Currently an idea to discuss, so no version supplied.
Possible prerequisite for #44779. We need a way to access context fields by name, other ones have associate VUT_*, maybe these ones also should. The problem here is that other objects do persist as long as a ruleset is loaded. Tiles persist during the game but we populate the map after we load the ruleset. Units and cities don't persist at all, but they have unique ids. Players for now don't even have any unique identifier within a game (probably they should, for many reasons). If we put requirements on these factors as we have e.g. for governments, we would need to put into them some function (leader name, id, coordinates...) instead of a pointer, that is a deviation from how we construct requirements now.
Also, I have a vague idea of AI reasoning of game objects within a labeled graph, see http://forum.freeciv.org/f/viewtopic.php?t=93031
Currently an idea to discuss, so no version supplied.
Possible prerequisite for #44779. We need a way to access context fields by name, other ones have associate VUT_*, maybe these ones also should. The problem here is that other objects do persist as long as a ruleset is loaded. Tiles persist during the game but we populate the map after we load the ruleset. Units and cities don't persist at all, but they have unique ids. Players for now don't even have any unique identifier within a game (probably they should, for many reasons). If we put requirements on these factors as we have e.g. for governments, we would need to put into them some function (leader name, id, coordinates...) instead of a pointer, that is a deviation from how we construct requirements now. Also, I have a vague idea of AI reasoning of game objects within a labeled graph, see http://forum.freeciv.org/f/viewtopic.php?t=93031