1 |
-----BEGIN PGP SIGNED MESSAGE----- |
2 |
Hash: SHA256 |
3 |
|
4 |
On 30/08/12 09:14 AM, Rich Freeman wrote: |
5 |
> On Thu, Aug 30, 2012 at 9:04 AM, Ian Stakenvicius <axs@g.o> |
6 |
> wrote: |
7 |
>> |
8 |
>> The primary benefit to the policy that dev's should bump EAPI |
9 |
>> when bumping ebuilds is so that older inferior EAPIs can be |
10 |
>> deprecated and eventually removed from the tree. |
11 |
> |
12 |
> What is the benefit from removing the old EAPIs? |
13 |
|
14 |
Simpler eclasses is the first thing that comes to mind. Consistency |
15 |
in the way the dependency tree is built would be another (when newer |
16 |
EAPIs address such things, that is) |
17 |
|
18 |
> |
19 |
> Simply bumping an ebuild to EAPI=5 doesn't even guarantee that |
20 |
> either of those features would be used anyway. |
21 |
|
22 |
..although this could technicaly be true, I think most developers |
23 |
would assume that bumping EAPI doesn't mean changing the EAPI variable |
24 |
from whatever to 5 and that's it; the "bump" should involve |
25 |
integrating any applicable features that the new EAPI offers. |
26 |
|
27 |
> |
28 |
> If there is a benefit from some specific practice, then let's |
29 |
> adopt it. However, I don't think that is the same as just bumping |
30 |
> EAPIs for their own sake. |
31 |
> |
32 |
> When there is a benefit to adopting a new EAPI of course |
33 |
> maintainers should try to take advantage of it. If there are |
34 |
> specific changes we want to try to make tree-wide let's try to do |
35 |
> that too. But, why bump ebuilds from 0 to 1 to 2 to 3 to 4 to 5 |
36 |
> when your only example of an end-user benefit would have been |
37 |
> achieved if we just bumped from 0 to 5 in one step? |
38 |
|
39 |
|
40 |
We used to have a policy that the oldest EAPI that implements all |
41 |
features needed for an ebuild is the one that should be used. Now (as |
42 |
of when EAPI=4 was made official i think?) it's the reverse -- the |
43 |
newest (official) EAPI is the one that should be used. All this |
44 |
policy change suggestion scarabeus provided is doing, is recommending |
45 |
that developers bump EAPI when they bump their ebuilds, as compared to |
46 |
only when writing new ones (which is all the current policy would |
47 |
require us to do). |
48 |
|
49 |
if you want another example, i'm sure end-user benefits would have |
50 |
ensued if many existing packages that die in pkg_setup were bumped to |
51 |
EAPI=4 right away and their checks moved to pkg_pretend. Examples |
52 |
prior to EAPI=4 are irrelevent as the policy was different then. |
53 |
|
54 |
-----BEGIN PGP SIGNATURE----- |
55 |
Version: GnuPG v2.0.19 (GNU/Linux) |
56 |
|
57 |
iF4EAREIAAYFAlA/a6sACgkQ2ugaI38ACPBO5wD+JSBmTT3j0MFc1GMjIDatRLnV |
58 |
J7Oj3rQWjS3GKpU8pQgBALsVg7R1QGGjETv0KS3j9yxBlflr4PlKECboVXqoRupL |
59 |
=+zxo |
60 |
-----END PGP SIGNATURE----- |