1 |
William Hubbs posted on Mon, 24 Feb 2014 02:14:54 -0600 as excerpted: |
2 |
|
3 |
>> Meanwhile, based on a decade plus of gentoo user experience, in |
4 |
>> practice, emerge --config is obscure enough that using it in an ebuild |
5 |
>> tends to hide it. |
6 |
|
7 |
> Think of /etc/runit/runsvdir as being similar to /etc/runlevels in |
8 |
> OpenRc. |
9 |
> It is a directory that gets set up once, then only the admin should mess |
10 |
> around in there by adding or removing directories (runlevels) and |
11 |
> symbolic links to services. |
12 |
|
13 |
Hmm... Then pkg_config does seem an appropriate "hiding" and the parallel |
14 |
to bind is even more striking. IIRC, when it massively reorganized the |
15 |
chroot, it put an ewarn in pkg_postinst with some instructions, including |
16 |
asking people to re-run emerge --config bind. IIRC it did NOT run that |
17 |
reconfigure itself. |
18 |
|
19 |
Now making it a USE flag is a bit different, but the idea of making the |
20 |
function independent and simply calling it from both pkg_config and |
21 |
(conditionally on the use flag) pkg_postinst or whatever does seem |
22 |
reasonable to me. |
23 |
|
24 |
And (as you may well be aware) I'd very likely disable that USE flag here |
25 |
(well, USE=-* and not enable it elsewhere), because my init setups tend |
26 |
to make rather complex use of advanced out of the ordinary corner-case |
27 |
features, thus the reason I ultimately ended up on openrc-9999 and |
28 |
generally git-logging every single change before I even install the |
29 |
update, let alone reboot. But hey, I've found and reported a number of |
30 |
openrc bugs so they could be fixed before they ever appeared in a release |
31 |
over the years as a result, too. =:^) |
32 |
|
33 |
-- |
34 |
Duncan - List replies preferred. No HTML msgs. |
35 |
"Every nonfree program has a lord, a master -- |
36 |
and if you use the program, he is your master." Richard Stallman |