When I tested my V2 Gertboard this week with the Python programs, I noticed a strange behaviour with the ADC and DAC programs. This has now been identified and the bug squashed.
Basically what was happening was if you ran the leds program, it used the spi ports in normal GPIO mode. A quirk of py-spidev, which texy and bgreat happened to pick up on, within a couple of days of me noticing this issue, means that in order for the spi ports to function properly (in the same session) after they have been used as "normal" GPIO ports, the drivers need to be reloaded.
So the code to achieve this has been incorporated into the dad.py, dtoa.py and atod.py programs.
This fixes the fact that all three of these programs would fail to function correctly if you had previously run the leds program in the same session (I'd never previously run all the tests in one session). Once I've fully tested the fix I will push out a new update.
Moral of the story? It pays to look in on threads you're not directly involved with.