please, someone could you help me to understand the problem pointed out by Rich in http://dev.clojure.org/jira/browse/CLJ-1766? I didn't get how changing hash and hasheq could hold some reference or avoid seq being GC'ed

(the only other stuff changed in the patch was the deletion of commented out code and the addition of a small test)

@andrewhr He may have misread the comment

I don't see an impl change

Dropped a note in there

@ghadi: thanks a lot!