| 1 |
On Tue, Jun 24, 2008 at 01:53:55AM +0200, Robert Buchholz wrote: |
| 2 |
> Hi, |
| 3 |
> |
| 4 |
> I've stumbled upon an inconsitency between package managers the other |
| 5 |
> day [1], which was due to both an ebuild and an eclass defining |
| 6 |
> inconsisting KEYWORDS. |
| 7 |
> |
| 8 |
> bla-1.ebuild: |
| 9 |
> inherit myeclass |
| 10 |
> KEYWORDS="~arch" |
| 11 |
> |
| 12 |
> myeclass.eclass: |
| 13 |
> KEYWORDS="arch" |
| 14 |
> |
| 15 |
> Portage will resolve this by overwriting the variable, so the last |
| 16 |
> (~arch) wins. Paludis, on the other hand, merges the variables, so it |
| 17 |
> is KEYWORDS="~arch arch". |
| 18 |
> |
| 19 |
> The PMS draft [2] defines that "IUSE, DEPEND, RDEPEND and PDEPEND" |
| 20 |
> variables be merged when defined in both eclass and ebuild (Section |
| 21 |
> 7.2), but only says "May be defined in an eclass" about KEYWORDS |
| 22 |
> (Section 8.2). |
| 23 |
> |
| 24 |
> Anyone up to toss a coin whose bug it is, and maybe we can have a more |
| 25 |
> specific wording in the PMS? |
| 26 |
|
| 27 |
Paludis bug; if you want KEYWORDS incremental, it'll need to be in |
| 28 |
>=eapi2, too nasty of a change to shoehorn into existing (in use) |
| 29 |
eapis. |
| 30 |
|
| 31 |
Cheers, |
| 32 |
~harring |
Archives