| 1 |
On 05/01/20 15:45, Ulrich Mueller wrote: |
| 2 |
>>>>>> On Sun, 05 Jan 2020, Piotr Karbowski wrote: |
| 3 |
>> I'd like to request Council to define rules regarding maintainership |
| 4 |
>> boundaries and provide guidance regarding under what conditions one is |
| 5 |
>> allowed to make changes to packages that are by metadata.xml maintained |
| 6 |
>> by another party. |
| 7 |
>> The current situation is land of undefined rules and double standards |
| 8 |
>> under disguise of 'common sense'. Although it does work for most part, |
| 9 |
>> it's not uncommon to come across people that are overly territorial, |
| 10 |
>> treating Gentoo packages as their own personal property, who openly |
| 11 |
>> prohibit others from joining them as maintainers on packages, with the |
| 12 |
>> solo reasoning that they feel territorial and do not want others |
| 13 |
>> touching it. |
| 14 |
>> This leads to a situations, where some bugs reported on bugzilla are not |
| 15 |
>> fixed in timely fashion, even when there are other developers that are |
| 16 |
>> willing to fix those bugs and deal with whatever aftermath of doing |
| 17 |
>> those changes would bring. |
| 18 |
>> Because those rules are unsanctioned, we have land of middle |
| 19 |
>> inconvenience where one can never be sure if by declaring maintainer |
| 20 |
>> fimeout and fixing a bug would not bring ComRel on him, for touching the |
| 21 |
>> package one does not maintain. By defining rules and guidelines, it |
| 22 |
>> would greatly benefit Gentoo as a whole as well as reduce the |
| 23 |
>> frustration that come from dealing with people who are gate keeping |
| 24 |
>> while being unable to provide a valid reason why they do not want anyone |
| 25 |
>> toucing their property. |
| 26 |
> This has been discussed a number of times in the past. I believe that |
| 27 |
> the best current practice is outlined in the devmanual: |
| 28 |
> |
| 29 |
> https://devmanual.gentoo.org/general-concepts/package-maintainers/#maintainer-authority |
| 30 |
> |
| 31 |
> Also, "request Council to define rules" normally isn't going to work, |
| 32 |
> because chances are that the Council won't see this as an actionable |
| 33 |
> item. So, if you want the rules to be changed (or better defined), |
| 34 |
> then it will work much better if you come up with a concrete proposal |
| 35 |
> and discuss it on the mailing list. |
| 36 |
> |
| 37 |
> Ulrich |
| 38 |
Agreed with Ulm here, fwiw, this does go back and forth. |
| 39 |
|
| 40 |
The only suggestion I could potentially think of (and may have arisen |
| 41 |
before) is to add some kind of field in metadata.xml about any particular |
| 42 |
salient issues that non-maintainers should be wary of touching, if |
| 43 |
relevant. Otherwise, perhaps another field in metadata.xml as a default |
| 44 |
reply/comment to any bugs opened, explaining that only 'best-effort' |
| 45 |
maintenance is available for this package, and responses may be |
| 46 |
particularly slow. |
| 47 |
|
| 48 |
What also might warrant some further discussion, is how to handle instances |
| 49 |
of devaway where maintainers haven't (most do..) set a policy for their |
| 50 |
packages. |
Archives