1 |
> I think remi was more speaking about incorrect deps (say misplaced in |
2 |
> RDEPEND) than problems concerning the package manager. |
3 |
> |
4 |
> In any case, openssl is the perfect example of what can go wrong because |
5 |
> of upstream's behavior. The problem is that program A compiled against |
6 |
> version X of openssl won't work with version Y>X. Currently we need to |
7 |
> keep X's libs around and run revdep-rebuild to fix this. |
8 |
|
9 |
Right, my example was mildly contrived. But I've run into this same issue with |
10 |
packages like ruby and glibc, where the build system had newer versions and you run |
11 |
into symbol issues (or errors like "invalid binary format") because you need to |
12 |
upgrade underlying libraries. |
13 |
|
14 |
> Most librairies don't cause this problem though so I don't really see |
15 |
> this as a bug on the gentoo side even if it's annoying. |
16 |
|
17 |
It's not really Gentoo's fault, no, but it's a problem that could be somewhat fixed. |
18 |
|
19 |
> Anyway, to keep machines using binary in sync without much headache, my |
20 |
> current solution is to use a squashfsed portage tree with --deep. It |
21 |
> works pretty well. |
22 |
|
23 |
Agreed, but the problem is (at least in my case) we're talking about production |
24 |
machines that are actively running, and the customer needs an upgrade of a package |
25 |
but we don't want to take a chance at ruining something else by upgrading --deep if |
26 |
we can help it. |
27 |
|
28 |
From their perspective, they just want it to work (and don't care about what has to |
29 |
be upgraded), but from a sysadmin perspective it's a difficult problem to solve over |
30 |
time, especially when you have 10+ other sysadmins, who all may not know that when |
31 |
you upgrade package X be sure to remember to also upgrade packages Y and Z at the |
32 |
same time or you'll run into problems. |
33 |
|
34 |
-- |
35 |
gentoo-dev@l.g.o mailing list |