Why did the merge doubles at point get removed? it was a core component.
Anyways this should fix the performance issue greatly and fix advanced mode literally doing nothing
Remove Zero Weights Improvements
- Add option to not remove 0 weight bones with children for things like skirts
- add option for listing zero weight bones rather than going ahead with deleting them
- add option for removing zero weight bones only if they have deform enabled.
- Added Options to preserve Parent Bones.
- Added List mode only where the user can select the bones there want to remove.
- Added Options to only target Deform bones only and non deform bones only.
This is complete, the UI needs a little cleanup but I do this in a UI cleanup nearer Alpha 2.
Not Complete Yet however this is almost ready for Prime time. This converts the basic human armature for righfy to a Unity standard.
This may work for the more advanced human rigfy armatures as well and I planning to improve this before Alpha 2.
- Fixes issue with addon registration which just randomly broke at some point
- Fixes issue where merge armatures decided to break due to me messing up with properties.
- Fixed issue where you still had to select the mesh in the 3D Scene for viseme creation even though we have a UI selector now.
- Improved typing in some areas.
- Improved code readability in some areas.
- Delete bone constraints would error out if the user is in edit mode, we now start in Object mode first.
- Fixed Eye tracking Ajust string not being in the translation files.
- There is now a selection box to select the mesh in the current active armature for viseme creation instead of the user having to select it in the 3D scene.
- Viseme preview mode won't allow you to start it if your in a other mode, you need to be in Object mode now.
- Combine Materials won't allow you to start it if your in a other mode, you need to be in Object mode now.
- Added Japanese and Korean UI Languages.
- The idea is to have several buttons which kinda mimic what Cats used to do.
- These are very basic, don't work very well, will improve before Alpha 1.
Added the Armature Validation modes now, we have Stritct, Basic and None, it will give a warning to the user in the panel if there have it set to basic or none.
Settings panel added, langauge change has been added back. Did some work on it to slightl improve the system.
Added dubug area, basically everything but autoload will use logging now, you be able to turn it on/off in debug settings.
Did other bits and bobs.
I decided to go through each function and UI section one by one, improving and overhauling things. Each function and section is going to be fully tested and not rushed out.
This is the best way to catch things, but also include the code base as much as possible.
- Re-wrote how the plugin registers itself.
- No longer need @register_wrapper classes get auto detected and added.
- The new Auto loader is much better then the old way, no longer need "if "bpy" not in locals():" this was an old way of doing things and wasn't really efficient.
using auto_load.py provides several advantages:
- It automatically discovers and loads all modules in the addon.
- It handles dependencies between classes correctly through topological sorting.
- It manages registration order automatically.
- It properly handles unregistration in the correct order.
This approach is much less error prone and I not had any issues so far. However it still needs testing fully.
I have also start to re-organise files into folders as well, this is going to be needed so we don't have a long list of files as Avatar Toolkit is getting larger then i originally planned.
- Added VRM to the list of files able to be imported.
- Added error handling for missing vrm importer, will popup to the user telling them to download vrm addon.
- Removed automatically url opening as we should have user consent to open a url, automatic url opening goes against blenders best practices and can be a security concern.
- Optimised the way we import multiple models, the old way just added unnecessary complexity, The try-except block was used to iterate over self.files and set is_multi to True if no exception occurs. However, this is an unnecessarily complex way to determine if the list is non-empty. Simply checking the length of the list is more straightforward and efficient. The way we were using the try-except box could of also lead to silent failures which is not user friendly.
Onan Chew
Yusarina