Le mer. 5 nov. 2025 à 16:24, Vladimir Sitnikov <[email protected]> a écrit :
> >depMgt and dependencies block > > Those are workarounds as they would imply every end user would have to > manually perform coordinated changes to deps/depMgt sections. > Which is exactly the issue I mention > Not at all, this is a clean solution since you have conflict in transitive pom - depMgt or not, the only one able to solve the conflict in a relevant manner is the integrator *by design*. Anything automatic will work or not depending the use case as you got it explained so these automatic mode relying on transitive data are workaround and "luck" based workarounds, do not reverse it please. > > > > > > *Claim (what I’m showing)With Maven 4 today, a library author cannot safely > split a monolithic artifact into several modules andrelease it as a > backward-compatible minor version (e.g., 1.1.0) without coordinated changes > in downstream applications.* > See https://lists.apache.org/thread/q0bn38qtxkdetdetv4orx6o9lhtonjcxkbtw5f > <https://lists.apache.org/thread/q0bn38qtxkv4orx6o9lhtonjcxkbtw5f> ...this is a claim not a proof and it is wrong, libraries did it - look at tomee to cite just a single one which does it multiple times over last 15 years, no issue, really. Can imply some dependency work in integrator poms sometimes but there is no way to make it right from transitive deps *by design*, you can play and assume you'll get luck. Only exception would be all the deps of your pom are semver and do respect it which is not the case in java. > > Vladimir >
