1 |
On Mon, 30 Apr 2012 15:50:18 -0700, Mark Knecht wrote: |
2 |
|
3 |
> > I'm not sure what the big deal is, so portasge skips emerging one |
4 |
> > package because it can't download the distfile. So what? The previous |
5 |
> > version worked OK the day before and won't suddenly break because an |
6 |
> > update is available. Just download and upgrade when you have the |
7 |
> > time, casting the appropriate curses for those that set the licence. |
8 |
> > |
9 |
> |
10 |
> I agree. To me it's not much of an issue. However sometime ago when |
11 |
> there was a conversation about how people update their machines I |
12 |
> mentioned that I always do an |
13 |
> |
14 |
> emerge -fDuN @world |
15 |
|
16 |
I do something similar, from a cron script that runs emerge --sync and a |
17 |
couple of other bits. |
18 |
|
19 |
> prior to kicking off the real emerge just to ensure that when the |
20 |
> build finally does start all the files are here and ready. This sort |
21 |
> of issue is precisely why I do that. |
22 |
|
23 |
This still isn't an issue, as long as you use --keep-going. The emerge |
24 |
world will proceed to update everything but the one affected package. |
25 |
|
26 |
> I know most people don't like calculating all the dependencies |
27 |
> multiple times but I prefer to do so, get the files and then pretty |
28 |
> much be guaranteed that the build proceeds without much attention from |
29 |
> me. |
30 |
|
31 |
It is a useful method of detecting potential problems, but this isn't |
32 |
really a problem. However, as my script emails me the output of emerge -p |
33 |
world (which means I actually calculate the dependencies three times, but |
34 |
I don't care as I'm asleep for the first two) I know about the fetch |
35 |
restriction as soon as I read my mail, and can decide whether to deal |
36 |
with it immediately or ignore it. |
37 |
|
38 |
|
39 |
-- |
40 |
Neil Bothwick |
41 |
|
42 |
Don't be humble, you're not that great. |