Ian and Machinbird
I don't know how applicable would be to the AF A330 scenario, but back during the original 777 flight testing, there was a story going around (from normally reputable sources) that an FAA pilot had inadvertently gotten a 777 into a deep stall at high altitude. The Boeing pilot promptly took over and was able to recover the airplane but the resultant altitude loss was "well over 10k". I can't personally vouch for the accuracy of the story (friend of a friend type stuff) - and assuming it did indeed happen it was kept pretty quiet (no news reports that I'm aware of).
BTW, again, not sure how applicable to how Brand A does things, but Boeing routinely does stall testing of its new aircraft, although not extreme deep stall scenario. Back in the late 1980's I was onboard for many hours of stall testing on a 767. Granted, fairly benign stalls (hold altitude, let the speed drop until the wing stalls and let it fall out of the stall and recover). Also, "windup turns" - used to test inlets at high airflow/high AOA (perform a constant altitude turn with engines set to the desired power, then keep pulling the turn tighter until the aircraft stalls and falls out of the turn). Such testing was always done in daylight in clear air - and during one preflight I recall a request that the flight test pilot "avoid" using the rudder when recovering from the stall, as large rudder swings could slice off the trailing cone which would end the test day. The pilot rather roughly responded that he was going to do whatever he felt necessary to recover the airplane
