No, not cocktail......
Not a big thing but I have spent the best part of an hour solving this so someone may benefit.
Working on some modules with the Xmega32E5 installed and one of them would not work even though it programmed correctly several times. The chip was otherwise dead, the old "finger on the grid trick" (audio hum) indicated the the DAC pin was high impedance but it should not be. After tracing everything I was at the point of replacing the chip when I saw that the Xtal pins were reading a low resistance value in one direction. WHHHAAT?
Then I remembered that all 500 boards had the crystal replaced by the assembly house because I returned the first 50 that had 11MHz crystals instead of 8MHz and things were very strange. Their excuse was that the suppliers had send them the wrong part in the correctly marked bag.
It turned out that a very small solder splash (invisible from my side of the board) was shorting out one of the crystal caps and it must have happened during the repairs.
So the Xmega would program and verify correctly (internal clock) but as soon as the code ran it would turn on the external crystal and wait for the clock to stabilise and that would never happen.