1 |
On Sunday 04 November 2007, Zac Medico wrote: |
2 |
> Fabian Groffen wrote: |
3 |
> > Just to have it clear and to be sure: |
4 |
> > |
5 |
> > On 04-11-2007 03:33:41 +0000, vapier@g.o wrote: |
6 |
> >> Modified: diffutils-2.8.7-r2.ebuild |
7 |
> >> Log: |
8 |
> >> do *not* include userland_GNU in IUSE |
9 |
> >> (Portage version: 2.1.3.16) |
10 |
> >> |
11 |
> >> |
12 |
> >> Index: diffutils-2.8.7-r2.ebuild |
13 |
> >> |
14 |
> >> -IUSE="nls static userland_GNU" |
15 |
> >> +IUSE="nls static" |
16 |
> > |
17 |
> > On 04-11-2007 08:10:29 +0000, Zac Medico wrote: |
18 |
> >> Author: zmedico |
19 |
> >> Date: 2007-11-04 08:10:29 +0000 (Sun, 04 Nov 2007) |
20 |
> >> New Revision: 8420 |
21 |
> >> |
22 |
> >> Modified: |
23 |
> >> main/trunk/pym/portage/dbapi/bintree.py |
24 |
> >> Log: |
25 |
> >> When evaluating *DEPEND conditionals for the Packages metadata |
26 |
> >> index, do not use IUSE to filter USE since there is currently |
27 |
> >> no guarantee that IUSE properly defines all of the necessary |
28 |
> >> flags. |
29 |
> > |
30 |
> > These two changes now mean that without having "userland_GNU" in IUSE |
31 |
> > |
32 |
> > DEPEND="!userland_GNU? ( some/package )" |
33 |
> > |
34 |
> > will correctly end up in the Packages file, such that Portage will |
35 |
> > properly calculate dependencies when reading from a binhost, right? |
36 |
> |
37 |
> Well, I consider my change to be a workaround for people behaving |
38 |
> like Mike and refusing to declare certain conditionals in IUSE. The |
39 |
> way that I see it, userland_GNU is a USE conditional, so it belongs |
40 |
> in IUSE just like any other USE conditional. Maybe we would be |
41 |
> better off if things like userland_GNU weren't in the USE |
42 |
> conditional space, but they are. |
43 |
|
44 |
userland_* and all other profile-expanded USE flags are "magical" and arent |
45 |
available for user consumption. that is how i view IUSE. it was my |
46 |
understanding that portage was going to get fixed to automatically include |
47 |
the profile-expanded ones and so adding anything to IUSE now for ebuilds is |
48 |
dumb when they're just going to get turned around and removed. the same goes |
49 |
for all implicit/automatic USE expanding things. portage can do this for us, |
50 |
so having developers track it themselves seems like a waste of time. |
51 |
-mike |