On 2022-02-02 at 11:21, Michael Stone wrote: > On Wed, Feb 02, 2022 at 09:39:02AM -0600, John Goerzen wrote: > >> On Tue, Feb 01 2022, Russ Allbery wrote: >> >>> I would hate to entirely lose the quality review that we get via >>> NEW, but I wonder if we could regain many those benefits by >>> setting up some sort of peer review system for new packages that >>> is less formal and less bottlenecked on a single team than the >>> current NEW processing setup. >> >> This is a fantastic idea. >> >> In fact, it wouldn't have to bottleneck packages at all. I mean, >> if a quality issue is found in NEW, wouldn't the same be an RC bug >> preventing a transition to testing? > > I'm not sure "nobody ever looked at this" is a suitable criteria for > inclusion in a stable release. We sort of have that problem now in > crusty corners of the archive if someone uploads a bad change, but at > least there's been one review at some point in the package's > lifetime.
Doesn't that, then, lead to the suggestion that any package entering unstable without having undergone NEW review (which, in the revised model, might be every new package) should automatically have a bug filed against it requesting suitable review, and that bug should be treated as a blocker for entering testing? That wouldn't help the "someone uploads a bad change" problem for already-accepted packages, but it would seem to avoid the "nobody ever looked at this" situation. It would also increase the number of automatically-filed bugs by quite a lot, I suspect, which would itself be some degree of downside... -- The Wanderer The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man. -- George Bernard Shaw
signature.asc
Description: OpenPGP digital signature