Date: Thu, 10 Mar 2005 19:34:47 -0600 (CST) From: Mark Linimon <linimon@lonesome.com> To: David O'Brien <obrien@FreeBSD.org> Cc: ports-committers@FreeBSD.org Subject: Re: cvs commit: ports/net/rdesktop pkg-plist Message-ID: <Pine.LNX.4.44.0503101916360.12714-100000@pancho> In-Reply-To: <20050310235755.GB58785@dragon.nuxi.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 10 Mar 2005, David O'Brien wrote: > Yet portmgr can't address issues of port stealing? No portmgr replied > to the thread when elk hijacked the 'bash' port from me. What portmgr can't do is to arbitrate every single conflict between every single committer. portmgr can only ask that the committers work together. Only in cases where this completely and utterly fails, do the people on portmgr feel compelled to step in. If you want a system in place where every single decision that's made has to be justified, then progress is going to start becoming much slower, not much faster -- and 'slowness' is the major complaint people seem to have with portmgr. In the meantime there are two rules for commits: the two-week timeout for maintainer commits and the three-month timeout for maintainer resets. These are compromises to try to satisfy as many people as possible. It's hopeless to expect that they will satisfy everyone -- or, in this particular case, anyone at all. Of course, if all our committers were maintaining their ports and closing their assigned PRs promptly, then this would never be a problem. Finally, 'maintainership' is a privilege, not a right. What ought to be our view, as ports committers and maintainers, is that the users are the most important. To the extent that some individual wants to commit patches, handle PRs, and answer questions from users, then they deserve to be listed as a maintainer. If not, then, IMHO, they don't -- despite whatever their past work on that port has been. > It boggles my mind that a comment fix to bsd.port.mk has to be run > thru an experimental build. Not that I think you'll believe me, but I put forward the point of view on the portmgr@ list that we were indeed being too conservative in this regard. However, my view didn't carry the day. portmgr is not a monolithic entity and works by 'best consensus'. If the ports tree were branched then it's probably true that commits to bsd.port.mk could be loosened. But since it isn't, and hundreds if not thousands of people using every single release of FreeBSD rely on it never, ever, breaking, then IMHO we need to continue to be hyper-conservative with it. mcl
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.LNX.4.44.0503101916360.12714-100000>