> > Commits onto a git repo can be reversed easily (via `git revert` > > or forced push, if one really wants to), but problematic uploaded > > packages to the archive are irreversible and might cause broader > > damages. That is why people are okay with allowing random git commits > > to Debian group but are not okay with uncoordinated uploads. > > Yes, I fully agree with that.
++1 Please don't suggest that people should lower their mental barrier to upload to Debian archives. Not only is it likely to cause extra work to the package's original maintainer, it is also more likely to break stuff in e.g. unstable and then cause extra work to everyone doing Debian packaging. For example, if apt or grep has a regression, a massive amount of builds and installs of other packages will immediately stop working. What if we had a culture that people file Merge Requests on Salsa against the package with whatever improvements they have to suggest and then wait to see if the maintainer is a) active b) in agreement with the change.? The reviewer/+1 does not even have to be the original maintainer, it can be anyone in with Salsa access who discovers the MR and has interest in it. As an added benefit the MR should also run Salsa CI to show that there are no testable regressions. If there is no response for multiple weeks, then the people who have open MRs against that package could coordinate and do the upload instead of the (absent) original maintainer.
