View Issue Details
|ID||Project||Category||Date Submitted||Last Update|
|0019443||AI War 2||[All Projects] Bug - Gameplay||Feb 17, 2018 11:43 pm||Jul 8, 2019 10:50 am|
|Fixed in Version|
|Summary||0019443: Poor collision detection for Golems|
|Description||So I have a Golem. I can fly a bunch of ships inside it, like it's not bothering to do collision detection at all. Here's a test chamber you can use to replicate it.|
I had the same issue with the original Dyson Sphere, so this is not a new issue.
|Tags||No tags attached.|
||Are you referring to them flying through it, or referring to them stopping inside it? The game doesn't do any collision detection for movement; neither did the first game.|
||Things can fly through the golem like it's not there. which is okay for smaller things but looks really weird for large things. Also the Golem is large enough that you can fit a dozen squads in it, so you can look at your force and think "Oh, there's the golem but where's everything else?!"|
Keith, I think that we're going to have to add forcefield-style collision detection that can be turned on on a per-squad-type basis, based on what he's saying there. I suppose this would apply to king units, Golems, and spire units.
I had at one point been thinking of having these units drop a bit lower down as a way of preventing this from looking strange, but these are just too large for that to work, and it would cause other problems. Golems and spire and king units are rare enough that this shouldn't eat the CPU too much in the sim I assume, but I'm not thrilled if we have to go to starships and guardians of course. I guess it still wouldn't be the absolute worst thing in the world considering some of the past savegames I've seen with tons of forcefields on one planet working fine, but hopefully we can avoid going there.
Feb 18, 2018 1:03 pm
Example.xml (1,054 bytes)
||Added Trello card.|
This is low priority compared to the AI behaviors and the performance stuff, Keith, but in general the third thing I'd like you to focus on is some manner of getting ships to spread out more. I don't know if we should maybe treat the biggest ships as if they are forcefields-to-everything in terms of blocking ships from running through them. That could get messy, expensive, and could lead to exploits like parking big ships on wormholes.
So something softer would be nice, but we don't have the CPU available for "graph untangling" types of spring logic that pushes and pulls ships away (not to mention that would look strange). I'm at a loss, presently, on good algorithms for handling this in an unorganized way. My main thought is that potentially a background thread could run through, process things that are moving in a similar direction in a similar area, and temporarily divert them to make them not stacked up so much. If things pass through each other, that's fine, but avoiding that bunching-up-while-moving thing would be the big winner here.
That's a whole complicated thing on its own, but it's at least theoretically tractable on the CPU.
This will at least partly help:
* Fixed a bug that was letting small ships sit inside large ones. Mainly affecting things like Golems and Arks having other ships sitting inside them.
** The current fix is untested but should probably work.
|Feb 17, 2018 11:43 pm||BadgerBadger||New Issue|
|Feb 18, 2018 12:41 pm||x4000Bughunter||Note Added: 0047001|
|Feb 18, 2018 12:43 pm||BadgerBadger||Note Added: 0047002|
|Feb 18, 2018 12:44 pm||x4000Bughunter||Assigned To||=> keith.lamothe|
|Feb 18, 2018 12:44 pm||x4000Bughunter||Status||new => assigned|
|Feb 18, 2018 12:48 pm||x4000Bughunter||Note Added: 0047003|
|Feb 18, 2018 1:03 pm||BadgerBadger||File Added: Example.xml|
|Feb 19, 2018 10:03 am||Dune||Note Added: 0047007|
|Aug 29, 2018 8:05 pm||x4000Bughunter||Note Added: 0048559|
|Apr 18, 2019 12:58 pm||BadgerBadger||Relationship added||related to 0021035|
|Jul 8, 2019 10:47 am||x4000Bughunter||Assigned To||keith.lamothe => x4000Bughunter|
|Jul 8, 2019 10:47 am||x4000Bughunter||Relationship added||related to 0021240|
|Jul 8, 2019 10:47 am||x4000Bughunter||Relationship added||related to 0021318|
|Jul 8, 2019 10:50 am||x4000Bughunter||Note Added: 0052075|