projectA.aggregate(projectB, projectC) will run tasks on all aggregated projects projectA.dependsOn(projectB, projectC) for classpath dependencies (also will run compile tasks in order)

https://github.com/sbt/sbt-release

release

1. Checks there are no outstanding changes
2. Asks to confirm current release version and next development version
3. Runs test:test
4. Writes release version to version.sbt
5. Commit changes to version.sbt and tags commit
6. publish
7. Updates version in ThisBuild := "nextVersion" in version.sbt
8. Commits version.sbt changes

https://github.com/rtimush/sbt-updates Running dependencyUpdates will show which dependencies can be updated from Maven for each project.

https://github.com/sbt/sbt-buildinfo Add a BuildInfo object to src-managed containing project name, version and anything custom put onto buildInfoKeys. Change namespace of BuildInfo object using buildInfoPackage := "name" Change object name with buildInfoObject := "MyBuildInfo"

buildInfoKeys ++= Seq[BuildInfoKey](
  "xxx" -> 68,
  BuildInfoKey.action("buildTime") {
    System.currentTimeMillis
  },
  libraryDependencies
)   

Why don't unions work? Avro should be looking up a class by name for each of them. Just force the return type to be Any and force the generated field type to be T <: GenericRecord. We can then do pattern matching.

With scala annotation macros you could stub out a class / trait hierarchy and have the Avro fields filled in?

trait Base

@FromSchema("Msg1.avsc")
class Msg1 extends Base

But you still need the stub classes. Unless you macro annotate a package or parent object or something.