1 |
On Mon, 16 Nov 2015 10:29:43 +0100 |
2 |
"Justin Lecher (jlec)" <jlec@g.o> wrote: |
3 |
|
4 |
> -----BEGIN PGP SIGNED MESSAGE----- |
5 |
> Hash: SHA512 |
6 |
> |
7 |
> On 16/11/15 10:14, Alexis Ballier wrote: |
8 |
> > Probably those that want to ban it should fix the(ir) tree so that |
9 |
> > developers have no pain in bumping to eapi6? |
10 |
> |
11 |
> Versioned APIs are made to have incompatible changes. What do you like |
12 |
> to see? |
13 |
|
14 |
deprecation warnings for some time or.. |
15 |
|
16 |
> Someone dropping all usages of that eclass from all ebuilds |
17 |
> which are using it so that the maintainer can bump without thinking? |
18 |
|
19 |
this would be preferred :) |
20 |
|
21 |
[...] |
22 |
> > While I agree we should move away from those eclasses, the "I |
23 |
> > decided to throw the crap at other developers with eapi6 without |
24 |
> > deprecation period" is a bit hard to grasp. Esp. when these |
25 |
> > eclasses were advertised as the way to go not so long ago... |
26 |
> > |
27 |
> |
28 |
> I don't really understand what deprecation you like to see? |
29 |
|
30 |
|
31 |
RepoMan scours the neighborhood... |
32 |
inherit.deprecated 1 |
33 |
x11-wm/xmonad-contrib/xmonad-contrib-0.11.2.ebuild: please migrate |
34 |
from 'base' (no replacement) on line: 10 |
35 |
|
36 |
|
37 |
these warnings have been there for ages and for several eclasses |
38 |
|
39 |
> We cannot |
40 |
> use EAPI=6 right now and when it starts to exist, nothing will be |
41 |
> broken. So you have some to time to adopt your thinking until you |
42 |
> write your first ebuild in EAPI=6. |
43 |
> |
44 |
> At which particular point do you seen problems coming up? What do you |
45 |
> think will make maintainers struggle with that change? |
46 |
|
47 |
Right. With our always decreasing, soon to be negative, number of |
48 |
open bugs, all those shiny areas where fellow developers spend more time |
49 |
looking for something to do rather than doing it, we should be thankful |
50 |
to have at least some ebuild rewrite to do ! :) |
51 |
|
52 |
More seriously, the problem is not in the technical way it is done |
53 |
(changing eclass API with EAPI is a nice proper way), but in adding |
54 |
useless burden. |