* - Added a `IMessageBase` as an interface to `MessageBase`. Existing messages which inherit `MessageBase` (including ones which do not implement Serialize/Deserialize and rely on the weaver) will continue to work in exactly the same way as before!
- Modified any methods which were generic `T where T : MessageBase` to be `T where T : IMessageBase`. this does not change things at all for all existing usages of the method.
These two changes allow end users to implement IMessageBase as a `struct` instead of `MessageBase` as a `class`, so they are no longer forced to allocate message objects in their own code.
* Update .gitignore
* Update Messages.cs
* Update MessageBaseTests.cs
* Update MessageBaseTests.cs
* Update MessageBaseTests.cs
Previously, if you wanted a synclist of structures, you would do: SyncListSTRUCT<MyStruct>
Now you can use SyncList<MyStruct>
SyncListSTRUCT is left there, but obsolete
This allows you to leave rotation out of the NetworkTransform if you only need position.
Right now if you don't need sync rotation and set compression to Lots, it can move your object in ways not expected. This forces you to use no compression as a fix. Using more bandwidth for something you don't need. I think this is all that is needed, tested it in my game and it works.
* Semantic versioning
* Access env variable correctly
* Invoke semantic release
* Not an npm package
* fix(pencil): Testing making a patch release
* c: support feature and breaking as tags
The name SyncObjectProcessor is confusing because it does not process syncobjects. Instead it initializes syncobjects in a NetworkBehaviour.
this avoids the confusion with SyncListStructProcessor, that does process synclists to add serialization methods
SyncListProcessor is a very confusing name, it does not process synclists at all, instead it initializes synclists in the network behaviour.
This is more confusing especially since SyncListStructProcessor does process the synclists to add serialization methods, and both of them normally work
in tandem in different parts of the code
* Fix Scene Handling
Line 361: Passing false prevents the switch to offline scene...pass true instead to forceReload paramater.
Line 588: Never assume Network Manager is in Scene 0 of Build Settings...it often is not.
* Reverted GetActiveScene
Will make separate PR
* Added code to unwind DontDestroyOnLoad
* Fixed typo in comment
* Removed the unwinding of DDOL
