Follow-up Comment #2, patch #4926 (project freeciv):
> It was already the case before since bug #20361
I thought we changed it to not be SINGLE_MOVE again with attack_move_cost() in
> Where do you see that the tile is set to TB_IGNORE?
Yes, sorry, the correct value for my comment is "TB_DONT_LEAVE", rather than
"TB_IGNORE". My apologies for the confusion.
> I am against allowing path-finding to make intermediate actions
In that case, rather than calculating a result, let's assert that an action
always consumes the entirety of the remaining moves. For deciding what to do,
this is enough to know that one can reach a tile, or take an action at a tile.
When something is actually done, after the action is completed, the unit can
do something else.
There's a fair bit of AI code that seems to assume pathfinding is magic and
just does everything that probably needs adjusting: when real units are really
moving, it probably makes sense to pathfind in a loop with various objectives
until there are no moves left (after having actually performed the move,
rather than guessing at moves left in pathfinding). This applies not just for
fighters/escorts, but also for city defenders normal attackers, ferries, etc.
Do you think it makes sense to repurpose this ticket to change the action cost
to the remainder of moves remaining in the turn for pathfinding in all cases,
or just mark it "Won't Do"? I no longer think it useful to apply the attached
patch, and would rather fix the individual bits of the AI in a series of
separate patches than trying to do it all here.
Reply to this item at:
Message sent via/by Gna!
Freeciv-dev mailing list