Assert enabled by PF_WAIT_DEBUG, added since 3.0.0-beta1 and 2.6.4, is currently failing a lot ( -> #41940). While it seems to be finding a genuine error, that error does not seem like a regression. So if there's any regression, it's that the existing error has now been turned more serious with an assert. More importantly, the assert might be overzealous with some rulesets even when the actual bug gets fixed. Things may change during the turn so that unit_move_rate(punit) is currently smaller than at the beginning of turn when moves_left was set. Similar for other variables asserted.
Assert enabled by PF_WAIT_DEBUG, added since 3.0.0-beta1 and 2.6.4, is currently failing a lot ( -> #41940). While it seems to be finding a genuine error, that error does not seem like a regression. So if there's any regression, it's that the existing error has now been turned more serious with an assert. More importantly, the assert might be overzealous with some rulesets even when the actual bug gets fixed. Things may change during the turn so that unit_move_rate(punit) is currently smaller than at the beginning of turn when moves_left was set. Similar for other variables asserted.
I propose disabling PF_WAIT_DEBUG by default.