1 |
maillog: 01/11/2005-11:45:08(+0100): Jakub Moc types |
2 |
> 1.11.2005, 11:00:22, Thierry Carrez wrote: |
3 |
> |
4 |
> > Aren't those messages displayed after the damage is done ? Typical use : |
5 |
> |
6 |
> > - emerge --sync run as a daily cron job |
7 |
> > - emerge -a mysql |
8 |
> > - great, a new version is there. Typing "Yes" |
9 |
> > - system gets borken |
10 |
> > - emerge spits out message saying 14 files need updating and there is 1 |
11 |
> > unread news item |
12 |
> |
13 |
> > I'm probably missing something here. Please elaborate on how this GLEP |
14 |
> > meets the "Preemptive" design goal... |
15 |
> |
16 |
> |
17 |
> I'm probably missing something obvious here, because I can't see why *existing* |
18 |
> emerge --changelog code cannot be recycled for this feature to display upgrade |
19 |
> messages when running emerge -uDav world... |
20 |
|
21 |
That reminds me of the idea to stick tags in the ChangeLog: |
22 |
http://groups.google.com/group/linux.gentoo.dev/msg/8f2dc84619be5c5b?fwc=1 |
23 |
|
24 |
But still, I'm guessing the idea of "--news" is to tell people that they |
25 |
need to do something A.S.A.P. This means as soon as the news are |
26 |
obtained, and the users are nagged about the news on *every invocation |
27 |
of emerge*, similar to the /etc messages, and not only when they decide |
28 |
to install some package, which is when --changelog kicks in. |
29 |
|
30 |
And then, I am not sure why glsa-check cannot do the same job... |
31 |
|
32 |
-- |
33 |
() Georgi Georgiev () Computers are unreliable, but humans are () |
34 |
() chutz@×××.net () even more unreliable. Any system which () |
35 |
() http://www.gg3.net/ () depends on human reliability is () |
36 |
() ------------------- () unreliable. -- Gilb () |