heyguosheng/Addressables-Sample

Demo project using Addressables package

Addressables-Sample

Demo project using Addressables package

These samples are broken up into projects based on high level functionality. These are intended as jumping off points for your own development. These have not been tested, and are not guarenteed to work in your situation. They are just examples, to make some concepts easier to understand, or easier to replicate in your own project. Use at your own risk.

Projects

Basic/Basic AssetReference

Several sample scenes to display functionality surrounding the asset reference class.

• Scenes/BasicReference
  • Simplest example using a refrence to spawn and destroy a game object
  • Each object is instantiated directly via the AssetReference which will increment the ref count.
  • Each object spawned has a script on it that will cause it to release itself after a certain amount of time. This destroys the object and decrements the ref count.
  • Any instances still around when the scene closes will automatically be released (decrementing ref count) by the scene closing process (even if their self-destruct script were disabled).
• Scenes/ListOfReferences
  • Showcases using references within a list.
  • Key feature: once an AssetReference is loaded it keeps a member callled .Asset. In this example, you would not want to use the on-complete callback to save off the asset as the loads may not complete in the same order as they were triggered. Thus it's useful that the reference keeps up with its own loaded asset.
  • Here the objects are instantiated via the traditional GameObject.Instantiate which will not increment the Addressables ref count. These objects still call into Addressables to release themselves, but since they were not instantiated via Addressables, the release only destroys the object, and does not decrement the ref count.
  • The manager of these AssetReferences must release them in OnDestroy or the ref count will survive the closing of the scene.
• Scenes/FilteredReferences
  • Showcases utilizing the various filtering options on the AssetReference class.
  • This scene also shows an alternative loading patter to the one used in other scenes. It shows how you can utilize the Asset property. It is recommended that you only use the Asset for ease of load. You could theoretically also use it to poll completion, but you would never find out about errors in that usage.
  • This sample shows loading via the AssetReference but instantiating via Unity's built in method. This will only increment the ref count once (for the load).
  • Currently, the objects created are being destroyed with Addressables.ReleaseInstance even though they were not created that way. As of version 0.8, this will throw a warning, but still delete the asset. In the future, our intent is to make this method not destroy the asset, or print a warning. Instead it will return a boolean so you can destroy manually if needed.

Basic/Scene Loading

The ins and outs of scene loading.

• Scenes/Bootstrap
  • This is the scene to start with. From this one you can transition to the other scenes.
  • "Transition to Next Scene" will open a new scene (non-additively), which will close the current scene.
  • "Add Object" will instantiate an addressable prefab into the scene. The point of this button is to show that these items do not need ReleaseInstance called on them. Should you use the Profiler, you will see they get cleaned up on scene close.
• Scenes/Foundation
  • This is the scene you transition to from Bootstrap.
  • "Load *" buttons will open other scenes additively.
  • "Unload *" buttons will unload scenes that have been additively loaded.
• Scenes/ItemScenes/*
  • These scenes just contain items with no code. Their purpose is to be additively loaded by the Foundation scene.

Basic/ComponentReference

This example creates an AssetReference that is restricted to having a specific Component.

• Scenes/SampleScene
  • This scene has a Spawner game object that alternates between spawning a direct reference prefab and an addressable one.
  • Both the direct reference and the addressable ComponentReference can only be set to one of the prefabs with the component ColorChanger on it.
• Scripts/ComponentReference - ComponentReference
  • This is the class that inhertits from AssetReference. It is generic and does not specify which Components it might care about. A concrete child of this class is required for serialization to work.
  • At edit-time it validates that the asset set on it is a GameObject with the requried Component.
  • At runtime it can load/instantiate the GameObject, then return the desired component. API matches base class (LoadAssetAsync & InstantiateAsync).
• Scripts/ColorChanger - ColorChanger & ComponentReferenceColorChanger
  • The component type we chose to care about.
  • Note that this file includes a concrete version of the ComponentReference. This is needed because if your game code just specified a ComponentReference it could not serialize or show up in the inspector. This ComponentReferenceColorChanger is what makes serialization and the inspector UI work.
  • Releasing a ComponentReference should be done through ReleaseInstance() in the ComponentReference class. To release an instance directly, see our implementation of ReleaseInstance to understand the requirements.

Basic/Sprite Land

2019.3.0a11+ - Sprite demo is back. There was an engine bug causing a crash when loading out of sprite sheets that caused us to remove this demo. This is in 2019.3 alpha, and is being backported to 2019.2 and 2018.4. If you use this demo, and your game crashes, or you get warnings about "gfx" not on main thread, you don't have a fixed vesrion of the platform. There are three sprite access methods currently demo'd in this sample. The on-screen button will change functionality with each click, so you can demo the methods in order. We have some additional mechanisms coming in Addressables 1.2+. We will update this demo once that is out.

• Scenes/SampleScene
  • First is having an AssetReference directly to a single sprite. Since this sprite is a single entry, we can reference the asset directly and get the sprite. This is the most simple case.
  • Second is accessing a sprite from within a sprite sheet. This is the one that was causing a crash, but should be fixed now. Here we load the sprite sheet asset as type IList. This tells addressables to load all the sub-objects that are sprites, and return them as a list.
  • Third is accessing a sprite from within an atals. In this case, you have to use addressables to load the sprite atlas, then use the normal atlas API to load a given sprite from it. This example also shows extending the AssetReference to provide a typed refernce that Addressables doesn't come with (AssetReferenceT in this case).
• All code is in Scripts/SpriteControlTest.cs

Basic/Space Shooter

A very simple Unity tutorial that we converted to use addressables. The main code file to look at would be Done_GameController.cs, but in general, this is just meant as a simple project to explore.

