1 |
On Tue, 2006-10-17 at 21:19 -0400, Preston Cody wrote: |
2 |
> Quickstart will become the primary method for automated installations. |
3 |
> It is designed to be able to be netbooted and will be the installer |
4 |
> integrated into Scire. It may also be added to the minimal livecd at |
5 |
> some point (why not, right?). Quickstart works with blank drives |
6 |
> only, and is not too user-friendly. Currently no frontend or |
7 |
> configurator exists though I have plans for a web-based one that will |
8 |
> plug in to Scire nicely. |
9 |
|
10 |
Sounds good. It would be nice to have at least one interface that isn't |
11 |
web-based, but that's not vital. |
12 |
|
13 |
> GLI will be refactored to become slightly more interactive. There |
14 |
> will probably be about four main steps/stopping points/actions that |
15 |
> will occur. |
16 |
|
17 |
Hrrmn... |
18 |
|
19 |
> 1. Setup networking, logfiles, chroot_dir and such.. i.e. all the |
20 |
> client_configuration stuff. This already is done interactively from |
21 |
> the two main frontends. We can thus chuck the entire client_profile |
22 |
> (yea!!!!) |
23 |
|
24 |
Are we talking things like net-setup? What exactly are we talking about |
25 |
using here? |
26 |
|
27 |
> 2. Drop to cfdisk or diskdruid to let the user do partitioning and |
28 |
> probably filesystem formatting too. This lets the user deal with |
29 |
> partitioning themselves and thus we don't get the blame when they |
30 |
> screw it up! :) |
31 |
|
32 |
That's probably a good idea, but really sucks for the people that don't |
33 |
know what they want. That was the nice thing about the current GLI, |
34 |
being able to click that "Recommended" button. Of course, you could |
35 |
limit the "Recommended" concept to only work with known-good |
36 |
partitioning schemes. |
37 |
|
38 |
For example: |
39 |
|
40 |
blank drive - we know we can partition this |
41 |
1 primary partition (and free space) - we know we can handle this |
42 |
|
43 |
...and that's about it. You would cover a vast majority of the users |
44 |
with this option. I would think (using the GTK+ installer as an |
45 |
example) that having something like the current partitioning screen + an |
46 |
"Advanced" button, which drops to a shell to allow the user to partition |
47 |
manually, using pretty much anything that they want (LVM, EVMS, DMRAID, |
48 |
MDRAID) would cover all the bases. I think having it as "opt-in" on the |
49 |
"easy" screen and only allowing known-good configurations should resolve |
50 |
most of the issues with partitioning. |
51 |
|
52 |
> 3. Define mountpoints.. these will replace the partition layout in the |
53 |
> install profile. Then we fetch/install the stage tarball and/or |
54 |
> portage. We will still keep the dynamic stage3 (as well as making |
55 |
> command-line tools for its use), because the installer-livecd isn't |
56 |
> changing. Having this step separate lets us choose profiles and |
57 |
> gather updated info about things like USE flags. |
58 |
|
59 |
I like it. |
60 |
|
61 |
> 4. Everything else. There really isn't much benefit to separate out |
62 |
> the rest of the steps. We'll let people drop to a shell to make their |
63 |
> own kernel config, but I don't see much beyond that and I still like |
64 |
> the ability to configure at once and then let it go for a while. |
65 |
|
66 |
Absolutely. I think the "configure and forget" aspect of the installer |
67 |
is one of its greatest strengths. |
68 |
|
69 |
> These changes will drastically simplify the partitioning code and make |
70 |
> GLI much easier to port to other architectures. |
71 |
|
72 |
One thing that I have never understood is why all of GLI is in python. |
73 |
Take catalyst as an example. It uses python for the places where python |
74 |
does best and uses bash for the rest. It makes it very easy to |
75 |
understand when it calls out to smaller, interchangeable scripts for |
76 |
sub-tasks. I'm not saying that this is what the GLI needs to do, but it |
77 |
could definitely make things easier, and also reduce the memory |
78 |
footprint of GLI, when loaded. |
79 |
|
80 |
> The plan is to release a version of GLI like it is now with bugfixes |
81 |
> only for 2007.0 and target the new GLI (god forbid, should we call it |
82 |
> GLI 2.0? <cringe>) for 2007.1. If enough progress has been made, we |
83 |
> can try to put an experimental copy on the livecd and hide it (doesn't |
84 |
> make much sence to have an entirely different /experimental livecd |
85 |
> just for that). The new GLI will be written first and foremost with |
86 |
> gli-dialog as the frontend. The GTK frontend is quite bloated and a |
87 |
> pain to change and maintain, so it may disappear entirely and that is |
88 |
> ok by the installer devs. |
89 |
|
90 |
That's OK by me, too. We'll get some flak for it, I'm sure, but at this |
91 |
point, I don't think we can do *anything* without getting flak, so our |
92 |
best course of action is to just do what we think is best. |
93 |
|
94 |
> That's about it, if you have any questions/comments, feel free to post |
95 |
> them. This list has like no traffic so it could use some discussion |
96 |
> every once and awhile. |
97 |
|
98 |
Well, if you jokers didn't discuss everything on IRC... ;] |
99 |
|
100 |
-- |
101 |
Chris Gianelloni |
102 |
Release Engineering Strategic Lead |
103 |
Alpha/AMD64/x86 Architecture Teams |
104 |
Games Developer/Council Member/Foundation Trustee |
105 |
Gentoo Foundation |