currently we do a repo rule that downloads the file, then use java_import to wrap that, then use java_library to build the dependency graph.
This is because in the past, java_import didn't allow exports or runtime_deps so we couldn't set up the dependency graph.
That's no longer the case, so we could just directly use java_import and scala_import and reduce the number of targets and also have a side benefit of avoiding producing the empty jars that are produced by java_library.
currently we do a repo rule that downloads the file, then use java_import to wrap that, then use java_library to build the dependency graph.
This is because in the past, java_import didn't allow exports or runtime_deps so we couldn't set up the dependency graph.
That's no longer the case, so we could just directly use java_import and scala_import and reduce the number of targets and also have a side benefit of avoiding producing the empty jars that are produced by java_library.