I wonder if the headless server will eventually only use the collision data or will it also load (but probably not use) the meshes? I am currently developing a client-server solution and I have the server nmot yet set to headless because I want to be able to actually see what’s going on on the server so that I can determine if positioning problems are on the server, the client or in between. Later I will switch to headless operation.
My original approach was to replace complex items with simple geometric approximations so that the server would not have to load all the meshes and textures that it doesn’t use anyway later but I am beginning to doubt that this extra work is necessary. If I could be sure that only the collision data is loaded later then I wouldn’t have to do all the extra work of creating a simplified geometry for my items and it would safe me time and work.
Does anybody know for sure how the server will handle this?
PS: There isn’t a forum “Engine Topics” so I picked the forum that most closely relates to the core of my question.