1 |
On 01/05/18 18:50, allan gottlieb wrote: |
2 |
> On Fri, Jan 05 2018, Daniel Frey wrote: |
3 |
> |
4 |
>> On 01/05/18 13:15, Neil Bothwick wrote: |
5 |
>>> On Fri, 05 Jan 2018 13:00:20 -0500, allan gottlieb wrote: |
6 |
>>> |
7 |
>>>> I am finally moving my production machine the the 17.0 profile. |
8 |
>>>> Currently running is |
9 |
>>>> |
10 |
>>>> emerge -e --keep-going @world |
11 |
>>>> |
12 |
>>>> So far there is one failure (libcryptui, the fix is easy). |
13 |
>>>> |
14 |
>>>> Am I correct that when the above emerge completes, I should run |
15 |
>>>> simply |
16 |
>>>> |
17 |
>>>> emerge --resume |
18 |
>>> |
19 |
>>> If the emerge has completed, there is nothing to resume. Just make a note |
20 |
>>> of the packages that failed to build ans emerge --oneshot them, with any |
21 |
>>> necessary fixes. |
22 |
>>> |
23 |
>>> |
24 |
>> |
25 |
>> You may also want to check for dependencies (as in: `equery depends |
26 |
>> <package>` and rebuild the dependencies of the failed package as well. |
27 |
>> |
28 |
>> Dan |
29 |
> |
30 |
> I don't understand. If the dependencies didn't fail, why should I |
31 |
> rebuild them? |
32 |
> |
33 |
> thanks, |
34 |
> allan |
35 |
> |
36 |
|
37 |
If you're transitioning over to the new profile, that means the |
38 |
dependencies would be built and linked to the old profile's binaries. |
39 |
This was merely a suggestion, as I've had some mighty strange things |
40 |
happen after upgrades to gcc in the past, all solved by `emerge -e |
41 |
world`, although... I didn't have packages fail back then. |
42 |
|
43 |
When I switched over to the 17 profile I had one rebuild failure, and it |
44 |
didn't have any direct dependencies fail. |
45 |
|
46 |
Dan |