1 |
Am Samstag, 16. August 2003 16:31 Paul de Vrieze wrote: |
2 |
> Well, for a user it works, but it does not work in terms of providing a |
3 |
> cache. Normally a cache is built. One thing needed for that is that we can |
4 |
> parse dependencies. |
5 |
|
6 |
Oh, I see. As far as I understand now, it works in this way: |
7 |
emerge rsyncs and then generates the cache in /var/vache/edb/dep. |
8 |
At this time, no ebuild is executed as a script and therefore the dependencies |
9 |
cannot be calculated. That I did not know, thanks for enlighten me, Paul. |
10 |
|
11 |
So, if I understand things right, it has to go this way: |
12 |
On the rsync-Servers there should be a file containing GLSAs. This file |
13 |
contains for example the following lines: |
14 |
VERSION 1.0 |
15 |
nfs-utils >=nfs-utils-1.0.4 |
16 |
gnupg >=gnupg-1.2.2-r1 |
17 |
|
18 |
*After* rsyncing, but *before* building the cache, emerge should create a |
19 |
statical version of app-admin/security/security-VERSION.ebuild with the |
20 |
dependencies corresponding to the current system and the correct version. |
21 |
After creating this ebuild, the cache could be build and an emerge -p security |
22 |
will show the needed updates. That seems to me a few lines of python code and |
23 |
an introduction of the new GLSA file. |
24 |
|
25 |
How is that? |
26 |
|
27 |
Karsten |
28 |
|
29 |
|
30 |
|
31 |
|
32 |
|
33 |
-- |
34 |
gentoo-dev@g.o mailing list |