边缘世界 RimWorld

边缘世界 RimWorld

Raider Swarm Compression
93 条留言
s_m_w  [作者] 9 月 23 日 下午 11:45 
Never feel bad about delivering feedback! It has been a great help
Skeleton Man 9 月 23 日 下午 11:12 
I genuinely appreciate you circling back and updating this as you have time. I hate to constantly point out perceived issues, but despite them I still play with this installed. I'll see what I can find in the coming days.
s_m_w  [作者] 9 月 23 日 下午 12:35 
Just pushed an update that should fix fake pawns being rendered onto incorrect maps. Also added something that may fix the issue of pawns exiting maps too early. Let me know if you see more odd behavior
Skeleton Man 9 月 5 日 下午 9:44 
I also seem to be having a consistent issue where fake squad members are being rendered on the world map and on other maps they're not actually present on.
s_m_w  [作者] 9 月 3 日 上午 5:15 
Odd. Thanks for the report again, I'll look into it in more detail.
Skeleton Man 8 月 31 日 上午 11:35 
I can confirm at least with sapper tribals that sometimes the squad tries to leave the map after its leader dies even if the map is well within their safe temperature. I think somewhere they end up losing what their job is supposed to be and they end up leaving.
Skeleton Man 8 月 30 日 下午 12:05 
Still having a strange issue i've produced a few times with human enemies. After a few of the squad members die, the squad starts moving towards the edge of the map to despawn. They're not full on fleeing like the raid strategy would do (and im using a mod called No One Ever Flees to prevent automatic fleeing). I used the devmode's inspector to look a bit at one of the fleeing squads and it seems as though somewhere their current job gets set to null, and they end up defaulting to leaving the map.
This could also be due to an unsafe temperature behavior; I'm on an ice cap so its not unreasonable that they'd automatically try to leave the map due to temperature issues. I'll keep testing it out as I play.
s_m_w  [作者] 8 月 28 日 上午 10:08 
AI behavior is really quite hard to test. Everything looks okay as far as I can tell, but please do share if you continue to see unexpected behavior
s_m_w  [作者] 8 月 28 日 上午 10:08 
Pushed another update.

- Raids should no longer flee early
- Invisible raiders should no longer have visible squadmates
- Raider behavior should once again follow the incident's strategy
- Added a message that mentions just how many raiders were compressed
Skeleton Man 8 月 28 日 上午 7:38 
One other reproducable AI issue I could tell you is mechanoids that are dropped with a ship part will stop assaulting the colony even if the ship part is destroyed. they’ll instead wander around the crash site as if it is still there and only attack anything that draws too close. Might be easier to test.
s_m_w  [作者] 8 月 28 日 上午 12:11 
Thanks for the feedback. I didn't observe pawns forgetting to attack in my tests, but I can see how that could happen... especially heavily scripted attacks like Horax cultists could likely be affected. I'm looking into it, especially the low fleeing threshold. Both issues are likely somewhat related
Riven 8 月 27 日 下午 4:27 
I wonder if it could work with the incidents from Dynamic Diplomacy, would be really cool to watch huge armies fight each other
Skeleton Man 8 月 27 日 下午 12:09 
Similar issue, but only played one raid so far. Raiders break and flee very easily even when at high squad sizes.
Some ideas I had to help combat this while preserving performance:
1) Possibly transfering the job from the deceased squadlead to the new one, negating the need to entirely recalculate the job’s requirements? Otherwise dying enemes would trigger new job requests which could get messy.
2) Raider fleeing potentially based on overall unit size? Previously defeating enough entire squads would cause them to break but maybe possible to base it on the total number of attacking enemies including fakes.
3) Haven’t tried it yet, but if an enemy gets downed by a hediff or pain thats not lethal, you think it could be a possible config option to drop that pawn from the group as an injured pawn instead of a dead one?
Hyndis 8 月 27 日 上午 1:01 
While the new update does fix crashes and is perfectly stable and zero crashes since, it seems that stacked pawns forget their orders when the first one dies.

If there's a stack of 10 pawns, you kill 1, the remaining stack of 9 pawns often stop attacking. The stack may wander aimlessly or slowly walk away off the map. This is even true for factions that do not retreat normally, such as horax cultists. It also seems to trigger the raid retreat threshold quickly, counting only the initial pawns in the stack as downed so it thinks the raid has failed and the survivors should retreat even though most pawns are still alive in stacks.

Is there a way to tell stacked pawns to be more aggressive even after they lose the first pawn in the stack? Maybe no retreat until they're down to the last pawn?
s_m_w  [作者] 8 月 26 日 上午 12:00 
Thanks for the report! Just pushed an update to squash that bug. Your reports are very helpful, thank you very much
Skeleton Man 8 月 25 日 下午 5:53 
Hi, me again. I have one new issue to report:
[Pawnname] set to dead while already dead.
Error while killing [Pawnname] during phase 5: System.ArgumentOutOfrangeException: Index was out of range. Must be non-negative and less than the size of the collection.

