Just read the report and as a non-airbus specialist I do not quite follow what happenned in the last minute : why so many things apparently failed or were they disconected on purpose for the test ? (or a mixure of both ) .
If somebody with more knowledge on airbus laws could answer that would help (me at least) .
A (small ) contributing factor is the ATC reaction to the original PLN. The plan was standard IFR one with only mention in field 18 of Training/ferry flight.
The reaction not to allow the required requests ( involving serie of 360s) probably forced the crew to shorten their programme , and perhaps pushed the crew to do certain things in this low altitude approach phase. ( the partial CVR release shows clear pressure to the XL captain from the NZ crew to perform those low speed tests)
There I agree with the BEA recomendation on the adequate planning of such flights. I would even add that it is more than wise to advise ATC BEFORE of what you intend to do. A simple phone call to the ACC Bordeaux prior departure , would have probably eased, if not solved the problem.
I had to deal with such requests many times, and we always tried to find a solution , sometimes involving another route/area or time to perform the tests. Of course when possible, making a proper request 3 days before via fax would guarantee a specific area reserved for this for the duration of the test. While I agree this is not always possible, I think in this case, as it involved the positioning of NZ crews, the 3 days notice could have most probably been done.