On Mon, 26 Oct 2026 16:30:32 +0000 Lexi Winter <[email protected]> wrote:
> Harry Schmalzbauer wrote in <[email protected]>: > > On 2026-10-26 15:44, Lexi Winter wrote: > > > it's been broken for five months. who cares? if the answer is anything > > > other than "no one", the person who cares can resurrect the port and > > > maintain it. > >[...] > > > > Is there really an improvement for the ports tree after this removal? > > yes: we used to have a port that didn't build. now we don't have a port > that doesn't build. the quality of the ports tree has improved. > > > From a users point of view, I consider almost any removal as loss, not as an > > improvement. > > what did you lose? the port doesn't build. > > > Somebody put significant efforts into providing this port. > > someone put a lot of effort into porting 4.3BSD to the Tahoe, but we > don't ship that in FreeBSD, because it's obsolete. > > > Hiding from user visible surface doesn't improve anything imho. > > no one is hiding anything, anyone can see the Git history. > > > Further cleanup should require approval by the maintainer imho. > > what maintainer? the port has been broken for five months. it is > clearly not maintained. Maintainers are natural human. So situations like below can always happen. *The maintainer was urgently taken to hospital with some reason like traffic accident and kept 6 months without Internet access. *Luckily back to home and see the maintainer's ports is/are removed without marking as BROKEN, then marked DEPRECATED. How the maintainer feels? I think this is the enough reason to always take described steps before removal. > > If (former) maintainer doesn't respond within a reasonable timeframe > > (like 8 weeks maybe) set expiration date (another 4 weeks). > > it's been five months. that's far longer than your proposed timeline. > > > People have another chance to notice about it and maybe adopt it. > > if people are really using this port, they would have noticed that it > cannot be built, and they would have fixed it -- if they cared. > > clearly these theoretical users do not actually exist. The first post by Harry should be indicating the fact. Isn't it? > > This tiny bureaucracy is justified by the resulting transparency in my > > opinion. > > the real bureaucracy is being unable to improve the ports tree due to > objections on behalf of users who don't actually exist. Agreed with this specific point. But non-maintainer users may lock the pkg (if they know it's possible and how-to) if pkg(8) attempt to delete it as there's no built pkg in the repo and forget about it unless already-installed one still work for them. They unlikely to "officially" report. Some limited fractures of them "could" post to forums.freebsd.org "Why is pkg want to remove ***?!", but it can be overlooked. Posts (toots / tweets) to SNS would more easier NOT to be noticed by devs. And unfortunately, registering at bugs.freebsd.org as a user seems NOT as straightforward as when I've registered, as of spammers. Newbies could be stuck there and simply resign with filing a PR. So, following steps and warn users via pkg (8) on each upgrade runs could give more chances for potential users to notice the situation with BROKEN/DEPRECATED messages. Isn't it? Regards. -- Tomoaki AOKI <[email protected]>
