1 |
This is one reason you should not do auto-updates via cron. It adds to |
2 |
the fun in tracking down such problems. |
3 |
|
4 |
I also wish that emerges would stop whenever a message like this is |
5 |
printed to the screen, but people have been cursing/moaning and getting |
6 |
caught with dead systems by this bug for literally years - check out the |
7 |
bugs for 2002 (einfo has some of them) for instance. Some are marked |
8 |
resolved, but its still broken - unless they never told anyone how to |
9 |
enable the fix. |
10 |
|
11 |
BillK |
12 |
|
13 |
|
14 |
On Fri, 2005-07-08 at 14:28 +0800, Ow Mun Heng wrote: |
15 |
> [Changed Subject. , should be com_err and not ss] |
16 |
> |
17 |
> On Thu, 2005-07-07 at 23:04 -0700, Zac Medico wrote: |
18 |
> > Ow Mun Heng wrote: |
19 |
> > > If you guys did a emerge -uDav world anytime these past few days. I'm |
20 |
> > > sure you'll be hit by the libcom_err.so.3 error. |
21 |
> > > |
22 |
> > > This was really bad. Ruined half my working day since I can't get Evo to |
23 |
> > > come up. |
24 |
> > > |
25 |
> > > Anyway, the ebuild _does_ mentioned breaking compatibility and stuffs. |
26 |
> > > |
27 |
> > > |
28 |
> > > $grep einfo /usr/portage/sys-libs/com_err/*.ebuild |
29 |
> > > |
30 |
> > > einfo "PLEASE PLEASE take note of this" |
31 |
> > > einfo "Please make *sure* to run revdep-rebuild now" |
32 |
> > > einfo "Certain things on your system may have linked against a" |
33 |
> > > einfo "different version of com_err -- those things need to be" |
34 |
> > > einfo "recompiled. Sorry for the inconvenience" |
35 |
> > > |
36 |
> > > Which is a huge bummer since most of us don't go checking the ebuilds |
37 |
> > > for these specific lines. How I wish It would just abort and BLAST the |
38 |
> > > Einfo on the screen for me to see the next day and decide to do it |
39 |
> > > manually. |
40 |
> > > |
41 |
> > > Oh well... It's fixed. |
42 |
> > > |
43 |
> > |
44 |
> > Shouldn't revdep-rebuild solve issues like that? I always run revdep-rebuild after updates and I don't recall having any problem with this com_err update ( I know lots of others did though :-( ). |
45 |
> |
46 |
> It didn't catch everything. Right now I'm recompiling xscreensaver. |
47 |
> |
48 |
> The thing which I'm talking about is, if there are _specific_ caveats |
49 |
> such as with this new app, it shouldn't even go ahead and |
50 |
> compile/upgrade/update. It should just shoot the ebuild einfo and then |
51 |
> force manual intervention. |
52 |
> |
53 |
> Updates are usually done at night. Output is not seen. That's why. If I |
54 |
> knew that it might break anything, I wouldn't leave it overnight to |
55 |
> update and then go to work and find Evo(or important "X" app) wouldn't |
56 |
> work and cause undue frustration. |
57 |
> |
58 |
> |
59 |
> -- |
60 |
> Ow Mun Heng |
61 |
> Gentoo/Linux on DELL D600 1.4Ghz 1.5GB RAM |
62 |
> 98% Microsoft(tm) Free!! |
63 |
> Neuromancer 14:25:41 up 22:22, 9 users, load average: 1.94, 1.73, 1.22 |
64 |
> |
65 |
> |
66 |
|
67 |
-- |
68 |
gentoo-user@g.o mailing list |