I ran into a roadblock on this that I'm trying to work through with the assistance of the AntiSamy team.
The simple update to AntiSamy 1.7.4 caused test failures in 3 existing tests where Validator.getValidSafeHTML returned different cleansed output (although it looks safe to me in all the cases).
I'm working through it, but the result seems unexpected by the AntiSamy folks.
-kevin