 |
This is because the name and model key MUST be the same for such entities, otherwise they are considered actual models (those using actual .lwo or .ase files). There is no clean distinction between model entities and ordinary entities, that's why this behaviour seems ugly. What is the intention of the rename operation? What did you want to achieve? |
 |
I could see why mappers might want a proper name, such as "balcony_stairs" rather than "func_static_234". I guess some sort of "name changed callback" would be needed to automatically update the model key as well. This is kind of ugly, because normally keys should be considered orthogonal, but in this case the original requirement is ugly so a workaround is the only real solution. |
|
|
Then we'll have to make sure that the model key is locked to have the same value as the name if the latter is renamed. I'll look into that when I have time. :) |
|
|
I just fiddled around with this and it seems to be non-trivial. Updating the "model" spawnarg when the "name" value is changed triggers a couple of updates and in the end a bunch of "assertions: keyobservers still attached" messages appear on shutdown. This is probably because during map load the entity key values are set and "name" key changes trigger this update as well. I could of course think of some nasty hacks to disable the "model" updating during map load, but I won't do this. Maybe we can think of another way. |
|
|
Rename operations of func_YYY entities is caught now and the "model" key is adapted accordingly. |
 |
That's all very scary sounding. But it seems to work. :) |
- Anonymous
