The code was (probably typical) contractor spaghetti code, with plenty of potential bugs. But, no conclusive reproduction steps were ever found - much less any that would cause it to continue acceleration with the brakes on, transmission disengaged, or key off.
That picture of the ETCS board has just reminded me of a time my friend gave me a tour of the Mclaren F1 workshop – he worked on the testing team.
They'd just spent a couple of weeks tracking down an issue where the engine had misfired during one of the races (like, once or twice). Turns out that the tracks on one of the control circuit boards were a tiny bit too close and there had been an arc. I couldn't believe how much work they had put into figuring it out.
Since they are limited to 5 engines per season[0], they were probably very worried that the fault could lead to the loss of an engine. When I did racing, for even the most minor of faults/errors/"that's odd", large amounts of effort were put into correcting the fault if it was on the list of "If this broke, we lost".