Implemented in this PTR patch.

Those are interesting. I've been wondering about how best to defend remote miners. At the moment they're defenceless - when invaders appear they will try to run away and wait for a defender to come and clear the room, but they don't move very fast so they usually die. Because it's ~100,000 energy between invasions it feels like it's probably more efficient to let them die, but it feels unsatisfying. This is what got me thinking about unusual part combinations.

Hi. Try to use server from github, not from steam.

Issue seems to have reappeared.

Updated after more consideration and lots of time. This has been an updated article on steam for a while, but I couldn't have the outdated version living on the web without amendments.

I have this problem too. I claimed second room. Unclaim from game menu. After this action: in game I can see when room is empty. Owner: none. But If I try claim again, I got error ERR_GCL_NOT_ENOUGH. I check DB: storage.db['users'].find({username: 'Ufo'}); [ { steam: { id: '76561198335008629' }, cpu: 100, cpuAvailable: 10000, registeredDate: '2019-06-15T13:01:48.873Z', credits: 0, gcl: 2952517, _id: 'aae5b3acdcaa3e6', meta: { revision: 1652367, created: 1560603708874, version: 0, updated: 1560870813412 }, '$loki': 8, active: 10000, username: 'Ufo', usernameLower: 'ufo', badge: { type: 12, color1: '#ebade7', color2: '#d9cf26', color3: '#808080', param: 24, flip: false }, rooms: [ 'W8N3', 'W7N3' ], lastUsedCpu: 2, lastUsedDirtyTime: 4, lastNotifyDate: 1560870034799 } ] and second query: storage.db['rooms.objects'].find({type: 'controller'}); { room: 'W8N3', type: 'controller', x: 23, y: 42, level: 6, _id: '1bc30772347c388', meta: { revision: 608019, created: 1477476557639, version: 0, updated: 1560871483268 }, '$loki': 42, user: 'aae5b3acdcaa3e6', progress: 1174475, downgradeTime: 1005116, safeMode: 20146, safeModeAvailable: 5 }, { room: 'W7N3', type: 'controller', x: 36, y: 18, level: 0, _id: 'a2db077296e87b8', meta: { revision: 2, created: 1477476559214, version: 0, updated: 1560865943559 }, '$loki': 66, user: null, progress: 0, downgradeTime: null, reservation: null, safeMode: null, safeModeAvailable: 0, safeModeCooldown: 907427, isPowerEnabled: false }, // many records cut's How to fix this ?

@demawi All changes to the market will be deployed as a separate patch.

I suspect this is him here. Creeps cannot be on the same tile at the same time. That's really their only rule in this regard. Your first problem is the creeps that got energy from the source aren't trying to move away from the source. Make them move! If you don't use same variant of move (Like 'moveTo(destination);') it won't try to continue it's journey. your second problem is you have way too many creeps in there. Make less creeps! With less creeps, it will be more clear why they aren't moving away from the source. Creeps swapping places: What actually happens is each creep jumps from tile to tile, like a chess piece.They don't have a 'collision radius' and propagate from place to place. If the move was probably legit when the creep.move(dir); (It always distills down to that call.) call was made, it returns '0' and registers the 'intent' to jump to that tile. If the simulation can move the creeps when it runs, the creeps jump to the next tile. On the client, they are moved smoothly around, so it looks like they go through each other.

@the_duck Thanks, we'll take a look

@deft-code said in Big Changes in the Market: I can't imagine getting to GPL 52 with only 2 rooms. How long would that take even with 100% power processing up time? Assuming a shard with 3-second ticks and a sufficient supply of power: Total power = 52521000 = 2704000 Ticks = 2704000/2 = 1352000 (2 power units per tick, since you have 2 rooms) That's just under 47 days on a shard with 3-second ticks.

How much is .energyAvailable in this room?

I would really like this to be fixed. As others have suggested, it would still work conveniently if it was linked to the room.toString representation, but would allow us to easily avoid the auto-linking if we don't want it. @artch, at the very least, if this really isn't going to be fixed please could you point us in the direction of the source code which does this so we can create the best workaround possible. Thanks.

@jbyoshi Thanks for the help; I'm kinda ashamed, seems like the error wasn't caused by _pos_a, but by _pos_b, in my actual code I use the variable "_closest" to find a RoomPosition which i pass to PathFinder.search, i also recently added another variable with the same name wich calculated distance, so in this case _pos_a was correct, but _pos_b was a integral value. I'm dumb, I stayed up too long, programmers need sleep too I guess. Thanks for your help it works again