Have you found some time to have a look at this problem?
Too bad this does not receive any attention. That makes it even harder to rebuild complex jobs in pipeline when you additionally have to abandon the workspace cloning.