1 |
>>>>> On Tue, 14 Dec 2010, Jorge Manuel B S Vicetto wrote: |
2 |
|
3 |
> If you have anything you'd like to push to the council for |
4 |
> discussion, feel free to reply to this thread. |
5 |
|
6 |
I have two points: |
7 |
|
8 |
1. In your November meeting [1] you had asked me to prepare a PMS |
9 |
patch for the last missing features of EAPI 4. That series of patches |
10 |
is ready [2] except for the following: |
11 |
|
12 |
It has been suggested that the REQUIRED_USE variable should be |
13 |
accompanied by a new phase function that would be called if |
14 |
REQUIRED_USE assertions are not met. It would be similar to |
15 |
pkg_nofetch(). Specification of the function would be as follows: |
16 |
|
17 |
,---- |
18 |
| pkg_required_use |
19 |
| ================ |
20 |
| |
21 |
| For EAPIs listed in table [..] as supporting REQUIRED_USE, the |
22 |
| pkg_required_use function may be called when at least one assertion |
23 |
| of the REQUIRED_USE variable is not met. The function should output |
24 |
| an appropriate explanation why these assertions failed. It may also |
25 |
| suggest USE flags to be adjusted by the user. |
26 |
| |
27 |
| pkg_required_use must not write to the filesystem. |
28 |
`---- |
29 |
|
30 |
This is the last thing missing for EAPI 4, which I would really like |
31 |
to see finalised in 2010. (As the older ones of us may remember, ;-) |
32 |
its feature set was originally approved in April 2009, still called |
33 |
EAPI 3 at the time.) |
34 |
|
35 |
So, could you please decide in the upcoming council meeting if we |
36 |
should have a pkg_required_use function, or not? See Bug 347353 [3] |
37 |
for the current status of discussion. |
38 |
|
39 |
2. To avoid such long delays as we have seen with EAPI 4, I would also |
40 |
like to suggest the following: New EAPI dependent features shall only |
41 |
be accepted if a preliminary patch for portage exists. |
42 |
|
43 |
Ulrich |
44 |
|
45 |
[1] <http://www.gentoo.org/proj/en/council/meeting-logs/20101130-summary.txt> |
46 |
[2] MERGE_TYPE: <http://bugs.gentoo.org/attachment.cgi?id=256075> |
47 |
REQUIRED_USE: <http://bugs.gentoo.org/attachment.cgi?id=256799> |
48 |
No slot operator dependencies, no profile IUSE injection: |
49 |
<http://archives.gentoo.org/gentoo-pms/msg_4b6e62d644717d42a2b25919889d46ee.xml> |
50 |
[3] <http://bugs.gentoo.org/show_bug.cgi?id=347353> |