On Fri, Nov 23, 2012 at 4:34 AM, Markos Chandras <hwoar...@gentoo.org> wrote: > On Fri, Nov 23, 2012 at 4:45 AM, Patrick Lauer <patr...@gentoo.org> wrote: >> I dislike this territorialism. Why add a single point of failure to >> package maintenance? (What if you or Pavel "disappear" for any reason?) >> > Sadly ( or not ;-) ) these are the rules. You have to _at_least_ let > the maintainer know you are going to touch his package. >
Agreed. Pinging a maintainer isn't creating a single point of failure. When I'm eager for a bump and confirm that it is working locally I'll file a bug, including in it a note that a simple bump appears to work, or attaching an ebuild. There might be nuances involved that somebody other than the maintainer might not be aware of. Perhaps upstream has a history of making some particular mistake/variance/etc and the maintainer checks for it on each bump. Maybe bumps to the package should be coordinated with some other package. Maybe a large number of users of the package have expressed some preference to the maintainer about version upgrades and as a result it gets bumped on a schedule. Maybe the maintainer has certain coding preferences and since they're the one in it for the long term they should be respected (think EAPI bumps/etc). Maintainers shouldn't be territorial. However, there are legitimate reasons for having one person or a coordinated team maintaining packages. That's why we have maintainers to begin with. If you say "I'd like to help maintain this" and are told "go away" - that's being territorial. If a random dev makes a commit to a package without at least pinging the maintainer, then a complaint is doing nothing more than pointing out the rule that was broken. Rich
