1 |
On Tue, 05 May 2020 22:19:59 +0200 |
2 |
Michał Górny <mgorny@g.o> wrote: |
3 |
|
4 |
> Hi, |
5 |
> |
6 |
> TL;DR: should NATTkA reject request to keyword on arch if the ebuild has |
7 |
> '-arch' (or '-*') in KEYWORDS already? |
8 |
> |
9 |
> |
10 |
> Background: I've recently been rekeywording two packages that gained |
11 |
> dependency on gevent. When I was mass-requesting rekeywording, it |
12 |
> escaped my attention that gevent is explicitly marked '-ia64'. The arch |
13 |
> team apparently got mad at me and added gevent to their package.mask to |
14 |
> make its breakage more explicit. |
15 |
> |
16 |
> I think it would make sense if NATTkA detected '-ia64' there and told me |
17 |
> that the package is keyword-masked on ia64. |
18 |
> |
19 |
> The flip side is that it would prevent people from using NATTkA to |
20 |
> restore keywords that were marked '-arch' before. Of course, if this |
21 |
> would ever be necessary it could easily be resolved via removing '-arch' |
22 |
> first or adding some extra hack. |
23 |
> |
24 |
> WDYT? |
25 |
|
26 |
Play it safe. -* is frequently used for binary packages where an arch |
27 |
will simply either work or it won't, with little likelihood of the |
28 |
situation changing. -arch is so rare that I don't recall ever seeing |
29 |
it. In either case, restoring an arch should be an explicit action. |
30 |
|
31 |
-- |
32 |
James Le Cuirot (chewi) |
33 |
Gentoo Linux Developer |