Yes, the simulation should be correct. Calling the sink_ship function is calling the sink_ship function, and it doesn't matter if the caller is a Lua script or a UI button. Both will send off the same player command. If they don't, it's a bug.
And yes, we do want Widelands not to get into an undefined state when a user performs an action, even if it's sinking a ship or destroying a port. We allow those user actions, so they need to be supported. I have been thinking about this some more, changes in the test suite should be OK in the following 2 cases: 1. The new algorithm needs a bit more time to get back into a well-defined state than the old algorithm. In this case, add a sleep statement to the test suite. 2. The new algorithm gets back into a well-defined state which is different from the well-defined state that the old algorithm used to have. Change the asserts to reflect the new well-defined state. In all other cases, the new algorithm needs fixing. -- https://code.launchpad.net/~widelands-dev/widelands/ship_scheduling_2/+merge/355510 Your team Widelands Developers is subscribed to branch lp:~widelands-dev/widelands/ship_scheduling_2. _______________________________________________ Mailing list: https://launchpad.net/~widelands-dev Post to : widelands-dev@lists.launchpad.net Unsubscribe : https://launchpad.net/~widelands-dev More help : https://help.launchpad.net/ListHelp
