I discovered this minor bug because ChicagoPianoTuner implemented my idea to have weights 78.455653%.
The post is here: https://lalescu.ro/liviu/fet/forum/index.php?topic=4151.msg22440#msg22440
Everything is working correctly in the generation. But when reporting the conflicts, FET will write: 0.784557, because it only considers here 6 decimal digits after the decimal point (it should have written 0.78455653).
Also the total sum of the conflicts has the same problem.
They should both (each line of conflict and the total sum of conflicts) write 8 decimal digits after the decimal dot.