1 |
Alan McKinnon wrote: |
2 |
> On Tuesday 16 September 2008 20:43:06 Dale wrote: |
3 |
> |
4 |
>> I have used --skipfirst before but I also don't think it is a good |
5 |
>> idea. For a idiot like me to say that must mean something. I guess it |
6 |
>> would depend on the package as to whether it could be skipped or not. |
7 |
>> If I do a emerge -e world, I prefer nothing to fail else what is the point? |
8 |
>> |
9 |
> |
10 |
> Sometimes stuff does fail - you can't fully know beforehand if an ebuild will |
11 |
> succeed or not. --skipfirst is perfectly safe if you have looked at the |
12 |
> package list and know that there's no dependency issues. The danger is |
13 |
> blindly resuming without understanding the consequences. |
14 |
> |
15 |
> For example, lets assume there's a package called alan-fonts-meta which |
16 |
> depends on 20 fancy wingding-type fonts. Lets also assume that I was a dork |
17 |
> when coding one font ebuild at 4am and botched the SRC_URI. That ebuild will |
18 |
> fail, but you know for sure that you can safely --resume --skipfirst |
19 |
> (followed by a scathing email with a blunt instrument as payload). |
20 |
> |
21 |
> Other examples exist. OTOH, if the expat upgrade of some months ago failed, a |
22 |
> blind resume would likely have caused endless troubles. |
23 |
> |
24 |
> |
25 |
|
26 |
That was my point in a round about way. If you don't know, don't use |
27 |
--skipfirst since it could mess up some stuff and turn your puter into a |
28 |
mess after logging out or a reboot. I'm thinking some of the more |
29 |
serious system packages with that. |
30 |
|
31 |
Me, I don't even claim to know so I either ask or fix what borked. LOL |
32 |
|
33 |
Dale |
34 |
|
35 |
:-) :-) |