Hi Mark, thanks for the swift reply. My Googling did not find this issue, but it indeed describes it fully. I would suggest that you add a bit of documentation to how Gradle handles Maven snapshots, as the Gradle docs do some explaining, but it gets confusing when mixing this with ‘dynamic’ versions and for multiple repo’s.
Sadly this does not solve our issue, as these work-arounds are not transparant, the user will always need to remember to clean Maven local cache himself, or to provide a flag to avoid local Maven usage when desired. This is hard to remember, snapshots may be built weeks ago and for many projects.
I am afraid we will have to avoid local repo completely, and only enable it when the user explicitly asks because he bult a local maven project to be used as dependency. That of course will be slower, as local Maven repo is a lot faster than remote for snapshots. But I also guess that’s why you don’t do multiple repo searches in first place on changing libs. The speed increase from porting Maven to Gradle is huge, but we do face caveats in resolving ldependencies ike this 