1 |
Quoting Sebastian Pipping (2012-09-09 20:02:10) |
2 |
> On 09/09/2012 07:10 PM, Richard Yao wrote: |
3 |
> > We could just refactor genkernel by doing the following: |
4 |
> > |
5 |
> > 1) Copying /usr/bin/genkernel to /usr/bin/geninitramfs. |
6 |
> |
7 |
> At that place I would add a step: |
8 |
> |
9 |
> 1.1: Make a new pass-through wrapper /usr/bin/genkernel that forwards |
10 |
> all valid optiuons to geninitramfs. Over time we can "move" features |
11 |
> from Bash-based geninitramfs to the Python-based genkernel. That way |
12 |
> we save ourselves a big bang problem, several branches. I'm |
13 |
> experiemnting with that right now. With luck I have something in a |
14 |
> few hours. |
15 |
|
16 |
There are not going to be any branches. Genkernel4 is going to have new |
17 |
repo, because it is being written from scratch. And geninitramfs is |
18 |
going to be moved to new repo, because of name change from genkernel to |
19 |
geninitramfs. Separate repos make life easier, because geninitramfs can |
20 |
live its own life and genkernel4 can live on its own, too. After |
21 |
helping with moving genkernel to geninitramfs I'd like to focus on |
22 |
Genkernel4 and Dracut only and not interfere in geninitramfs in any way. |
23 |
|
24 |
Also please don't waste your time for writing wrappers (or doing |
25 |
anything ad-hoc). It is not necessary, as I have described in "[RFC] |
26 |
Genkernel4 and Geninitramfs plan". |
27 |
|
28 |
|
29 |
-- |
30 |
Amadeusz Żołnowski |