|
@@ -1,8 +1,5 @@
|
|
|
--------- ALL LONG TERM TASKS / FIXES BELONG TO GOOGLE DOCS: ImplementationTODO OR PossibleImprovements ----------
|
|
--------- ALL LONG TERM TASKS / FIXES BELONG TO GOOGLE DOCS: ImplementationTODO OR PossibleImprovements ----------
|
|
|
|
|
|
|
|
- New issues:
|
|
|
|
|
- - Since Mesh refactor when I select/deselect a gizmo the entire render texture flashes white for one frame (might be related to RT refactor instead)
|
|
|
|
|
-
|
|
|
|
|
1. Make sure shader create() methods returns a HShader
|
|
1. Make sure shader create() methods returns a HShader
|
|
|
2. Make sure material uses HShader
|
|
2. Make sure material uses HShader
|
|
|
3. Make sure material listens for HShader load event (derives from IResourceListener)
|
|
3. Make sure material listens for HShader load event (derives from IResourceListener)
|
|
@@ -12,12 +9,26 @@
|
|
|
|
|
|
|
|
See GDrive/Resources doc for resources refactor
|
|
See GDrive/Resources doc for resources refactor
|
|
|
|
|
|
|
|
|
|
+ManagedComponent deserialization IS WRONG
|
|
|
|
|
+ - I register the component with ScriptGameObjectManager during deserialization when its ID is wrong
|
|
|
|
|
+
|
|
|
|
|
+Add a special mode to GameObjectManager that is active during deserialization
|
|
|
|
|
+ - Any handle or GameObject deserialized when its active will use a special mapping for resolving/updating IDs
|
|
|
|
|
+ - THIS doesn't really work because I might need to resolve a handle before the object is deserialized
|
|
|
|
|
+ - But could I add entries to the map from both handle and object regardless of which is resolved first?
|
|
|
|
|
+ - YES - This gets rid of the need for post-processing of GameObject, solves my ScriptComponent deserialization issue and allows me to deserialize managed fields on demand
|
|
|
|
|
+
|
|
|
|
|
+Modify managed serialization so that deserialization/serialization happens per field instead of in deserialize/serialize methods during pre- and post-processing step.
|
|
|
|
|
+ - Make sure to fix the GameObjectManager delayed deserialization issue first
|
|
|
|
|
+ - When generating SerializableFieldInfo generate mFieldGlobalIdx (index that takes into consideration any base classes)
|
|
|
|
|
+ - Then when de/serializing use that index to get field info and call get/setFieldData accordingly in RTTI getter/setter directly
|
|
|
|
|
+ - How to use it when deserializing though?
|
|
|
|
|
+ - Ensure that ManagedObjectInfo is deserialized first. Since this is an unclear dependency make sure to add asserts so when it fails it is clear why.
|
|
|
|
|
+ - Then use that to look up field during runtime
|
|
|
|
|
+
|
|
|
I can get mono errors by checking g_print calls in goutput.c
|
|
I can get mono errors by checking g_print calls in goutput.c
|
|
|
- Calling thunks incorrectly can cause those weird errors with no real callstack
|
|
- Calling thunks incorrectly can cause those weird errors with no real callstack
|
|
|
|
|
|
|
|
-Mouse wrap doesn't seem to work the first time
|
|
|
|
|
-If you release mouse outside the window the handle doesn't get deselected and wrap remains active
|
|
|
|
|
-
|
|
|
|
|
Other:
|
|
Other:
|
|
|
Window resize end callback
|
|
Window resize end callback
|
|
|
Add cutoff plane when rendering discs for rotation handle.
|
|
Add cutoff plane when rendering discs for rotation handle.
|
|
@@ -29,11 +40,6 @@ Add ProjectWindow and HierarchyWindow to C#
|
|
|
- Make TreeViews a C# element?
|
|
- Make TreeViews a C# element?
|
|
|
Set up a default layout and save it
|
|
Set up a default layout and save it
|
|
|
|
|
|
|
|
-Test:
|
|
|
|
|
- - If cursor wrap works
|
|
|
|
|
- - If handles work when dragged outside of scene window
|
|
|
|
|
- - If handle is properly released if mouse up is pressed outside
|
|
|
|
|
-
|
|
|
|
|
-----------------
|
|
-----------------
|
|
|
|
|
|
|
|
Need a way to drag and drop items from Scene tree view to Scene view
|
|
Need a way to drag and drop items from Scene tree view to Scene view
|