| 1 |
>>>>> On Wed, 17 Jun 2020, Michael Lienhardt wrote: |
| 2 |
|
| 3 |
> But maybe, "error" here in the PMS mean "the cpvs without the use flag |
| 4 |
> does not match that dependency and a warning should be raised to |
| 5 |
> improve compatibility in the future". In that case, it would be |
| 6 |
> clearer for me to change 'error' in the PMS into something like |
| 7 |
> "results in a no match, |
| 8 |
|
| 9 |
IMHO we cannot assume that. If the flag is not in the dependency's |
| 10 |
IUSE_EFFECTIVE then behaviour is undefined. |
| 11 |
|
| 12 |
> but should be avoided". That way, it is explicitly stated that missing |
| 13 |
> use flags for a 2-style USE dependency is accepted (which is the |
| 14 |
> current behavior of emerge) but frown upon, without forcing any |
| 15 |
> specific error handling, like Michał accurately pointed out. |
| 16 |
|
| 17 |
The real problem is that we don't have a good procedure for removing |
| 18 |
flags from ebuilds with reverse (2-style) use dependencies. (And even |
| 19 |
with 4-style use dependencies the problem remains that one cannot know |
| 20 |
in advance whether removal of the flag implies that the feature is now |
| 21 |
unconditionally enabled, or that it is disabled.) |
| 22 |
|
| 23 |
Ulrich |
Archives