Jump to content

historic_bruno

WFG Retired
  • Posts

    2.755
  • Joined

  • Last visited

  • Days Won

    47

Everything posted by historic_bruno

  1. Working on paths in random maps, here they are in action (rivers and roads)
  2. Hi, hopefully this will be fixed in Alpha 7
  3. It's too late to include in Alpha 6, but if you want to upload the files somewhere, I'm sure people would like to play a new map. Also here is some info about the map types: http://trac.wildfiregames.com/wiki/Design_Document#Maps
  4. Interesting info, it will be useful for map makers However, a Google search did reveal one seemingly well-known oasis in Iran.
  5. Maybe your version of gcc doesn't like their code? Just a shot in the dark, but I can't imagine why you'd get these errors.
  6. 1. I agree with this, it's likely just a bug (they should move in formation to the target and then spread out). 2. I've just raised the limit from 64 to 200, for the next release, we can test that and see how it works. 3. Not sure if this fits in the concept of 0AD as a RTS game, but we do have heroes and we'll have diplomacy eventually which sort of implies having leaders (the player performs all these roles). 4. Yeah the AIs are still in their early stages and will need improvements 5. Technologies and research are planned. 6. I like this idea too, something like Age of Kings' chat commands (31-attack an enemy now, 34-build a navy, etc.), only more explicit.
  7. Strange. Which map were you playing? (I'm assuming you don't get this every time you win)
  8. Hi, there was a problem with the autostart commands for random maps, which has been fixed in r9699. (Also make sure you have a hyphen "-" before each autostart )
  9. I may be missing something obvious, but isn't it intended that units will always be able to pass through a forest, so trees should not even create an obstruction? (Except to siege units, if I remember correctly) Forests are probably the most illustrative example of pathfinder bugs, but I'm not sure we should tailor the algorithms to them if the above is true. Related: I think it would be nice if there was a higher terrain cost applied for passing through a forest, and most of the time units would want to go around if possible.
  10. Thanks to Philip, we've discovered the memory error is not occurring in the map generator context, but when the data is returned - so changing that constant wouldn't help. I'm guessing you use a 64-bit build of 0AD on 64-bit Linux, because Philip was able to reproduce this (I was not, I only have a 32-bit build on 64-bit Windows). I've committed some changes to reduce the size of the random map data. Try the latest revision (r9664) and let us know how it works
  11. On second thought, it can't go next to the unit portrait because when you have groups of many units selected that space can fill up (the individual unit info disappears as well, so we'd have to change that).
  12. You have to use Shift to add enemies to a selection of your own units, then you have control of them all as a group.
  13. I think the stance icons should always be shown, if possible. Either lined up next to the unit portrait or grouped with the formation panel (there's empty space there, too).
  14. If you'd like to try compiling the source code yourself, the constant RMS_RUNTIME_SIZE is defined in \source\graphics\MapGenerator.cpp. Otherwise, I can try gradually increasing the limit but you'd have to wait for the next dev package to test whether it works or not.
  15. That's odd, it seems to always work for me. I also have 8GB RAM on a 64-bit system. Which version of the game are you playing?
  16. Does it always fail to generate a very large or giant map, or only occasionally?
  17. I tried it out today. By the way, you can rename your version of the common-api folder to something like rootbot-api, and then include it in _init.js, that way it's easy to test changes Few comments on the behavior tree: 1. Is the priority ordered from left to right (for nodes of equal depth)? 2. I noticed females were stuck standing at the civil centre and only used for building - maybe they have been flagged as builders and not changed back to gatherers? There's no advantage though to having idle units, so women at least should always be gathering or building something. 3. "if (poor) Economy" seems vague and covers many possibilities, what is a poor economy? Maybe it should be broken up into smaller tasks like the Attack tree, so the behavior is more explicit, for example: if (pop limit reached) -> build houses if (need workers) -> create more workers if (need resources) -> if (need food) -> if (domestic animal near) -> hunt -> if (fruit tree near) -> forage -> if (farm near) -> farm -> else -> build farm -> else -> if (resource site near) -> gather -> else -> build dropsite if (idle workers) -> give orders else -> balance economy for higher level plans Although there may be a reason why the tree was arranged that way, this is just a programmer's point of view, and not an AI designer's 4. Also it may not be clear, but "citizen-soldiers" can gather too. So, let's say you need resources fast and have some idle or defending citizen-soldiers, then you could order them to gather. Their templates use the CitizenSoldier class rather than Worker. As a general comment, the AI causes the game to lag severely on Acropolis map (2v2 AIs). Most of the time the lag is related to "economy update", particularly GatheringManager.update (you can see what's going on by pressing F11 during the game, then a number shown on the left side of the screen to expand the profiling info. The AI stuff is inside of 3-Simulation Update. It's hierarchical, so you can use Engine.ProfileStart/Stop inside of an existing profiling block to get more specific timings.) The resource gathering needs to be more balanced, they should probably be weighted so that there's a calculated minimum number of workers on each type of resource but where the weighting changes depending on various factors (stage of the game, the player's civ, etc.) What I'm seeing is suddenly there's 0 food, so send all workers to a farm, then suddenly we have 0 wood after building houses, so send all the workers to a forest, etc. I didn't see any military activity so that's why I commented only on economic behavior for now. Which command did you try?
  18. What version of the game? I think it has been fixed as of r9595, meaning it will be included in the next release.
  19. Assuming you mean Engine.QueryInterface? That command accesses the interface of a simulation component, or the part of it that's been exposed to simulation scripts. There's no single place that documents all of them as far as I know, they're scattered through scripts (simulation\components) and C++ (source\simulation2\components). Components can be defined in C++ or JS along with their interface. But I believe they have this in common that they can only be accessed by other simulation code, so the AI for instance, must go through AIProxy. Looks like simulation\ai\common-api\base.js is responsible for this, at least for a few of the possible events (the rest like "Attacked" are seemingly not implemented yet). I can't really give specific guidance on that part, as it was designed by Philip/Ykkrosh. For specifics, you might want to find him in IRC The command line arguments should be identical on Windows and Linux.
  20. I don't think it has to do with RAM but rather the amount of memory set aside for random map scripting (which is hard coded to something like 96MB right now).
  21. Just sounds like developing strategy to me, nothing a human player couldn't do
×
×
  • Create New...