Suggestion: Implement Builder Pattern for SaveManager Initialization

[km19809]

Hi team,

I’m new here and have a question regarding the documentation. The docs mention that some functions of the SaveManager struct should not be called twice and that it must be initialized before any access.

Would it be possible to introduce a builder pattern for this?

Here’s a suggestion:

1. SaveManagerBuilder could have init_* functions, with each function returning a specific type of SaveManager.
2. Only types implementing the SaveManager trait would have access and access_with_timer methods, ensuring the save manager is properly initialized.

I would love to hear your thoughts on this.

[gwilymk]

It would probably make sense to. I don't think I've personally been able to get saving correct first try yet 😂.

I'm wondering if we want a higher level API to be the one that's exposed eventually. Maybe it even abstracts away everything and you just ask it to save something Serializable and load something Deserializable and that's it.

Our current focus is on objects and backgrounds, but we'll probably take a look at saving after that since we've gotten it wrong many times.