Gentoo Archives: gentoo-council

From: Luca Barbato <lu_zero@g.o>
To: Ryan Hill <dirtyepic@g.o>
Cc: gentoo-council@g.o
Subject: Re: [gentoo-council] Re: Preliminary Meeting-Topics for 12 February 2009
Date: Sun, 22 Feb 2009 23:37:53
Message-Id: 49A1E1CB.1000806@gentoo.org
In Reply to: [gentoo-council] Re: Preliminary Meeting-Topics for 12 February 2009 by Ryan Hill
1 Ryan Hill wrote:
2 > On Fri, 20 Feb 2009 00:11:32 +0300
3 > Peter Volkov <pva@g.o> wrote:
4 >
5 >> That said, technically there are other solutions for this problem,
6 >> e.g. 1) it is possible to read one line of defined format from any
7 >> file 2) it is possible to make eapi inside ebuild name
8 >> (foo-1.0-eapi2.ebuild), but not as extension. Any solution, even
9 >> breaking compatibility solution, we could already start using if we
10 >> had forgotten about GLEP 55 long time ago...
11 >
12 > I really don't understand why foo-0.1.eapi3.ebuild is considered an
13 > acceptable solution and foo-0.1.ebuild.eapi3 is not.
14
15 I guess that principle of the least surprise counts there.
16
17 > They have the same advantages, arguments about aesthetics aside, and
18 > seem to be a much cleaner solution than any other that has been
19 > proposed.
20
21 Using either manifests and or switch sync path is even less invasive if
22 you consider that point raised against the proposal to switch extensions
23 every time something changes in the ebuild format is that is misleading.
24
25 > But the former has one distinct disadvantage that the latter
26 > does not: any currently released version of portage does not work
27 > correctly with ebuilds having version suffixes it does not recognize.
28 > For example, you cannot currently create a Manifest in a package
29 > directory containing a file named package-1.0.eapi3.ebuild.
30
31 Portage should warn/die if stray files are present. So the whole thing
32 looks to me as a way to harness a bug.
33
34 > We can modify portage, of course, to recognize this suffix, and then
35 > wait long enough for that release to trickle down. But later, if we
36 > want to add another suffix, -scm perhaps, or something we
37 > haven't considered yet, we again have to go through the same process. I
38 > thought the reason we introduced EAPI was to free us from this.
39
40 As stated before this eapi had been considered a ugly solution looking
41 for problems to solve.
42
43 > With a format such as .ebuild.eapiX we would avoid these issues.
44
45 Using manifest to have portage validate/invalidate ebuilds works as well
46 and is completely transparent.
47
48 > Portage (without any modifications) will not recognize these files as
49 > ebuilds (which is exactly the behaviour we want if it doesn't
50 > understand the EAPI), so the format of the version string is moot. We
51 > could introduce -scm (or whatever) in EAPI X, and be able to use it
52 > immediately.
53
54
55 > I'm sorry, but until you come up with a better reason than "it's
56 > tradition", I'm afraid this will continue to be submitted indefinitely.
57 > You will have to provide technical objections why this approach is
58 > unacceptable before anyone can come up with something that is.
59
60 Usually in order to get something changed is the burden of the
61 proponents make it worthy for everybody else. Moreover if the change
62 causes any annoyance, its usefulness has to be considered superior to
63 the damage. We got people that annoyed about this proposal that they
64 stated they'll quit if it is passed.
65
66 > Here is an example, to get you started:
67 > If a Manifest is generated using a portage version that supports an
68 > EAPI and recognizes a package atom containing a version suffix that
69 > was added in that EAPI as an ebuild and thus categorizes it as EBUILD in
70 > the Manifest, do portage versions that do not support the EAPI have
71 > trouble when they see the entry marked EBUILD for a package atom they
72 > think is invalid?
73
74 Manifest2 is backward compatible to manifest1 by ignoring lines it
75 couldn't parse, so if we have portage embed the eapi information there
76 we'd archive the same result being completely transparent.
77
78 This proposal is in the migration-paths document, why we shouldn't use a
79 less invasive approach, that is using pretty much the same principle but
80 doesn't have the shortcoming the extension rename ?
81
82 lu
83
84 --
85
86 Luca Barbato
87 Gentoo Council Member
88 Gentoo/linux Gentoo/PPC
89 http://dev.gentoo.org/~lu_zero

Replies

Subject Author
Re: [gentoo-council] Re: Preliminary Meeting-Topics for 12 February 2009 Ciaran McCreesh <ciaran.mccreesh@××××××××××.com>