1 |
On Sat, 29 Sep 2012 09:12:38 -0700 |
2 |
Zac Medico <zmedico@g.o> wrote: |
3 |
|
4 |
> On 09/29/2012 02:42 AM, Michał Górny wrote: |
5 |
> > Hello, |
6 |
> > |
7 |
> > Following the late discussion, I have updated GLEP-62. It no longer is |
8 |
> > designed to be 'backwards compatible' and instead it was suited for |
9 |
> > addition in a new EAPI. |
10 |
> > |
11 |
> > Thus, IUSE_RUNTIME is now independent of IUSE, and runtime dependencies |
12 |
> > can be expressed in SDEPEND only. |
13 |
> |
14 |
> Thanks, these changes make it much more manageable. Trying to do |
15 |
> something like this retroactively for existing EAPIs is just a mess. |
16 |
> |
17 |
> > There's still a case of REQUIRED_USE. I'm not really convinced to |
18 |
> > create a REQUIRED_RUNTIME_USE especially for it. Also, it may be |
19 |
> > actually better to put all IUSE_RUNTIME flags to IUSE as well -- since |
20 |
> > the package manager will after all be required to concatenate them |
21 |
> > anyway. |
22 |
> |
23 |
> Will runtime flags be able to interact with buildtime flags then? It |
24 |
> seems like it could be useful, so it we should probably allow it. Sure, |
25 |
> people could write some expressions that don't make sense, but that's |
26 |
> already the case with REQUIRED_USE. |
27 |
|
28 |
Yes, that's the intent, e.g. when a particular runtime feature requires |
29 |
another build-time feature. |
30 |
|
31 |
-- |
32 |
Best regards, |
33 |
Michał Górny |