• 0 Posts
  • 69 Comments
Joined 2 years ago
cake
Cake day: July 10th, 2023

help-circle
rss
  • This was a great blog post. I love Rust and Bevy, but I can definitely see why you made the switch.

    The primary issue with your decision to use Rust/Bevy, for me, was that you were taking on the task of getting others to work in a difficult language for novice developers. I would never suggest Rust as someone’s first language, coupling that with a regularly-changing library like Bevy.

    I would love to know what the pros and cons were between Unity and Godot. If you were going to switch to C# anyway, Godot seems like the next logic choice to me, so I’m curious about what your team’s evaluation was for that engine.



  • Basically, you can generalize your trait types into their parent (super) traits for situations when functionality is specific to those supertrait objects.

    As an example, if you have a trait CanBark and it is a super trait for the trait IsDog, you can coerce your references of &dyn IsDog into a &dyn CanBark. You can then work with other trait types that share a super trait.

    trait CanBark {
        fn bark(&self);
    }
    trait IsSeal: CanBark { }
    trait IsDog: CanBark { }
    
    fn bark_as_group(barkers: &Vec<&dyn CanBark>) {
        for barker in barkers {
            barker.bark();
        }
    }
    
    let spot: &dyn IsDog = get_spot();
    let seal: &dyn IsSeal = get_seal();
    let barkers: Vec<&dyn CanBark> = Vec::new();
    barkers.push(spot);  // coerced
    barkers.push(seal);  // coerced
    bark_as_group(&barkers);
    

    At least, I hope this is possible now. If it’s purely “you can return a coerced type from a function”, that is less useful.




  • There are a few different ways to solve this problem without using unsafe and I’m not sure what would be considered idiomatic. Another option is to ultimately encapsulate all of the nodes in a reference-counted box like Rc or Arc and specify the type of your parent/child/sibling references to the same reference-counted box type. With that, you just share cloned instances around as needed.

    The primary drawback here is that for mutable access you end up having to perform a lock every time on an underlying Mutex (or something similar). You also no longer have direct access to the singular instance of the node.

    There are pros and cons to every approach.


  • One way of solving this is to structure all of your nodes into a HashMap with the node ID as the key and the node type as the value. The underlying node type could have node IDs for referencing purposes. You lose the ability to reference the parent/child/sibling directly, but you avoid direct circular dependencies. That said, now you need to manage dangling references for when the node is removed from the main HashMap collection.







  • That’s an interesting take. I didn’t know that the tool screen even existed, so I had no idea that it would be nice to have in my toolbelt for future needs. A README also helps those that may already know a lot about screen and want to know the differences between screen and their implementation.

    There is nothing better in open source than a thorough, well-written README at the root of the project. Wanting others who don’t understand the source of the inspiration to be completely clueless is unfriendly (at best).