Ghost Boundary

[DESTROYGIRL]

Description

See "Last-Held-Stationary-Position Teleport" in the design doc

A brush entity that can prevent the ghost from becoming unplayable. When touched by a ghost or a player carrying the ghost, the ghost will be teleported back to where the player was last touching the ground.

For reviewers: Download this map
Spawn a bot, try getting the ghost into the pit. You cannot

neoAssets

Toolchain

• Windows MSVC VS2022

Linked Issues

• fixes There should be a smarter way of blocking the ghost #1175

[DESTROYGIRL]

Magical linux build fail that wasn't my fault?

[Rainyan]

Magical linux build fail that wasn't my fault?

Looks like the CI machine internet just ate dust for whatever reason. Probably not code related 😄

[Rainyan]

Also, this is a complete nit, but could we rename this thing "boundry" -> "boundary"? Because I think it's a typo(?)

[DESTROYGIRL]

Also, this is a complete nit, but could we rename this thing "boundry" -> "boundary"? Because I think it's a typo(?)

Please don't embarass me like that

[Rainyan]

Also, this is a complete nit, but could we rename this thing "boundry" -> "boundary"? Because I think it's a typo(?)

Please don't embarass me like that

Well, it wasn't me who even noticed it but my computer with its red squiggly underlines. Who even knows how English works, it is the hardest programming language of them all

[DESTROYGIRL]

I've changed the test map to match the classname change so you'll need to download it again if testing with it

[Rainyan]

LGTM.

One thing I noticed is that there's a slight difference in ghost teleport location depending on how it goes OOB:

• If carried by a ghoster, it seems to:
  • Teleport to last valid standing pos
  • Have physics disabled
• If shot OOB:
  • Teleports back to the ghost spawn
  • Does not have physics disabled

It would be nice if there's a way to simplify this down to one set of OOB behaviours, so it's more predictable to the player. But maybe this is something to iterate on for some future PR.

(This also somewhat becomes a design-doc thing, so not really relevant to fine-tuning as part of an implementation PR probably.)

[DESTROYGIRL]

It would be nice if there's a way to simplify this down to one set of OOB behaviours, so it's more predictable to the player. But maybe this is something to iterate on for some future PR.

I considered this to be a very rare scenario where this would happen, because it would mean players could manipulate the ghost off an edge from where it spawned without ever holding it. I don't think any maps have ghost spawns this close to an edge, maybe saitama(?), because when a ghost spawns it isn't physics disabled, so I thought the behaviour should be the same. Otherwise the ghost would probably be floating in mid air, or wherever the mapper placed the ghost spawn

[Rainyan]

I considered this to be a very rare scenario where this would happen, because it would mean players could manipulate the ghost off an edge from where it spawned without ever holding it. I don't think any maps have ghost spawns this close to an edge, maybe saitama(?), because when a ghost spawns it isn't physics disabled, so I thought the behaviour should be the same. Otherwise the ghost would probably be floating in mid air, or wherever the mapper placed the ghost spawn

Good points. Mostly thinking in terms of generalising the behaviour, not just for vanilla maps, but for any hypothetical map X with whatever geometry they might come up with. There's also things like a grenade hitting the ghost to make it fly into some pit. And it's also noteworthy that ghost won't only remain at spawnpoints, but it can be dropped at any point in the map either intentionally or by the carrier dying.

[DESTROYGIRL]

There's also things like a grenade hitting the ghost to make it fly into some pit. And it's also noteworthy that ghost won't only remain at spawnpoints, but it can be dropped at any point in the map either intentionally or by the carrier dying.

Not sure I'm following your logic, but if the carrier dies and it falls into a pit, it's just going to teleport back to where the carrier was before they died
The difference here shouldn't be noticeable during regular gameplay, as it will only ever happen if the ghost was pitted before ever being picked up, which means none of the players have actually taken intrest in using it yet. In this scenario, it's as if the ghost had never been knocked into the pit in the first place. If we disable physics on it, players are more likley to notice that

I guess there is the discrepency of how the ghost might land on the ground when recalled back to the spawn VS how it landed originally at the start of the round but that sounds like the mapper placed the spawn too high if it's causing an issue
And if the ghost is spawning and falling into a pit on it's own, I consider that a map issue too

[Rainyan]

I meant this in reference to:

I don't think any maps have ghost spawns this close to an edge, maybe saitama(?)

since even if there's no mapper-designed ghostspawns close to a cliff, player actions could result in a situation like that nonetheless. Since players are free to toss the ghost anywhere, and ghosters could die with the ghost anywhere.

Regardless, this is likely off-topic for this specific PR, so I'll leave it at that. Since this should probably go to Discord, or design-docs. I'm happy to merge what we've got here.

[sunzenshen]

Updates LGTM. Downloaded the test map and checked that the behavior seems reasonable.