1 |
People seem to have started using blockers with package moves recently. |
2 |
For example, if cat/a is being moved to cat/b, people have started |
3 |
putting !cat/a as a dependency in cat/b. This is bad, for two reasons. |
4 |
|
5 |
First, you shouldn't have to do that. If package moves aren't working, |
6 |
we've got bigger problems, and throwing in some blockers won't help |
7 |
there. |
8 |
|
9 |
Second, when performing updates, Paludis also rewrites dependencies of |
10 |
installed packages to use the names. This means that your block on |
11 |
cat/a will be rewritten to a block on cat/b, which means the package |
12 |
ends up blocking itself. |
13 |
|
14 |
We've got two options here. |
15 |
|
16 |
Option the first is that people stop writing stupid blockers on package |
17 |
moves. Unless someone can come up with a convincing reason to keep them |
18 |
there, this is the option that should be taken, and repoman should |
19 |
enforce it. |
20 |
|
21 |
Option the second is that I make Paludis stop rewriting blockers for |
22 |
package moves. This is bad, because it means legitimate blockers which |
23 |
should be honoured will end up disappearing. |
24 |
|
25 |
Does anyone care to justify their "block the old name" habits? |
26 |
|
27 |
-- |
28 |
Ciaran McCreesh |