On Wed, 2018-08-01 at 12:42:35 +0800, David Bremner wrote: > Guillem Jover <guil...@debian.org> writes: > > Some packages might not show activity for longish periods of time, > > because maintainers batch changes, for example to do at least one > > upload per release, with general packaging and QA updates/improvements, > > etc. > > > > Also there might be bugs open that are difficutl to fix (with no > > patch), etc, that might show no activity for a long time. > > > > So I'd probably qualify the requirement above. I'm not entirely sure > > how though. I mean [c] kind of covers it superficially, but I'm not > > sure that's clear enough or if the intention was something along these > > lines. For example, if there's a difficult bug open, and then a patch > > is sent and gets no reply, that could count as inactivity in my book, > > but not otherwise. > > Would it be a big hassle for you to reply to a hypothetical salvaging > bug saying essentially what you just wrote, or maybe some short form?
For me personally, I guess not, mostly because I don't maintain many packages which follow this pattern. But this was more of a global view comment, not as how it might affect me. > As > far as I understand that would be enough to stop any salvaging process. > I agree that it's a bit of busywork for the maintainer, but maybe that > bug could be left open and marked wontfix to avoid too much opening and > closing of bugs. I guess the root problem is that everything else in the proposal seems so concrete and clear-cut, all other conditionals for triggering the process are very explicit, except for this one, which seems fuzzy and loop-holey, which gives the impression that in contrast to the other points there's much room for the initiation process on conditions that might not be appropriate. Also while this (dealing with such bugs) might be fine for someone who maintains few packages, consider, say, something like a huge team, like perl :), that might batch changes for many packages and perform uploads just once per release or similar. Also my feeling is that if we have to keep bugs open as a notice for something that might actually be a normal pattern, that looks like the conditions are not correct. In comparison, what gregor proposed (the simple version) in the other thread, while less explicit, seems apparently more internally consistent and less loop-holey! Because it just gives some rough guidelines and delegates to the initiator's good judgement. I'm not sure that's better, though because in cases like this, where these might be contentious issues perhaps we'd better be more explicit than implicit, dunno. Hope that this helps explain the issue I see. :) Thanks, Guillem
