concurrent access in scan_saving

Following the discussion in !2402 (closed) I made a test that illustrates the current problem with scans if a custom scan_saving is provided on scan init. Do you think scenario in the test is a valid use-case @sebastien.petitdemange, @matias.guijarro ?

I see two possible solutions to this: a) fixing concurrent access in Scan_Saving b) adding a sort of clone method to Scan_Saving that is called systematically in the init of each scan.