Hi Stefano, I need to retract my previous mail. Ironically, it was based on a careless misread of the proposed policy change diff.
On 2024-03-03 00:07, Stefano Rivera wrote: > Now that we have Salsa with Merge Requests, it's hard for me to see much > benefit from having packages in the team with the weak team membership > (uploader). I fully support removing weak team membership. I have always considered having DPT in Uploaders to be pointless. What I misunderstood in the diff was that the proposed policy change of "anyone can commit" to effectively mean that Uploaders is not relevant anymore, but that was not what was proposed. > I don't think any of the things you describe there are acceptable for > team maintenance. > > Disabling tests or docs may be necessary in the short term. Or if they > will never be able to work again. But I don't think those are OK to do, > upload and walk away. > > If the tests are broken and you can't figure it out, you talk to the > people who know the package better. > > We could spell these things out more clearly in the team rules. > We can also push back on team members who behave like this. Repeatedly > doing the things you describe, when asked not to, should lead to being > kicked out of the team. I like that suggestion a lot! And, to be more constructive this time, I want to make it clear that contributions are always truly welcome, that's why it's under DPT. And while diligence (or lack thereof) can be criticized, I'm aware that often, contributors are unaware that what they are doing is wrong. For example, I also often skip tests -- it's just that I do it under conditions that I'm happy to defend (cause isolated, reported upstream, etc.), but others may not be aware of that. Spelling this out more clearly would improve this. > Picking up a bug and realising you are in over year head is something > that will happen to new contributors to Debian. Having a team there to > help out when someone does make a mess like that is useful. > > A few lines in a README.source about what makes a package complex is > probably also useful to your collaborators (although, yes, of course > writing this is work). I'd like that. But to be effective, it must really be in-grained that for anything DPT, one should check README.source first. This may come naturally to many of us, but it's certainly not universal. >> I see Uploaders as a signal of "these are the regular maintainers, I >> should check with these people before doing any *major* changes". And I >> argue that this is reasonable. > > I'd say it's best practice to do that whenever a package has people who > seem to be caring about it. Agreed. > That's not the case for many packages in the team. Even ones listed with > the team in Uploaders and a human as Maintainer. Also agreed. But in that case, this inter-personal conflict usually does not arise, for lack persons. Best, Christian
