1 |
On Wed, 2004-01-21 at 03:32, Pieter Van den Abeele wrote: |
2 |
> Hey, |
3 |
> |
4 |
> I tried building ppc livecds with catalyst and genkernel yesterday. I |
5 |
> have some remarks. |
6 |
|
7 |
OK, I just committed a couple of improvements to the code: |
8 |
|
9 |
I added a new file_locate() function to modules/catalyst_support.py |
10 |
which provides a generalized way to try to see if a file exists, and if |
11 |
not, an exception is raised. file_locate() is now called for the |
12 |
snapshot, source_path and the distdir. catalyst now aborts right away if |
13 |
any of these files cannot be found. |
14 |
|
15 |
Also, livecd-stage1 no longer gets tarred up. livecd-stage2 does a cp -a |
16 |
from /var/tmp/catalyst/tmp/<profile>/livecd-stage1-<stamp>. The cp -a is |
17 |
necessary because livecd-stage2 cleans and preps the chroot area, and we |
18 |
need to keep the pristine livecd-stage1 version if we need to tweak |
19 |
anything in the livecd-stage2 process. |
20 |
|
21 |
In theory, livecd-stage1 and livecd-stage2 could be combined, and the |
22 |
recreation of the pristine livecd-stage1 environment could happen by |
23 |
unpacking the stage3 and using the livecd-stage1 package cache. This |
24 |
would save up to a GB or so of disk storage. And reduce the number of |
25 |
livecd spec files from 2 to 1. If people are finding the disk usage to |
26 |
still be excessive, I can do this. |
27 |
|
28 |
But the current changes I just committed should save, say, 500-700MB of |
29 |
disk usage and a lot of tar.bz2 compress/uncompress time. We'll need to |
30 |
see if this is enough. |
31 |
|
32 |
Regards, |
33 |
|
34 |
Daniel |
35 |
|
36 |
|
37 |
-- |
38 |
gentoo-releng@g.o mailing list |