1 |
On 27-02-2008 13:56:51 +0000, Roy Marples wrote: |
2 |
> |
3 |
> On Wed, 27 Feb 2008 13:29:15 +0100, Fabian Groffen <grobian@g.o> |
4 |
> wrote: |
5 |
> > Well... that's great! But a jail or a (ch)root is in general not the |
6 |
> > same as a "prefix". |
7 |
> |
8 |
> No, but it's the same kettle of fish as chroots, jails and vps systems - |
9 |
> basically |
10 |
> there is a need to disable dependencies that provide what the host already |
11 |
> does. |
12 |
|
13 |
Ok, the host will for instance do "net", so "need net" should indeed not |
14 |
fail. However I could imagine that "need net" would just get satisfied |
15 |
or something, like by a dummy. |
16 |
|
17 |
> We current have nojail for FreeBSD jails, novps for VServer/OpenVZ systems |
18 |
> and |
19 |
> a few others. I would be trivial to add another no for prefix :) |
20 |
|
21 |
I just need the machinery of "runscript" as first thing, I suppose. If |
22 |
we need a dozen of no* things for that, it probably indicates some |
23 |
problem, but could work for me. I want a framework to start and stop |
24 |
daemons in Prefix, and it feels obvious that we can reuse existing code |
25 |
for that. |
26 |
|
27 |
> > I have to look more closely at what openrc does |
28 |
> > these days, but for the (ancient) version of baselayout we have in |
29 |
> > prefix now, I recall that: |
30 |
> > a) most of it didn't compile on Darwin and Solaris |
31 |
> |
32 |
> It compiles and works on Linux/glibc/uclibc, FreeBSD-6 and NetBSD-4. |
33 |
> So it stands a fair chance of working on Darwin for sure. |
34 |
|
35 |
Well... I've some experience here, and I'm not as sure as you ;) |
36 |
Anyway, I concur the codebase has changed dramatically since, and |
37 |
probably in favour of portability. |
38 |
|
39 |
> I have no idea about Solaris, but it should work as it sports libkvm which |
40 |
> we use to find processes. |
41 |
|
42 |
Part of the summer of code project to me would be to 1) evaluate to what |
43 |
extent this is all necessary in the Prefix equivalent and 2) create/fix |
44 |
the code. |
45 |
|
46 |
> > And maybe even a sort of init-level stuff, such that one can start all |
47 |
> > services in the Prefix and stop them as well. That basically gets quite |
48 |
> > useful once Prefix goes "privileged" and you could start sshd, slapd, |
49 |
> > apache2, etc, etc. on privileged ports, and you really would like those |
50 |
> > to be started as well in some correct order (on e.g. Solaris). |
51 |
> |
52 |
> If OpenRC compiles and /bin/sh points to a POSIX shell it should work as it |
53 |
> stands. |
54 |
|
55 |
Ok, then we already fail here. |
56 |
/bin/sh is no way POSIX, it is just bourne, so that's where we come in |
57 |
and simply use /usr/bin/env {sh,bash,posix-sh} or a full path to make |
58 |
your assumption true. |
59 |
|
60 |
> At present there is no need for the default interpreter to be changed, but |
61 |
> there may |
62 |
> be the need for Prefix. |
63 |
|
64 |
See above. But that's trivial work, that we do all the time. For the |
65 |
GSoC I see more challenges in the rest of the job and to make some |
66 |
obvious examples. |
67 |
|
68 |
But then again, it was just a mere suggestion. If everything is already |
69 |
there then fine, but we still need someone (Google code or not) to do |
70 |
it, as it's currently not. I'm not sure how far OpenRC actually can |
71 |
deal with unprivileged installs, so that are just things we have to find |
72 |
out along the way. |
73 |
|
74 |
|
75 |
-- |
76 |
Fabian Groffen |
77 |
Gentoo on a different level |
78 |
-- |
79 |
gentoo-dev@l.g.o mailing list |