Am Thu, Aug 14, 2025 at 12:02:19PM -0700 schrieb Russ Allbery: > > For the record, I will not use MRs routinely for my own packages. I do not > like this workflow for my own changes, and I am going to continue to push > directly to the primary branch unless I need an MR for some reason (such > as triggering checks of a change that I think might fail checks).
Very same here and nearly all teams I'm working in share this habit. We usually have to few potential reviewers and development would be heavily stalled by waiting for reviews. > I don't care if occasional mistakes get into my repositories. +1 > Worrying > about such things excessively is paralyzing and demoralizing and is a > recipe for doing even less volunteer work. +1 > Or it results in the work not happening at all because people are too > annoyed by being told how they should do development to have fun, and > therefore they find something else more enjoyable to do with their free > time. While I generally agree it might make sense to suggest a MR based workflow for devscripts since it contains quite different things where different people are working on. I for myself (as a very rare contributor to devscripts) would feel more safe in this specific case. Kind regards Andreas. -- https://fam-tille.de
