| 1 |
On Wed, 22 Jun 2016 14:57:50 -0500, Dan Douglas wrote: |
| 2 |
|
| 3 |
> >> --keep-going is in EMERGE_DEFAULT_OPTS so the problem is only when |
| 4 |
> >> that fails for whatever reason, --resume (with or without |
| 5 |
> >> --skip-first) always fails too. |
| 6 |
> > On 06/22/2016 12:31 PM, Neil Bothwick wrote: |
| 7 |
> > That makes sense, --keep-going has already made sure that all updates |
| 8 |
> > that are not dependant on the failing one have emerged, so there's |
| 9 |
> > nothing left to emerge until you fix the broken package. |
| 10 |
> |
| 11 |
> |
| 12 |
> That's what it should do but it clearly doesn't quite work that way. |
| 13 |
> It's easy to prove it's broken by finding any package on the resume list |
| 14 |
> that can merge on its own without pulling in a previously failed |
| 15 |
> package. |
| 16 |
> |
| 17 |
> I've had completely up-to-date systems where no possible failure could |
| 18 |
> result in an unsatisfied dependency and portage refuses to resume an |
| 19 |
> `emerge -e @world` with hundreds of possible packages on the resume list |
| 20 |
> that would work in isolation. That's the problem. |
| 21 |
|
| 22 |
Without specifics it's not possible to say any more. |
| 23 |
|
| 24 |
|
| 25 |
-- |
| 26 |
Neil Bothwick |
| 27 |
|
| 28 |
"There are two ways of constructing a software design: One way is to make |
| 29 |
it so simple that there are obviously no deficiencies, and the other way |
| 30 |
is to make it so complicated that there are no obvious deficiencies. |
| 31 |
The first method is far more difficult" -C.A.R. Hoare |
Archives