1 |
On Saturday, September 11, 2010 22:51:23 Ryan Hill wrote: |
2 |
> On Sun, 12 Sep 2010 20:59:25 +1200 Alistair Bush wrote: |
3 |
> > There should be nothing stopping a user from running a mixed arch/~arch |
4 |
> > system. Those problems just point to our dependency information not |
5 |
> > being recorded correctly. It might be understandable that this info |
6 |
> > can be incredibly hard to get correct but that doesn't mean it isn't a |
7 |
> > valid bug. |
8 |
> |
9 |
> It's invalid as soon as you bring system-set packages into the mix, which |
10 |
> falls outside of dependency correctness. |
11 |
> |
12 |
> The trouble is knowing if this applies in the situation you're looking |
13 |
> into. |
14 |
|
15 |
indeed. for people who disagree, feel free to go through every single stable |
16 |
package in the tree that breaks with glibc-2.12, or linux-headers-2.6.35, or |
17 |
make-3.82, or xxx and add a blocker against the newer versions of these |
18 |
packages. and then do it every time we get a new version. |
19 |
|
20 |
or, let's keep our sanity and continue doing what we've been doing over the |
21 |
years -- stabilize newer versions of packages that do work with the stable & |
22 |
unstable versions of the packages they build against. |
23 |
|
24 |
people who install unstable core system packages way before we're interested |
25 |
in stabilizing and then file bugs that only apply to stable get INVALID->their |
26 |
problem. especially when a simple bugzilla search would have told them that |
27 |
their issue is already fixed in the unstable version of whatever package is |
28 |
failing for them. |
29 |
-mike |