| 1 |
On Mon, Feb 02, 2004 at 09:03:17PM +0100 or thereabouts, Paul de Vrieze wrote: |
| 2 |
> Could you explain why you think that architecture specific stable keywords are |
| 3 |
> necessary? Would that not create a too big strain on the arch developers. If |
| 4 |
> a package is stable on an arch shouldn't it also be automatically a candidate |
| 5 |
> for the stable tree? |
| 6 |
|
| 7 |
Well, let's take BerkDB as an example. |
| 8 |
|
| 9 |
Currently, db-4.1.25_p1-r3 has the following keywords: |
| 10 |
|
| 11 |
KEYWORDS="ia64 ~x86 ppc ~alpha amd64 ~sparc ~mips ppc64 hppa" |
| 12 |
|
| 13 |
Then, db-4.0.14-r2.ebuild has: |
| 14 |
|
| 15 |
KEYWORDS="x86 ~ppc sparc alpha mips hppa ~arm amd64 ia64" |
| 16 |
|
| 17 |
So which one would be marked stable? The first is ~x86 but ppc, the second is |
| 18 |
~ppc but x86. |
| 19 |
|
| 20 |
Also, what happens with ~x86 ebuilds that are marked stable on other arches |
| 21 |
(such as db-4.1.25) Should AMD64, Itanium, PPC and HPPA users all be |
| 22 |
forced to use an older version of db? Or do we force x86 users to use a |
| 23 |
version of db that is currently ~masked? |
| 24 |
|
| 25 |
--kurt |
Archives