akom Yep I couldn't agree more that it's absolutely unnecessary to set custom build numbers on feature/bugfix/hotfix jobs, but what about release branches? How should they be managed by design in Multibranch pipeline projects? Could you please suggest/shed some light on the intended workflow for them?
The whole multibranch pipeline feature is super hot and we'd absolutely love to switch to it entirely, but still we'd like to keep the build number part of our workflow. Let me describe what it's like and why we want it. Assume our current master build number is N. We branch each release out of master and bump master build numbers to N + M, where M is a "window" that should be enough for this release branch lifetime. We do not build releases on each commit but every 4 hours, which makes it easy to make a fine estimate for M. This scheme makes build comparison straight-forward, by their build numbers. We can track where the branching has happened and determine where in commit history the particular build belongs (approx.) just by looking at its build number. Convenient because you can see if a patch was included into that particular build.
Also, as brydenoliver has mentioned, build numbers might be visible to customers (which is true in our case as well, but we don't set them to something like '43p1'; we just bump the standard numerical build numbers in Jenkins), so I suppose many would like an opportunity to control them.
Please, for the sake of never failing build, help.