Currently trace history is stored in "Resume" context. I'm not aware of any good reason for this but it results in problems during debugging. The point is that one cannot extend currently active trace history using ":trace <varId>" while stopped at breakpoint. The nested trace history is extended instead. But the nested history is freed when the ":trace <varId>" command finishes so there is no access to the trace data. This makes it hard to investigate why the output of the trace command looks the way it was printed: it is not possible to investigate values of variables which contributed to the <varId> value.

A discussion about the reasons for global trace history is here: http://www.haskell.org/pipermail/glasgow-haskell-users/2009-January/016436.html