Doesn't concern the release directly, but I ran into it while testing and
I don't think it's worth its own bugzilla issue.
http://www.digitalmars.com/d/2.0/operatoroverloading.html#OpAssign
"a op= b
is rewritten as:
a.opOpAssign!("op")(b)"
is misleading. I deduced that op doesn't include "=" but it does.