Advanced/Texture Variations

An example project to show one use case or workflow for creating "variants". The new build pipeline (Scriptable Build Pipeline) upon which Addressables is built, does not support asset bundle variants. This old mechanism was useful in many instances, so this sample is meant to show how to accomplish similar results for one of those instances. There are other purpose for variants not shown here. Some will be coming in future samples.

• Scenes/SampleScene
  • In the scene, there's a prefab with an existing texture that can load alternate textures based on button input. (VariationController.cs)
  • The project only has one instance of the texture in question (Texture/tree2.png). It is marked as addressable, with the address "tree" and no labels
  • The group containing "tree" has a custom schema added to it (TextureVariationSchema). This defines the label for the provided texture, and a scale and label for alternate textures.
  • For "Fast Mode" in the player, run with the play mode script of "Vary Fast". This will look at all the label variations defined in the schema, and apply all labels to the "tree". This will then go into play mode in the normal Fast Mode setup of loading from AssetDatabase, but will fake having an "HD", "SD", etc. version of the texture.
  • For "Virtual Mode" in the player, run with the play mode script of "Virtual Variety". This will do the same things as the "Vary Fast" script above. Note, this is not a very accurate virtual mode right now because it does not emulate the fact that each variety should be in its own bundle.
  • With the build script of "Pack Variations" selected, building the player content will:
    • Find all groups with the TextureVariationSchema
    • Loop over all size/label pairs, and copy the textures on-disk.
    • Change the import settings on the created textures so as to scale them.
    • Build all bundles
    • Remove the extra files/groups that were created.
  • After building with "Pack Variations", you can enter play mode using the standard "Packed Play Mode" script.

Advanced/Sync Addressables

Synchronous Addressables! What a crazy thing. The value of exploring this demo can be broken into two categories. One is looking at what would be involved in making addressables synchronous. The other is looking at creating custom providers. On the synchronous front, this can be used as a starting point for making your own project support synchronous loading. As you can see in the code, there are a lot of fail cases, but if you can know that things are on-device and ready to go, it should work.
For custom providers, this project has a couple examples. Custom providers are a really good way to extend addressables. They are relatively easy to create and set up, but open up a lot of opportunity to inject logic during your load or instantiate.

Why don't we put these sync methods in Addressables itself? The best way to understand that is to look at SyncAddressables/SyncAddressables.cs and search for "throw". The code is very specific about how it needs to be used, and will cause pain for the caller if not used in the right way at the right time. That being said, if you want to create a game built on sync interfaces, you can copy this code, and run with it. If you are using it, all the existing async methods would still work, so you are capable of doing a mix & match in your game, if you are willing to accept the constraints when doing things sync. Note that the group schema is what associates a given asset group with either the sync providers or the regular ones. So you could not mix & match within a group.

One common workflow not shown here would have been to set things up to support async loading, but sync instantiation. This would only work if the game always instantiated after loading was complete. That complicates the game-code, but is a simplified version of this demo from the addressables standpoint.

Not all play modes done. Packed content (for play mode, or the player) needs no custom builders. Fast mode and Virtual mode on the other hand do. At this point, we have only implemented a sample script for Fast Mode.

• Scenes/SampleScene
  • This scene waits until the SyncAddressables system has been initialized, and then starts spawning a cube every 60 fixed-update calls.
• SyncAddressables code
  • SyncAddressables.cs - A class that simply calls into Addressables and adds some synchronous guards. Contains methods for:
    • Ready() - True if the main addressables has finished initializing.
    • LoadAsset<>() & Instantiate() - Calls the addressables version of the method, returning the result if things were ready, throwing exceptions if not.
  • SyncBundleProvider.cs - Loads the asset bundle into memory using synchronous methods. If the bundle is online this will fail. Also note, in it's current form, this will fail on Android as loading there is a little more complex. It can load sync, we just didn't have time to add that support to this demo. This is the most likely point in the flow for there to be an issue in the sync process. If this were used in production, it would probably need extended error checking.
  • SyncBundledAssetProvider.cs - Loads from an asset bundle using the synchronous methods. This is unlikely to be a failure point, as it isn't called until the bundle is loaded successfully.
  • Editor/SyncFastModeBuild.cs - Since fast mode does not load from bundles, the default fast mode script has to inject it's own provider for all assets. This custom script just replaces that standard provider with a sync one.
  • SyncAssetDatabaseProvider.cs - An overriden provider to do asset database loads immediately.
  • No Change Needed: SyncBuildScriptPackedMode or SyncBuildScriptPackedPlayMode. Since the group schema allows you to specify provider, the standard build script works as is.
  • Missing - the two main things missing from this demo are Virtual mode and the ability to load from Resources using the sync interfaces.

Advanced/Custom Analyze Rule

This sample shows how to create custom AnalyzeRules for use within the Analyze window. Both rules follow the recommended pattern for adding themselves to the UI. There are no scenes to look at in this project, just analyze code.

• Editor/AddressHasC
  • This is a non-fixable rule (meaning it will not fix itself).
  • When run, it checks that all addresses have a capital C in them. Any that do not are flagged as errors.
  • A rule like this would be useful if your studio enforced some sort of naming convention on addresses. (though it would probably be best if it could fix itself)
• Editor/PathAddressIsPath
  • This is a fixable rule. Running fix on it will change addresses to comply with the rule.
  • When run, it first identifies all addresses that seem to be paths. Of those, it makes sure that the address actually matches the path of the asset.
  • This would be useful if you primarily left the addresses of your assets as the path (which is the default when marking an asset addressable). If the asset is moved within the project, then the address no longer maps to where it is. This rule could fix that.