[BUG][TM2020] `VehicleState::GetVis` can return the wrong CSceneVehicleVis

[AmazingBeu]

In TM2020, if you open the Debug Window in the "Player States", you will see that players can have the same VehicleState values.

Note that the VehicleState::GetAllVis function works well (as you can see in the "All state" tab)

[AmazingBeu]

after investigating a bit, VehicleState::GetPlayerVehicleID can't find the VehicleId and returns 0 except for myself

[AmazingBeu]

Looks like 0x20 is 0x14 now here: https://github.com/openplanet-nl/vehiclestate/blob/3f57bdf0901e70be5e9cfd134f4d0f468edee6bd/Internal/Vehicle/VehicleNext.as#L30C98-L30C98

Tested with the 2023-12-20 update. Do you want i open a pull request for 2 characters? 😅

[codecat]

Thanks! I'm taking a look now.

[codecat]

Looks like this was changed in #2 by @Xandaros (see commit 1e0d97e).

Perhaps this change is no longer valid, or I verified it wrong? Not sure, but a negative offset of 0x14 does seem to give the correct ID.

[Xandaros]

Is this just going to be a recurring thing, where the offset changes every now and again? I can tell you that we had this exact issue with 0x14 and 0x20 fixed it.

I haven't played in a while, so never noticed anything breaking, but I'm assuming the magic number just changed again and broke things. Maybe nadeo undid their change there and so we are back to the old offset? I have no idea.

Either way, if there are issues with the vehicle state, at least we know where to look.

[codecat]

Perhaps. It could also be that it's a buffer of vehicle ID's, as there are more than 1 in that range already, and we have to read the "first" in the buffer. At negative 0x20 for example there is often just 0.