| 1 |
Michał Górny <mgorny@g.o> wrote: |
| 2 |
> |
| 3 |
>> Ciaran McCreesh <ciaran.mccreesh@××××××××××.com> wrote: |
| 4 |
>> > Martin Vaeth <martin@×××××.de> wrote: |
| 5 |
>> >> The idea is to act "as usual", just to skip unnecessary phases... |
| 6 |
>> > |
| 7 |
>> > So someone adds optional selinux support to a package, and then you end |
| 8 |
>> > up with selinux being "on", despite not having it, and then another |
| 9 |
>> > package depends upon your package with [selinux], and the dependency is |
| 10 |
>> > mistakenly treated as met... |
| 11 |
>> If the developer has added IUSE=3Dselinux and bumps from -r1 to -r1.1, |
| 12 |
>> he has of course verified that this USE-change does not require |
| 13 |
>> recompilation either way, since otherwise he would not have been |
| 14 |
>> allowed to explicitly say the package manager that recompilation |
| 15 |
>> is unnecessary. |
| 16 |
>> So the dependency is *correctly* treated as met. |
| 17 |
> |
| 18 |
> Excuse me but are we talking about updating *DEPEND or IUSE? |
| 19 |
|
| 20 |
You are right, according to my original suggestion, the developer |
| 21 |
should not even have been allowed to bump only a minor revision. |
| 22 |
This is already discussing some possible extension/misuse |
| 23 |
of the feature. |
Archives