Without going into the complete technical details, we’ve spent more time reproducing things with Quality Assurance, found more problems, and tried more code tweaks to fix the things we’ve found since the initial hotfix attempts. We thought we fixed the problem, we were wrong, and so we went back to the drawing board. On behalf of all of us at Blizzard I wanted to apologize that we set kind of a bad expectation earlier in the thread.
Unfortunately, the problem was deeper than what we initially had identified. That was what we tried to fix with the hotfix we mentioned above. When we got the initial report of this bug, we had our Quality Assurance folks look into it, and they helped us find some scenarios where the crashes happened.