1 |
Brian Harring wrote: |
2 |
|
3 |
> On Tue, Jun 24, 2008 at 01:53:55AM +0200, Robert Buchholz wrote: |
4 |
>> Hi, |
5 |
>> |
6 |
>> I've stumbled upon an inconsitency between package managers the other |
7 |
>> day [1], which was due to both an ebuild and an eclass defining |
8 |
>> inconsisting KEYWORDS. |
9 |
>> |
10 |
>> bla-1.ebuild: |
11 |
>> inherit myeclass |
12 |
>> KEYWORDS="~arch" |
13 |
>> |
14 |
>> myeclass.eclass: |
15 |
>> KEYWORDS="arch" |
16 |
>> |
17 |
>> Portage will resolve this by overwriting the variable, so the last |
18 |
>> (~arch) wins. Paludis, on the other hand, merges the variables, so it |
19 |
>> is KEYWORDS="~arch arch". |
20 |
>> |
21 |
>> The PMS draft [2] defines that "IUSE, DEPEND, RDEPEND and PDEPEND" |
22 |
>> variables be merged when defined in both eclass and ebuild (Section |
23 |
>> 7.2), but only says "May be de?ned in an eclass" about KEYWORDS |
24 |
>> (Section 8.2). |
25 |
>> |
26 |
>> Anyone up to toss a coin whose bug it is, and maybe we can have a more |
27 |
>> specific wording in the PMS? |
28 |
> |
29 |
> Paludis bug; if you want KEYWORDS incremental, it'll need to be in |
30 |
>>=eapi2, too nasty of a change to shoehorn into existing (in use) |
31 |
> eapis. |
32 |
|
33 |
hmm, the program you use for posting should really have a delay function in |
34 |
case you respond too fast (1:25 according to my news reader, gmane and the |
35 |
assumption clocks are in sync). |
36 |
|
37 |
well, if the PMS doesn't say anything about it, it's a lack of specification |
38 |
and not a bug of a package manager. |
39 |
|
40 |
Can you please give more info why this is "too nasty"? |
41 |
|
42 |
Cheers, |
43 |
Tiziano |
44 |
|
45 |
|
46 |
-- |
47 |
gentoo-dev@l.g.o mailing list |