Re Fallersleben, I think you're confusing Fallersleben (2002) and Leiferde (2022).

The former was an actual safety-critical error in the LZB software, leading to the rear train getting an erroneous movement authority after a restart of the LZB control centre. This was fixed with a corresponding software update.

The latter was indeed somewhat comparable to the current situation – a train operating under LZB in partial block mode dropped out of LZB, and the signaller didn't properly verify whether the line ahead was clear before authorising the train to restart and did lead to on-sight running being mandated for that kind of situation (dropping out of LZB in partial block mode).