I think if you kill a pawn too hard, sometimes their mirrorimage hediff gets to squadsize 0 and it can't select a new pawn to take their place.
Skeleton Man 8 月 25 日 下午 1:20 
Was just trying to reload a save that was crashing during a raid, but that update seems to have done the trick. Nice work.
s_m_w  [作者] 8 月 25 日 上午 11:20 
I have just pushed an update that should address all known issues:

- Instead of healing the main pawn and removing a fake, instead lets the main pawn die/be downed and spawns in one of the fakes as the new squad leader
-- Long story short: This should fix pawns losing their weapons as well as letting an entire squad use all their consumable weapons/items
- Found the origin of the crash
- Made all patches significantly less invasive. May (hopefully) lead to even better compatibility with modded content than before, though I have not tested truly exotic enemies, like the Cinders content etc.
s_m_w  [作者] 8 月 25 日 上午 7:30 
Worrying reports. I have an idea what could be happening, but will take me a little bit.
fossette 8 月 25 日 上午 5:19 
When the raid points exceed 50,000 and a mechanoid raid occurs, the game crashes during the battle. The crash timing varies each time, and occasionally the game doesn't crash at all (this happened once out of eight tests). That raid included only Scythers and Lancers. There were no issues during regular human raids(tested four times with raid for me mod).
Skeleton Man 8 月 17 日 上午 9:14 
The CTD's don't generate logs, i've been scouring my own stuff trying to see what's causing it. It might actually be related to explosions after all, but there's always such a mosh pit going on its hard to tell. If I very carefully make sure only bullets are being used, it seems like everything functions as normal.

For the visual bugs, I think it happens when the squad leader has damage that carries over from the next squad member. I found a scyther that had its head blown off but were still active, so all of the other squad members were trying to use a horizontal sprite.

The only common factor I can say has been happening is the targets are squadded mechanoids, and explosives (with CE) seem to be the only consistent contributing factor. Sometimes explosives are fine, sometimes they cause a CTD. I'm still playing with it so i'll let you know.
s_m_w  [作者] 8 月 17 日 上午 1:49 
@Skeleton Man
Thanks for the testing! I thought I had squashed the bug of dropped weapons causing unarmed ghosts, but it appears not quite.. Do you have a log for the crash? I have not seen crashes, nor the visual issues you were reporting that at all during my own testing.
Skeleton Man 8 月 16 日 下午 2:21 
I sometimes noticed a weird issue where trying to forbid the human corpses dropped from squads would instead visually toggle the forbidding of weapons and equipment nearby that were carried by them, instead of the corpses. Maybe a visual issue? Not sure.

I played for a handful of big raids before I got another crash to desktop: this time with mechanoids on the main base. I haven't been using explosives since I suspected that was the crash cause in the first place, so this appeared to happen without correlating to an explosion going off.
Skeleton Man 8 月 16 日 下午 2:21 
Hey, couldn't resist toying with this some more. Just a few interesting things I wanted to report:
I've noticed in the mosh pit sometimes the rendered ghosts or the "real" pawn have some incorrect display. One time I noticed a squad of scyther ghost was horizontal while its squad leader was upright, and another time I saw a squad leader tribal with no head running around. Purely visual issue.

In regards to weapons - This could be a CE or a vanilla thing, but I think if pawns get their arms or hands blown off or have manipulation reduced to 0, they may drop their equipped weapon while still being alive. I think this is causing your subsequent ghosts to appear unarmed. This is particularly apparent with durable but amputatable enemies like Lancers.
億KiotheCloud四 8 月 11 日 上午 7:34 
and cinder raids i think its the tiny turrets that they spawn with thats causing it.
億KiotheCloud四 8 月 11 日 上午 7:34 
It was a mech psychic ship when i finally woke them up crash.
s_m_w  [作者] 8 月 11 日 上午 3:33 
@れPURE い.億KiotheCloud四
Thanks, that log looks helpful. Could you describe what raid it was?
Damion Wolf 8 月 10 日 下午 11:43 
This is really cool. I wish the best for you in squashing buts and your other endeavors. I love seeing this level of creativity.
億KiotheCloud四 8 月 10 日 下午 3:11 
Yeah this mod for some reason just outright crashes my game and its consistently on the same raid/mechs and after removing it i dont get a crash from the raid i dont have the hublib logs but i can provide u the player-prev file hopefully it helps but for the meantime i wont be using it

Its an outright crash no loading just game closes so idk how much this will help but here

https://pastebin.com/A59gEUFa
s_m_w  [作者] 8 月 10 日 上午 2:47 
Added CE and Cinders to the description as potentially being incompatible until I can look into them.

