1 |
On 08/13/2017 07:00 PM, M. J. Everitt wrote: |
2 |
> Interesting .. I'm sure I shied away from that option for some reason |
3 |
> ... wonder if zmedico can shed some light on the difference between the |
4 |
> new options and the old, apart from some added flexibility ... |
5 |
|
6 |
The --autounmask-keep-keywords option allows you to adjust the the way |
7 |
that decisions are made during dependency resolution. It's like a |
8 |
selective version of --autounmask=n that targets decisions involving |
9 |
dependencies that cannot be satisfied without changes to |
10 |
package.accept_keywords. |
11 |
|
12 |
Changes in decision making behavior have a large impact on the resulting |
13 |
dependency calculation. It can mean the difference between a successful |
14 |
calculation, and one that produces useless results. |
15 |
|
16 |
The --autounmask-write=n option has no influence on the decisions made, |
17 |
so it --autounmask was making a bunch of keyword changes that you didn't |
18 |
want, then your only choice was to use --autounmask=n if it was causing |
19 |
the dependency calculation was going too far off course. |
20 |
|
21 |
The --autounmask-keep-keywords option gives finer-grained control. This |
22 |
finer-grained control is only useful in cases where --autounmask=n would |
23 |
prevent useful configuration changes from being made. |
24 |
-- |
25 |
Thanks, |
26 |
Zac |