Getting started (Rust)

Package	Cargo Latest	Downloads
bebop
bebop-tools

There are two packages to be aware of, the Rust runtime bebop and the compiler-helper bebop-tools.

So to get started, add the bebop package to your dependencies and add bebop-tools to your build-dependencies if you plan to use a cargo build script for generating the Rust code. Next, add the necessary steps to your build script to build the schema, this is an example buildscript:

build.rs

use bebop_tools as bebop;

fn main() {
    // download the bebop binary automatically and cache it into your target directory
    // it will automatically download the same version as the package you installed
    bebop::download_bebopc(
        PathBuf::from("target").join("bebopc"),
    );
    // build all `.bop` schemas in `schemas` dir and make a new module `generated` in `src` with all of them.
    bebop::build_schema_dir("schemas", "src/generated");
}

Right now bebop-tools is pretty bare-bones as far as what it supports, but at least with auto-downloading it should help with automation—if not do it for you.

Now assuming you have the schema

schemas/mystruct.bop

struct MyStruct {
  int32 a;
  int32 b;
}

To actually build the generated file run cargo build and you should see src/generated/mystruct.rs is created automatically. You may want to add the generated directory to your git ignore. Also be aware that any changes to files in generated will be wiped out on every single build. Or if you have an IDE that is constantly running in the background like I do, after any change you save…

Also src/generated/mod.rs is generated automatically for you and will publicly declare all the schemas as modules.

In your code, you can now write something like this:

main.rs

pub mod generated;

use bebop::prelude::*;
use generated::mystruct::MyStruct;

fn main() {
    let mut buf = Vec::with_capacity(MyStruct::SERIALIZED_SIZE);
    let s1 = MyStruct {
        a: 123,
        b: -6232
    };
    s1.serialize(&mut buf).unwrap();
    let s2 = MyStruct::deserialize(&buf).unwrap();
    assert_eq!(s1, s2);
}

Pro-tip: Check out the owned module which is generated behind a feature flag. If you enable bebop-owned-all or the appropriate namespace flag (only if you generated with a namespace) you can then access versions of all Records which do not borrow the data but simply take ownership/clone the data. These can be useful if you want to use bebop types throughout your codebase and not just for serialization.

That should be enough to get you kick-started. Checkout the crate docs for the runtime here and pay extra close attention to the Record trait which is the main bebop serialization definition.