +0200, b.n. wrote: |
| 2 |
|
| 3 |
> > Now it makes sense. If you have not modified conf.d/net since the last |
| 4 |
> > baselayout emerge, portage considers the file to be part of the old |
| 5 |
> > package and removes it. That's why only some machines are affected. It |
| 6 |
> > also shows that this is not a bug with the new baselayout but a time |
| 7 |
> > bomb in the 1.x ebuilds. |
| 8 |
> |
| 9 |
> Err, how can it make sense? |
| 10 |
> Does it make sense to have portage *remove* (or substitute silently) |
| 11 |
> files in /etc? Maybe if I don't modify conf.d/net is because I don't |
| 12 |
> had the need to modify it... |
| 13 |
|
| 14 |
I mean it makes sense how it happens, not that it is sensible to do. It's |
| 15 |
not that you haven't modified it, in that case it doesn't matter that the |
| 16 |
1.x default is replaced with the 2.0 default. But the way this explains |
| 17 |
the 1.x ebuild working means that if you do modify the file under 1.x, |
| 18 |
then emerge baselayout 1.x again, the modified file is considered to have |
| 19 |
been installed by portage and safe to replace with a later default, |
| 20 |
although even that logic is flawed. |
| 21 |
|
| 22 |
It's all academic now, as the bug has been uncovered and fixed, which is |
| 23 |
exactly what the testing arches are for. |
| 24 |
|
| 25 |
|
| 26 |
-- |
| 27 |
Neil Bothwick |
| 28 |
|
| 29 |
Is fire supposed to shoot out of it like that? |
Archives