| 1 |
On Fri, 23 Feb 2007 22:56:19 +0900 Jason Stubbs <jstubbs@g.o> |
| 2 |
wrote: |
| 3 |
| Disallowing it would be the cleaner in terms of package manager |
| 4 |
| responsibilities, but ... |
| 5 |
|
| 6 |
Well, I looked through the tree. |
| 7 |
|
| 8 |
There is exactly one package using this construct that doesn't get it |
| 9 |
wrong. That package is openoffice, and it uses it to do this: |
| 10 |
|
| 11 |
java? ( || ( !amd64? ( =virtual/jdk-1.5* ) =virtual/jdk-1.4* ) |
| 12 |
dev-java/ant-core ) |
| 13 |
|
| 14 |
The other fourteen packages using it are either making the mistake |
| 15 |
described in the original email in the thread, or using it where no use |
| 16 |
flag is required at all. |
| 17 |
|
| 18 |
Given that the one legitimate case can easily be rewritten in another |
| 19 |
way with no loss of functionality, is there really a justification for |
| 20 |
keeping this? |
| 21 |
|
| 22 |
-- |
| 23 |
Ciaran McCreesh |
| 24 |
Mail : ciaranm at ciaranm.org |
| 25 |
Web : http://ciaranm.org/ |
| 26 |
Paludis, the secure package manager : http://paludis.pioto.org/ |
Archives