| 1 |
Santiago M. Mola wrote: |
| 2 |
> |
| 3 |
> I've been talking about it with some users and everyone agrees that |
| 4 |
> they would like to have such an interface... |
| 5 |
> |
| 6 |
> What do you think about? Would it be easy to integrate it with |
| 7 |
> packages.g.o or should it belong somewhere else? Do you think this is |
| 8 |
> a suitable project for SoC? |
| 9 |
|
| 10 |
I like the idea, although it is a bit redundant with bugzilla. One |
| 11 |
thing that would be nice is better workflow management. Right now it |
| 12 |
would be nice as an arch dev to be able to get a list of all stable |
| 13 |
requests that have been checked by an arch tester on my arch - that |
| 14 |
isn't simple now. We used to keyword bugs STABLE or TESTED but I don't |
| 15 |
think that anybody is doing that any more - and it breaks down when you |
| 16 |
have 7 archs CC'ed on a bug anyway (which one is tested?). |
| 17 |
|
| 18 |
The fundamental issue, though, is that keywording obscure packages is |
| 19 |
not trivial. I cringe when I see a stable request for some dialup |
| 20 |
networking package - I doubt many devs even own modems these days. |
| 21 |
|
| 22 |
A tool like the one proposed could even raise questions about how more |
| 23 |
obscure packages should be maintained. Maybe all interested users could |
| 24 |
subscribe to a package and then vote on when they will go stable. If |
| 25 |
66% of users interested in a package vote that a package is stable then |
| 26 |
a dev would have discretion to just keyword it without any testing at |
| 27 |
all (obviously this would not be done with critical packages, but the |
| 28 |
world isn't going to end if autopano-sift breaks down on some edge |
| 29 |
case). It also gets users more involved in the QA process and is a |
| 30 |
little less "cathedral" like... |
| 31 |
-- |
| 32 |
gentoo-dev@l.g.o mailing list |
Archives