1 |
On Thu, 2006-07-20 at 11:11 +0000, Alex wrote: |
2 |
> > There was also a bug in catalyst that caused "xdm" to be added to the |
3 |
> > default runlevel for "generic-livecd" which was never how it was |
4 |
> > designed. Instead, the specific session was *supposed* to be started |
5 |
> > via "startx" as the user. |
6 |
> |
7 |
> so I do not understand the sense livecd/xdm option. Does it only change |
8 |
> the variable in /etc/rc.conf? |
9 |
|
10 |
Correct. |
11 |
|
12 |
> > What if you have more than one user and don't want any to login? |
13 |
> |
14 |
> Well, I just get irritated by the template-specfile. But if it would |
15 |
> automatically login by default, I would just keep the livecd/xdm option |
16 |
> blank in this case. |
17 |
|
18 |
Then file a bug with suggested changes. People who get "irritated" but |
19 |
don't bother to take a few minutes to try to *fix* the problem don't |
20 |
help us any. The truth is that we code catalyst for Release |
21 |
Engineering, first and foremost. The only reason pretty much *anything* |
22 |
works for non-releases is because interested users have supplied patches |
23 |
or given good ideas. |
24 |
|
25 |
Doing catalyst support for non-releases takes a significant enough |
26 |
portion of my time that I sometimes wish we'd never bothered to make an |
27 |
ebuild for it. ;] |
28 |
|
29 |
> > What if you don't want a particular session to be the default and want |
30 |
> > the user to choose? |
31 |
> > |
32 |
> |
33 |
> Just specify no xsession (that would by _my_ way) |
34 |
|
35 |
That would break horribly considering how the automatic login code is |
36 |
written. Instead, you'd get twm, rather than sitting at the display |
37 |
manager. Remember that things built with catalyst act *just* like a |
38 |
real Gentoo system does, configuration and all. |
39 |
|
40 |
I think that this is the thing lost on catalyst users the most. We |
41 |
expect *you* to make it do what you want, rather than for us to make it |
42 |
do what you want. In the old days of catalyst, we required people to |
43 |
make their own archscript/runscript to make changes, because catalyst |
44 |
only had a single function, to build releases. Things have changed |
45 |
quite a bit, but the underlying principle is still very similar. If you |
46 |
want something changed, write a patch. If it isn't a "bug" in the code, |
47 |
we're not very likely to work on it unless we find some spare time, |
48 |
which is rare, especially for Release Engineering. |
49 |
|
50 |
> I like the flexibility of gentoo and its sub-projects, without it I |
51 |
> probably wouldn't use it. I just get irritated by the comments in the |
52 |
> template-specfile, but you have said it is out of date, so I think this |
53 |
> is clarified (for me), except the sense of livecd/xdm and |
54 |
> livecd/xsession (I don't understand it). |
55 |
|
56 |
That wasn't exactly what I meant. It was out of date with the current |
57 |
state of the code at that time. The code was the problem. The template |
58 |
comments are 100% spot-on to what the design should be. It was a code |
59 |
refactoring done during this release cycle that *ever* caused "xdm" to |
60 |
be added to the default runlevel for non-release builds. It was a bug. |
61 |
The livecd/xdm and livecd/xsession only set the defaults for these |
62 |
variables on non-release builds. It is up to you to add "xdm" to your |
63 |
default runlevel if you want it to start at boot, as well as perform any |
64 |
configuration on the display manager and any X sessions that you wish |
65 |
customized. |
66 |
|
67 |
The general rule of thumb with catalyst is you have to do it yourself. |
68 |
Don't assume that anything is done for you. Unfortunately, more and |
69 |
more non-developers have been using catalyst. By that, I don't mean |
70 |
users, I mean people who don't look at the code to see what it does and |
71 |
doesn't do on their own. Right now, there is exactly one person |
72 |
actively maintaining catalyst. There are only two maintainers total. |
73 |
Because of this, the code changes much faster than any documentation |
74 |
does, except for the templates, which are kept (mostly) in sync with the |
75 |
actual code. In cases where the templates and code are not in sync, it |
76 |
is likely a bug in the code, and not the template. |
77 |
|
78 |
-- |
79 |
Chris Gianelloni |
80 |
Release Engineering - Strategic Lead |
81 |
x86 Architecture Team |
82 |
Games - Developer |
83 |
Gentoo Linux |