On Mon, Jan 11, 2021 at 11:37 AM Adam Williamson <adamw...@fedoraproject.org> wrote: > > On Mon, 2021-01-11 at 15:41 +0100, Kevin Kofler via devel wrote: > > Matthew Miller wrote: > > > And, also hopefully also a rare occasion, but if this were enabled (and > > > the definitions up to date), problems like > > > > > https://lists.fedoraproject.org/archives/list/us...@lists.fedoraproject.org/thread/CAS6KHTZLR6LUNWEVK3BOIO6HVNQDETZ/#N5HJDKMTGOTL44BT2HZ43LE6Q23345IQ > > > would be caught before they hit users. > > > > This issue was a result of using autopush. See also: > > https://bodhi.fedoraproject.org/updates/FEDORA-2021-5da509fd8e#comment-1807700 > > A maintainer would normally not manually push an update before the library > > update it depends on, the autopush logic was what made that accident happen. > > Well, but maintainers *did*, back before we had autopush. > > You might just as well say "a maintainer would not normally enable > autopush if they're submitting an update that depends on another update > that is not yet stable", because they shouldn't. But they do. All the > time. > > Maintainers get stuff wrong sometimes. We've tweaked the mechanisms in > various ways for more than a decade and that hasn't stopped happening > yet. Even if we changed Bodhi to precisely your preferred > configuration, I'm gonna posit that maintainers would still manage to > get stuff wrong sometimes. :)
Just to be clear: We're working under the assumption that autopush and the other policies we have in place now *reduces* the rate at which mistakes are made with the lowest amount of maintainer overhead. Tweaks can continue to be made, but we're trying to balance convenience with accuracy (while not introducing so much process that people ignore or bypass it). Adam, I think your response came out a little more like "perfect things aren't possible, so why change anything?" rather than the "no solution is perfect, so we're trying to get as close as we can" that I think you actually meant. To rephrase their statements in my own words (and correct me if I get them wrong): Kevin is suggesting that he believes that maintainers should be the sole arbiters of when a package is pushed, not that maintainers are infallible. Adam is saying that we have tried Kevin's approach previously and found it to be less successful on the whole than what we are doing right now. In general, I would probably agree with Kevin *if* all maintainers were as dedicated as Kevin. Unfortunately, experience has shown that this is not the case. Outside of those paid to work on Fedora packaging as part of their day-jobs, most maintainers are hobbyists who work on Fedora whenever they have time (or the mood takes them). As we are essentially *all* volunteers, we cannot load them up with too much process or too many hoops to jump through, lest they decide it's not fun anymore to be part of our community. At the same time, we *do* want to deliver the best Fedora that we can. The best way to do this is, I think, what we have been doing: experimenting with small tweaks in the process over time to determine which pieces have a better convenience:accuracy ratio and then keep iterating. _______________________________________________ devel mailing list -- devel@lists.fedoraproject.org To unsubscribe send an email to devel-le...@lists.fedoraproject.org Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/ List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org
