- Apparently, one should not request a reset after connecting. Removing that helped a lot. Well, turned out later that this was a symptom of not restarting the housekeeping session (see V0.9.5).
- MCUs with Stuck-at-1-bits are identified.
- Added code to honor the fact that the ATmega88/168/328 pretend to be P-versions when in debugWIRE mode
- Set the EEARH field in the activation record to EEARL+1. This seems to be the right choice when following the datasheet for the ATmega48. And it led to success in starting the MCU.