Version: 42.19 Mode: Multiplayer Server settings: Dedicated Mods: Tested with No mods & Mods enabled Save: New Save Zombie's seem to be respawning in claimed safehouse zones, even when the safehouses are fully fenced off. On top of this the setting RespawnUnseenHours does not seem to function as the zombies are respawning where players have been. Our zombie respawn settings are as follows:         ZombieRespawn = 2         RespawnHours = 24.0         RespawnUnseenHours = 336.0         RespawnMultiplier = 0.25 On testing with same safehouse: - Zombies will respawn in numbers of 2 at timed intervils, consistent with our respawn settings - Zombie will respawn when the server is restarted - Zombies will respawn if the house is fully fenced and not fully fenced. - Zombies will respawn INSIDE the house

So far I've only encountered it the once to my knowledge on the first day of the server, but as mentioned it's very easy to overlook since the only sign of the zombies is an inexplicable panic moodle.  At the time the issue occurred I was the only player on the server.  If I manage to notice it happening again I'll try to capture a video clip of it.

Thanks so much for your support and feedback. It's very appreciated. ☺️🙏   Oh yes the ARF has been used internally for a while and it is pure gold I can tell you. ☺️ Especially so with difficult to repro edge-cases. Oftentimes it will tell us exactly why that particular action/anim happened when 99/100 it just doesn't.   Othertimes, like in this case, the clues are in what is NOT there. ☺️ From the data, we can see that the Player's graph is not even attempting a MedicalCheck, is why it's not in the data at all. This eliminates 2-3 layers of logic from the list of suspects, and narrows it down to just the ActionQueue. 🎯   The console gets a bit spammy while the ARF is recording. Sorry about that. Currently it clones the output to console while it's going to the ARF. I'm working on making it just pipe it into the ARF data without spamming console as well. 😅   Imma go rummage in the ActionQueue for why it's refusing to cooperate. Probably gonna add some data probes to it so it'll add any errors/warnings to the ARF as it goes. Should come in handy for debugging in future. ☺️

Doing kind of a necro here....but there are still [B42.19.0] sometimes VHS - Retail tapes around which cannot be watched (cannot even be put into the TV). Not sure if this also occurs in pure Vanilla games. 

@gm.outside Ooof, I see your point. 🤔   The ControllerStateCache has two states being bass-pedalled. The Active state and the Polling state.   The Controller has a gamepadState, which is being read from when we do things like CharacterJoypadButton.Melee etc. Each ControllerState has an array of GamepadState and an array of Controllers.   During ControllerState.set, it assigns all its Controllers the corresponding GamepadState. This is where it's going bad. 😅 Because both bass-pedals have the same pointers to the same Controllers, and each is fighting the other for which gamepadState's are set to the same Controllers.   During the ControllerStateCache.swap, it calls pollingState.set(activeState), which goes and sets all Controllers' gamepadState to that of the Polling state. 😩 Exactly like you said, we are reading from the gamepadState memory that is being actively polled.   Alright. I've gone through and given the Input, ControllerState, and ControllerStateCache a good tidy up. The ControllerState.controllers is gone entirely. We already have arrays of Controllers in a couple places, and the extra was causing exactly this confusion. The ControllerStateCache notifies its bass-pedals when each becomes the Active vs Polling state. ie.  - Active state, sets all Controllers' gamepadState to its own.  - Polled state, just copies its internal state data from the Active state, and gets ready for polling.   I've submitted my changes to QA and code review. They're gonna give it a thorough test and it'll hopefully be in a hotfix soon. Thanks again for your patience. ☺️🙏