Ah, I didn't know this bug was present before I started working on the code (post 2.0 release). It makes me feel a little better - I thought it was something I had done.

Please try the latest beta build (ie: from sep 16). I have made radical changes to the way segments are reassembled and I have not seen any of this strangeness.

The way I saw files getting beyond 100% was when new data was coming in while segments were being merged. This is prevented in the latest beta, as well as a bunch of other things.