1 |
On Tue, 20 Jan 2015 12:12:32 -0800 |
2 |
Zac Medico <zmedico@g.o> wrote: |
3 |
> >> Regardless of whether or not (or how) we choose to apply |
4 |
> >> REQUIRED_USE to various cases, I think we should keep REQUIRED_USE |
5 |
> >> around, since having a machine-readable representation of these |
6 |
> >> constraints can potentially be extremely useful to dependency |
7 |
> >> resolvers. |
8 |
> > |
9 |
> > [evidence needed] |
10 |
> |
11 |
> REQUIRED_USE is already successfully used to apply constraints that |
12 |
> would otherwise need to be enforced by executing shell code in |
13 |
> pkg_pretend. I think it's obvious that the declarative and |
14 |
> machine-readable nature of REQUIRED_USE make if preferable to using |
15 |
> pkg_pretend in many cases. |
16 |
|
17 |
It's not obvious at all. The evidence so far suggests that good human |
18 |
readable messages are important, and that if we're ever going to get |
19 |
automated resolution then we'll need a new not-SAT-based format anyway. |
20 |
|
21 |
-- |
22 |
Ciaran McCreesh |