1 |
Am Sonntag, den 17.05.2009, 11:11 -0600 schrieb Ryan Hill: |
2 |
> On Fri, 15 May 2009 23:31:25 +0200 |
3 |
> Tiziano Müller <dev-zero@g.o> wrote: |
4 |
> |
5 |
> > Wrong. For example: |
6 |
> > - stuff like docompress may change the content being installed depending |
7 |
> > on the package manager |
8 |
> > - --disable-static (maybe in a later EAPI) changes content |
9 |
> > - slot-dep-operators change the rdepend of installed packages, so it |
10 |
> > changes how the package manager has to handle reverse packages on |
11 |
> > uninstall (in EAPI 3) |
12 |
> |
13 |
> None of which are a problem when changing the EAPI from 0 to 1, which is the |
14 |
> situation here. The first two examples fall under the currently established |
15 |
> guideline of revbumping when content changes (and I emphasize guideline). I |
16 |
> don't see how the third example is any different than adding or removing |
17 |
> dependencies, which also does not require a revbump. |
18 |
Which is mostly wrong as well since a change in dependency means that |
19 |
the currently installed stuff may break if a package (the new dependency |
20 |
for example) gets removed and since the package you changed does not |
21 |
reference it, it gets broken (for example if you had a magic dep before |
22 |
and add it now as an explicit dep). So, unless you're doing a pkgmove |
23 |
it's a dangerous thing since the PM can't reliably track reverse deps |
24 |
when doing uninstalls since it has to use the vdb entry for that, |
25 |
doesn't it? |
26 |
|
27 |
> So I'd argue that an |
28 |
> EAPI change does not require a revision bump in and of itself. |
29 |
EAPI may imply a decent implicit change to the ebuild and therefore |
30 |
needs a rev-bump as per the manual. |
31 |
|
32 |
> That's not to |
33 |
> say it shouldn't be done if the situation allows, or you have any doubts, or |
34 |
> just because you want to. But unconditionally putting an ebuild through full |
35 |
> ~arch to stable cycle because you added a simple SLOT dependency or a + to a |
36 |
> USE flag is bureaucratic nonsense. |
37 |
> |
38 |
> > And we also always said that a rev bump should be done on non trivial |
39 |
> > changes or non-build-fixes and changing the EAPI is technical seen |
40 |
> > mostly a non-trivial change. |
41 |
> |
42 |
> As above, it depends on the situation. 0 -> 1 is a trivial change. |
43 |
> |
44 |
> > Do we want to document the following? (do we have already?) |
45 |
> > - When is it allowed to use an EAPI in the tree (given as offset to the |
46 |
> > release of portage supporting that eapi) |
47 |
> > - When is it allowed to use an EAPI in the stable tree (given as offset |
48 |
> > of when a portage version supporting that EAPI got stable) |
49 |
> |
50 |
> As soon as a version of portage supporting that EAPI is available. |
51 |
And how much time a portage with a new EAPI got stabled? |
52 |
(see for example early python eapi bumps) |
53 |
|
54 |
> |
55 |
> This is the entire point of the EAPI, that we don't have to wait X amount of |
56 |
> time before using new features. If the user hasn't updated portage yet, they |
57 |
> simply won't see ebuilds which use the new EAPI. |
58 |
> |
59 |
> We may want to document a suggested waiting time before removing ebuilds |
60 |
> using older EAPI's. For example, should we always keep an EAPI 0 ebuild in |
61 |
> stable as a fallback? Or if the user tries to install or update a package |
62 |
> where all versions are masked by EAPI, should (does?) portage suggest updating |
63 |
> itself? |
64 |
It would maybe suggest to update to an unstable version of portage, not |
65 |
so good then? |
66 |
|
67 |
-- |
68 |
Tiziano Müller |
69 |
Gentoo Linux Developer, Council Member |
70 |
Areas of responsibility: |
71 |
Samba, PostgreSQL, CPP, Python, sysadmin, GLEP Editor |
72 |
E-Mail : dev-zero@g.o |
73 |
GnuPG FP : F327 283A E769 2E36 18D5 4DE2 1B05 6A63 AE9C 1E30 |