1 |
>>>>> On Tue, 15 Dec 2009, Fabian Groffen wrote: |
2 |
|
3 |
> With the current route where EAPI=3 will simply be EAPI=2 + |
4 |
> offset-prefix support, |
5 |
|
6 |
That's not entirely right, as EAPI 3 will also include mtime |
7 |
preservation. |
8 |
|
9 |
> Should an ebuild using an EAPI that has offset-prefix support make the |
10 |
> use of that support mandatory or optional? |
11 |
|
12 |
Optional. |
13 |
|
14 |
> In other words, one can perfectly fine write an ebuild EAPI=3 that |
15 |
> will not work in an offset-prefix install, due to improper absence |
16 |
> of EPREFIX, ED and EROOT. Should we allow this formally, or not? |
17 |
|
18 |
I'd say you can only claim Prefix support for an ebuild, when that |
19 |
ebuild has been tested under Prefix, i.e. when KEYWORDS contain at |
20 |
least one Prefix architecture. |
21 |
|
22 |
> The pros for forcing ebuilds to be offset-prefix aware are: |
23 |
> - an ebuild having EAPI >= 3 (as it looks now) is supposed to work |
24 |
> for Prefix users |
25 |
|
26 |
Again, we have KEYWORDS for this. |
27 |
|
28 |
> - hence also obviously is (supposed to be) checked for Prefix |
29 |
|
30 |
I don't see this as a pro. At least I don't want to delay any updates |
31 |
to EAPI 3 (which I need for mtime preservation), only to ensure that |
32 |
the ebuild is also working in Prefix. Most of my ebuilds in question |
33 |
aren't even in the Prefix overlay. |
34 |
|
35 |
Also there are some packages that are Linux-only and will never run on |
36 |
Prefix. Certainly we don't want to restrict them to EAPI <= 2 forever? |
37 |
|
38 |
Ulrich |