Update ModuleCoPFollowTransform #7
Conversation
- Updated ModuleCoPFollowTransform to make better use of part prefabs while still supporting B9PS OnLoad calls.
|
Hey @JonnyOThan, |
| if (followTransform == null || followTransform.name != transformName) | ||
| { | ||
| if (transformName != null) followTransform = part.FindModelTransform(transformName); | ||
| if (followTransform == null) Debug.LogError($"[{MODULENAME}] transformName was empty or does not exist."); |
There was a problem hiding this comment.
would be a good idea to include the transform name that was searched for in the error message.
Oh, right, I can just do that.
| if (followTransform == null) | ||
| { | ||
| // this may be important if someone is swapping out versions of this module with B9PS | ||
| part.CoPOffset = Vector3.zero; |
There was a problem hiding this comment.
This makes sense but I feel like we might not want to set it to zero but to whatever it was in the part config itself, assuming there's a good way to get this at this point.
There was a problem hiding this comment.
This actually isn’t a KSPField so can’t be configured in the cfg. But maybe grabbing the value from the prefab is better
There was a problem hiding this comment.
Huh i swear i’ve seen it configured in the partconfig before, but the prefab should be exactly that i guess.
edit: yeah there’s like a bajillion parts on github here that define a CoPOffset in the part cfg. I assume that works.
There was a problem hiding this comment.
I don't see anything in Assembly-CSharp that would set it from the cfg. Maybe it's in FirstPass? Maybe it's done by reflection somehow.
Oooh yeah it is reflection. Didn't realize parts did that. Any top-level key in the PART cfg will bind to a public field on the Part class even if it doesn't have [KSPField] or [Persistent]
Given stock code, this should always be 0 for everything except parts with ModuleProceduralFairing. But if some mod poked a different value into the prefab, this should be more correct. And if a mod poked a different module into the part instance, then we're kinda screwed anyway.
Resolves #2