Re: [ports discussion] enforcing maintainership?
I think that, for one port that I'm maintainer or, there are probably 10+ ports I originated/did major changes on. One thing Solene is right about: what does port maintainership means ? It means there's one person who cares deeply enough about the port that he's going to keep it (more or less) up-to-date and will be the go-to person for changes to that port. As it stands, I don't think "multi-maintainer" ports are a solution. I also don't think keeping everything squeaky up-to-date is the solution either. And finally (surprise!) I don't think that removing ports without recent activity is the solution either. The way Unix works, we got 5000+ ports that don't need much activity nor maintainership. Some of them might be somewhat out-of-date but this doesn't have huge implications, security-wise. The remaining 5000 split among stuff that's trivial to maintain (meaning: it's in good enough shape/portable enough that an update can be done by anyone) and stuff that's trickier. I think the main goal we should have is to get more people on-board. Face it: we are lagging behind badly when it comes to gettin more ports aboard. Yeah, there's the question of quality and all that. Let me wonder a bit. We've got all kinds of tools to catch bad mistakes these days. What's wrong with opening the Ivory Tower a wee little bit more and getting more ports people ? Newcomers may make mistakes. We got more tools to catch those mistakes. And we need tooling for that. More people to work with -> more time to work on stuff that needs working on.
Re: [ports discussion] enforcing maintainership?
On 2022/05/01 01:01, Marc Espie wrote: > On Sat, Apr 30, 2022 at 08:43:13PM +0100, Stuart Henderson wrote: > > On 2022/04/30 16:37, Solène Rapenne wrote: > > > Hi, > > > > > > I don't know if it has been debated before but I'd like to propose a > > > change, or at least discuss about it. > > > > > > - people submitting new ports must become maintainer > > > > > > - ports without a maintainer must find a maintainer > > > > > > - ports where the maintainer is getting removed or not answering MUST > > > find a new maintainer > > > > strong dislike. > > > I'm with Stuart on this, but I will elaborate. > > Our current problem is having enough developers to handle the flow of ports. > > As far as porting new stuff/updating stuff goes, sometimes people will do > the work, but do not want to commit to maintainership. If we try to enforce > that, we may actually push some people away. > > I don't see any real problem this would solve. > > In an ideal world, we would have three times as many dedicated people curating > the tree. But we want relevant ports to keep coming. > Also, a port with an unresponsive maintainer is a block to someone else doing work in ports (either the port itself, or another one which requires changes in some port). You might say this covers it: > > > - ports where the maintainer is getting removed or not answering MUST > > > find a new maintainer But then what if nobody wants to take maintainer for something which is a dependency of a bunch of other ports? Remove them all? Who is going to do that? It's actually quite delicate work. For ports which I have written, I tend to take maintainer if I want to review changes to that port (either something I use fairly often, maybe in production) or know that it's a tricky one, otherwise I often won't because I don't want to block other people working on it.
Re: [ports discussion] enforcing maintainership?
On Sat, Apr 30, 2022 at 08:43:13PM +0100, Stuart Henderson wrote: > On 2022/04/30 16:37, Solène Rapenne wrote: > > Hi, > > > > I don't know if it has been debated before but I'd like to propose a > > change, or at least discuss about it. > > > > - people submitting new ports must become maintainer > > > > - ports without a maintainer must find a maintainer > > > > - ports where the maintainer is getting removed or not answering MUST > > find a new maintainer > > strong dislike. I'm with Stuart on this, but I will elaborate. Our current problem is having enough developers to handle the flow of ports. As far as porting new stuff/updating stuff goes, sometimes people will do the work, but do not want to commit to maintainership. If we try to enforce that, we may actually push some people away. I don't see any real problem this would solve. In an ideal world, we would have three times as many dedicated people curating the tree. But we want relevant ports to keep coming.
Re: [ports discussion] enforcing maintainership?
On 2022/04/30 16:37, Solène Rapenne wrote: > Hi, > > I don't know if it has been debated before but I'd like to propose a > change, or at least discuss about it. > > - people submitting new ports must become maintainer > > - ports without a maintainer must find a maintainer > > - ports where the maintainer is getting removed or not answering MUST > find a new maintainer strong dislike.
Re: [ports discussion] enforcing maintainership?
On Sat Apr 30, 2022 at 04:37:12PM +0200, Solène Rapenne wrote: > Hi, > > I don't know if it has been debated before but I'd like to propose a > change, or at least discuss about it. > > - people submitting new ports must become maintainer > > - ports without a maintainer must find a maintainer > > - ports where the maintainer is getting removed or not answering MUST > find a new maintainer > > To make this more feasable, maybe we could create some small groups of > maintainers for specific areas like language specific ports (ghc, > python etc..). MAINTAINER could be a single person or a list of person, > or refer to a maintainer group defined somewhere else. > > Like, if you join the python maintainer group, you are fine dealing > with all the python packages that doesn't have a maintainer (this would > be set to the group). > What problem are you trying to address?
