No, there was one deanonymisation problem I posited late today which proved to be incorrect (that users can trivially and unwittingly deanonymise themselves, through change addresses, when sending successive post-mixing transactions). Thus far I believe the rest of my assertions to be correct.