1 |
On Sun, 17 May 2009 21:03:46 +0200 |
2 |
Tiziano Müller <dev-zero@g.o> wrote: |
3 |
|
4 |
> Am Sonntag, den 17.05.2009, 11:11 -0600 schrieb Ryan Hill: |
5 |
> > On Fri, 15 May 2009 23:31:25 +0200 |
6 |
> > Tiziano Müller <dev-zero@g.o> wrote: |
7 |
> > |
8 |
> > > Wrong. For example: |
9 |
> > > - stuff like docompress may change the content being installed depending |
10 |
> > > on the package manager |
11 |
> > > - --disable-static (maybe in a later EAPI) changes content |
12 |
> > > - slot-dep-operators change the rdepend of installed packages, so it |
13 |
> > > changes how the package manager has to handle reverse packages on |
14 |
> > > uninstall (in EAPI 3) |
15 |
> > |
16 |
> > None of which are a problem when changing the EAPI from 0 to 1, which is the |
17 |
> > situation here. The first two examples fall under the currently established |
18 |
> > guideline of revbumping when content changes (and I emphasize guideline). I |
19 |
> > don't see how the third example is any different than adding or removing |
20 |
> > dependencies, which also does not require a revbump. |
21 |
> |
22 |
> Which is mostly wrong as well since a change in dependency means that |
23 |
> the currently installed stuff may break if a package (the new dependency |
24 |
> for example) gets removed and since the package you changed does not |
25 |
> reference it, it gets broken (for example if you had a magic dep before |
26 |
> and add it now as an explicit dep). |
27 |
|
28 |
I don't understand. Removing a runtime dependency of a package will break it |
29 |
regardless of whether or not it's referenced in the package's VDB. We don't |
30 |
prevent uninstalling dependencies, so how does having it referenced prevent |
31 |
breakage? The only case I can think of is depclean, but it already checks to |
32 |
see if removing a package will break the linkage map of another installed |
33 |
package. |
34 |
|
35 |
> So, unless you're doing a pkgmove |
36 |
> it's a dangerous thing since the PM can't reliably track reverse deps |
37 |
> when doing uninstalls since it has to use the vdb entry for that, |
38 |
> doesn't it? |
39 |
|
40 |
Since when do we track reverse deps for uninstalls? |
41 |
|
42 |
> > So I'd argue that an |
43 |
> > EAPI change does not require a revision bump in and of itself. |
44 |
> |
45 |
> EAPI may imply a decent implicit change to the ebuild and therefore |
46 |
> needs a rev-bump as per the manual. |
47 |
|
48 |
Exactly. :) |
49 |
|
50 |
It's not the EAPI itself that requires the revbump, it's the changes the EAPI |
51 |
makes. That's all I'm saying. And in the case of going from EAPI 0 to EAPI |
52 |
1, the changes are not those that require a revbump. If I were going from |
53 |
EAPI 1 to 2, and I'm using the default src_compile, then yes, a revbump is in |
54 |
order. |
55 |
|
56 |
> > We may want to document a suggested waiting time before removing ebuilds |
57 |
> > using older EAPI's. For example, should we always keep an EAPI 0 ebuild in |
58 |
> > stable as a fallback? Or if the user tries to install or update a package |
59 |
> > where all versions are masked by EAPI, should (does?) portage suggest updating |
60 |
> > itself? |
61 |
> It would maybe suggest to update to an unstable version of portage, not |
62 |
> so good then? |
63 |
|
64 |
If the user is installing a package that doesn't have a stable version with |
65 |
an EAPI that their package manager supports, then it's no different than if |
66 |
they are installing a package that doesn't have a stable version with their |
67 |
KEYWORDS. And when unmasking ~arch packages, you often have to unmask ~arch |
68 |
dependencies. Portage is just another of these dependencies. |
69 |
|
70 |
|
71 |
-- |
72 |
gcc-porting, by design, by neglect |
73 |
treecleaner, for a fact or just for effect |
74 |
wxwidgets @ gentoo EFFD 380E 047A 4B51 D2BD C64F 8AA8 8346 F9A4 0662 |