Gentoo Archives: gentoo-dev

From: Zac Medico <zmedico@g.o>
To: gentoo-dev@l.g.o
Cc: Ian Stakenvicius <axs@g.o>
Subject: Re: [gentoo-dev] Let's populate IUSE_IMPLICIT in the base profile
Date: Thu, 27 Sep 2012 17:58:36
Message-Id: 5064935F.3050107@gentoo.org
In Reply to: Re: [gentoo-dev] Let's populate IUSE_IMPLICIT in the base profile by Ian Stakenvicius
1 On 09/27/2012 10:45 AM, Ian Stakenvicius wrote:
2 > On 27/09/12 12:02 PM, Ian Stakenvicius wrote:
3 >> On 27/09/12 11:57 AM, Mike Gilbert wrote:
4 >>> On Thu, Sep 13, 2012 at 1:40 AM, Zac Medico <zmedico@g.o>
5 >>> wrote:
6 >>>> Hi,
7 >>>>
8 >>>> The council has approved [1] "Profile IUSE injection" [2] for
9 >>>> inclusion in EAPI 5, and in latest Portage we have
10 >>>> experimental EAPI 5_pre2 [3] which implements all of the
11 >>>> approved features. So, now would be a good time to start
12 >>>> populating IUSE_IMPLICIT with whatever values may be
13 >>>> appropriate.
14 >>>>
15 >>>> What values belong there? Some of the flags that appear in
16 >>>> profiles/base/use.mask might make good candidates, such as
17 >>>> prefix and selinux. How about other special flags like
18 >>>> bootstrap, build, and test?
19 >>>>
20 >
21 >>> prefix and test make sense to me. I'm not so familiar with the
22 >>> others.
23 >
24 >
25 >> build is specifically for catalyst and/or for building the stages,
26 >> right? If so, this one makes sense to me to add.
27 >
28 >> bootstrap I would guess is similar? Unsure how that one is used
29 >> at present. If IUSE_IMPLICIT would still allow the boostrapping
30 >> tool to set the use flag, i see no issues having it in the list.
31 >
32 >
33 > For the purposes of EAPI5 testing (overlays etc), would it make sense
34 > to start with this list of flags within IUSE_IMPLICIT on
35 > base/make.defaults now, and then based on consensus that list can be
36 > trimmed or appended?
37
38 I would recommend to stay on the conservative side and only add ones
39 that we're sure we need for specific ebuilds. We can remove flags later,
40 but it's better if can avoid adding unneeded ones in the first place.
41
42 > floppym's already requested 'prefix' so that his chromium tests with
43 > EAPI5 don't fail or need an explicit 'prefix' in IUSE, for instance
44
45 I think it's pretty clear that the 'prefix' flag is special, so I would
46 go ahead and add it.
47 --
48 Thanks,
49 Zac