1 |
On Fri, 13 Nov 2009 17:06:02 -0800, Mark Knecht wrote: |
2 |
|
3 |
> > I think it is being over-cautious, which results in packages being |
4 |
> > rebuilt multiple time unnecessarily, but I's rather give it the |
5 |
> > chance to fix itself. That said, I've never had a list anything like |
6 |
> > 50 packages long, but I do update frequently. |
7 |
> > |
8 |
> It wasn't that I had 50 packages in the emerge -DuN @world. That was |
9 |
> something like 10. It was after that finished and I ran |
10 |
> @preserved-rebuild that it said 50 packages were effected by something |
11 |
> it found, but those 50 were all dependent on just one or two packages |
12 |
> that Alan was suggesting to me are held in the preserved database |
13 |
> file, or so I think. |
14 |
|
15 |
I realised it was 50-odd in the rebuild list, but in my experience |
16 |
multiple runs gradually reduces that number. Maybe portage could be more |
17 |
intelligent about the order in which it re-emerges these packages, but |
18 |
running it enough times always works for me. |
19 |
|
20 |
Removing the registry is potentially risky because you could still have |
21 |
packages linked to a library that is not managed by portage, and that |
22 |
will never update. If someone finds a security hole in that library, you |
23 |
could be in trouble. |
24 |
|
25 |
"Fixing" the problem by deleting the registry is akin to fixing low oil |
26 |
pressure in your car by disconnecting the warning light. |
27 |
|
28 |
|
29 |
-- |
30 |
Neil Bothwick |
31 |
|
32 |
We are THOR of Borg... your RFC compliant mailbox has been assimilated |