I definitely agree that showing true enemy count would be great, will look into how to do that best.
Latex Santa 8 月 9 日 下午 7:58 
So, Cinders of the Embergarden and Combat Extended have bugs when used with this mod right now.

Good to know.
Skeleton Man 8 月 9 日 下午 2:43 
huge fan of the concept of this mod. i've had a few hours to toy with it in my massive modlist. Unfortunately I have been running into an issue where attacking a group of merged lancers with a CE 90mm flak cannon using explosive shells causes a crash to desktop fairly consistently. I'm not precisely sure what's causing it since it isn't generating any logs for me to look at. But removing this mod and testing the same situation hasn't caused a crash yet. I'll eagerly be waiting out for any updates.

As a QOL suggestion if you can do it, it would be nice to be able to tell the true enemy count somehow. Normally I use allow tool to select all the enemies of a type to figure what i'm up against, but this obviously doesn't work when they're merged with a hediff.
億KiotheCloud四 8 月 8 日 下午 4:12 
I think i found a bug that just crashes the game out right everytime the enemy just "take what they can and leave" specifically from cinders their turrets if compress would become stationary but the moment any of the enemy uproots it cause its valuable enough the game would hard freeze and just crash (tried it without the mod it works fine with this mod instant hard crash) also same turrets even when compress the moment they root them selves they would become one thing which i think isnt intended.
Skeleton Man 8 月 8 日 下午 3:05 
Cant wait to try this when I get home, but I assume if a mod touches the number of points used for incident generation (but is still generating a regular incident) this mod can still take effecf?
s_m_w  [作者] 8 月 8 日 上午 10:06 
@Falco
Err, no. That sounds like an oversight I need to address.
Falco 8 月 8 日 上午 3:43 
Is it normal behavior for Sightstealer duplicates to be perfectly visible whilst the actual pawn isn't? Because that is happening for me right now.
Latex Santa 8 月 7 日 下午 12:40 
@s_m_w
I'll just have to test it on my own, then.
Thanks for your time in answering and making a very promising mod.
s_m_w  [作者] 8 月 7 日 下午 12:35 
@Latex Santa
I wouldn't expect any conflicts, but since I don't know those mods, I can't guarantee it either
Latex Santa 8 月 7 日 下午 12:15 
@s_m_w
How does this mod interact with Smarter Raider AI, or indeed any such improved combat AI mods?
NUT 7 月 29 日 下午 8:09 
Think it would be cool to have this with settlers to i maybe when you recruit them you can turn them into squads or have teams that can work like this.
s_m_w  [作者] 7 月 29 日 下午 1:29 
Just pushed a new update with a large number of bugfixes regarding manhunters:

- A manhunter losing a squadmate by being killed no longer heals its own Scaria in the process (leading to revenge message spam)
- Properly add scaria to squadmates so their graphics get the nice scaria overlay
- Improve pawn rendering for large pawns. Anything larger than a bear had its graphic cut off
- Fixed an bug that led to more squadmates disappearing per death than intended
s_m_w  [作者] 7 月 29 日 上午 1:32 
That does sound annoying. I'll look into it
億KiotheCloud四 7 月 28 日 上午 9:07 
Im having an issue where manhunter packs constantly getting the "revenge" message everytime on of the compress animals dies cause another one would spawn
s_m_w  [作者] 7 月 26 日 下午 1:42 
@Hyndis
Oh that is a funny bug. I'll look into it!
Hyndis 7 月 26 日 下午 12:34 
I love the mod, thank you, its brilliantly clever! I did find a hilarious bug though, visual only. If you have two maps active (main base and a ship visiting somewhere) and a big raid happens on one map you get ghosts on the second map. Purely visual, and a hundred tribal ghosts running through an asteroid is certainly visually striking. Ghosts might almost be a feature though.
Wednesday 7 月 24 日 上午 3:21 
Thanks :)
s_m_w  [作者] 7 月 24 日 上午 1:40 
@Wednesday
Fixed!
s_m_w  [作者] 7 月 24 日 上午 12:32 
@Wednesday
Thank you, I'll take a look!
Wednesday 7 月 23 日 下午 6:43 
Hello,

Found a bug, I could reproduce it in dev quick test.:

1 - Set always replace 20% of raiders with fakes.
2 - Generate a maximum point 'distress call' quest

When trying to land with a gravship in a anomaly's 'distress call' quest map, the rendering breaks and a lot of errors appear.

https://gist.github.com/HugsLibRecordKeeper/92b894b3813df4140e7f28a996ae794f

When entering with a caravan in an anomaly's 'distress call', the game throws some errors, the location despawns and the caravan stays on the world map.

https://gist.github.com/HugsLibRecordKeeper/72925805831f6061d29da352bdeb2084
s_m_w  [作者] 7 月 13 日 下午 11:42 
Found and fixed the issue! Was a very silly mistake on my part..